Process Systems

Engineering for Biofuels

Development
 Wiley Series
in
Renewable Resources
Series Editor:
Christian V. Stevens, Faculty of Bioscience Engineering, Ghent University, Belgium

Titles in the Series:

Wood Modification: Chemical, Thermal and Other Processes
Callum A. S. Hill
Renewables-Based Technology: Sustainability Assessment
Jo Dewulf, Herman Van Langenhove
Biofuels
Wim Soetaert, Erik Vandamme
Handbook of Natural Colorants
Thomas Bechtold, Rita Mussak
Surfactants from Renewable Resources
Mikael Kjellin, Ingegärd Johansson
Industrial Applications of Natural Fibres: Structure, Properties and Technical Applications
Jörg Müssig
Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power
Robert C. Brown
Biorefinery Co-Products: Phytochemicals, Primary Metabolites and Value-Added Biomass
Processing
Chantal Bergeron, Danielle Julie Carrier, Shri Ramaswamy
Aqueous Pretreatment of Plant Biomass for Biological and Chemical Conversion to Fuels and
Chemicals
Charles E. Wyman
Bio-Based Plastics: Materials and Applications
Stephan Kabasci
Introduction to Wood and Natural Fiber Composites
Douglas D. Stokke, Qinglin Wu, Guangping Han
Cellulosic Energy Cropping Systems
Douglas L. Karlen
Introduction to Chemicals from Biomass, 2nd Edition
James H. Clark, Fabien Deswarte
Lignin and Lignans as Renewable Raw Materials: Chemistry, Technology and Applications
Francisco G. Calvo-Flores, Jose A. Dobado, Joaquín Isac-García, Francisco J. Martin-Martínez
Sustainability Assessment of Renewables-Based Products: Methods and Case Studies
Jo Dewulf, Steven De Meester, Rodrigo A. F. Alvarenga
Cellulose Nanocrystals: Properties, Production and Applications
Wadood Hamad
Fuels, Chemicals and Materials from the Oceans and Aquatic Sources
Francesca M. Kerton, Ning Yan
Bio-Based Solvents
François Jérôme and Rafael Luque
Nanoporous Catalysts for Biomass Conversion
Feng-Shou Xiao and Liang Wang
Thermochemical Processing of Biomass: Conversion into Fuels, Chemicals and Power, 2nd
Edition
Robert C. Brown
The Chemical Biology of Plant Biostimulants
Danny Geelen, Lin Xu
Chitin and Chitosan: Properties and Applications
Lambertus A.M. van den Broek and Carmen G. Boeriu
Biorefinery of Inorganics: Recovering Mineral Nutrients from Biomass and Organic Waste
Erik Meers, Evi Michels, Rene Rietra, Gerard Velthof
Process Systems Engineering for Biofuels Development
Adrián Bonilla-Petriciolet, Gade Pandu Rangaiah
Forthcoming Titles:
Waste Valorization: Waste Streams in a Circular Economy
Carol Sze Ki Lin, Chong Li, Guneet Kaur, Xiaofeng Yang
Biobased Packaging: Material, Environmental and Economic Aspects
Mohd Sapuan Salit, Rushdan Ahmad Ilyas
High-Performance Materials from Bio-based Feedstocks
Andrew J. Hunt, Nontipa Supanchaiyamat, Kaewta Jetsrisuparb, Jesper T.N. Knijnenburg
 Process Systems
Engineering for Biofuels
Development

Edited by

ADRIÁN BONILLA-PETRICIOLET
Instituto Tecnológico de Aguascalientes, México

GADE PANDU RANGAIAH

National University of Singapore, Singapore
and
Vellore Institute of Technology, India
This edition first published 2020
© 2020 John Wiley & Sons Ltd.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any
form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice
on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah to be identified as the authors of the editorial material
in this work has been asserted in accordance with law.

Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Office
The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

For details of our global editorial offices, customer services, and more information about Wiley products visit us at
www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in
standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of Warranty

In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of
information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and
evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent,
or device for, among other things, any changes in the instructions or indication of usage and for added warnings and
precautions. While the publisher and authors have used their best efforts in preparing this work, they make no
representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically
disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular
purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional
statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or
potential source of further information does not mean that the publisher and authors endorse the information or services
the organization, website, or product may provide or recommendations it may make. This work is sold with the
understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained
herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers
should be aware that websites listed in this work may have changed or disappeared between when this work was written
and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial
damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging-in-Publication Data

Names: Bonilla-Petriciolet, Adrián, editor. | Rangaiah, Gade Pandu, editor.

Title: Process systems engineering for biofuels development / edited by

Adrián Bonilla-Petriciolet, Dept. Chemical Engineering, Instituto
Tecnológico de Aguascalientes, Aguascalientes, México, Gade Pandu Rangaiah,
National University of Singapore, Singapore.
Description: First edition. | Hoboken, NJ : John Wiley & Sons, Inc., [2020]
| Series: Wiley series in renewable resources | Includes bibliographical references and index.
Identifiers: LCCN 2020016306 (print) | LCCN 2020016307 (ebook) | ISBN
9781119580270 (cloth) | ISBN 9781119580317 (adobe pdf) | ISBN
9781119580331 (epub)
Subjects: LCSH: Biomass energy. | Chemical processes. | Systems engineering.
Classification: LCC TP339 .P753 2020 (print) | LCC TP339 (ebook) | DDC
662/.88–dc23
LC record available at https://lccn.loc.gov/2020016306
LC ebook record available at https://lccn.loc.gov/2020016307

Cover Design: Wiley

Cover Image: © Jim Barber/Shutterstock

Set in 10/12pt, TimesLTStd by SPi Global, Chennai, India.

Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY

10 9 8 7 6 5 4 3 2 1
 Contents
List of Contributors xiii
Series Preface xv
Preface xvii

1 Introduction 1
Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah
1.1 Importance of Biofuels and Overview of their Production 1
1.2 Significance of Process Systems Engineering for Biofuels
Production 3
1.2.1 Modeling of Physicochemical Properties of
Thermodynamic Systems Related to Biofuels 4
1.2.2 Intensification of the Biomass Transformation Routes for
the Production of Biofuels 5
1.2.3 Computer-Aided Methodologies for Process Modeling,
Design, Optimization, and Control Including Supply Chain
and Life Cycle Analyses 7
1.3 Overview of this Book 9
References 11

2 Waste Biomass Suitable as Feedstock for Biofuels Production 15

Maria Papadaki
2.1 Introduction 15
2.1.1 The Need for Biofuels 15
2.1.2 Problem Definition 17
2.1.3 The Biomass Pool 18
2.2 Kinds of Feedstock 20
2.2.1 Spent Coffee Grounds 21
2.2.2 Lignocellulose Biomass 22
2.2.3 Palm, Olive, Coconut, Avocado, and Argan Oil Production
Residues 25
2.2.4 Citrus 33
2.2.5 Grape Marc 36
viii Contents

2.2.6 Waste Oil and Cooking Oil 37

2.2.7 Additional Sources 38
2.3 Conclusions 40
Acknowledgment 40
References 40

3 Multiscale Analysis for the Exploitation of Bioresources: From Reactor

Design to Supply Chain Analysis 49
Antonio Sánchez, Borja Hernández, and Mariano Martín
3.1 Introduction 49
3.2 Unit Level 50
3.2.1 Short Cut Methods 50
3.2.2 Mechanistic Models 51
3.2.3 Rules of Thumb 56
3.2.4 Dimensionless Analysis 56
3.2.5 Surrogate Models 56
3.2.6 Experimental Correlations 59
3.3 Process Synthesis 60
3.3.1 Heuristic Based 60
3.3.2 Supestructure Optimization 61
3.3.3 Environmental Impact Metrics 65
3.3.4 Safety Considerations 66
3.4 The Product Design Problem 66
3.4.1 Product Design: Engineering Biomass 66
3.4.2 Blending Problems 68
3.5 Supply Chain Level 68
3.5.1 Introduction 68
3.5.2 Modeling Issues 70
3.6 Multiscale Links and Considerations 71
Acknowledgment 74
Nomenclature 74
References 75

4 Challenges in the Modeling of Thermodynamic Properties and Phase

Equilibrium Calculations for Biofuels Process Design 85
Roumiana P. Stateva and Georgi St. Cholakov
4.1 Introduction 85
4.2 Thermodynamic Modeling Framework: Elements, Structure,
and Organization 86
4.3 Thermodynamics of Biofuel Systems 88
4.3.1 Phase Equilibria 88
4.3.2 Thermodynamic Models 90
4.4 Sources of Data for Biofuels Process Design 98
4.5 Methods for Predicting Data for Biofuels Process Design 102
4.5.1 Group Contribution Methods for Biofuels Process Design 103
4.5.2 Quantitative Structure–Property Relationships for Biofuels
Process Design 105
 Contents ix

4.6 Challenges for the Biofuels Process Design Methods 109

4.7 Influence of Uncertainties in Thermophysical Properties of Pure
Compounds on the Phase Behavior of Biofuel Systems 112
4.8 Conclusions 114
Acknowledgment 114
Exercises 114
References 115

5 Up-grading of Waste Oil: A Key Step in the Future of Biofuel

Production 121
Luigi di Bitonto and Carlo Pastore
5.1 Introduction 121
5.2 Physicochemical Pretreatments of Waste Oils: Removal of
Contaminants 124
5.3 Direct Treatment and Conversion of FFAs into Methyl Esters 125
5.3.1 Homogeneous Catalysis: Brønsted and Lewis
Acids 125
5.3.2 Heterogeneous Catalysis 127
5.3.3 Enzymatic Biodiesel Production 128
5.3.4 ILs Biodiesel Production 130
5.3.5 Use of Metal Hydrated Salts 133
5.4 Future Trends of the Pretreatments of Waste Oils 139
5.5 Conclusions 140
Acknowledgment 141
Abbreviations 141
References 142

6 Production of Biojet Fuel from Waste Raw Materials: A Review 149

Ana Laura Moreno-Gómez, Claudia Gutiérrez-Antonio, Fernando Israel
Gómez-Castro, and Salvador Hernández
6.1 Introduction 149
6.2 Waste Triglyceride Feedstock 150
6.3 Waste Lignocellulosic Feedstock 159
6.4 Waste Sugar and Starchy Feedstock 164
6.5 Main Challenges and Future Trends 165
6.6 Conclusions 167
Acknowledgments 167
References 167

7 Computer-Aided Design for Genetic Modulation to Improve Biofuel

Production 173
Feng-Sheng Wang and Wu-Hsiung Wu
7.1 Introduction 173
7.2 Method 175
7.2.1 Flux Balance Analysis 175
7.2.2 Flux Variability Analysis 176
7.2.3 Minimization of Metabolic Adjustment 176
x Contents

7.2.4 Regulatory On-Off Minimization 177

7.2.5 Optimal Strain Design Problem 177
7.3 Computer-Aided Strain Design Tool 179
7.4 Examples 181
7.4.1 E. coli Core Model 181
7.4.2 Genome-Scale Metabolic Model of E. coli iAF1260 183
7.5 Conclusions 185
Appendix 7.A: The SBP Program 187
References 187

8 Implementation of Biodiesel Production Process Using

Enzyme-Catalyzed Routes 191
Thalles Allan Andrade, Massimiliano Errico, and Knud Villy Christensen
8.1 Introduction 191
8.2 Biodiesel Production Routes: Chemical versus Enzymatic Catalysts 194
8.2.1 Chemical Catalysts 195
8.2.2 Enzymatic Catalysts 196
8.3 Optimal Reaction Conditions and Kinetic Modeling 198
8.3.1 Evaluation of the Reaction Conditions 199
8.3.2 Kinetic Modeling 201
8.4 Process Simulation and Economic Evaluation 205
8.5 Reuse of Enzyme for the Transesterification Reaction 210
8.5.1 Recovery of Eversa Transform by Means of Centrifugation 210
8.5.2 Recovery of Eversa Transform by Means of Ceramic
Membranes 211
8.6 Environmental Impact and Final Remarks 215
Acknowledgments 217
Nomenclature 217
References 217

9 Process Analysis of Biodiesel Production – Kinetic Modeling,

Simulation, and Process Design 221
Bruna Ricetti Margarida, Wanderson Rogerio Giacomin-Junior,
Luiz Fernando de Lima Luz Junior, Fernando Augusto Pedersen Voll,
and Marcos Lucio Corazza
9.1 Introduction 221
9.1.1 Homogeneous-Based Reactions 222
9.1.2 Heterogeneous-Based Reactions 223
9.1.3 Enzyme-Catalyzed Reactions 224
9.1.4 Supercritical Route Reactions 224
9.1.5 Methanol or Ethanol for Biodiesel Synthesis 224
9.2 Getting Started with Aspen Plus V10 224
9.2.1 Pure Compounds 225
9.2.2 Mixture Parameters 229
9.3 Kinetic Study 232
9.3.1 Esterification Reaction 232
 Contents xi

9.3.2 Experimental Reaction Data Regression 234

9.3.3 Transesterification Reaction 236
9.3.4 Supercritical Route 238
9.4 Process Design 239
9.4.1 Esterification Reaction 239
9.4.2 Methanol Recycling 243
9.4.3 Transesterification Reaction 244
9.4.4 Biodiesel Purification 245
9.4.5 Additional Resources 248
9.5 Energy and Economic Analysis 252
9.6 Concluding Remarks 254
Acknowledgment 255
Exercises 255
References 256

10 Process Development, Design and Analysis of Microalgal Biodiesel

Production Aided by Microwave and Ultrasonication 259
Dipesh S. Patle, Savyasachi Shrikhande, and Gade Pandu Rangaiah
10.1 Introduction 259
10.2 Process Development and Modeling 262
10.3 Sizing and Cost Analysis 272
10.4 Comparison with the WCO-Based Process of the Same Capacity 277
10.4.1 Biodiesel Process Using WCO as Raw Material 277
10.4.2 Comparative Analysis 277
10.5 Comparison with the Microalgae-Based Processes 280
10.6 Conclusions 280
Acknowledgment 281
Appendix 10.A 281
Exercises 282
References 282

11 Thermochemical Processes for the Transformation of Biomass into

Biofuels 285
Carlos J. Durán-Valle
11.1 Introduction 285
11.2 Biomass and Biofuels 288
11.3 Combustion 289
11.4 Gasification 290
11.4.1 Fixed Bed Gasification 291
11.4.2 Fluidized Bed Gasification 292
11.4.3 Dual Fluidized Bed Gasification 292
11.4.4 Hydrothermal Gasification 293
11.4.5 Supercritical Water Gasification 294
11.4.6 Plasma Gasification 294
11.4.7 Catalyzed Gasification 295
11.4.8 Fischer–Tropsch Synthesis 295
xii Contents

11.5 Liquefaction 296

11.6 Pyrolysis 296
11.6.1 Slow Pyrolysis 297
11.6.2 Fast Pyrolysis 297
11.6.3 Flash Pyrolysis 297
11.6.4 Catalytic Biomass Pyrolysis 303
11.6.5 Microwave Heating 304
11.6.6 Product Separation 304
11.7 Carbonization 305
11.8 Conclusions 308
Acknowledgments 309
References 309

12 Intensified Purification Alternative for Methyl Ethyl Ketone

Production: Economic, Environmental, Safety and Control Issues 311
Eduardo Sánchez-Ramírez, Juan José Quiroz-Ramírez, and
Juan Gabriel Segovia-Hernández
12.1 Introduction 311
12.2 Problem Statement and Case Study 316
12.3 Evaluation Indexes and Optimization Problem 317
12.3.1 Total Annual Cost Calculation 319
12.3.2 Environmental Index Calculation 319
12.3.3 Individual Risk Index 320
12.3.4 Controllability Index Calculation 322
12.3.5 Multi-Objective Optimization Problem 323
12.4 Global Optimization Methodology 324
12.5 Results 325
12.6 Conclusions 335
Acknowledgments 335
Notation 335
References 336

13 Present and Future of Biofuels 341

Juan Gabriel Segovia-Hernández, César Ramírez-Márquez, and
Eduardo Sánchez-Ramírez
13.1 Introduction 341
13.2 Some Representative Biofuels 344
13.2.1 Bioethanol 344
13.2.2 Biodiesel 347
13.2.3 Biobutanol 348
13.2.4 Biojet Fuel 349
13.2.5 Biogas 351
13.3 Perspectives and Future of Biofuels 352
References 354

Index 357
 List of Contributors
Thalles Allan Andrade Department of Chemical Engineering, Biotechnology and
Environmental Technology, University of Southern Denmark, Odense M, Denmark
Luigi di Bitonto Istituto di Ricerca Sulle Acque (IRSA), Consiglio Nazionale delle
Ricerche (CNR), Bari, Italy
Adrián Bonilla-Petriciolet Instituto Tecnológico de Aguascalientes, Aguascalientes,
Mexico
Georgi St. Cholakov Department of Organic Synthesis and Fuels, University of
Chemical Technology and Metallurgy, Sofia, Bulgaria
Knud Villy Christensen Department of Chemical Engineering, Biotechnology and
Environmental Technology, University of Southern Denmark, Odense M, Denmark
Marcos Lucio Corazza Department of Chemical Engineering, Federal University of
Paraná, Polytechnic Center (DEQ/UFPR), Curitiba, Brazil
Carlos J. Durán-Valle Departamento de Química Orgánica e Inorgánica, Universidad
de Extremadura, Badajoz, Spain
Massimiliano Errico Department of Chemical Engineering, Biotechnology and
Environmental Technology, University of Southern Denmark, Odense M, Denmark
Wanderson Rogerio Giacomin-Junior Department of Chemical Engineering, Federal
University of Paraná, Polytechnic Center (DEQ/UFPR), Curitiba, Brazil
Fernando Israel Gómez-Castro Departamento de Ingeniería Química, Universidad de
Guanajuato, Guanajuato, Guanajuato, México
Claudia Gutiérrez-Antonio Facultad de Química, Universidad Autónoma de Querétaro,
Querétaro, Querétaro, México
Borja Hernández Department of Chemical Engineering, University of Salamanca,
Salamanca, Spain
Salvador Hernández Departamento de Ingeniería Química, Universidad de Guanajuato,
Guanajuato, Guanajuato, México
xiv List of Contributors

Luiz Fernando de Lima Luz Junior Department of Chemical Engineering, Federal

University of Paraná, Polytechnic Center (DEQ/UFPR), Curitiba, Brazil
Mariano Martín Department of Chemical Engineering, University of Salamanca,
Salamanca, Spain
Ana Laura Moreno-Gómez Facultad de Química, Universidad Autónoma de Querétaro,
Querétaro, Querétaro, México
Maria Papadaki Department of Environmental Engineering, University of Patras,
Agrinio, Greece
Carlo Pastore Istituto di Ricerca Sulle Acque (IRSA), Consiglio Nazionale delle
Ricerche (CNR), Bari, Italy
Dipesh S. Patle Chemical Engineering Department, Motilal Nehru National Institute of
Technology, Allahabad, India
Fernando Augusto Pedersen Voll Department of Chemical Engineering, Federal
University of Paraná, Polytechnic Center (DEQ/UFPR), Curitiba, Brazil
Juan José Quiroz-Ramírez CONACyT – CIATEC A.C. Centro de Innovación Aplicada
en Tecnologías Competitivas, León, México
Bruna Ricetti Margarida Department of Chemical Engineering, Federal University of
Paraná, Polytechnic Center (DEQ/UFPR), Curitiba, Brazil
César Ramírez-Márquez Departamento de Ingeniería Química, Universidad de
Guanajuato, Guanajuato, México
Gade Pandu Rangaiah Department of Chemical and Biomolecular Engineering,
National University of Singapore, Singapore and School of Chemical Engineering, Vellore
Institute of Technology, Vellore, India
Antonio Sánchez Department of Chemical Engineering, University of Salamanca,
Salamanca, Spain
Eduardo Sánchez-Ramírez Departamento de Ingeniería Química, Universidad de
Guanajuato, Guanajuato, México
Juan Gabriel Segovia-Hernández Departamento de Ingeniería Química, Universidad
de Guanajuato, Guanajuato, México
Savyasachi Shrikhande School of Chemical Engineering, VIT, Vellore, India
Roumiana P. Stateva Institute of Chemical Engineering, Bulgarian Academy of
Sciences, Sofia, Bulgaria
Feng-Sheng Wang Department of Chemical Engineering, National Chung Cheng
University, Chiya, Taiwan
Wu-Hsiung Wu Department of Chemical Engineering, National Chung Cheng
University, Chiya, Taiwan
 Series Preface
Renewable resources, their use and modification are involved in a multitude of important
processes with a major influence on our everyday lives. Applications can be found in the
energy sector; paints and coatings; and the chemical, pharmaceutical, and textile industry,
to name but a few.
The area interconnects several scientific disciplines (agriculture, biochemistry, chem-
istry, technology, environmental sciences, forestry), which makes it very difficult to have
an expert view on the complicated interaction. Therefore, the idea to create a series of sci-
entific books, focusing on specific topics concerning renewable resources, has been very
opportune and can help to clarify some of the underlying connections in this area.
In a very fast-changing world, trends are not only characteristic of fashion and politi-
cal standpoints; science too is not free from hypes and buzzwords. The use of renewable
resources is again more important nowadays; however, it is not part of a hype or a fashion.
As the lively discussions among scientists continue about how many years we will still be
able to use fossil fuels – opinions ranging from 50 to 500 years – they do agree that the
reserve is limited, and that it is essential not only to search for new energy carriers but also
for new material sources.
In this respect, the field of renewable resources is a crucial area in the search for alterna-
tives for fossil-based raw materials and energy. In the field of energy supply, biomass- and
renewables-based resources will be part of the solution alongside other alternatives such as
solar energy, wind energy, hydraulic power, hydrogen technology and nuclear energy. In the
field of material sciences, the impact of renewable resources will probably be even bigger.
Integral utilization of crops and the use of waste streams in certain industries will grow in
importance, leading to a more sustainable way of producing materials. Although our soci-
ety was much more (almost exclusively) based on renewable resources centuries ago, this
disappeared in the Western world in the nineteenth century. Now it is time to focus again
on this field of research. However, it should not mean a “retour à la nature,” but should be
a multidisciplinary effort on a highly technological level to perform research toward new
opportunities, to develop new crops and products from renewable resources. This will be
essential to guarantee an acceptable level of comfort for the growing number of people liv-
ing on our planet. It is “the” challenge for the coming generations of scientists to develop
more sustainable ways to create prosperity and to fight poverty and hunger in the world. A
global approach is certainly favored.
xvi Series Preface

This challenge can only be dealt with if scientists are attracted to this area and are rec-
ognized for their efforts in this interdisciplinary field. It is, therefore, also essential that
consumers recognize the fate of renewable resources in a number of products. Further-
more, scientists do need to communicate and discuss the relevance of their work. The use
and modification of renewable resources may not follow the path of the genetic engineering
concept in view of consumer acceptance in Europe. Related to this aspect, the series will
certainly help to increase the visibility of the importance of renewable resources. Being
convinced of the value of the renewables approach for the industrial world, as well as for
developing countries, I was myself delighted to collaborate on this series of books focusing
on the different aspects of renewable resources. I hope that readers become aware of the
complexity, the interaction and interconnections, and the challenges of this field, and that
they will help to communicate on the importance of renewable resources.
I certainly want to thank the people of Wiley’s Chichester office, especially David
Hughes, Jenny Cossham and Lyn Roberts, in seeing the need for such a series of books on
renewable resources, for initiating and supporting it, and for helping to carry the project to
the end.
Last, but not least, I want to thank my family, especially my wife Hilde and children
Paulien and Pieter-Jan, for their patience, and for giving me the time to work on the series
when other activities seemed to be more inviting.

Christian V. Stevens,
Faculty of Bioscience Engineering Ghent University, Belgium
Series Editor, “Renewable Resources”
June 2005
 Preface
Biofuels (e.g. biodiesel, bioalcohols, and biojet fuel) are alternative energy solutions to the
environmental and safety problems related to the use of petroleum-based fuels. This renew-
able energy can be generated from a wide variety of low-cost feedstocks and transformation
routes that also imply a spectrum of process units based on different technologies. During
the past two decades, significant developments and improvements have been achieved to
increase the commercial production of biofuels worldwide. However, the creation and oper-
ation of sustainable biofuel production chains have imposed new challenges to the field of
Process Systems Engineering (PSE). The analysis, modeling, design, optimization, intensi-
fication, and control of individual units (e.g. reactors and separators) and the entire facilities
to produce biofuels have generated drivers for PSE research and development, which should
be addressed via theoretical, computational, and experimental studies.
The PSE of biofuel production schemes demands advances and novel contributions
to handle the challenges associated with the diversity of physicochemical properties of
available feedstocks, biofuel processing routes, operating conditions, and characteristics
of technologies applied in pretreatment units, reactors, separators, and other process
equipment. The opportunities of PSE in the production of renewable biofuels include
(i) development of a reliable thermodynamic framework for estimating the properties
of pure components and mixtures that are required in the design, control, and intensi-
fication of biomass transformation routes; (ii) intensification and optimization of the
processing routes to handle a wide variety of feedstocks for obtaining biofuels and other
high-value-added by-products; (iii) implementation of realistic and proper models for PSE
analysis; (iv) application of reliable global and multiobjective optimization techniques for
solving design problems and improving the performance of biofuel production schemes;
and (v) utilization of computer-aided methodologies for process controllability, mass
and energy integration, and other tasks associated with PSE. Therefore, theoretical,
computational, and experimental studies in these and other topics are required to develop
a sustainable biofuel production chain.
The present book is the first one specifically devoted to PSE for the production of bio-
fuels. It covers a wide range of topics associated with the process engineering of biofuel
production including the thermodynamic modeling, process design and control, reaction
engineering, separation, and purification of biofuels obtained from different biomass feed-
stocks and transformation routes. In all, this book contains 13 chapters devoted to PSE
xviii Preface

for biofuel production. It provides an overview of the subject and covers the portfolio of
available biomass feedstocks for biofuel production, multiscale analysis of bioresources,
challenges in modeling thermodynamic properties and phase equilibrium calculations, the
production and separation of biofuels, computer-aided design, enzyme-catalyzed biodiesel
production, process analysis of biodiesel production (including kinetic modeling and simu-
lation), and the use of ultrasonification in biodiesel production, as well as thermochemical
processes for biomass transformation and production of alternative biofuels. It is a collec-
tion of contributions from leading researchers in PSE and biofuels. Every chapter in this
book has been reviewed anonymously by at least two experts and then thoroughly revised by
the respective contributors. This review process has been attempted to provide high-quality
and educational value for all chapters.
This book will provide researchers and postgraduate students with an overview of the
recent developments and applications of some state-of-the-art technologies and PSE for bio-
fuel production. We consider that this book is a useful resource for researchers in renewable
energies and practitioners working on the production of biofuels.
We are grateful to all the contributors and reviewers of the chapters for their cooperation
to meet the requirements and schedule to finalize this book. We would like to thank the
book publishing team of John Wiley & Sons, Ltd, for their support and assistance during
the preparation of this book.

Adrián Bonilla-Petriciolet
Instituto Tecnológico de Aguascalientes, México
Gade Pandu Rangaiah
National University of Singapore, Singapore
June 2020
 1
Introduction
Adrián Bonilla-Petriciolet1 and Gade Pandu Rangaiah2,3
1
Instituto Tecnológico de Aguascalientes, Aguascalientes 20256, Mexico
2 Department of Chemical and Biomolecular Engineering, National University of Singapore,117585,
Singapore
3
School of Chemical Engineering, Vellore Institute of Technology, Vellore 632014, India

1.1 Importance of Biofuels and Overview of their Production

The relevance and importance of biofuels are recognized worldwide, mainly due to the
problems caused by fossil fuel depletion and environmental pollution (e.g. climate change)
arising from the generation and consumption of traditional energy sources (Li et al. 2019;
Raud et al. 2019; Quiroz-Perez et al. 2019). Biofuels belong to the category of sustainable
energy that can be obtained from biological (e.g. anaerobic digestion and fermentation),
physicochemical (e.g. transesterification), and thermochemical (e.g. liquefaction, gasifica-
tion, and pyrolysis) processing routes, which can involve the application of conventional
and intensified technologies (Gutierrez-Antonio et al. 2017; Li et al. 2019; Quiroz-Perez
et al. 2019). Several researchers have concluded that biomasses can be regarded as a pri-
mary source for obtaining green and renewable energy because they are distributed and
generated worldwide (Li et al. 2019; Quiroz-Perez et al. 2019; Wei et al. 2019). In fact, it
has been estimated that the biomass-based fuel sources can account for 70% of all renewable
energy production (Raud et al. 2019).
Diverse processes have been studied and implemented to perform the transforma-
tion of biomass-based feedstocks to solid, liquid and/or gaseous products that contain
energy-enriched chemicals (Quiroz-Perez et al. 2019). Lignocellulosic materials, food
crops, urban wastes, animal fats, vegetable oils, starch-rich compounds and non-edible

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
2 Process Systems Engineering for Biofuels Development

Table 1.1 Classification of biofuels based on the biomass feedstock and its transformation route.

Type of biofuels according to their processing routes

Primary Secondary
1st generation 2nd generation 3rd generation 4th generation
Firewood, wood, Bioethanol or Bioethanol, Biodiesel or Biofuels produced
pellets, chips, butanol from biobutanol or bioethanol from using genetically
forest and fermentation of synthesized biofuels microalgae, modified
agricultural starch or sugars made from seaweed or microalgae or
residues, gas. contained in food non-food microorganisms. microorganisms.
crops. lignocellulosic
biomass.

Source: Raud et al. 2019. Reproduced with permission of Elsevier.

biomasses like algae and microorganisms (with and without genetic modifications) can
be utilized as feedstocks to produce renewable fuels (Sawangkeaw and Ngamprasertsith
2013; Loman and Ju 2016; Stephen and Periyasamy 2018). Biofuels include end products
known as biodiesel (a mixture of long-chain alkyl esters), biojet fuel (a mixture of C8–C16
alkanes, iso-alkanes, naphthenic derivatives, and aromatic compounds), biogasoline
(C6–C12 hydrocarbons), and bioalcohols (e.g. bioethanol and biobutanol) (Hassan et al.
2015; Gutierrez-Antonio et al. 2017; Wei et al. 2019). Table 1.1 shows a common and
simple classification of biofuels based on the biomass used as the starting material and its
processing route (Raud et al. 2019).
The production of biofuels comprises several process units that should be analyzed, mod-
eled, designed, optimized, intensified, and controlled. In general, conventional processes
employed in biofuels production rely on unit operations that are performed independently
without mass and/or energy integration, whose process conditions are not optimized and the
tradeoff between process efficiency and cost may not be the best (Quiroz-Perez et al. 2019).
On the other hand, intensified process operations outperform their conventional counter-
parts in terms of energy consumption, profitability, and effectiveness. Process intensifica-
tion generally reduces the equipment number, sizes and/or energy consumption, to increase
the productivity and to enhance other performance metrics via the synergy obtained from
multifunctional phenomena at different spatial and time scales (Stankiewicz and Moulijn
2000; Tian et al. 2018). It allows the integration of two or more operations in multitask-
ing units, the development of alternative configurations and design of process equipment,
besides the application of optimization tools and reliable process synthesis methodologies
to improve the pathways for obtaining biofuels (Nasir et al. 2013; Quiroz-Perez et al. 2019).
Sustainable development of biofuels supply chains from the variety of available feed-
stocks and process routes imply new challenges for Chemical Engineering. In particular,
there are key process design aspects of biofuels production to be improved and intensified
(Nasir et al. 2013; Oh et al. 2018; Raud et al. 2019). They include the collection (harvest-
ing/production or recovery) of the biomass, feedstock pretreatment, biomass transformation
routes, end-products separation and purification, and the corresponding logistic tasks that
are linked to the elements of the supply chain. All these factors impact the economic fea-
sibility of the specific pathway to produce the biofuel. For instance, some authors have
highlighted that the production of 4th generation renewable fuels could imply expensive
 Introduction 3

and energy intensive operations thus limiting its current commercialization (Darda et al.
2019). The application of sustainable technologies in each process stage is paramount to
reach the goal of a green and feasible large-scale production of bioenergy. In terms of pro-
cess modeling, there is also the necessity of improving the thermodynamic framework and
conceptual design approaches employed in the biofuels process engineering.
It is clear that biofuels production creates new applications for process system engineer-
ing (PSE) in terms of biomass valorization, green chemistry, thermodynamics, catalysts,
reaction engineering, separation units, process modeling, optimization, design, and con-
trol. Although several developments have been achieved in this direction, there are still
technical limitations and barriers to be overcome with the objective of minimizing costs
and energy requirements of commercial biofuels production facilities utilizing affordable
feedstocks and consequent protection of the environment via energy efficiency and waste
reduction. This book aims to contribute to the development of sustainable production of
renewable biofuels. Specifically, it covers different topics associated with PSE of biofuels
production. The remainder of this chapter is organized as follows: Section 1.2 provides an
overview of relevant issues of PSE associated with biofuels production. Examples of gaps
and current challenges in the production of biofuels are briefly discussed. Finally, Section
1.3 outlines the scope of all the chapters in this book.

1.2 Significance of Process Systems Engineering for Biofuels Production

PSE is devoted to analyzing the elements associated with the creation and operation of
chemical supply chains (Grossmann and Westerberg 2000). This implies the development of
systematic procedures that can be applied in the discovery, design, manufacture and distri-
bution of chemical products starting from the microsystem level until reaching the industrial
scale applications (Grossmann and Westerberg 2000); see Figure 1.1. Undoubtedly, these

Time scale
Month Enterprise

Week Site

Day Plants
Hour
Process units
Minute Single and
multiphase systems
Second Chemical scale
Particle, thin film
ms Small
Molecule
cluster Intermediate
ns
Molecules Large
ps

1 pm 1 nm 1 μm 1 mm 1m 1 km Length scale

Figure 1.1 Conceptual description of a chemical supply chain considering the time, length and chemical
scales. Source: Grossmann and Westerberg 2000. Reproduced with permission of John Wiley & Sons.
4 Process Systems Engineering for Biofuels Development

PSE elements can be extrapolated to the development of biofuels supply chains, and they
include theoretical, computational and experimental studies.
As stated by Grossmann and Westerberg (2000), research and development in PSE com-
prise the process and product design, process modeling, integration, control and operation,
supporting design methods and numerical tools. The feasible and environmentally friendly
production of biofuels also need advances in these PSE areas (Nasir et al. 2013). The exis-
tence of diverse processing routes for the biomass transformation and the incorporation
of novel technologies with the corresponding discovery of alternative feedstocks are the
main drivers of PSE research in biofuels production. This section provides an overview of
opportunities of PSE areas for the production of renewable fuels. Many of these topics are
analyzed in detail in the other chapters in this book.

1.2.1 Modeling of Physicochemical Properties of Thermodynamic Systems Related

to Biofuels
Thermodynamic modeling of properties of pure components and their mixtures, includ-
ing the prediction of the phase equilibrium behavior, is paramount for the engineering of
biofuels production because it is the basis of process design. Reliable prediction of thermo-
dynamic properties is fundamental to calculate the type and size of different equipment,
profiles of state variables (e.g. concentrations and temperature) of separation units and
energy consumption for separation and purification tasks, and also to identify optimal oper-
ating conditions of reaction systems and other process units. For example, knowledge of
density and viscosity of a given thermodynamic system is relevant for vessel design, piping,
and calculation of mass transfer rates.
As stated, feedstocks for the production of biofuels include biomass by-products (e.g.
forest residues, sugar cane bagasse, cereal straw), urban wastes (e.g. organic compounds
present in industrial and municipal solid and liquid wastes), animal fats, vegetable oils,
insect lipids and dedicated materials as energy crops (Sawangkeaw and Ngamprasertsith
2013; Loman and Ju 2016; Stephen and Periyasamy 2018; Kumar et al. 2020). Conse-
quently, mixtures involved in the process operations to synthetize biofuels are characterized
by the presence of a wide spectrum of organic (e.g. lipids, dyes, aromatic hydrocarbons and
biopolymers) and inorganic compounds (e.g. electrolytes and heavy metals). This complex
composition imposes different challenges in the development and application of suitable
thermodynamic models to predict correctly physicochemical behavior. For instance, the
fatty acid profile of feedstocks can affect the physicochemical properties of biodiesel, and
this profile could change substantially depending on biomass origin (Sawangkeaw and
Ngamprasertsith 2013).
On the other hand, mixtures present in biofuels production usually show non-ideal phase
behavior with complex phase diagrams that could be very sensitive to changes in pres-
sure, temperature and composition. Consequently, thermodynamic calculations required
to predict phase behavior/diagrams of biofuels-based systems usually pose computational
challenges. These calculations involve multivariable and nonlinear problems that are char-
acterized by the potential of multiple solutions due to the complexity of thermodynamic
models. Phase equilibrium calculations must be performed numerous times in the process
simulators for the design, optimization and control of process units. These include Gibbs
free energy minimization to estimate phase compositions, phase stability analysis to verify
 Introduction 5

the reliability of solutions obtained for phase equilibrium problems, prediction of bubble
and dew points, critical conditions, azeotropic points, etc.
Reliable determination of parameters of thermodynamic models employed in phase
equilibrium calculations is an additional issue that should be resolved. These adjustable
parameters can be obtained from the regression analysis of experimental data, whose
(un)availability limits the implementation of some thermodynamic models for the
study of biofuels-related systems. Therefore, the application of predictive models and
computer-aided methodologies is necessary to estimate the required physicochemical
properties. Fortunately, there are scientific databanks of experimental physical and
chemical properties of many compounds (Su et al. 2017). However, they usually contain
limited information for the molecules involved in the mixtures associated with biofuels
systems. This issue also highlights the importance of developing a robust thermodynamic
framework for the process design and modeling of the biofuels supply chain.
Computational chemistry approaches, group contribution methods and equation of states
can be utilized to estimate the properties required in biofuels process design at different
modeling scales (i.e. atomic, group, and molecular) (Su et al. 2017). The conventional
thermodynamic models (e.g. cubic equations of state or local composition models) could
fail to predict the physicochemical behavior of biofuels-based systems (Reynel-Avila et al.
2019). Consequently, reliable predictive methods are required to calculate the physical and
chemical properties of pure components and their mixtures in the processing routes of bio-
fuels. Application of artificial intelligence tools such as artificial neural networks and deep
learning can be an interesting option to improve the available models for predicting the
physicochemical performance of biofuels systems (Reynel-Avila et al. 2019). Reliable and
improved numerical methods for solving nonlinear equations and global optimization prob-
lems should be developed to resolve, robustly and efficiently, the mathematical problems
arising in the phase equilibrium modeling of biofuels. In summary, development of robust
and flexible models with improved capabilities, effective solution methods and software
tools for predicting the thermophysical behavior and properties of biofuels-related sys-
tems (from molecular to macroscopic level) is one of the challenges in PSE for biofuels
production.

1.2.2 Intensification of the Biomass Transformation Routes for the Production

of Biofuels
Process intensification is a relevant area of PSE to enhance the performance of biofuels
production routes (Nasir et al. 2013; Quiroz-Perez et al. 2019). Classical schemes for bio-
fuels production imply the operation of process units that work independently without the
integration of mass and energy, where their performance metrics are usually not optimum.
Strategies to intensify the biofuels processing routes have increased substantially allowing
significant reductions in the production cost and environmental impact. Overall, process
intensification principles have been applied in different stages of the pathways for the
transformation of biomasses to biofuels (Nasir et al. 2013; Quiroz-Perez et al. 2019; Wong
et al. 2019).
The diversity of transformation routes for biofuels production has promoted advances
in catalytic and non-catalytic processes, biotechnology, separation and reaction technolo-
gies. For instance, catalyst-based transformation routes are very common to obtain biofuels
(Wong et al. 2019). Transesterification-based processes can be used to convert edible and
6 Process Systems Engineering for Biofuels Development

non-edible fats and oils into biodiesel, where homogeneous and heterogeneous (acid, base,
or enzymatic) catalysts are employed (Rezania et al. 2019). This processing route may
require a pretreatment stage (e.g. esterification reaction) if the feedstock contains high fatty
acids (Nasir et al. 2013). The need to reduce costs in these processes has led to the synthe-
sis and application of novel catalysts (Trombettoni et al. 2018), the study of novel reaction
media such as supercritical fluids (Deshpande et al. 2010) and the proposal of alternative
reactor technologies (Tabatabaei et al. 2019; Wong et al. 2019).
On the other hand, some authors have concluded that microbial fermentation for obtain-
ing bioalcohols is a simple and promising approach to produce bioenergy (Bhatia et al.
2017). In particular, alcohols with two or more carbon atoms (e.g. ethanol and butanol)
have been considered as interesting alternatives to conventional petroleum-based fuels.
However, fermentation processes utilized in the production of these alcohols have several
disadvantages that limit their large-scale industrial applications. The process intensification
of this route should address the inhibition of competitive pathways that affect the alcohol
productivity due to by-products formed, the genomic adaptation of strains to enhance the
substrate utilization capability to use low cost feedstocks (e.g. lignocellulosic wastes), the
genetic diversification of microbes with improved alcohol producing capabilities to inten-
sify specific metabolic performance for obtaining the desired end-products and to design
synthetic biofuels pathways (Shanmugam et al. 2020). Indeed, advances and developments
in metabolic engineering have contributed to the process intensification of biofuels pro-
duction via the optimization of bioprocess yields and productivities (Shanmugam et al.
2020). Microbial genome engineering can be utilized to maximize the efficiency of fermen-
tation processes via the improvement of the genomic characteristics of biofuels producing
microorganisms to direct the metabolic flux toward the generation of desirable bioproducts
(Shanmugam et al. 2020). Several authors have analyzed and discussed these and other
advances in metabolic engineering and synthetic biology for biofuels production (e.g. Bilal
et al. 2018; Majidian et al. 2018).
Separation units also represent an important area for process intensification in the pro-
duction of biofuels. Separation technologies are utilized in the pretreatment and preprocess-
ing stages of biomass transformation due to the heterogeneous composition of feedstocks
and in the purification of process streams to recover biofuels and their by-products. Both
non-intensive and intensive energy separation methods have been applied in biofuels pro-
duction. Distillation, extraction, adsorption and membrane-based methods are part of the
spectrum of technologies for obtaining renewable fuels (Atadashi et al. 2011; Levario et al.
2012; Abdehagh et al. 2014; Li et al. 2019). The application of intensified non-reactive sep-
arations such as heat-integrated and membrane-based distillation, has been explored in the
production of biofuels (Diaz and Tost 2017; Kumar et al. 2019). Also, intensified schemes
that combine reaction and separation units (e.g. reactive distillation and extraction) (Plesu
et al. 2015; Poddar et al. 2017; Gor et al. 2020), and purification systems assisted with
microwave, ultrasound and supercritical fluids (Patil et al. 2018; Li et al. 2019; Mahmood
et al. 2019) have been reported to produce biofuels.
Separation and purification methods applied in biofuels production show different limita-
tions and advantages in terms of energy consumption and product(s) recovery. For example,
extraction techniques are relevant for biofuels processing that usually require low energy
consumption (Li et al. 2019). Extraction is a key step to carry out the recovery of the
desired bioproducts and to reduce the content of undesired substances in the intermediate
 Introduction 7

stream to be processed. Fatty acids, hydrocarbons, lipids and biosolids can be extracted from
extractable feedstocks for biofuels production such as animal fats, energy crops, agricul-
tural residues and microalgae. The selection of the extraction technique is constrained by the
characteristics of the feedstock to be processed and the specific components to be recovered
or concentrated, which impact the separation efficacy and selectivity. Mechanical, physical
and chemical extraction methods have been utilized in the production of different genera-
tion biofuels (Li et al. 2019). Extraction techniques can be intensified via the application
of microwave, ultrasound and supercritical fluids. Also, novel extractive agents such as
ionic liquids and green solvents have been explored to intensify the recovery of the tar-
get compound(s). Li et al. (2019) have analyzed in detail the advantages and limitations
of extraction techniques utilized in biofuels production. These extraction processes may
generate residues that could cause health hazards and environmental pollution, which is an
issue to be resolved as part of PSE challenges.
With respect to energy intensive separation methods, distillation is the primary method in
chemical process industries but its application in the recovery of biofuels depends signifi-
cantly on the characteristics of the streams to be purified. However, conventional distillation
is not an effective approach for the purification of bioalcohols from fermentation broths due
to the occurrence of homogeneous azeotropes (Abdehagh et al. 2014). Therefore, hybrid
and intensified distillation schemes have been applied to recover these and other biofuels.
For example, Nagy et al. (2015) reported that the combination of distillation and pervapo-
ration can decrease the energy demand for downstream separation of fermentation broths.
Several studies have also reported the application of reactive distillation for the production
of biofuels. Reactive distillation allows simultaneous transesterification and separation of
products within the same equipment (Poddar et al. 2017). Several improvements to this
reactive separation scheme to produce different renewable fuels have also been reported
(Gutierrez-Antonio et al. 2018; Gao et al. 2019). See Singh and Rangaiah (2017) for a
review of advances in separation processes for bioethanol recovery and dehyration.
Overall, it is required to develop improved process units that should be flexible and
robust for the transformation of feedstocks with changing physicochemical characteristics
to biofuels. Advanced and less energy-intensity separation techniques are needed to
increase the sustainability of biofuels production. The development of green technologies
for the purification and recovery of biofuels and by-products is considered a relevant
PSE issue. The application of intensification technologies based on supercritical fluids,
microwave, ultrasound, and ionic liquids opens new opportunities for the development
of improved processes for biofuels production. Research on these technologies should
be increased to establish their benefits and limitations for industrial applications. Efforts
should also be focused on the recovery and use of value-added compounds generated
during biomass transformation such as glycerol. These and other shortcomings should be
addressed with the aim of developing cost-effective separation and purification schemes
for the production of biofuels.

1.2.3 Computer-Aided Methodologies for Process Modeling, Design, Optimization,

and Control Including Supply Chain and Life Cycle Analyses
Process design of biofuels production facilities should consider performance metrics
and objectives related to environment, economics, and safety. In particular, current and
8 Process Systems Engineering for Biofuels Development

anticipated regulations for environmental protection impose additional restrictions to

this design stage. Biofuels process design requires the application of proper models
that accurately represent the characteristics and properties of the systems, units and all
elements involved in the supply chain ranging from the microscopic to macroscopic level
(Figure 1.1). Therefore, development of realistic models to be used in process design is
an important PSE challenge for biofuels production. Note that high level physicochemical
description in the design problem formulation is challenging. For instance, Quiroz-Perez
et al. (2019) have highlighted the importance of computational fluid dynamics for process
design and modeling of equipment involved in biofuels production where transport
phenomena are paramount to ensure correct scale up and industrial operation. Also, kinetic
and thermodynamic data of the reacting systems involved in biomass transformation routes
are fundamental for reliable design of reactors including fermenters.
Processing routes for biofuels production can be optimized via the formulation of design
problems with one or more objectives to be minimized or maximized simultaneously.
Indeed, multi-objective optimization (MOO) has found numerous applications in chemical
engineering and related areas (Rangaiah and Bonilla-Petriciolet 2013; Rangaiah et al.
2015; Madoumier et al. 2019). Optimization can be employed to improve the performance
of specific process units and the entire processing route for producing biofuels. Biofuels
process optimization is not an easy task and robust numerical methods are required to solve
the design problems, which are usually multivariate, nonlinear and with equality/inequality
constraints. Deterministic and stochastic optimizers have been applied to solve design
problems in the biofuels production. In particular, stochastic optimizers (metaheuristics)
have shown several advantages for solving both global optimization and MOO problems of
biofuels production due to their easy implementation, computational efficiency and ability
to handle both discrete and continuous design variables. Optimization has been used for
the design of intensified separation sequences for biofuels purification (Sanchez-Ramirez
et al. 2019; Gor et al. 2020), for the improvement of processing routes to obtain biofuels
(Woinaroschy 2014), for the integrated design of biorefineries to produce biodiesel from
different feedstocks (Prieto et al. 2017), to identify processing paths for obtaining biofuels
from different feedstocks (Eason and Cremaschi 2014), and for biodiesel plant design
(Patle et al. 2014a; Alvaraes et al. 2019).
Biofuels production facilities comprise a large set of operating variables that should be
manipulated and regulated. Therefore, process controllability is an important issue for the
implementation, operation and safety of biofuel production. The control problem of a com-
plete biofuels production process is large with nonlinear functions of states, many inputs
and outputs, and a reduced number of degrees of freedom (Bildea and Kiss 2011; Prunescu
et al. 2017). Consequently, nonlinear control concepts and plantwide control are required to
achieve flexible and stable operation of process units in biofuels production. For example,
plantwide control has been studied for a complete biodiesel plant (Patle et al. 2014b).
Life cycle analysis (LCA) is desirable for comprehensive assessment of the sustainability
of biofuels processes in terms of environmental, social, energetic and economic indicators
(Collotta et al. 2019). Several authors have reported LCA of the production of renewable
fuels using different levels of details, methodologies, analytical boundaries, and impact
metrics (e.g. Mu et al. 2017; Liu et al. 2018). However, it has been pointed out that the stan-
dardization of methodologies utilized for LCA of biofuels, including life cycle inventory
data, is a relevant issue to be addressed for performing reliable comparison and supporting
 Introduction 9

the decision-making process to identify the best options for biofuels production (Mayer
et al. 2020).
Finally, biofuels supply chains include all the activities related to the transformation of
biomasses into renewable fuels and their delivery to the end-users (An et al. 2011; Awudu
and Zhang 2012). The biofuels supply chain is affected by several uncertainties in terms
of prices, demand and supply of feedstocks and end-products, transportation and storage
issues, performance of processing facilities, among other factors (Awudu and Zhang 2012).
Consequently, the design of a reliable and sustainable biofuels supply chain requires the
application of the latest computer-aided methodologies to optimize the operational, tactical
and strategic decisions.
In summary, PSE contributions and developments are fundamental to consolidate, opti-
mize and operate the biofuels supply chains to achieve the economic, environmental and
social benefits of this type of renewable energy.

1.3 Overview of this Book

After this chapter, this book contains 12 chapters that describe and analyze different applica-
tions of PSE for biofuels production. Chapters 2–13 are briefly summarized in this section.
Chapter 2 provides an overview of different biomasses that can be utilized for biofuels
production. It highlights the relevance of feedstock composition for biofuels production.
Biomasses analyzed in this chapter include spent coffee grounds, different lignocellulosic
materials, residues of oil production from palm, olive, coconut, avocado and argan, residues
from crops such as citrus and grapes, and waste oil and waste cooking oil. This chapter ends
stating the importance of developing new methods and technologies to exploit the variety
of available feedstocks for producing biofuels.
In Chapter 3, analysis and discussion of PSE contributions for the process design of
biorefineries and biomass-based infrastructure are presented. This chapter describes meth-
ods for the design of process units and approaches for process synthesis. The product design
problem for biomass processing, supply chain modeling and the importance of multiscale
analysis are also discussed.
The challenges of thermodynamic properties and phase equilibrium calculations in bio-
fuels process design are covered in Chapter 4. Elements of the thermodynamic model-
ing framework for the prediction of properties required for process design of biofuels are
described. The formulation of phase equilibrium problems and a survey of available thermo-
dynamic models for phase equilibrium calculations are presented. A brief analysis of prop-
erty databanks for biofuels process design and the impact of uncertainties of thermophysical
properties are also provided in this chapter. Finally, some methods for the prediction of
thermodynamic properties of compounds involved in biofuels production are described.
Chapters 5 reports pretreatment methods and processing routes to transform waste oil
into biodiesel. Capabilities and limitations of homogeneous and heterogeneous catalysis,
enzymatic-, ionic liquid- and hydrated salts-based conversions for processing waste oils
are discussed. The authors of this chapter have highlighted the technical limitations and
challenges to intensify biodiesel production from waste oil.
Biojet fuel production from wastes and residues is reviewed in Chapter 6. This chapter
discusses the importance of biofuels development for the aviation sector. It contains the
10 Process Systems Engineering for Biofuels Development

state of the art in the processing of triglyceride-containing wastes, lignocellulosic materials,

sugar and starchy residues for the production of biojet fuel. The authors of this chapter have
analyzed the challenges and future trends to potentiate this renewable fuel for the aviation
industry.
Computer-aided design is important to develop new processes for the production of
biofuels. Therefore, Chapter 7 focuses on the development of a simulation platform for
biological models associated with biofuels systems. The modeling approach involves an
optimization problem subject to specific constraints. Two examples are reported to show
the application of this modeling approach. Results show that this approach can be used to
assist the industrial production of biofuels via mutated strains.
Different aspects of PSE of biodiesel production via enzyme-catalyzed routes are ana-
lyzed in Chapter 8. A comparison of the biodiesel production routes with chemical and
enzymatic catalysts is performed. The authors have discussed the optimal reaction condi-
tions and the kinetic modeling in biodiesel production routes catalyzed by liquid enzymes.
Details of process simulation and economic evaluation of this type of transformation route,
including the reuse of the enzymes, are also included in this chapter.
Chapter 9 deals with simulation and design of process scenarios for biodiesel production.
In particular, this chapter describes the application of the Aspen Plus® simulator to model
the biodiesel process. Examples are described for the calculation of thermodynamic prop-
erties of both pure components and mixtures, required for process design. Also, the authors
have discussed some aspects of utilization of Aspen Plus to model reactions involved in
biodiesel production. Case studies related to the process design of esterification and trans-
esterification reactions with different reactor models are described. Finally, use of Aspen
Plus for energy and economic analyses is illustrated. This chapter provides a simple and
handy guide for students and practitioners in the use of Aspen Plus for biofuels process
design.
Chapter 10 also describes the modeling and simulation of a continuous biodiesel pro-
cess from microalgae using Aspen Plus. For this, process parameters and reaction kinetic
data were based on reported experimental results. A sensitivity analysis was performed to
analyze the impact of some design variables of process units. The sizing and cost analysis
of the equipment utilized in the biodiesel process simulation were carried out. The authors
compared the performance of this biodiesel process from microalgae with the results for a
biodiesel process using waste cooking oil. Based on this, some research topics to reduce
the cost of biodiesel production from wet microalgae are suggested.
A state of the art of thermochemical methods for the production of renewable fuels
is given in Chapter 11. Thermochemical methods that are utilized to obtain solid, liquid
and gaseous biofuels are described in this chapter. First, a simple classification of ther-
mochemical methods is provided. Combustion, gasification, liquefaction, pyrolysis, and
carbonization are analyzed. Advantages, limitations, energy requirements and equipment
used in these thermochemical methods are also covered in this chapter.
A perspective of the present and future of biofuels is presented in Chapter 12. The
importance, implications and advantages of utilizing biomass to produce renewable energy
are analyzed. Characteristics of some biofuels feedstocks and their processing routes are
provided. This chapter includes a detailed discussion of bioethanol, biodiesel, biobutanol,
biojet fuel, and biogas. It concludes that biofuels production from some specific feedstocks
will be commercially attractive in the next decade.
 Introduction 11

Finally, Chapter 13 deals with the design of intensified purification options to produce
methyl ethyl ketone, which has been suggested as a biofuel that can be produced by the fer-
mentation route. Purification of mixtures that contain this biofuel is challenging due to the
presence of azeotropes. Hence, this chapter analyzes some intensified schemes to improve
methyl ethyl ketone purification. Separation schemes based on distillation and liquid–liquid
extraction were designed via the MOO approach considering economic, environmental,
controllability and safety indexes. Results show that the intensified process requires lower
energy compared with the separation scheme based on distillation alone.
In short, the contents of this book expand and cover developments and contributions of
PSE for biofuels production. Students of Chemical Engineering, Environmental Engineer-
ing, Energy Engineering and related areas will find the chapters useful in their studies on
biofuels. The editors and authors of this book hope that its contents will contribute to further
research and development of PSE for biofuels production in both academia and industrial
practice. As stated, biofuels are at the forefront of new energy solutions to the environmental
and safety problems related to the use of petroleum-based fuels. Consequently, consoli-
dation of biofuels production and supply chains are important to support the sustainable
human development of future generations.

References
Abdehagh, N., Tezel, F.H., and Thibault, J. (2014). Separation techniques in butanol production: challenges
and developments. Biomass and Bioenergy 60: 222–246.
Alvaraes, A.O., Prata, D.M., and Santos, L.S. (2019). Simulation and optimization of a continuous biodiesel
plant using nonlinear programming. Energy 189: 116305.
An, H., Wilhelm, W.E., and Searcy, S.W. (2011). Biofuel and petroleum-based fuel supply chain research:
a literature review. Biomass and Bioenergy 35: 3763–3774.
Atadashi, I.M., Aroua, M.K., and Aziz, A.A. (2011). Biodiesel separation and purification: a review. Renew-
able Energy 36: 437–443.
Awudu, I. and Zhang, J. (2012). Uncertainties and sustainability concepts in biofuel supply chain manage-
ment: a review. Renewable and Sustainable Energy Reviews 16: 1359–1368.
Bhatia, S.K., Kim, S.H., Yoon, J.J., and Yang, Y.H. (2017). Current status and strategies for second genera-
tion biofuel production using microbial systems. Energy Conversion and Management 148: 1142–1156.
Bilal, M., Iqbal, H.M.N., Hu, H. et al. (2018). Metabolic engineering and enzyme-mediated processing:
a biotechnological venture towards biofuel production – a review. Renewable and Sustainable Energy
Reviews 82: 436–447.
Bildea, C.S. and Kiss, A.A. (2011). Dynamics and control of a biodiesel process by reactive absorption.
Chemical Engineering Research and Design 89: 187–196.
Collotta, M., Champagne, P., Tomasoni, G. et al. (2019). Critical indicators of sustainability for biofuels: an
analysis through a life cycle sustainability assessment perspective. Renewable and Sustainable Energy
Reviews 115: 109358.
Darda, S., Papalas, T., and Zabaniotou, A. (2019). Biofuels journey in Europe: currently the way to low
carbon economy sustainability is still a challenge. Journal of Cleaner Production 208: 575–588.
Deshpande, A., Anitescu, G., Rice, P.A., and Tavlarides, L.L. (2010). Supercritical biodiesel produc-
tion and power cogeneration: technical and economic feasibilities. Bioresource Technology 101:
1834–1843.
Diaz, V.H.G. and Tost, G.O. (2017). Energy efficiency of a new distillation process for isopropanol, butanol
and ethanol (IBE) dehydration. Chemical Engineering and Processing: Process Intensification 112:
56–61.
12 Process Systems Engineering for Biofuels Development

Eason, J.P. and Cremaschi, S. (2014). A multi-objective superstructure optimization approach to

biofeedstocks-to-biofuels systems design. Biomass and Bioenergy 63: 64–75.
Gao, X., Tu, S., Li, T., and Li, H. (2019). Feasibility evaluation of reactive distillation process for the
production of fuel ethanol from methyl acetate hydrotreating. Chemical Engineering and Processing:
Process Intensification 139: 34–43.
Gor, N.K., Mali, N.A., and Joshi, S.S. (2020). Intensified reactive distillation configurations for production
of dimethyl ether. Chemical Engineering and Processing: Process Intensification 149: 107824.
Grossmann, I.E. and Westerberg, A.W. (2000). Research challenges in process systems engineering. AIChE
Journal 46: 1700–1703.
Gutierrez-Antonio, C., Gomez-Castro, F.I., de Lira-Flores, J.A., and Hernandez, S. (2017). A review on the
production processes of renewable jet fuel. Renewable and Sustainable Energy Reviews 79: 709–729.
Gutierrez-Antonio, C., Ornelas, M.L.S., Gomez-Castro, F.I., and Hernandez, S. (2018). Intensification of
the hydrotreating to produce renewable aviation fuel through reactive distillation. Chemical Engineering
and Processing: Process Intensification 124: 122–130.
Hassan, S.N., Sani, Y.M., Abdul Aziz, A.R. et al. (2015). Biogasoline: an out-of-the-box solution to the
food-for-fuel and land-use competitions. Energy Conversion and Management 89: 349–367.
Kumar, M., Sun, Y., Rathour, R. et al. (2020). Algae as potential feedstock for the production of biofuels
and value-added products: opportunities and challenges. Science of the Total Environment 716: 137116.
Kumar, R., Ghosh, A.K., and Pal, P. (2019). Fermentative ethanol production from madhuca indica flow-
ers using immobilized yeast cells coupled with solar driven direct contact membrane distillation with
commercial hydrophobic membranes. Energy Conversion and Management 181: 593–607.
Levario, T.J., Dai, M., Yuan, W. et al. (2012). Rapid adsorption of alcohol biofuels by high surface area
mesoporous carbons. Microporous and Mesoporous Materials 148: 107–114.
Li, P., Sakuragi, K., and Makino, H. (2019). Extraction techniques in sustainable biofuel production: a
concise review. Fuel Processing Technology 193: 295–303.
Liu, H., Huang, Y., Yuan, H. et al. (2018). Life cycle assessment of biofuels in China: status and challenges.
Renewable and Sustainable Energy Reviews 97: 301–322.
Loman, A.A. and Ju, L.K. (2016). Soybean carbohydrate as fermentation feedstock for production of bio-
fuels and value-added chemicals. Process Biochemistry 51: 1046–1057.
Madoumier, M., Trystram, G., Sebastian, P., and Collignan, A. (2019). Towards a holistic approach for
multi-objective optimization of food processes: a critical review. Trends in Food Science & Technology
86: 1–15.
Mahmood, H., Moniruzzaman, M., Iqbal, T., and Khan, M.J. (2019). Recent advances in the pretreatment
of lignocellulosic biomass for biofuels and value-added products. Current Opinion in Green and Sus-
tainable Chemistry 20: 18–24.
Majidian, P., Tabatabaei, M., Zeinolabedini, M. et al. (2018). Metabolic engineering of microorganisms for
biofuel production. Renewable and Sustainable Energy Reviews 82: 3863–3885.
Mayer, F.D., Brondani, M., Carillo, M.C.V. et al. (2020). Revisiting energy efficiency, renewability and
sustainability indicators in biofuels life cycle: analysis and standardization proposal. Journal of Cleaner
Production 252: 119850.
Mu, D., Ruan, R., Addy, M. et al. (2017). Life cycle assessment and nutrient analysis of various processing
pathways in algal biofuel production. Bioresource Technology 230: 33–42.
Nagy, E., Mizsey, P., Hancsok, J. et al. (2015). Analysis of energy saving by combination of distillation and
pervaporation for biofuel production. Chemical Engineering and Processing: Process Intensification 98:
86–94.
Nasir, N.F., Daud, W.R.W., Kamarudin, S.K., and Yaakob, Z. (2013). Process system engineering in
biodiesel production: a review. Renewable and Sustainable Energy Reviews 22: 631–639.
Oh, Y.K., Hwang, K.R., Kim, C. et al. (2018). Recent developments and key barriers to advanced biofuels:
a short review. Bioresource Technology 257: 320–333.
 Introduction 13

Patil, P.D., Dandamudi, K.P.R., Wang, J. et al. (2018). Extraction of bio-oils from algae with supercritical
carbon dioxide and co-solvents. Journal of Supercritical Fluids 135: 60–68.
Patle, D.S., Sharma, S., Ahmad, Z., and Rangaiah, G.P. (2014a). Multi-objective optimization of two alkali
catalyzed processes for biodiesel from waste cooking oil. Energy Conversion and Management 85:
361–372.
Patle, D.S., Ahmad, Z., and Rangaiah, G.P. (2014b). Plantwide control of biodiesel production from waste
cooking oil using integrated framework of simulation and heuristics. Industrial and Engineering Chem-
istry Research 53: 14408–14418.
Plesu, V., Puigcasas, J.S., Surroca, G.B. et al. (2015). Process intensification in biodiesel production with
energy reduction by pinch analysis. Energy 79: 273–287.
Poddar, T., Jagannath, A., and Almansoori, A. (2017). Use of reactive distillation in biodiesel production:
a simulation-based comparison of energy requirements and profitability indicators. Applied Energy 185:
985–997.
Prieto, C.V.G., Ramos, F.D., Estrada, V. et al. (2017). Optimization of an integrated algae-based biorefinery
for the production of biodiesel, astaxanthin and PHB. Energy 139: 1159–1172.
Prunescu, R.M., Blanke, M., Jakobsen, J.G., and Sin, G. (2017). Model-based plantwide optimization of
large scale lignocellulosis bioetanol plants. Biochemical Engineering Journal 124: 13–25.
Quiroz-Perez, E., Gutierrez-Antonio, C., and Vázquez-Roman, R. (2019). Modelling of production pro-
cesses for liquid biofuels through CFD: a review of conventional and intensified technologies. Chemical
Engineering and Processing: Process Intensification 143: 107629.
Rangaiah, G.P. and Bonilla-Petriciolet, A. (eds.) (2013). Multi-Objective Optimization in Chemical Engi-
neering: Developments and Applications. Wiley.
Rangaiah, G.P., Sharma, S., and Sreepathi, B.K. (2015). Multi-objective optimization for the design and
operation of energy efficient chemical processes and power generation. Current Opinion in Chemical
Engineering 10: 49–62.
Raud, M., Kikas, T., Sippula, O., and Shurpali, N.J. (2019). Potentials and challenges in lignocellulosic
biofuel production technology. Renewable and Sustainable Energy Reviews 111: 44–56.
Reynel-Avila, H.E., Bonilla-Petriciolet, A., and Tapia-Picazo, J.C. (2019). An artificial neural
network-based NRTL model for simulating liquid-liquid equilibria of systems present in biofuels
production. Fluid Phase Equilibria 483: 153–164.
Rezania, S., Oryani, B., Park, J. et al. (2019). Review on transesterification of non-edible sources for
biodiesel production with a focus on economic aspects, fuel properties and by-product applications.
Energy Conversion and Management 201: 112155.
Sanchez-Ramirez, E., Quiroz-Ramirez, J.J., Hernandez, S. et al. (2019). Synthesis, design and optimization
of alternatives to purify 2,2-butanediol considering economic environmental and safety issues. Sustain-
able Production Consumption 17: 282–295.
Sawangkeaw, R. and Ngamprasertsith, S. (2013). A review of lipid-based biomasses as feedstocks for bio-
fuels production. Renewable and Sustainable Energy Reviews 25: 97–108.
Shanmugam, S., Ngo, H.H., and Wu, Y.R. (2020). Advanced CRISPR/Cas-based genome editing tools for
microbial biofuels production: a review. Renewable Energy 149: 1107–1119.
Singh, A. and Rangaiah, G.P. (2017). Review of Technological Advances in Bioethanol Recovery and
Dehyration. Industrial and Engineering Chemistry Research 56: 5147–5163.
Stankiewicz, A.I. and Moulijn, J.A. (2000). Process intensification: transforming chemical engineering.
Chemical Engineering Progress 96: 22–33.
Stephen, J.L. and Periyasamy, B. (2018). Innovative developments in biofuels production from organic
waste materials: a review. Fuel 214: 623–633.
Su, W., Zhao, L., and Deng, S. (2017). Group contribution methods in thermodynamic cycles: physical
properties of pure working fluids. Renewable and Sustainable Energy Reviews 79: 984–1001.
Tabatabaei, M., Aghbashlo, M., Dehhaghi, M. et al. (2019). Reactor technologies for biodiesel production
and processing: a review. Progress in Energy and Combustion Science 74: 239–303.
14 Process Systems Engineering for Biofuels Development

Tian, Y., Demirel, S.E., Hasan, M.M.F., and Pistikopoulos, E.N. (2018). An overview of process systems
engineering approaches for process intensification: state of the art. Chemical Engineering and Process-
ing: Process Intensification 133: 160–210.
Trombettoni, V., Lanari, D., Prinsen, P. et al. (2018). Recent advances in sulfonated resin catalysts for
efficient biodiesel and bio-derived additives production. Progress in Energy and Combustion Science 65:
136–162.
Wei, H., Liu, W., Chen, X. et al. (2019). Renewable bio-jet fuel production for aviation: a review. Fuel 254:
115599.
Woinaroschy, A. (2014). Multiobjective optimal design for biodiesel sustainable production. Fuel 135:
393–405.
Wong, K.Y., Ng, J.H., Chong, C.T. et al. (2019). Biodiesel process intensification through catalytic enhance-
ment and emerging reactor design: a critical review. Renewable and Sustainable Energy Reviews 116:
109399.
 2
Waste Biomass Suitable as
Feedstock for Biofuels Production
Maria Papadaki
Department of Environmental Engineering, University of Patras, Agrinio, 30100, Greece

2.1 Introduction
2.1.1 The Need for Biofuels
Babu (2008) defines biomass as a term used to describe all Earth’s living matter. It is a
general term for material derived from growing plants or from animal manure (which is
effectively a processed form of plant material), while, according to Jessup (2009), biofuels
are solid, liquid, or gaseous energy sources derived from renewable biomass sources. On
the other hand, in article 2 of the European Community Directive “On the promotion of the
use of biofuels” specific definitions of a narrower spectrum are given. As such, the term
“biomass” is used to express the biodegradable fraction of products, waste and residues
from agriculture (including vegetal and animal substances), forestry and related indus-
tries, as well as the biodegradable fraction of industrial and municipal waste. The term
“biofuels” is exclusively referred to liquid or gaseous fuel for transport, i.e. the directive
focuses in fuels which can partially replace fossil-origin fuels employed in transport. It
describes specific characteristics of “bioethanol,” “biodiesel,” “biogas,” “biomethanol,”
“biodimethylether,” “bio-ethyl tert-butyl ether,” “bio- methyl tert-butyl ether,” “synthetic
biofuels,” “biohydrogen,” “pure vegetable oil” which are the referred biofuels, which natu-
rally originate from processed biomass. In this work, the broader definitions are preferred,
although a great deal of the chapter focuses on biomass which can provide liquid and gas
biofuels.

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
16 Process Systems Engineering for Biofuels Development

Biomass has always been used for the production of energy. Chemicals and pharma-
ceutical products have been obtained from biomass; biomass has been burnt to produce
energy since fire was harnessed. Since the energy crisis of 1973, considerable interest has
developed in biomass use toward meeting the energy needs of the world. Furthermore, the
interest in valorization of biomass was awakened as the awareness of the finite nature of
fossil liquid and gaseous hydrocarbons and their ultimate depletion rose. There was much
argument over when this would occur (Goldstein 2018). However, in the mind of society
the search for alternative energy sources had begun.
A drop in oil prices in the 1990s kept bioenergy markets apart from the production. But
the reliance on fuel supplies from a few major producers and the need to reduce greenhouse
emissions imposed the search for renewable sources of energy; valorization of biomass was
a good option. Special attention was given to the replacement of transport fuels as their
numbers were steadily rising (Karp and Halford 2010). The European Union in particular
in 2008 decided that by 2020, one fifth of the totally consumed energy should be obtained
from a renewable energy reservoir. Great weight was also given to the energy consumed
for transportation. As such, biofuels were expected and resolved to comprise at least 10%
of the energy used in that sector by the end of the second decade of the twenty-first century
(Rechberger and Lötjönen 2009). Renewable forests and field crops appeared to be attrac-
tive and suitable sources of biomass in processes aiming at generating biofuels. Moreover,
it was expected that existing technology could function efficiently and that it was a feasible
solution toward meeting set targets of gradual fossil fuel replacement. In fact, by the end
of the first decade of the twenty-first century a high production of biofuels was reached
in both North and South America. Brazil, Argentina, and the United States produced and
consumed large amounts of bioethanol. A very high percentage of liquid biofuels mainly
in the form of bioethanol was also employed in Australia, China, and Canada. Europe was
predominantly producing and utilizing biodiesel (Azadia et al. 2017).
Appropriate for biofuel production biomass includes in principle, food-crops containing
sugars and starch or crops which have high oil content; those can be converted to appropri-
ate fuels via a sequence of transformation and reaction paths. Biomass of lignocellulosic
content is also a very good candidate. It can be either obtained as the side or waste product
of other processes, such as agricultural and forestry residues. It can also be obtained via pur-
posefully cultivated plants in the form of the so-called energy crops. Additional sources for
biofuel production are: used and waste oils, other organic wastes, algae (micro- or macroal-
gae, encountered or grown in open or pond waters or in photobioreactors) (Stafford et al.
2017). Every year, the plants on Earth produce and store up to four times the energy needed
by humanity per annum (Guo et al. 2015). As such, shrewd, systems-approach planning
for the valorization of the available biomass can substantially reduce our dependency on
fossil fuels. However, to reach such a goal, a number of parameters have to be accounted
for prior to investment in new processes. The aim of producing replacement fuels has to
be harmonically enlaced with objectives such as CO2 reduction, independence from a few
energy providers in the world, that fossil fuels replacement does not impact on food produc-
tion and availability, that deforestation is not a consequence, etc. In other words, the urge
for biofuel production will not be shifting the problems associated with energy elsewhere,
thus forming alternative types and kind of dependencies. Moreover, new, fit-for-purpose
technologies have to be developed and the industrial size production of large quantities
of desired commodities may have to be questioned and probably partially shifted or be
 Waste Biomass Suitable as Feedstock for Biofuels Production 17

replaced by smaller, specialized-goods production. Alternative methods of production that

are mobile, transformable, or of extendable or shrinkable capacities may need to be invented
and evolve. Alternative products of a functionality similar to the one offered by existing
commodities, or entities with completely new properties and potential uses may also arise
as side or by-products of the alternative energy production processes.
The selection of good or adequately suitable feedstocks for biofuels production is a ques-
tion with multiple answers. Many of the answers which have been given so far have created
more important questions. Much progress is being made however, thanks to the substan-
tial research already conducted and which is continuously expanding. In the search of such
feedstocks, one has to wind through a multi-stage and multi-scale labyrinthine pathway
in order to first identify the primary qualities required in the feedstock, to subsequently
devise the processes which have the potential to provide expected outputs and to finally
implement a multi-objective optimization process in order to guarantee the viability of the
venture. There are multiple issues which have to be addressed and the primary ones lie
in the definition of the problem which is to be solved. The top-to-bottom approach is tra-
ditionally implemented by the chemical engineers during the conceptual process design.
The problem is relatively static, with well defined raw materials and desired products; the
engineer searches for the best route to take from the starting to the end point. Biomass
valorization and the production of biofuels from biomass form an unprecedented opportu-
nity for the implementation and development of the-top-to-bottom, bottom-to-top coupled
method, described in detail by Pham and El-Halwagi (2012). The method can be also
coupled with the atomic targeting and design (El-Halwagi 2017). Furthermore, in their
book Sengupta and Pike (2012) present how different chemicals including biofuels can
be obtained from waste biomass. They focus on processes which can contribute toward a
sustainable economy. They present the advances of chemistry and technologies which can
be used toward biomass valorization and how non-renewable feedstocks can be replaced
by renewable ones.
Furthermore, given the dynamics and the volatility of the market in the globalized world
an embedded flexibility in the kind of the products, the methods of production, and the
capacity may counterbalance any costs which may affect the economics and the viability of
the process. Thus, space should be sought for versatile designs to replace traditional rigid
requirements where stability of the selected feedstock and product prices are the key factors
governing the process economics.

2.1.2 Problem Definition

Until the end of the first decade of the twenty-first century the raw materials for biofuel pro-
duction were carefully selected and they were obtained through cultivations dedicated to
that very purpose. Associated with this objective, agricultural cultivation involved starchy,
sugary and oily plants which would otherwise have been used for the production of food.
Moreover, the land employed for their growth would have normally been used for the pro-
duction of vegetables for feeding people. This had its impact on agricultural products, rais-
ing their prices, with tropical deforestation due to the expansion of food and non-food crop
cultivation in those areas. Thus, the apparent solution of the energy problem was causing a
number of new problems. As such, the net result was not moving the situation forward; the
18 Process Systems Engineering for Biofuels Development

energy availability problem was just transformed to another entity. However, the technolo-
gies which have been developed, the research which has been conducted, and the wisdom
which has been acquired can be implemented in the valorization of biomass of a differ-
ent origin, such as waste and residual biomass and perennial grasses. Such an application
will provide additional economic and environmental benefits. Naturally, the exploitation of
these alternative sources of biomass will not be as straightforward as the collection of the
waste biomass; the variability of its composition, its potential degradation before treatment,
and discontinuity in availability pose different technological challenges. However, sustain-
ability is a parameter which always has to be taken into consideration. Cyclic economy is
also gaining ground in the engineering thinking and objectives for future development. As
such, the value of bioenergy, instead of being measured merely in terms of the quantity of
the replaced fossil fuels, should be holistically assessed and the impact that it has on the
food production, on the forests and the potential deforestation, on the spent water resources
and their pollution, on its effects on the wildlife, the soils, and society should be “measured”
and accounted for (Union of Concerned Scientists 2012).
Furthermore, the initial approach of employing corn, and other crops rich in sugars and
starch, traditionally used to cover needs of human and animal food, is proven incapable
of reaching the required needs for bioethanol as transport fuel (Sarkar et al. 2012). On the
other hand, there is a great quantity of organic wastes produced and accumulated via human
activities (city, kitchen, agricultural, animal-farm wastes, wastes of lignocellulosic origin
such as forest residues or residual ligocellulosic mass from other processes) which can be
exploited for the production of biofuels via a number of appropriately designed processes
(Stephen and Periyasamy 2018). Amongst those, excellent quality lignocellulosic waste
biomass can be accumulated from the residuals of crops production (wheat, corn, barley,
rye, oats, and others), from a number of different agricultural activities as well as from the
residual biomass of food industry branches which employ processing of freshly collected
fruits and vegetables, such as the juice industry for instance. The option of plant-origin
waste biomass also consists of an environmentally friendlier way of disposal (Rivas-Cantu
et al. 2013).

2.1.3 The Biomass Pool

The waste biomass which can be used for biofuels production ranges from anything which
can be burnt, producing thermal energy as a solid fuel, to more sophisticated specialized
waste, which can provide high efficiency and high quality liquid biofuels. Waste biomass
which can be used for the production of biofuels can be:
1. Waste biomass which is generated as a by-product or residual product following agricul-
tural activities for the production of food. Typical examples are branches of fruit trees
after pruning, fruit tree leaves, husks of legumes and their stems and leaves, wild vege-
tation trimmings, corn residues, like rice, wheat, barley, oat, rye, millet straws, sorghum
stalks, cassava peels and stalks, yam straws, and nut shells.
2. Forest material which involves branches and leaves of forest trees, willow, poplar, bam-
boo, canes, sawdusts, firewood, and woodchips.
 Waste Biomass Suitable as Feedstock for Biofuels Production 19

3. Uncontrolled growth wild type vegetation like perennial grasses and shrubs such as
switchgrass, miscanthus, jatropha, algae, micro- and macroalgae and vascular land
plants.
4. Waste energy cane. Energy cane refers to high biomass sugarcane.
5. Natural cosmetics industries waste biomass like seeds, roots, stems and peels of fruits
and vegetables.
6. Food industry waste: remains of plant or animal origin, fish remains, bones, skins, fer-
mentations waste biomass, frying oils and other fats, corn cobs, coffee and other bever-
ages of plant origin waste biomass, sugarcane and sugar beet waste, kernels and shells
of nuts, etc.
7. Waste from human activities, city and general waste, plastics, paper, waste food, waste
tires, greenhouses plastic sheets, and crop protection plastic nets.
8. Other organic waste like animal and poultry manure, and energy industry waste of oil
or lignocellulosic origin (see Figure 2.1).
Such biomass feedstocks can be used to produce biofuels via a number of treatment
methods. The methods of treatment and their value as biofuel feedstocks depends on their
chemical composition and on their content in cellulosic compounds, sugars and carbo-
hydrates and their content in oils. However, a great variety of those also contain small
quantities of compounds which have valuable properties as natural medicinal, food or cos-
metic agents. Furthermore, their potential use as biofuel feedstock often solves the problem
of their disposal as a waste.
There are different ways that this biomass can be treated so as to provide valuable prod-
ucts. The main methods of biofuel production employ processes such as anaerobic fer-
mentation, pyrolysis and co-pyrolysis, gasification, transesterification, fermentation, acid
or base hydrolysis, solvent extraction supercritical or not, simple mechanical processes of
grinding and pelletizing, simple thermal treatment such as drying, direct biomass combus-
tion, or combinations of the above.

Figure 2.1 Sugar canes. Source: Truncated photo from https://www.pexels.com/search/agricultural

%20waste.
20 Process Systems Engineering for Biofuels Development

The chemical composition of the biomass is a very important factor when seeking the
most appropriate methods of treatment which could potentially be employed. However, the
uniformity of biomass in terms of composition, the variation of its availability during the
year, the potential need of collection, transportation and storage are factors which define
the kind of process which will make its exploitation economically viable and the quantity
and quality of the products which can be obtained. As such, before selecting a specific type
of biomass, the following type of questions have to be answered.
What secondary biomass is already available? Is the selected biomass accessible? In
what quantities? What are their qualities (characteristics)? How is its distribution through
time? What is its spatial distribution? How long can it be stored for before degrading or
suffering an alteration of its composition? In which form should it be stored? Does it need
to be transported? How easy will that be? Can transport fuels be obtained? Can they be
used locally? Is their quality better or worse than that of the fossil fuels (i.e. what is their
nitrogen and/or sulfur content?).
Furthermore, the importance of its valorization as a biofuel has to be assessed in addi-
tional terms which involve safety, environment, and society. For example, does its use
for biofuel production solve any other problems? Does it result in a substantial reduc-
tion of the environmental impact that it would otherwise provoke as waste? Does it con-
tribute toward local or national energy independence? Does it reduce the energy footprint
by reducing transportation of fossil fuels in the area? Does it produce a safer alternative
to employed energy sources? Does it contribute to the economic development of the area
without shifting problems to other places in the world? Does it help the preservation of
biodiversity?
In a preliminary evaluation of the biomass valorization, the geography and the pop-
ulation of the area of the industrial process are of crucial importance for this type of
enterprise and the collection and transportation of residual biomass can constitute an
important expense while the long term storage of biomass is of crucial importance for
its quality. As such, mobile units of treatment or multiple establishments of small units
versus one large facility may be worth considering. Moreover, the potential of extraction
of high added value compounds prior to biorefining may substantially increase the
financial potential of the process. This often implies that the biomass under question has
already been well and reliably characterized. In the following paragraphs a number of
promising feedstocks is presented. Amongst the huge multitude of potential feedstocks
a few selected biomasses are discussed. Their primary common characteristic is their
relative abundance and a substantial amount of completed research on their characteristics.
The reason for the selection of each specific biomass is explained in the relevant sections
(Figure 2.2).

2.2 Kinds of Feedstock

A number of potential feedstocks are presented in the following sections. Each feedstock
is accompanied by a description which intends to answer as many of the above questions
as possible.
 Waste Biomass Suitable as Feedstock for Biofuels Production 21

Figure 2.2 Equisetum. A plant with important medicinal and pesticide properties, abundantly encoun-
tered in wet soils. Potential biofuels precursor following extraction of added value compounds.

2.2.1 Spent Coffee Grounds

According to the European Coffee Federation (ECF 2016), a few billions of coffee cups are
consumed daily all over the world thus making it the most extensively preferred beverage
worldwide: the USA consumed about one and a half million tons of coffee each year for the
period 2011–2013. Mexico consumes two hundred thousand tons, Brazil over one million
tons, while the EU consumes about two and a half million tons of coffee; however, it imports
around three million tons, part of which is re-exported. The amount of coffee produced in
2015 was approximately nine million tons; from each kilogram of coffee, 0.91 kg of solid
waste is produced, thus the importance of a further valorization of this residue becomes
obvious. As Murthy and Naidu (2012) confirm, coffee is a very popular drink and as such
its trade is extensive.
In a comprehensive review, Campos-Vega et al. (2015) report a detailed catalog of com-
pounds which can be exploited by spent coffee grounds, as the cited research articles in
the review suggest. In the mentioned review, it is proposed to use spent coffee grounds in
order to extract several added value products. More specifically they state that the resid-
ual mass after the initial coffee extraction contains numerous organics like different kinds
of polysaccharides and fatty acids, which can be extracted following further treatment. As
presented in their review, research is being conducted toward biofuel generation from the
exploitation of coffee residues. Moreover, its capacity as a source of sugars, as absorbent
of pollutants and in particular heavy metals, and as a primary source for the production of
activated carbons are also being investigated.
At least 10–15% lipids, measured on a dry basis, have been found to remain in the used
coffee grounds. Different coffee residual biomass samples have been analyzed and approx-
imately 90% lipids were found to remain in the solid residue after coffee extraction. Spent
coffee biomass oil consists mainly of triglycerides and small amounts of diglycerides, free
fatty acids, terpenes, sterols, and tocopherols. Therefore, it represents an important source
of raw material for a variety of products including biofuels. Cholakov et al. (2013) applied
22 Process Systems Engineering for Biofuels Development

extraction methods on spent coffee grounds to evaluate their potential as a source of bio-
fuels and activated carbon absorbents. Their results were promising in both aspects but not
conclusive, so they recommended the need for further research. Coelho et al. (2018) exam-
ined the influence of three co-solvents, namely ethanol, isopropanol and ethyl lactate, on
the yield and composition of the oil extracted from the above spent coffee grounds. The
highest yield (12.4%) was obtained at a temperature of 333.2 K with 5% ethyl lactate as a
co-solvent (Georgieva et al. 2018).
A great proportion of the overall, worldwide spent coffee waste biomass production is
generated at the industrial sites where instant coffee is manufactured. Therefore, a further
treatment for added value compounds, for biofuels or any additional application will be free
of costs of collection and transportation of the raw material, and their management. The
weight that these costs carry in processes which target the valorization of waste biomass
constitute a major issue; they can be such that they can actually make or break the respec-
tive processes (Iervolino et al. 2018). Consequently, expansion of existing instant coffee
industries to incorporate processes for the extraction of added value compounds from the
remains of the spent coffee may be a venture worth considering.

2.2.2 Lignocellulose Biomass

Sindhu et al. (2016) present an overview of lignocellulosic biomass, which is high in cel-
lulose, hemicellulose, lignin and as such it can serve as an excellent source for biofuel
generation. And of course, an important virtue of this type of biomass is that it is renewable.
Second generation lignocellulosic biomass in particular, in addition to being an excellent
candidate for bioethanol production, is an amply available and low cost feedstock, which
is often accumulated in places favoring its treatment and as such, minimalizing the other-
wise high costs of transport and collection process as analyzed in the previous subsection
for spent coffee grounds. Such biomass can be the straw of cereals cultivated for human or
animal consumption, i.e. field residues or process residues such as residual biomass pro-
duced during the collection of wheat, rye, barley, oats, corn and other cereals, which are
traditionally used for animal feed, as natural absorbent of water, and as insulation mate-
rial in certain constructions. The collection of cereals takes place in the summer months
and as such this type of biomass can be collected in a dry form and can be transported to
biorefineries simultaneously with the collection of the edible crop. Such an option reduces
the management needs associated with the residual biomass collection. Husks of legumes,
rice and corns can be also used in a similar way and the same methods can be applied for
their collection. Additional source of lignocellulosic origin is the bagasse disposed of in
bioethanol production units, which with further treatment can produce additional biofu-
els. Moreover, they are readily available in the sites of bioethanol production and unless
they are further treated for further added value products generation, they form a waste
which has to be disposed of according to environmental requirements at additional costs
(Figure 2.3).
According to Eurostat, cereals in the EU are aiming to be used as food or food deriva-
tives by humans and/or animals. Cereals are extensively produced in Europe and all over the
world. During 2015 cereals and rice production in Europe reached around three hundred and
twenty million tons, half of which were different kinds of wheat and approximately 20%
 Waste Biomass Suitable as Feedstock for Biofuels Production 23

Figure 2.3 Hay residues. Source: Truncated photo by Petar Starčević (https://www.pexels.com/photo/
hay-field-under-clear-sky-2389122).

was barley (KFE 2017). In addition to cereals, Europe cultivates plants with oil-generating
seeds, such as turnip-rape, rape, soya, and sunflower, the production of which was approx-
imately 30 million tons in 2016 with rapes accounting for two thirds of the quantity and
sunflower approaching one third. Soya production, which was only two and a half mil-
lion tons at the time, has increased (EAF&F 2017). These seeds are primarily used for oil
and/or bioethanol production. Therefore, the waste biomass produced after the treatment of
those seeds could always be further utilized for the extraction or formation of value-added
products, while appropriate further treatment can provide additional biofuels.
According to Zabed et al. (2016), the world availability of such biomass is approximately
three to four billion tons per year while 2–5 tons of such biomass could produce approx-
imately 1–2 m3 of ethanol. In their review, amongst others, they provide the content of
cereals of different origin (i.e. from rice or wheat, or barley or oat) in cellulose, hemicel-
lulose and lignin as measured by Saini et al. (2015), Ludueña et al. (2011) (rice husk), and
Sánchez (2009) (oat and rye straw).
Bioethanol production has well advanced over the last decade. The processes for its pro-
duction were designed to employ primary biomass produced for that very purpose. As such,
the composition and properties of the raw materials were assumed to be adequately similar.
However, the biomass properties are subject to numerous conditions such as variety of the
crop, season, temperature variations during growth, soil properties, irrigation frequency,
type of fertilizer and quantities used. Additionally, as these were industrial size continuous
processes, the constant supply of biomass of the same quantity and composition is of key
importance for their efficient operation which also ensures good and within the required
specification product quality (Abraham et al. 2016). Second generation biomass and gen-
erally waste biomass of a varying origin and composition will require inventive actions and
the design of processes versatile enough to adapt to the expected variations. This may be
proven a very challenging task. Abraham et al. (2016) discuss the potential of rice straw
as a second-generation source of biofuel so that this waste with a high annual volume can
be exploited. While this type of biomass can be used for the production of bioethanol, the
solid waste of the process can be further treated for the production of other type of biofuels.
The formation of biopolymers can also be feasible from this feedstock. According to Saini
24 Process Systems Engineering for Biofuels Development

et al. (2015), the annual universal production of rice residues is over 700 tons. Useful waste
biomass from rice straw consists of the rice stems, the leaf-sheaths and their blades. Much
research has been conducted on this waste biomass and data on its composition are available
in the literature.
Wheat straw is the residue of harvested wheat with an estimated annual yield of 1–3 tons
per acre. Talebnia et al. (2010) report research on wheat straw as a biofuel source. This
raw material, as many others of its kind, mainly contains cellulose (of strains very highly
attached to each other), hemicellulose and lignin in mass proportions of around one third
to 40% for the former, one fifth to one forth hemicellulose, and 15–20% for the latter, as
well as a smaller proportion of extractives. This kind of biomass resource is encountered
in ample quantities everywhere in the world. Those remains of agricultural produce after
all useful food parts have been collected can be exploited for direct bioethanol production
without the need for any additional investment. The United States alone generates over
400 million tons of biomass which comes from agricultural wastes usable for bioethanol
production, while an additional amount of approximately 0.4 billion tons of energy crops
is also available (Saini et al. 2015).
The production of corn, wheat, rice and sugar cane can provide the majority of agricul-
tural waste biomass which with appropriate management can form a precious raw material
for biofuel production (Kim and Dale 2004).
The production of biofuels from dedicated crops is straightforward because the employed
biomass composition is more or less the same. Waste biomass on the other hand has a wide
variability of composition. Consequently, its use as as an alternative feedstock for the pro-
duction of biofuels introduces several technical challenges. An effective way to resolve this
shortcoming is the implementation of appropriate pretreatment stages. Such stages can for
instance increase the biomass content in fermentable sugars. Alternatively, new fermen-
tation technologies can be developed to ensure viability and efficiency of the production
process (Sarkar et al. 2012).
Corn waste biomass, which consists of stalks and leaves, the empty cobs and the husks,
is a promising biomass for bioethanol production, because of its quantity (1 g per kg of
produced corn grains or about four tons of waste biomass per acre according to Kim and
Dale (2004)) and its composition (Sarkar et al. 2012).
A study conducted by Ayeni and Daramola (2017) considered the exploitation of corn
waste biomass for the production of biofuels and other products of everyday use. For that
purpose, they utilized different pretreatment methods involving, amongst others, alkaline
hydrolysis with or without hydrogen peroxide and dilute acid hydrolysis. They used these
processes in order to characterize the corncob, to enhance the cellulose, to remove the lignin,
to solubilize the hemicellulose and overall to evaluate the economics of the aforementioned
exploitation of this type of biomass on the basis of the above separations and characteri-
zations and they proposed methods which they demonstrated can ensure process viability
(Ayeni and Daramola 2017). Shariff et al. (2016) have also conducted research on corn-
cob residual biomass and in other kinds of feedstocks like palm wastes, rice husk, wheat
straw, wood sawdust, corncob (which is abundant in Malaysia throughout the year) for the
slow pyrolysis process. Corncob waste biomass was characterized for its cellulose, hemi-
celluloses and lignin content; the mass fractions of which were found to be approximately
0.46, 0.4, and over 0.11, respectively (Shariff et al. 2016). Corncob has low nitrogen and
 Waste Biomass Suitable as Feedstock for Biofuels Production 25

Figure 2.4 Perennial plants.

sulfur contents and a high proportion of volatile matter, thus it was demonstrated that slow
pyrolysis is a suitable method for its valorization.
Bagasse from a multiplicity of processes can form an excellent feedstock for the gen-
eration of biofuels, simultaneously solving the environmental issues associated with their
disposal or cutting costs from their further treatment. According to Sánchez (2009), the
quantity of bagasse estimated at the time from all over the world was over a third of a mil-
lion tons. Similarly, perennial biomass can be exploited for the same objective (Mantziaris
et al. 2017). However, for both categories of these biomasses, the variation in quantity, qual-
ity (composition of feedstock) and availability of feedstock are posing new technological
challenges imposing the need for versatile and inventive approaches for their valorization.
At the same time however, they can act as a make-up feedstock to counterbalance seasonal
and compositional variations of biomass (Figure 2.4).
In summary, lignocellulosic residual biomass is abundant on the planet, it is of low cost
and often a waste which has to be environmentally disposed of, concentrated and accessi-
ble. It has a composition which is suitable for biofuel production; much research has been
conducted already and technology for its treatment is available.

2.2.3 Palm, Olive, Coconut, Avocado, and Argan Oil Production Residues
Edible oil production from trees serves the biofuel biomass supply in different ways. One
is that the oil production trees have a considerable life span and as such they are provid-
ing yearly biomass from pruning and leaves. The trees producing oily fruits have a smaller
but non-negligible content of oils in their leaves and branches; as such, the valorization of
residual biomass from trimmings toward biodiesel production in particular, is appropriate.
They are evergreen trees and as such they produce a continuous leaf supply. Their leaves
have valuable compounds, the extraction of which can be considered a means of improv-
ing the economics of processes, thus valorizing their residual biomass for biofuel. And, of
course, the used cooking oil resulting at a later stage is an excellent biodiesel source.
26 Process Systems Engineering for Biofuels Development

For some, the residual biomass has been already well characterized and their exploitation
technologies have been developed or are in the process of development. Moreover, the need
for oil production from all the aforementioned plant fruits is in increasing demand and
plantations are rapidly increasing. Therefore, the future and present abundance of this type
of biomass is also an important factor for which this type of feedstock is worth considering.

2.2.3.1 Olive Oil Production Residues

The olive tree is a very long living tree. It is not uncommon to find trees as old as a few
or even several ages, while the existence of trees over a thousand years old can also be
found. They are evergreen trees and their fruit, the olive, is either mildly treated and its
flesh is eaten or it is crashed to provide its juice, olive oil. The great majority of olive oil
production comes from trees grown around the Mediterranean (Christoforou and Fokaides
2016). Recent global olive oil production amounts to approximately two million tons. Olive
oil production involves mechanical separation of the fruit from the leaves that are gath-
ered together with the fruit during collection, the mechanical crashing of the olives, the
separation of the oil from the solid mass with or without the addition of water as an assist-
ing extraction solvent, and the subsequent separation of the olive oil from the solid (olive
mill solid wastes) and the liquid wastewaters. Depending on the method of treatment, and
whether there has been addition of water during treatment or not, subsequent wastewaters
can be removed separately or together with the solid residue. Both those waste streams
are produced in very large quantities and their pollution potential is high. Christoforou and
Fokaides (2016) present a comprehensible review of techniques employed for the olive
mill solid waste to energy utilization. The treatment and disposal of olive mill wastew-
aters however, forms a much greater problem, the solution of which is a really pressing
necessity. The applied techniques have not given satisfactory results (Haddad et al. 2017).
In addition to the above, olive leaves, olive tree branches from pruning and bark can be
utilized toward that direction. Although extensive research has been done on the reduc-
tion of the pollution caused by olive mill wastewaters, substantially less work has been
conducted toward the valorization of the solid and liquid wastes associated with olive oil
production. There are however studies on added value products from olive leaves; studying
important effects of olive leaves as antihypertensive, anticarcinogenic, anti-inflammatory,
hypoglycemic, antimicrobial and hypocholesterolemic (Talhaoui et al. 2014). Additional
medicinal properties have been reported. Extracts of olive leaves can also serve as cosmetic
precursors and also as food preservatives (Figure 2.5).
Altogether, every year a great amount and variety of waste biomass is generated from
the olive cultivation (e.g. pruning of branches) and the olive treatment processes to obtain
edible olives and olive oil. These residues involve olive leaves, olive branches from prun-
ing, olive mill wastewaters and olive mill solid wastes. The latter two, as mentioned above,
form an undesirable and difficult to treat waste. The leaves of the former two are occasion-
ally used as animal food. Branches are usually burned on site as they are difficult to be
moved and transported (Talhaoui et al. 2015). In addition to being in the pruned branches,
olive leaves, are collected in large amounts in the olive mills. This is because leaves are
also falling during the collection of the olives. Leaves collected together with the olives are
approximately one third of the volume or one tenth of the weight of the collected olives
which are taken to the olive mill (Herrero et al. 2011). The leaves are separated from the
 Waste Biomass Suitable as Feedstock for Biofuels Production 27

(a) (b)

(c) (d)

Figure 2.5 Typical olive trees before pruning (a, c). Leaf-load of the tree before trimming (a). Typical
branch-load of the tree before trimming (c). Typical pruned trees (b, d).

olives by suction, immediately before milling. Erbay and Icier (2010) analyzed olive leaves
and found that they contained oil (about 7%), carbohydrates (about 30%), protein, fiber,
ash, and water. In a subsequent study, Erbay and Icier (2010) characterized olive leaves
and provided a list of typical compounds which can be found in and also extracted from
the leaves. These compounds are worth knowing and considering for improving the eco-
nomics of a facility which aims at using olive cultivation waste biomass for the production
of biofuels, as mentioned in Section 2.1.1. As Velazquez-Martı et al. (2011) and Talhaoui
et al. (2015) report in their review, the leaves from a typical olive tree pruning amount to
about 25% w/w of the trimmings; the thin branches constitute approximately half of the
total weight of the pruned biomass and the thicker branches or wood the remaining 25%.
Naturally, this is only an example and the proportions vary depending on a number of fac-
tors such as the variety of olives, the kind of cultivation and the protocols followed, the size
and the age of the trees, and the local pruning practice. However, rainfall during the year
may also be an important factor. The pruning of olive trees can produce around 10 ton/ha of
residual biomass depending on the country, the variety, the morphology of the area, as well
as the local climate conditions and cultivation practices. However, these figures can sub-
stantially change as pruning frequency and extent can be substantially altered because of
the transition of the farming activities to complementary ones (i.e. to complement agricul-
tural activities with animal pasture or with involvement in tourism). Velazquez-Martı et al.
(2011) give a comprehensive and detailed quantification of the pruning residual biomass as
a function of tree kind and conditions.
28 Process Systems Engineering for Biofuels Development

Olive bark and branches possess properties similar to those of the leaves and as such
they can be used for the extraction of valuable compounds. After the extraction of such
compounds, the solid waste can be further treated for biofuel production (Cara et al. 2007;
Ballesteros et al. 2011; Romero et al. 2010; Sequeiros and Labidi 2017); the waste leaves
can be also used for the same purpose.
Therefore, olive oil production wastes have the potential to be utilized for biofuel produc-
tion. Their production can be combined with that of olive oil and that of valuable products
from olive leaves. There are multiple methods to employ in order to minimize costs, increase
profitability and minimalize environmental impact on the way toward a cyclic economy and
sustainability. For example, Iervolino et al. (2018) have integrated the olive wood waste
supply chain with a thermocatalytic reforming process in which biofuels are produced and
optimally distributed. They have assessed the environmental gains of the venture using car-
bon dioxide equivalency as an indicator and they identified the conditions under which
such a process can be self-sustained and eco-friendly. Haddad et al. (2017) proposed a new
method for olive mill wastewaters valorization. Their method involved mixing olive mill
wastewaters with sawdust to produce biofuels and to transform a polluting entity to valuable
products.

2.2.3.2 Biomass Wastes from Palm Oil Production

Palm oil production comes mainly from Indonesia and Malaysia in particular. Smaller quan-
tities are produced in West Africa and in Latin America. The tree life span is 25–30 years,
they are harvested every 10 days and they are pruned twice a year; the waste biomass from
pruning is often left to decay and fertilize the soil of the farm.
According to the Department for Environment, Food and Rural Affairs (DEFRA) in the
UK (DEFRA 2012) palm oil is the most used vegetable oil in the world and it has a great
number of uses in addition to being inexpensive. Moreover, the palm tree is a very efficient
producer of oil, although its lifespan does not extend beyond 30 years (DEFRA 2012).
Palm oil is extracted from the flesh of the fruits of the oil palm tree. The quantity of
palm oil which can be produced from the flesh corresponds to approximately one third of
the mass of the fruit. The stone of the fruit is crashed and pressed to produce the palm
kernel oil. Those two oils are usually refined and further processed to produce a number
of derivatives for a number of different applications in the food sector (biscuits, margarine,
bakery products, frying oil), and also in the production of biodiesel, animal foods, soaps,
and cosmetics. Its use as a first-generation biofuel is also increasing.
Similarly to olive oil production, after the extraction of the palm oil there is a substan-
tial amount of biomass left, the management of which is very problematic. Consequently,
a great deal of effort is put toward its valorization via the production of fuels. However,
most of this waste mass is continuously produced and therefore abundantly available and
also concentrated at the place of oil production. Therefore, the economics of a subsequent
treatment will not be burdened by the costs of collection; moreover, if additional treatment
is planned in the same area via an extension of the oil production facility, transport costs
will be also negligible.
The energy content of those wastes can be exploited following different treatments
employing a number of existing technologies, ranging from direct combustion, combustion
to electricity generation, pyrolysis, esterification, torrefaction, gasification, up to anaerobic
 Waste Biomass Suitable as Feedstock for Biofuels Production 29

digestion and fermentation. Research is advancing in biomass characterization, factors

affecting economic viability, optimal existing technologies and improved methods of
treatment, economics, life cycle analysis, and others (International Institute for Sustainable
Development 2014).
Lee et al. (2017) investigated the slow pyrolysis products obtained by the relevant palm
oil production wastes (palm kernel shell, empty fruit bunch and palm oil sludge) as feed-
stock. Their findings implied that the former two had higher lignin contents; moreover they
appealed as precursors for biochars while displaying promising potential for biofuel gen-
eration.
As mentioned above, Malaysia is a leading country as a palm oil producer. Chiew and
Shimada (2013) have applied a life cycle analysis, in order to make a comparison of biofuel
production using different technologies and a waste biomass consisting of empty bunches
of palm-oil fruits as feedstock. The research also involved other potential uses of this waste.
They found that the best option of biofuel was for combined heat and power production.
In a later study, Kurnia et al. (2016) compare and discuss developed and in the process of
being developed technologies which are used for the valorization of palm oil wastes toward
biofuel production. They have furthermore discussed the challenges that future research
and development has to address. They classify the biomass treatment and conversion meth-
ods as thermochemical, biological or physical, each category having its pros and cons. The
thermochemical processes are more suitable for large scale applications, although they are
more energy intensive compared with the other two categories. The studies comparing the
different technologies back up their arguments by means of relevant life cycle analysis stud-
ies via which the sustainability aspects of the different scenarios are evaluated. In the whole
process of palm oil generation, the plantation stage appears to carry a heavy environmental
weight. However, in terms of costs and on the basis of economics, biofuels from the treat-
ment of palm oil waste do not appear attractive, especially when the cost of fossil fuels is
dropping or stays at the levels of 2016. Therefore, aspects which affect the economics of
waste treatment such as transport and distribution of biofuels, design of conversion meth-
ods, have to be improved. Better improved metrics for a more representative evaluation of
the environmental impact of the processes via life cycle analysis have to be devised. It is
worth mentioning here though, that the plantation costs and environmental impact have to
be considered in connection with the life span of the tree. Palm oils become productive four
years after they are planted, and they are productive for 20–30 years. Trees with a longer life
span may form a better alternative toward edible oil production and subsequent exploitation
of their residual biomass. As such, if a comparison was to be made, olive trees are superior
to palms, although they grow in different geographical areas.
The aforementioned findings regarding the methods of exploitation of energy using palm
oil waste biomass are also supported by a review conducted by Sukiran et al. (2017).
According to them, the hygroscopic nature, the low calorific value and the high quan-
tity of oxygen and moisture contained in those wastes restricts their potential as fuels. As
such, they propose to enhance the properties of those wastes toward the formation of fuels
employing torrefaction processes; they present an overview of factors associated with such a
treatment. They also show characterization results of these waste biomasses. They discuss
different views resulting from previous research (which is also projected to the present)
on oil palm solid wastes torrefaction; potential applications related to the yield obtained
via those methods of treatment were shown. Moreover, discussion was conducted around
30 Process Systems Engineering for Biofuels Development

the potentials and the evolution of methods for the valorization of palm oil wastes. They
conclude that the torrefaction technology can form an excellent method of biomass valoriza-
tion toward energy production. Finally, they evaluate the relevant costs of the technique and
they characterize it as potentially feasible and economically viable.
Asibey et al. (2018) have conducted research on the valorization of waste biomass in
Ghana toward the production of electricity. They acknowledge the impact which waste col-
lection and management has on the economics of waste valorization, but also the health
risks from the untreated waste, especially in African cities. Palm oil waste biomass is an
excellent feedstock for the production of electricity according to their research; it is also a
better source or raw material for combustion compared with the rest of the waste produced
by the local agro-industry. The volumes of waste biomass produced in the process of palm
oil production are very large and have already been used for the generation of electricity as
a means of energy sources for the needs of the relevant industrial processes; the potential of
such a valorization in sub-Saharan Africa is great. However, in their view, there is a hurdle
due to the existing local policies, which have to be reconsidered together with the current
institutional arrangements so as to take advantage of this potential. Their work also refers
to the need to give consideration to maintaining biodiversity and to increase the safety of
the processes (Asibey et al. 2018). Obviously, the appropriateness of biomass for biofuel
production strongly depends on the area, the alternative energy resources of the local com-
munity, the variety of potential feedstocks, and as such, valorization of palm oil wastes in
Africa is unlikely to have the same efficacy as in Malaysia or Indonesia or in South East
Asia in general. Consequently, studies which deal with the economics of such processes,
especially when based on life cycle analyses, have to refer to the local conditions, because
owing to the different prices and costs of labor, all such findings are area specific.
A study conducted in Malaysia by Hansen and Nygaard (2014) examined the operational
status and the evolution of a number of plants which were using or were planning to use
palm oil waste biomass for the production of energy; the study examined both completed
plants and plants under construction. Amongst a continuing increase of construction
of such plants, 39 were advancing toward full operation. Plants accounted for in the
study involved plants producing electricity either as stand alone or in connection with
others. The produced electricity was then directly used by the industries of the area, by
plants connected to the national grid, and by combined power and heat generation plants.
During 1990–2010, substantial pressure was exerted from the international community
because the industry of palm oil production was expanding in a way that could provoke
serious environmental problems in the future. As a consequence, Malaysia was forced to
adopt environmentally friendlier practices including the valorization of palm oil effluents
and waste biomass. A simultaneous fossil-fuel price rise encouraged energy-intensive
industries like those of cement and rubber gloves, refineries, and power generators to use
energy produced by sources like palm oil waste biomass to meet their respective needs.
However, as in the case of Ghana, progress on waste mass valorization had been hindered,
amongst others, by problems in the implementation of energy policy, cost increases, and
obstacles in network formation.
In addition to palm oil waste, date palm residues can be utilized for biofuel produc-
tion. Bensidhom et al. (2018) have studied this type of lignocellulosic biomass, which is
extremely abundant in Tunisia. For that purpose, they used a fixed-bed reactor to pyrolize
the respective biomass, i.e. date palm wood, leaves, empty bunches of fruit and similar
 Waste Biomass Suitable as Feedstock for Biofuels Production 31

remains; they produced bio-oil (yield 17–26%), biochar (yield up to 37%), and syngas
(yield over 40%). They characterized the respective biofuels and found that they have a
great potential to function as alternatives to the usual fossil fuels. This biomass however,
before being exploited for biofuel production, needs to be aggregated and transported to
appropriate places for treatment.

2.2.3.3 Biomass Wastes from Coconut Oil Production

Besides palm oil waste biomass, coconut oil residues are also an available source for biofuel
production. The pros and cons of the management of its valorization have analogies with
those of palm oil, as the issues of transportation and aggregation of biomass residues are of
a similar kind, while at the same time, they are produced in similar areas. In Malaysia for
example, coconut is extensively cultivated; however, after the extraction of coconut milk
the residual biomass is usually used as fertilizer or as animal feed or it is left to decay.
Research on the characterization of this biomass has been conducted by several researchers
and a brief account is given below.
Abigor et al. (2000) studied biofuel production by catalytic (PS30 lipase) transesterifi-
cation of palm stone oil and coconut oil and examined the effect of tert-butanol, 1-butanol,
n-propanol, and iso-propanol on the overall outcome. They characterized the oils and
reported the efficiency of the reaction in each case. They found that they could produce
fuel of a quality consistent with biodiesel specifications.
Biodiesel fulfilling the requirements of quality was also produced by Thushari and Babel
(2018), in a high efficiency process which employed methanol and a catalyst produced by
coconut meal waste in an open reflux reactor.
Sulaiman et al. (2013) considered the valorization of coconut waste by means of a reac-
tive extraction method recovering coconut milk from the coconut solid waste in order to
produce biodiesel. Different temperatures, potassium hydroxide concentration and agitation
speeds were employed in their experimental study. They then employed response surface
methodology to identify the range of optimum conditions (reported in their work) with a
yield of approximately 90% toward biodiesel.
In a subsequent study, Talha and Sulaiman (2018) produced biodiesel using solid coconut
waste and a CaO-waste derived catalyst, in a packed bed reactor; they conducted an in situ
transesterification reaction obtaining biodiesel yield of 95% at 61 ∘ C with 2.3% w/w catalyst
loading and methanol. The mass of methanol was 12 times that of the solid.
A refinery waste coconut oil employed as raw material, was treated with ethanol, with
and without water, for the production of biodiesel, by Oliveira et al. (2010). The conversion
was over 99% on a molar basis. They found that when water adsorption was simultane-
ously conducted with the esterification, a lower molar quantity of alcohol was needed, thus
improving the process costs and mass economy.

2.2.3.4 Biomass Wastes from Avocado Oil Production

The consumption of avocado is very high, and it has been increasing worldwide in recent
years. Avocado is eaten raw, but it is also extensively used in large quantities in the food
industry. It is also used in the production of cosmetics.
32 Process Systems Engineering for Biofuels Development

The seeds are high in starch and polyphenols. Avocado seeds were characterized in
terms of thermal and physical properties as well as chemical composition and technolo-
gies for their valorization as fuel are shown by Domínguez et al. (2014). Avocado is in
fact, a fruit which is extensively traded in the world and its world-wide production was
estimated to be nearly five million tons in 2014, which corresponded to an increase of
140% from 1995 to 2015. A great amount of valuable biomass is the residual products
(skin, stone) of guacamole manufacturing. This industry generates a large amount of such
waste biomass products which are good for energy production, because they possess high
calorific value (Perea-Moreno et al. 2016). Durak and Aysu (2014) studied experimentally,
by means of slow pyrolysis, the potential of avocado stones for biofuel production. Avocado
stones were used in a fixed-bed plug-flow reactor with and without catalyst; bio char, oil and
gas were obtained. Potassium hydroxide and alumina were used as individual catalysts, at
three temperatures ranging from 400 to 600 ∘ C. In addition, the same group also produced
bio-oil from the same biomass, employing ethanol and acetone with potassium hydrox-
ide and zinc chloride as individual catalysts, as well as without catalyst. Avocado stones
were put in an autoclave and were subjected to supercritical treatment using ethanol or
acetone as solvents. The processes were conducted under high pressure, at temperatures of
250, 270, and 290 ∘ C with or without the aforementioned catalysts. The highest conversion
(liquid+gaseous products) of nearly 77% was obtained at 290 ∘ C with acetone via a zinc
chloride 10% catalytic process (Aysu and Durak 2015). Furthermore, Díaz-Muñoz et al.
(2016) and Bazzo et al. (2015) have also researched valorization of avocado stones toward
absorbents and adsorbents formation. Perea-Moreno et al. (2016) described the enthalpy of
combustion properties of avocado stones. Their calorific value was similar to that of olive
stones and almond shells, and it was concluded that the avocado stones may be used as a
solid fuel for heating.
Rachimoellah et al. (2009) investigated the capacity of oil from avocado seeds to produce
biodiesel. They conducted a series of measurements at different temperatures with different
alcohols by changing their molar ratios to the employed quantity of oil; they also modified
the method of removing impurities from the end product. They used sodium hydroxide as
a catalyst (1% w/w). They found that at 60 ∘ C, using six times more alcohol than oil and a
subsequent product dry washing process, a biodiesel 85% in methyl ester was formed.

2.2.3.5 Biomass Wastes from Argan Oil Production

Argan oil is produced in small quantities at present and its main production is centered
at the country of its origin, i.e. Morocco. However, as its properties are gaining increas-
ing recognition and publicity, its production in large scale is advancing fast. The argan oil
exports from Morocco have dramatically increased in the last 20 years; importers employ
it for a number of uses, cosmetics production being the most common one. As Charrouf
and Guillaume (2008) report, argan oil is an advanced product with excellent, valuable
dietary compounds. For that reason, interest in it has been increasing and there is a con-
tinuous rise and expansion of the extent of plantations and of the trees producing this oil,
both in Morocco and in other countries where the trees can grow. In a subsequent publi-
cation of theirs, Charrouf and Guillaume (2018) report on the progress of a project which
holistically addresses production, social, environmental and other issues associated with
sustainable processes. In their work the seven stages in argan oil production are explained.
 Waste Biomass Suitable as Feedstock for Biofuels Production 33

Argan oil is produced via inefficient production methods which however do not harm its
physicochemical properties. More efficient processes are currently being developed, while
their application and industrial implementation are also advancing.
As already mentioned, argan oil production and argan oil tree plantations are rapidly
expanding. Given the experience from other oils presented in this section, where the prob-
lem of waste first became significant and then consideration was given to the development
of methods for its management, it would be prudent to consider valorization of waste
biomass prior to the industrialization of production so that integrated, energy independent,
self-sustained argan oil production units can be developed. Such units can make allowance
to incorporate the production of added value products and the valorization of their waste
in their original process design. Naturally, intensive research is required in this field as the
characteristics of the materials involved are far from well known.

2.2.4 Citrus
Citrus is the most important fruit crop in the world with a production estimated at approxi-
mately 90 million tons in 2014. According to the Food and Agriculture Organization of the
United Nations, in 2012, the production of orange, lemon and grapefruit worldwide was
around 95 million tons. Nearly 26% of citrus fruits are used for the production of juice.
The biomass residues from the juice production is estimated to be 15 million tons, and it
consists of seeds, peels, and pulp. Citrus peel is rich in antioxidants and multiple other valu-
able compounds with anti-inflammatory, anti-cancer, anti-proliferative, anti-viral and other
activities which may contribute to the prevention of disease (Mhiri et al. 2017; Geraci et al.
2017). The largest share of citrus juice production is orange juice, followed by grapefruit,
lemon and lime. Other citrus fruit such as mandarins, tangerines, and pomelos are produced
and traded in much smaller quantities. The food industry is expanding very fast, especially
in the most developed countries, where the time spent in food preparation by families and
individuals is continuously shrinking. Consequently, the food industry is expected to gener-
ate more by-products, which in many cases is merely waste biomass of plant origin, such as
fruit skins, vegetable stems, fruit stones and seeds, pomace, nut shells, etc. Amongst those,
the citrus juice industry is producing a great deal of waste biomass very high in added value
compounds (Sharma et al. 2016).
The orange (Citrus sinensis) is a very common and highly consumed crop. It is part of
the citrus family, and it is one of the highest worldwide crops produced. The production of
concentrate for the juice industry is one of the primary applications. The orange peel is a
part of the residual biomass generated by the processing of citrus. The quantity of orange
juice production residuals is estimated to be around 16 million tons per year (Ayala et al.
2017).
Oranges are extensively cultivated in America (Florida and California in the United
States, Mexico, and Brazil), in Asia (China, India, Pakistan, Iran, and Mediterranean coun-
tries), as well as in all European, Asian and African Mediterranean counties (Siles-López
et al. 2010). More than 50% of the world orange production is generated in the Ameri-
can continent, where Mexico makes a significant contribution. Mexican orange production
was over four million tons in 2013, which represents over 6% of the worldwide orange
production. Mexico accounts for 5% of the world grapefruit production and about 15%
34 Process Systems Engineering for Biofuels Development

of the lime–lemon global production. In Mexico, the citrus waste generation consists of
approximately 750 thousand tons of orange, 70 thousand tons of grapefruit, and around
350 thousand tons of lemon and lime (Ayala et al. 2017). According to the 2007 data of
the Statistical Database of the Food and Agriculture Organization of the United Nations,
Brazil produces over approximately 28% of the world orange production; it is followed
by the United States which produces over 10%, while Europe is in third place with a pro-
duction of just under 10%. Over 50% of the processed fruit becomes waste. This waste
biomass consists mainly of citrus peel (flavedo and albedo), seeds, membranes of the seg-
ments and the crushed endocarp. The peel can be used for extracting added value products
for preserves, to produce marmalade for food or the peels are even used to produce animal
food. However, in many cases it is just treated as a problematic waste and it has often just
been dumped in deserted areas close to the juice industries. In such cases, the decompo-
sition resulted in a serious pollution threat for the underground water, a health and safety
threat due to uncontrolled methane production, and a visual sore and odor-repellant for the
area. However, the valuable properties of the citrus peels and seeds have driven research
and technologies development aiming at valorization of this waste in multiple ways and
to a great range of applications from the extraction of pharmaceuticals, food preservatives
and flavoring agents to cosmetics and biofuels production. Of course, in order to develop
such alternative valorization methods a good knowledge of their composition is required;
Siles-López et al. (2010) report in their article of publications providing information and
characterization data of different varieties of citrus fruits. According to those, in citrus peel,
sugars, cellulose, hemicellulose, starches, lignin, organic acids and proteins are contained
in proportions which depend on variety, climate conditions, soil and other factors. More-
over, in this study, Siles-López et al. (2010) report how to produce methane from citrus
industry waste biomass; they also state the problems associated with such a production.
Erukainure et al. (2016) highlight that the use of orange peel waste can be utilized for the
extraction of high purity essential oils to decrease the purchase of high-cost raw materials
using water as a solvent. Orange and citrus peels in general can be used to produce numer-
ous quality added value products the production of which can compensate for the costs of
less profitable or even non-viable synergistic processes aiming at waste exploitation or at
environmentally oriented transformation processes. For example, Raga and Lopes (2003)
studied the chemical composition of a fivefold sweet orange oil obtained via vacuum distil-
lation. They found that the quantitative composition of the obtained products demonstrated
that by appropriate adjustment of the conditions of the fractionator, high quality oils could
be obtained.
John et al. (2017) conducted a detailed literature survey on bioethanol production using
citrus waste. They provide the composition of lignocellulose biomass of different feedstock
used for bioethanol production and compare it with that of citrus peel waste which contains
low lignin and high carbohydrates (fructose, sucrose, cellulose, hemicellulose, and pectin),
and as such it can favor bioethanol production. Citrus peels contain polysaccharides and
soluble sugars. They found that it is possible to make citrus peel waste more susceptible to
saccharification with enzymatic treatment. They employed pretreatment methods to reduce
the formation of compounds restricting its effectiveness. The pretreatment involved dilute
acid hydrolysis and steam explosion. A great incentive for the valorization of citrus peel
waste comes from the understanding that extraction of added value products prior to biofuel
production can greatly improve the economics of the venture, so scale up efforts are under
 Waste Biomass Suitable as Feedstock for Biofuels Production 35

way. Extraction of D-limonene as an added value product can also facilitate the bioethanol
production via saccharification because D-limonene inhibits the enzymatic fermentation.
The economics of the saccharification can be improved by the selection of enzyme type,
function or method of application. Options include the immobilization of the enzymes for
example, the use of cell-recycle reactors, and co-culture fermentation (John et al. 2017).
Bicu and Mustata (2011) worked with cellulose extraction from orange peel, the pulping
of which was achieved employing sodium sulfite and sodium metabisulfite. They evaluated
how the addition of the pulping precursor, the reaction residence time and the tempera-
ture affected the result. They found conditions at which good quality cellulose could be
recovered with over 40% w/w yield. The obtained cellulose was good for a number of uses,
such as water absorbents or fillers, including availability of cellulose as raw material for
production of further derivatives.
Lanfranchi (2012) conducted an evaluation of the economics of different methods of pro-
duction of energy using citrus waste biomass which resulted after removing the juice of the
fruits, i.e. what has been reported before as juice production residual biomass such as peel,
seeds, membranes of the segments, residual juice and pulp. Solid fuel can be produced via
drying of the waste and produce solid pellets for combustion. However, as previous experi-
ence has shown, it has to be done via very well-designed processes and employ only the less
moist parts of this residual. This biomass has occasionally been used as a fuel. However, it
was highly uneconomical. This is because its water content is very high (more than 90%).
In these applications, the residual biomass, as a whole, was dried before it was burnt. How-
ever, this incurred high energy costs. As a result, it ended up being a very inefficient energy
source, requiring more energy for its drying than the energy provided by its combustion.
Biofuels (bioethanol and biogas) production from those wastes is a much better option.
Rivas-Cantu et al. (2013) have conducted a study to present the best at the time tech-
nologies which can be employed for the valorization of citrus waste biomass toward the
production of energy, using laboratory data obtained from the pretreatment and enzymatic
hydrolysis of industrial waste. Their work involves results obtained from an extensive study
and long experience in the relevant work. They examined the production of sugars and
value-added compounds. They also studied how different methods of pretreatment, includ-
ing mechanical and chemical methods, could be used to increase the yields toward the
most valuable sugars. The key processing steps of the waste were the following: shred-
ding and grinding below 1.25 cm particles to form a suspension in water, which was par-
tially hydrolyzed by steam-, flash distillation or centrifugation for fermentation-inhibitor
limonene removal, enzymatic hydrolysis of the resulting slurry under a controlled tem-
perature and pH range, fermentation, or simultaneous saccharification and fermentation to
produce ethanol and finally distillation to separate the produced biofuel from the waste.
They also report problems associated with the implementation of this technology. Such
technologies have not yet found their way to an efficient large scale, although much progress
is being made.
In summary, citrus peel waste can form an excellent second-generation non-food ligno-
cellulose feedstock for biofuel production. Prior to its utilization as a biofuel raw material,
it can provide other added value products following extractive processes. All citrus wastes
are good for those applications but orange peel has a higher concentration of carbohydrates
and as such it can provide the desired products at higher yields (Joshi et al. 2015).
36 Process Systems Engineering for Biofuels Development

2.2.5 Grape Marc

According to the literature review conducted by Brunerova and Brozek (2017), the vine-
yards of the world extend to approximately 8 billion ha of land. The residual biomass from
grapes and wine production consists of pruning and trimming the vines producing an aver-
age of 5 kg per waste biomass per vine and later from the pomace, left over after the grapes
are crushed for the production of wine.
Corbin et al. (2015) report on the usage of grape pomace which is either further fermented
and distilled for the production of alcoholic drinks in the form of grapa, or is used as animal
feed, as a raw material for added value products like antioxidants or it can be used as a
promising source for biofuel (bioethanol or biogas) production. According to them, the
pomace contains a high portion of water soluble compound which upon fermentation can
provide ethanol equal to approximately 25% of the pomace waste or be subjected to acid
pretreatment followed by enzymatic hydrolysis to produce 0.4 l of ethanol per kilogram of
pomace.
The environmental impact of the burying of untreated grape pomace is highlighted by
Mendes et al. (2013), water contamination and soil erosion being amongst the most impor-
tant. In their review they report on the composition of the residual biomass following wine
production. High concentrations of phenolic compounds hinder the natural degradation of
this waste biomass; extraction of such compounds is something that should be aimed for
because they are very valuable and can find diverse applications such as for the formation of
food supplements or cosmetics or for other uses in the pharmaceutical industry, as adhesives
or as compounds whose range of applications can expand if the necessary research is con-
ducted. In their study, Mendes et al. (2013) focus on the grape skins and in particular on red
ones; after having assessed their composition, they evaluated their chemical structure with
attention on the main macromolecular constituents; the target was to develop strategies for
their valorization toward formation of compounds for biocomposite applications and insu-
lation materials. Typical composition identified was: cellulose, hemicelluloses, proteins,
tannins, sugars, aliphatic compounds, and ash in approximate percentages of 21, 12, 19,
14,12, 14, and 8%, respectively, a great proportion of which are water soluble. Following
this analysis, they considered it plausible to employ a multistage process for the treatment of
grape skins; they proposed pre-extraction of polar compounds, subsequent polar compound
extraction with hot water, and finally valorization of grape skin solid part as biocomposite.
In their work “Extraction and purification of high added value compounds from
by-products of the winemaking chain using alternative/nonconventional processes/
technologies,” Yammine et al. (2018) considered the valorization of winery waste prod-
ucts. Their literature review provided quantitative data (approximately 65 million tons
world production of grapes in 2014, around 50 million tons of which were used for wine
production with an estimated maximum of over 15 million tons of valuable waste biomass
left over), which can be used for the production of multiple value-added products. They
also compared different technologies for the extraction methods, but their comparison
could not provide any solid results (Yammine et al. 2018).
Prado et al. (2012) and Coelho et al. (2018) studied the supercritical fluid extraction
of grape seeds. The collated literature review of Prado et al. (2012) highlighted the ben-
efits of grape seeds for food supplement and their antimicrobial activity, which resulted
in the commercialization of relevant applications. The researchers focus on the extraction
 Waste Biomass Suitable as Feedstock for Biofuels Production 37

of those valuable compounds by means of supercritical carbon dioxide, but they do not
focus on any further treatment of the waste biomass from this process. Grapes employed
in wine production are crashed and the produced pomace is left to ferment. The fermented
pomace is subsequently subjected to distillation. The batch or semi-batch distillation of
those fermented residues provides pisco, a grapa-like alcoholic drink.
Farías-Campomanes et al. (2013) conducted an experimental study for the process of
extraction of polyphenolic compounds from such fermented grape bagasse. Bagasse of this
type contains approximately 50% seeds, while the remaining 50% is equally shared between
grape stems and skins. They employed supercritical carbon dioxide extraction with ethanol
and water to enhance extraction of polar compounds and they found that it is viable to
establish an industrial supercritical fluid extraction plant to recover phenolic compounds
from grape bagasse. They found that this is viable even for small scale production units.
Sanahuja-Parejo et al. (2019) have produced bio-oil via the co-pyrolysis of polystyrene
and grape seeds achieving high conversions and good oil properties, i.e. with a higher pH
value than the acidic bio-oils obtained from lignocellulosic mass, a lower content of phenols
and oxygen and an increased quantity of aromatics, owing to the hydrogen-donor contribu-
tion of polystyrene in the process.
Xu et al. (2009) subjected winery waste to pyrolytic treatment in a bubbling fluidized bed
reactor. The calorific value of the produced gases and bio-oil were evaluated and optimum
conditions for the process were identified in order to ensure a positive net energy operation.
Using self-produced biogas and bio-oil to cover the thermal needs of the pyrolysis, they
achieved sustainable bio-oil production.
Zhang et al. (2017) designed and compared two methods of valorization of grape pomace
toward energy uses; combustion for electricity and pyrolysis for biofuels, i.e. for bio-oil
and bio-gas. They found that the latter had a better positive impact on wineries in terms
of economics as well as environmental aspects. According to this study, small scale winer-
ies (<50 ton annual wine production) appeared to obtain a high economic advantage by
composting of this biomass, as opposed to the large scale ones, which did not benefit sub-
stantially from the addition of such processes for the pomace valorization.
Miralles et al. (2008) studied the grape stalks waste as low cost biosorbents: their work
established that grape stalk waste is an efficient sorbent in the removal of cadmium(II),
lead(II), nickel(II), and copper(II). The results obtained showed that grape stalks can be
used as sorbing material for those metal ions.

2.2.6 Waste Oil and Cooking Oil

This type of waste is found in great abundance. It is very harmful for soil and the aquatic
environment if disposed of. At the same time, it forms an important source of energy. Its
importance from an economic point of view becomes preponderant for countries which
import edible oil (Hajjari et al. 2017). The treatment that it requires to be transformed to
useful fuel is relatively limited. Therefore, there is ample scope in its large-scale transfor-
mation to biodiesel. The key difficulties in its valorization are primarily associated with its
collection and the substantial variability of its composition, which depends on a number
of factors. A great quantity of biodiesel is already being generated by means of process-
ing waste oil and cooking oil in smaller and larger scale using different oils and extensive
38 Process Systems Engineering for Biofuels Development

research is being conducted in order to enhance its transformation to good quality biodiesel.
Piker et al. (2016) have tested the efficacy of catalytic biodiesel production at ambient tem-
perature using appropriately treated waste egg shells for the formation of the employed
catalyst. The study was conducted in small scale using different alcohol to oil proportions,
stirring conditions, treatment length, and catalyst cycles. In all cases promising results were
obtained. Atabani et al. (2019) have combined waste cooking oil with spent coffee grounds
and higher alcohols to achieve good quality biodiesels with very good cold flow proper-
ties. Along the same lines, Thushari et al. (2019) developed and characterized catalysts
formed from coconut husk. Subsequently they employed these catalysts to convert waste
coconut oil to biodiesel. Using catalyst, methanol and oil in approximately 10:12:1 weight
proportions they obtained over 75% conversion of good quality, within standards, biodiesel.
They studied the effect of increasing temperature, catalyst content and reaction time and
achieved an optimum 90% conversion to biodiesel. Chicken fat and waste cooking oil were
catalytically transformed to good quality biodiesel in a homemade system by Soegiantoro
et al. (2019). Different temperatures, catalyst and alcohol types were tested, and efficiency
reached 91% under optimal conditions while the cost per unit product was lower than the
commercial biodiesel price.
As reported earlier, the collection of waste cooking oil is one of the important problems
associated with its valorization. In a comprehensive work, Hajjari et al. (2017) provide an
effective and practical method for the solution of this problem. On the other hand, Shein-
baum et al. (2015) conducted a dual-purpose study in Mexico city. One aspect of the study
was focusing on the methods and obstacles faced during waste cooking oil collection and the
other on the emissions produced by its use in public transport bus diesel engines. They found
that the collection and disposal had a great variation, while the use of biodiesel resulted in
reduced emissions only in the case where the exhaust gas of the buses employed a recircu-
lation system.
In conclusion, waste cooking oil is a promising feedstock for biodiesel production, but
research on this topic is imperative and versatile methods of its valorization still need to be
invented.

2.2.7 Additional Sources

Silva et al. (2016) and Souza et al. (2016) have studied the production of biodiesel using
the non-edible acidic Macauba (Acrocomia aculeata) oil with very promising results. City
waste or food waste can be subjected to anaerobic digestion for the production of biogas.
However, a substantial amount of residual mass has to be further treated and valorized (e.g.
Yang et al. 2016).
In addition to the above reported residual biomass biofuel feedstock, a multiplicity of
other sources exist. For example, wastes such as husks and kernels of seeds, which have a
high calorific value and can be used as solid biofuels via direct burning. However, they can
also be utilized as adsorbents of unwanted pollutants like dyes and heavy metals as reported
in Senthil Kumar et al. (2010, 2011a,b,c) and Vieira et al. (2012) (Figure 2.6).
Similarly as Reynel-Avila et al. (2016), Trevino-Cordero et al. (2013), and Vargas et al.
(2011) have published, flamboyant and mango biomass have the capacity to remove heavy
 Waste Biomass Suitable as Feedstock for Biofuels Production 39

Figure 2.6 Municipal waste. Source: Truncated photo by Kelly Lacy (https://www.pexels.com/photo/
photo-of-a-dump-truck-across-buildings-2382894).

metals from polluted waters. The husk of mango seed in particular can be used as an adsor-
bent for the removal of acid dyes in the monoazo and anthraquinone class (Davila-Jimenez
et al. 2009) and may be possibly employed for the adsorption and removal of dyes from
textile effluents (Shoukat et al. 2016). Similarly, dyes can be removed from wastewaters
using dried and powdered mango leaves. Such a powder can be easily formed in abundance
(Uddin et al. 2017). Moreover, Memon et al. (2009) report mango kernels can adsorb toxic
chemicals such as parathion and Kanjilala et al. (2014) have shown that mango seed and
husks can also adsorb heavy metals, whereas Reynel-Avila et al. (2016) mention that the
biomass of bush mango seeds in combination with the biomass of flamboyant can be used
for heavy metal removal from aqueous solutions. It is also worth mentioning that activated
carbons made from mango husks have been used as carbon dioxide adsorbents (Correia
et al. 2017).
Castor seed hull can remove heavy metals from aqueous solutions (Mohammod et al.
2011) while walnut shell can be used as a precursor for the absorption of benzene and
toluene from waste gases (Karimnezhad et al. 2014). Peach palm can also remove heavy
metals from wastewaters (Duranoglu et al. 2010; Salvado et al. 2012; Massocatto et al.
2015).
Peach stones can remove emerging contaminants from aqueous solutions such as stimu-
lants, anti-inflammatory and psychiatric drugs ibuprofen and tetracycline (Álvarez Torrellas
et al. 2015, 2016b), real pharmaceutical wastewaters, such as hospital effluents (Álvarez
Torrellas et al. 2016a), and phenol (Hannafi et al. 2008). Finally, they can also be used for
the removal of dyes (Attia et al. 2006; Álvarez Torrellas et al. 2016a).
Heavy metal dyes and pharmaceuticals can be also removed from wastewaters by means
of chicken feathers, a high-volume food biomass (Reynel-Avila et al. 2015a). Chicken
feathers can also be appropriately treated to be used in the process of cleaning wastew-
aters. They can remove organic material and ammoniacal nitrogen (Moon et al. 2017), and
phenol (Banat and Al-Asheh 2001). Activated feather carbon was an efficient absorbent
of amoxicillin in surrogate wastewater and other persistent antibiotic pollutants (Li et al.
2017). They have been also used as very effective absorbents of different types of dyes from
wastewaters (Mittal et al. 2012; Gupta et al. 2006; Cervantes-González et al. 2017).
40 Process Systems Engineering for Biofuels Development

Research has shown potential additional uses of other waste biomass. As such, waste
meat bones have been also used for the production of bone char adsorbents. It has
been shown that bone chars are good for the removal of pharmaceutical compounds
(Reynel-Avila et al. 2015b) or heavy metals (Rojas-Mayorga et al. 2016), or endotoxins
(Rezaee et al. 2009) from wastewaters, or formaldehyde from air (Rezaee et al. 2013).
Waste biomass can be also used for the production of pharmaceuticals as shown for
example by Tian et al. (2014) and Rodriguez et al. (2018).

2.3 Conclusions
The scientific community is very active in research on the valorization of residual biomass
in many different forms. There is extensive work being conducted on the characterization
of potential second-generation biomass which can be used as a source for different biofu-
els production. The objective is to develop new methods and technologies to exploit their
potential in the best possible way. In this chapter, we focused mainly on amply available,
sufficiently well characterized residual biomass, which has the highest potential for imme-
diate exploitation or for biomass, the feasibility of the exploitation of which requires the
need for substantial technological improvements. Anaerobic fermentation of wastes for bio-
gas production can be used as a simple form of mixed waste valorization for a broad extent
of waste. However, after biogas generation, a substantial residual waste mass remains. The
potential of more efficient and multipurpose methods of treatment are presently sought.
In addition to the issues of availability of biomass and the management of its collection
and transportation to the treatment facilities, the variability of the biomass quantity and
quality can pose significant technological challenges. Methods which can be used for the
pretreatment of different kinds of biomass have to be developed. Pretreatment can aim at
the production of selected key intermediates which can be further treated for the production
of biofuels. Such a breakdown of biomass treatment may be the solution to the variability
of feedstock quality. Different waste biomass sources can be used for the production of the
same intermediates and the temporal variability of the major sources can be topped up by
using smaller and varying alternative sources of a different origin.

Acknowledgment
This project has received funding from the European Union’s Horizon 2020 research and
innovation programme under the Marie Sklodowska-Curie grant agreement No. 778168.

References
Abigor, R.D., Uadia, P.O., Foglia, T.A. et al. (2000). Lipase-catalysed production of biodiesel fuel from
some Nigerian lauric oils. Biochemical Society Transactions 28: 979–981.
Abraham, A., Mathew, A.K., Sindhu, R. et al. (2016). Review potential of rice straw for bio-refining: an
overview. Bioresource Technology 215: 29–36.
Álvarez Torrellas, S., García, L.R., Escalona, N. et al. (2015). Chemical-activated carbons from peach stones
for the adsorption of emerging contaminants in aqueous solutions. Chemical Engineering Journal 279:
788–798.
 Waste Biomass Suitable as Feedstock for Biofuels Production 41

Álvarez Torrellas, S., Ovejero, G., Garcıa-Lovera, R. et al. (2016a). Synthesis of a mesoporous carbon
from peach stones for adsorption of basic dyes from wastewater: kinetics, modeling, and thermodynamic
studies. Clean Technologies and Environmental Policy 18: 1085–1096.
Álvarez Torrellas, S., Rodríguez, A., Ovejero, G., and García, J. (2016b). Comparative adsorption per-
formance of ibuprofen and tetracycline from aqueous solution by carbonaceous materials. Chemical
Engineering Journal 283: 936–947.
Asibey, M.O., Yeboah, V., and Adabor, E.K. (2018). Palm biomass waste as supplementary source of elec-
tricity generation in Ghana: case of the Juaben Oil Mills. Energy & Environment 29: 165–183.
Atabani, A.E., Shobana, S., Mohammed, M.N. et al. (2019). Integrated valorization of waste cooking oil
and spent coffee grounds for biodiesel production: blending with higher alcohols, FT-IR, TGA, DSC and
NMR characterizations. Fuel 244: 419–430.
Attia, A.A., Girgis, B.S., and Fathy, N.A. (2006). Removal of methylene blue by carbons derived from
peach stones by H3 PO4 activation: batch and column studies. Dyes and Pigments 76: 282–289.
Ayala, J.R., Montero, G., Campbell, H.E. et al. (2017). Extraction and characterization of Orange Peel
essential oil from Mexico and United States of America. Journal of Essential Oil Bearing Plants 20:
897–914.
Ayeni, A.O. and Daramola, M.O. (2017). Lignocellulosic biomass waste beneficiation: evaluation of oxida-
tive and non-oxidative pretreatment methodologies of South African corn cob. Journal of Environmental
Chemical Engineering 5: 1771–1779.
Aysu, T. and Durak, H. (2015). Assessment of avocado seeds (Persea americana) to produce bio-oil through
supercritical liquefaction. Biofuels, Bioproducts and Biorefining 9: 231–257.
Azadia, P., Malina, R., Barretta, S.R.H., and Kraft, M. (2017). The evolution of the biofuel science. Renew-
able and Sustainable Energy Reviews 76: 1479–1484.
Babu, B.V. (2008). Biomass pyrolysis: a state-of the-art review. Biofuels, Bioproducts and Biorefining 2:
393–414.
Ballesteros, I., Ballesteros, M., Cara, C. et al. (2011). Effect of water extraction on sugars recovery from
steam exploded olive tree pruning. Bioresource Technology 102: 6611–6616.
Banat, F.A. and Al-Asheh, S. (2001). The use of columns packed with chicken feathers for the removal of
phenol from aqueous solutions. Adsorption Science & Technology 19: 553–563.
Bazzo, A., Adebayo, M.A., Dias, S.L.P. et al. (2015). Avocado seed powder: characterization and its appli-
cation for crystal violet dye removal from aqueous solutions. Desalination and Water Treatment 57:
15873–15888.
Bensidhom, G., Hassen-Trabelsi, A.B., Alper, K. et al. (2018). Pyrolysis of date palm waste in a fixed-bed
reactor: characterization of pyrolytic products. Bioresource Technology 247: 363–369.
Bicu, I. and Mustata, F. (2011). Cellulose extraction from orange peel using sulfite digestion reagents.
Bioresource Technology 102: 10013–10019.
Brunerova, A. and Brozek, M. (2017). Is it advantageous to reuse fruit waste biomass from processing of
grapevine (Vitis vinifera L.) for briquette production? 16th International Scientific Conference Engineer-
ing for Rural Development, Jelgava. doi: https://doi.org/10.22616/ERDev2017.16.N109.
Campos-Vega, R., Loarca-Piña, G., Vergara-Castañeda, H.A., and Dave Oomah, B. (2015). Spent coffee
grounds: a review on current research and future prospects. Trends in Food Science & Technology 45:
24–36.
Cara, C., Romero, I., Oliva, J.M. et al. (2007). Liquid hot water pretreatment of olive tree pruning residues.
Applied Biochemistry and Biotechnology 379: 136–140.
Cervantes-González, E., Martínez-Gutiérrez, H., and Reyes-García, C.L. (2017). Optimization, kinetic,
equilibrium and thermodynamics parameters for Congo red adsorption from aqueous phase by untreated
chicken feathers. Desalination and Water Treatment 66: 291–298.
Charrouf, Z. and Guillaume, D. (2008). Argan oil: occurrence, composition and impact on human health.
European Journal of Lipid Science and Technology 110: 632–636.
42 Process Systems Engineering for Biofuels Development

Charrouf, Z. and Guillaume, D. (2018). The argan oil project: going from utopia to reality in 20 years.
Oilseed and fats, Crops and Liquids 25 (2): D209. https://doi.org/10.1051/ocl/2018006.
Chiew, Y.L. and Shimada, S. (2013). Current state and environmental impact assessment for utilizing oil
palm empty fruit bunches for fuel, fiber and fertilizer. A case study of Malaysia. Biomass and Bioenergy
51: 109–124.
Cholakov, G., Toteva, V., Nikolov, R. et al. (2013). Extracts from coffee by-products as potential raw,
materials for fuel additives and carbon. Journal of Chemical Technology and Metallurgy 48: 497–504.
Christoforou, E. and Fokaides, P.A. (2016). A review of olive mill solid wastes to energy utilization tech-
niques. Waste Management 49: 346–363.
Coelho, J.P., Filipe, R.M., Robalo, M.P., and Stateva, R.P. (2018). Recovering value from organic waste
materials: supercritical fluid extraction of oil from industrial grape seeds. The Journal of Supercritical
Fluids 141: 68–77.
Corbin, K.R., Hsieh, Y.S.Y., Betts, N.S. et al. (2015). Grape marc as a source of carbohydrates for
bioethanol: chemical composition, pre-treatment and saccharification. Bioresource Technology 193:
76–83.
Correia, L.B., Fiuza, R.A. Jr., Andrade, R.C., and Andrade, H.M.C. (2017). CO2 capture on activated
carbons derived from mango fruit (Mangifera indica L.) seed shells. Journal of Thermal Analysis and
Calorimetry 133: 337–354.
Davila-Jimenez, M.M., Elizalde-Gonzalez, M.P., and Hernandez-Montoya, V. (2009). Performance of
mango seed adsorbents in the adsorption of anthraquinone and azo acid dyes in single and binary
aqueous solutions. Bioresource Technology 100: 6199–6206.
DEFRA (2012). Sustainable production of palm oil, UK statement. https://www.gov.uk/government/
uploads/system/uploads/attachment_data/file/256254/pb13833-palm-oil-statement-1012.pdf (accessed
12 February 2019).
Díaz-Muñoz, L.L., Bonilla-Petriciolet, A., Reynel-Ávila, H.E., and Mendoza-Castillo, D.I. (2016). Sorp-
tion of heavy metal ions from aqueous solution using acid-treated avocado kernel seeds and its FTIR
spectroscopy characterization. Journal of Molecular Liquids 215: 555–564.
Domínguez, M.P., Araus, K., Bonert, P. et al. (2014). The avocado and its waste: an approach of fuel
potential/application. In: Environment, Energy and Climate Change II, The Handbook of Environmental
Chemistry, vol. 34 (eds. G. Lefebvre, E. Jiménez and B. Cabañas). Cham: Springer.
Durak, H. and Aysu, T. (2014). Effect of pyrolysis temperature and catalyst on production of bio-oil and
bio-char from avocado seeds. Research on Chemical Intermediates 41: 8067–8097.
Duranoglu, D., Trochimczuk, A.W., and Beker, U. (2010). A comparison study of peach stone and
acrylonitrile-divinylbenzene copolymer based activated carbons as chromium(VI) sorbents. Chemical
Engineering Journal 165: 56–63.
EAF&F (2017). Eurostat Agriculture, forestry and fishery statistics, 2017 edition. KS-FK-17-001-EN-N
.pdf, p. 69. https://ec.europa.eu/eurostat/web/products-statistical-books/-/KS-FK-17-001 (accessed 2
May 2018).
ECF (2016). ECF European Coffee Federation Market Overview, February 2016 (issue no. 14). https://
www.scribd.com/document/432390773/Coffee-market-overview-February-2016-pdf (accessed 19
April 2018).
El Halwagi, M. (2017). A shortcut approach to the multi-scale atomic targeting and design of C–H–O
symbiosis networks. Process Integration and Optimization for Sustainability 1: 3–13.
Erbay, Z. and Icier, F. (2010). The importance and potential uses of olive leaves. Food Reviews International
26: 319–334.
Erukainure, L.O., Osaretin, A.T., Ebuehi, M. et al. (2016). Orange peel extracts: chemical characterization,
antioxidant, antioxidative burst, and phytotoxic activities. Journal of Dietary Supplements 13: 585–594.
Farías-Campomanes, A.M., Rostagno, M.R., and Meireles, M.A.A. (2013). Production of polyphenol
extracts from grape bagasse using supercritical fluids: Yield, extract composition and economic
evaluation. Journal of Supercritical Fluids 77: 70–78.
 Waste Biomass Suitable as Feedstock for Biofuels Production 43

Georgieva, S., Coelho, J.A.P., Campos, F.C. et al. (2018). Green extraction of high added value substances
from spent coffee grounds: preliminary results. Journal of Chemical Technology and Metallurgy 53:
640–646.
Geraci, A., Di Stefano, V., Di Martino, E. et al. (2017). Essential oil components of orange peels and
antimicrobial activity. Natural Product Research 31: 653–659.
Goldstein, I.S. (ed.) (2018 reissued). Organic Chemicals from Biomass. Boca Raton, FL: CRC Press/Taylor
& Francis Group.
Guo, M., Song, W., and Buhain, J. (2015). Bioenergy and biofuels: history, status, and perspective. Renew-
able and Sustainable Energy Reviews 42: 712–725.
Gupta, V.K., Mittal, A., Kurup, L., and Mittal, J. (2006). Adsorption of a hazardous dye, erythrosine, over
hen feathers. Journal of Colloid and Interface Science 304: 52–57.
Haddad, K., Jeguirim, M., Jerbi, B. et al. (2017). Olive mill wastewater: from a pollutant to green fuels,
agricultural, water source and biofertilizer. ACS Sustainable Chemical Engineering 5: 8988–8996.
Hajjari, M., Tabatabaei, M., Aghbashlo, M., and Ghanavati, H. (2017). A review on the prospects of sus-
tainable biodiesel production: a global scenario with an emphasis on waste-oil biodiesel utilization.
Renewable and Sustainable Energy Reviews 72: 445–464.
Hannafi, N.E.L., Boumakhla, M.A., Berrama, T., and Bendjama, Z. (2008). Elimination of phenol by
adsorption on activated carbon prepared from the peach cores: modelling and optimization. Desalination
223: 264–268.
Hansen, U.E. and Nygaard, I. (2014). Sustainable energy transitions in emerging geconomies: the formation
of a palm oil biomass waste-to-energy niche in Malaysia 1990–2011. Energy Policy 66: 666–676.
Herrero, M., Temirzoda, T.N., Segura-Carretero, A. et al. (2011). New possibilities for the valorization of
olive oil by-products. Journal of Chromatography A 1218: 7511–7520.
Iervolino, G., Casson-Moreno, V., Tugnoli, A., and Cozzani, V. (2018). Carbon balance of a waste biomass
supply chain: the integration of a pyrolysis-based valorization process. Chemical Engineering Transac-
tions 65: 2283–9216.
International Institute for Sustainable Development (2014). Standards and Green Economy.
https://www.iisd.org/library/state-sustainability-initiatives-review-2014-standards-and-green-economy
Aroundthe world, vegetable oil, Chapter 11 (accessed 1 February 2019).
Jessup, R.W. (2009). Development and status of dedicated energy crops in the United States. In Vitro Cel-
lular & Developmental Biology – Plant 45: 282–290.
John, I., Muthukumar, K., and Arunagiri, A. (2017). A review on the potential of citrus waste for
D-limonene, pectin, and bioethanol production. International Journal of Green Energy 14: 599–612.
Joshi, S.M., Waghmare, J.S., Sonawane, K.D., and Waghmare, S.R. (2015). Bio-ethanol and bio-butanol
production from orange peel waste. Biofuels 6: 55–61.
Kanjilala, T., Babub, S., Biswasa, K. et al. (2014). Application of mango seed integuments as bio-adsorbent
in lead removal from industrial effluent. Desalination and Water Treatment 56: 984–996.
Karimnezhad, L., Haghighi, M., and Fatehifar, E. (2014). Adsorption of benzene and toluene from waste
gas using activated carbon activated by ZnCl2 . Frontiers of Environmental Science & Engineering 8:
835–844.
Karp, A. and Halford, N. (2010). Energy crops: introduction. In: Issues in Environmental Science and Tech-
nology, RSC Energy and Environment Series, No. 3. https://doi.org/10.1039/9781849732048-00001.
KFE (2017). Key figures on Europe, 2017 edition. KS-El-17-001-EN-N.pdf, p. 102. https://ec.europa.eu/
eurostat/web/products-statistical-books/-/KS-EI-17-001 (accessed 24 May 2018).
Kim, S. and Dale, B.E. (2004). Global potential bioethanol production from wasted crops and crop residues.
Biomass Bioenergy 26: 361–375.
Kurnia, J.C., Jangam, S.V., Akhtar, S. et al. (2016). Advances in biofuel production from oil palm and palm
oil processing wastes: a review. Biofuel Research Journal 9: 332–346.
Lanfranchi, M. (2012). Economic analysis on the enhancement of citrus waste for energy production. The
Journal of Essential Oil Research 24: 583–591.
44 Process Systems Engineering for Biofuels Development

Lee, X.J., Lee, L.Y., Gan, S. et al. (2017). Biochar potential evaluation of palm oil wastes through slow
pyrolysis: thermochemical characterization and pyrolytic kinetic studies. Bioresource Technology 236:
155–163.
Li, H., Hu, J., Wang, C., and Wang, X. (2017). Removal of amoxicillin in aqueous solution by a novel
chicken feather carbon: kinetic and equilibrium studies. Water Air and Soil Pollution 228 Article number:
201, DOI:https://doi.org/10.1007/s11270-017-3385-6.
Ludueña, L., Fasce, D., Alvarez, V.A., and Stefani, P.M. (2011). Nanocellulose from rice husk following
alkaline treatment to remove silica. Bioresources 6: 1440–1453.
Mantziaris, S., Iliopoulos, C., Theodorakopoulou, I., and Petropoulou, E. (2017). Perennial energy crops vs.
durum wheat in low input lands: economic analysis of a Greek case study. Renewable and Sustainable
Energy Reviews 80: 789–800.
Massocatto, C.L., Andrade, M., Honorato, A.C. et al. (2015). Biosorption of Pb2+ , Cr3+ , and Cu2+ by
peach palm sheath modified colonized by Agaricus Blazei. Desalination and Water Treatment 57:
19927–19938.
Memon, G.Z., Bhanger, M.I., Memon, J.R., and Akhtar, M. (2009). Adsorption of methyl parathion from
aqueous solutions using mango kernels: equilibrium, kinetic and thermodynamic studies. Bioremediation
Journal 13: 102–106.
Mendes, J.A.S., Prozil, S.O., Evtuguin, D.V., and Cruz-Lopes, L.P. (2013). Towards comprehensive utiliza-
tion of winemaking residues: characterization of grape skins from red grape pomaces of variety Touriga
Nacional. Industrial Crops and Products 43: 25–32.
Mhiri, N., Ioannou, I., Ghoul, M., and Mihoubi Boudhrioua, N. (2017). Phytochemical characteristics of
citrus peel and effect of conventional and nonconventional processing on phenolic compounds: a review.
Food Reviews International 33: 587–619.
Miralles, N., Martınez, M., Florido, A. et al. (2008). Grape stalks waste as low cost biosorbents: an alter-
native for metal removal from aqueous solutions. Solvent Extraction and Ion Exchange 26: 261–270.
Mittal, A., Thakur, V., and Gajbe, V. (2012). Evaluation of adsorption characteristics of an anionic azo dye
Brilliant Yellow onto hen feathers in aqueous solutions. Environmental Science and Pollution Research
19: 2438–2447.
Mohammod, M., Sen, T.K., Maitra, S., and Dutta, B.K. (2011). Removal of Zn2+ from aqueous solution
using castor seed hull. Water, Air, & Soil Pollution 215: 609–620.
Moon, W.Ch., Jbara, M.H., Palaniandy, P., and Yusoff, M.S. (2017). Shrimp pond wastewater treat-
ment using pyrolyzed chicken feather as adsorbent. Proceedings of the International Conference of
Global Network for Innovative Technology and AWAM International Conference in Civil Engineering
(IGNITE-AICCE’17). AIP Conference Proceedings 1892: 040022-1–040022-8. doi: 10.1063/1.5005702.
Murthy, P.S. and Naidu, M.M. (2012). Sustainable management of coffee industry by-products and value
addition. A review. Resources, Conservation and Recycling 66: 45–58.
Oliveira, J.F.G., Lucena, I.L., Saboya, R.M.A. et al. (2010). Biodiesel production from waste coconut oil by
esterification with ethanol: the effect of water removal by adsorption. Renewable Energy 35: 2581–2584.
Perea-Moreno, A.-J., Aguilera-Ureña, M.-J., and Manzano-Agugliaro, F. (2016). Fuel properties of avocado
stone. Fuel 186: 358–364.
Pham, V. and El-Halwagi, M. (2012). Process synthesis and optimization of biorefinery configurations.
AIChE Journal 58: 1212–1221.
Piker, A., Tabah, B., Perkas, N., and Gedanken, A. (2016). A green and low-cost room temperature biodiesel
production method from waste oil using egg shells as catalyst. Fuel 182: 34–41.
Prado, J.M., Dalmolin, I., Carareto, N.D.D. et al. (2012). Supercritical fluid extraction of grape seed: pro-
cess scale-up, extract chemical composition and economic evaluation. Journal of Food Engineering 109:
249–257.
Rachimoellah, H.M., Resti, D.A., Zibbeni, A., and Wayan, S.I. (2009). Production of biodiesel through
transesterification of avocado (Persea gratissima) seed oil using base catalyst. Jurnal Teknik Mesin 11:
85–90.
 Waste Biomass Suitable as Feedstock for Biofuels Production 45

Raga, C.A. and Lopes, D. (2003). Influence of vacuum distillation parameters on the chemical composition
of a five-fold sweet orange oil (Citrus sinensis Osbeck). Journal of Essential Oil Research 15: 408–411.
Rechberger, P. and Lötjönen, T. (2009). Energy from field energy crops –a handbook for energy producers.
Jyväskylä Innovation Oy & MTT Agrifood Research Finland.
Reynel-Avila, H.E., Bonilla-Petriciolet, A., and Rosa, G. (2015a). Analysis and modeling of multicompo-
nent sorption of heavy metals on chicken feathers using Taguchi’s experimental designs and artificial
neural networks. Desalination and Water Treatment 55: 1885–1899.
Reynel-Avila, H.E., Mendoza-Castillo, D.I., Bonilla-Petriciolet, A., and Silvestre-Albero, J. (2015b).
Assessment of naproxen adsorption on bone char in aqueous solutions using batch and fixed-bed
processes. Journal of Molecular Liquids 209: 187–195.
Reynel-Avila, H.E., Mendoza-Castillo, D.I., Olumid, A.A., and Bonilla-Petriciolet, A. (2016). A survey of
multi-component sorption models for the competitive removal of heavy metal ions using bush mango
and flamboyant biomasses. Journal of Molecular Liquids 224: 1041–1054.
Rezaee, A., Ghanizadeh, G., Behzadiyannejad, G. et al. (2009). Adsorption of endotoxin from aqueous
solution using bone char. Bulletin of Environmental Contamination and Toxicology 82: 732–737.
Rezaee, A., Rangkooy, H., Jonidi-Jafari, A., and Khavanin, A. (2013). Surface modification of bone char
for removal of formaldehyde from air. Applied Surface Science 286: 235–239.
Rivas-Cantu, R.C., Jones, K.D., and Mills, P.L. (2013). A citrus waste-based biorefinery as a source of
renewable energy: technical advances and analysis of engineering challenges. Waste Management &
Research 31: 413–420.
Rodriguez, J.M.F., de Souza, A.R.C., Krüger, R.L. et al. (2018). Kinetics, composition and antioxidant
activity of burdock (Arctium lappa) root extracts obtained with supercritical CO2 and co-solvent. Journal
of Supercritical Fluids 135: 25–33.
Rojas-Mayorga, C.K., Mendoza-Castillo, D.I., Bonilla-Petriciolet, A., and Silvestre-Albero, J. (2016). Tai-
loring the adsorption behavior of bone char for heavy metal removal from aqueous solution. Adsorption
Science & Technology 34: 368–387.
Romero, I., Ruiz, I., Castro, E., and Moya, M. (2010). Acid hydrolysis of olive tree biomass. Chemical
Engineering Research and Design 88: 633–640.
Saini, J.K., Saini, R., and Tewari, L. (2015). Lignocellulosic agriculture wastes as biomass feedstocks for
second-generation bioethanol production: concepts and recent developments. Biotech 5: 337–353.
Salvado, A.P.A., Campanholi, L.B., Fonseca, J.M. et al. (2012). Lead(II) adsorption by peach palm waste.
Desalination and Water Treatment 48: 335–343.
Sanahuja-Parejo, O., Veses, A., Navarro, M.V. et al. (2019). Drop-in biofuels from the co-pyrolysis of
grape seeds and polystyrene. Chemical Engineering Journal 377: 120246. https://doi.org/10.1016/j.cej
.2018.10.183.
Sánchez, C. (2009). Lignocellulosic residues: biodegradation and bioconversion by fungi. Biotechnology
Advances 27: 185–194.
Sarkar, N., Ghosh, S.K., Bannerjee, S., and Aikat, K. (2012). Bioethanol production from agricultural
wastes: an overview. Renewable Energy 37: 19–27.
Sengupta, D. and Pike, R.W. (2012). Chemicals from Biomass: Integrating Bioprocesses into Chemical
Production Complexes for Sustainable Development. CRC Press.
Senthil Kumar, P., Ramalingam, S., Senthamarai, C. et al. (2010). Adsorption of dye from aqueous solu-
tion by cashew nut shell: studies on equilibrium isotherm, kinetics and thermodynamics of interactions.
Desalination 261: 52–60.
Senthil Kumar, P., Ramalingam, S., Abhinaya, R.V. et al. (2011a). Lead(II) adsorption onto sulphuric acid
treated cashew nut shell. Separation Science and Technology 46: 2436–2449.
Senthil Kumar, P., Ramalingam, S., Dinesh Kirupha, S. et al. (2011b). Adsorption behavior of nickel(II)
onto cashew nut shell: equilibrium, thermodynamics, kinetics, mechanism and process design. Chemical
Engineering Journal 167: 122–131.
46 Process Systems Engineering for Biofuels Development

Senthil Kumar, P., Ramalingam, S., Sathyaselvabala, V. et al. (2011c). Removal of copper(II) ions from
aqueous solution by adsorption using cashew nut shell. Desalination 266: 63–71.
Sequeiros, A. and Labidi, J. (2017). Characterization and determination of the S/G ratio via Py-GC/MS of
agricultural and industrial residues. Industrial Crops and Products 97: 469–476.
Shariff, A., Aziz, N.S.M., Ismail, N.I., and Abdullah, N. (2016). Corn cob as a potential feedstock for slow
pyrolysis of biomass. Journal of Physical Science 27: 123–137.
Sharma, S.K., Bansal, S., Mangal, M. et al. (2016). Utilization of food processing by-products as dietary,
functional, and novel fiber: a review. Critical Reviews in Food Science and Nutrition 56: 1647–1661.
Sheinbaum, C., Balam, M.V., Robles, G. et al. (2015). Biodiesel from waste cooking oil in Mexico City.
Waste Management & Research 33: 730–739.
Shoukat, S., Bhatti, H.N., Iqbal, M., and Noreen, S. (2016). Mango stone biocomposite preparation and
application for crystal violet adsorption: a mechanistic study. Microporous and Mesoporous Materials
239: 180–189.
Siles-López, J.A., Li, Q., and Thompson, I.P. (2010). Biorefinery of waste orange peel. Critical Reviews in
Biotechnology 30: 63–69.
Silva, L.N., Cardoso, C.C., and Pasa, V.M.D. (2016). Production of cold-flow quality biodiesel from
high-acidity on-edible oils-esterification and transesterification of Macauba (Acrocomia aculeata) oil
using various alcohols. Bioenergy Research 9: 864–873.
Sindhu, R., Binod, P., and Pandey, A. (2016). Biological pretreatment of lignocellulosic biomass – an
overview. Bioresource Technology 199: 76–82.
Soegiantoro, G.H., Chang, J., Rahmawati, P. et al. (2019). Home-made eco green biodiesel from chicken
fat (CIAT) and waste cooking oil (PAIL). Energy Procedia 158: 1105–1109.
Souza, G.K., Scheufele, F.B., Pasa, T.L.B. et al. (2016). Synthesis of ethylesters from crude macauba oil
(Acrocomia aculeata) for biodiesel production. Fuel 165: 360–366.
Stafford, W., Lotter, A., Brent, A., and von Maltitz, G. (2017). Biofuels technology: a look forward.
WIDER working paper 2017/87. United Nations University World Institute for Development Economics
Research.
Stephen, J.L. and Periyasamy, B. (2018). Innovative developments in biofuels production from organic
waste materials: a review. Fuel 214: 623–633.
Sukiran, M.A., Abnisa, F., Ashri, W.M. et al. (2017). A review of torrefaction of oil palm solid wastes for
biofuel production. Energy Conversion and Management 149: 101–120.
Sulaiman, S., Aziz, A.R.A., and Aroua, M.K. (2013). Reactive extraction of solid coconut waste to produce
biodiesel. Journal of the Taiwan Institute of Chemical Engineers 44: 233–238.
Talebnia, F., Karakashev, D., and Angelidaki, I. (2010). Production of bioethanol from wheat straw: an
overview on pretreatment, hydrolysis and fermentation. Bioresource Technology 101: 4744–4753.
Talha, N.S. and Sulaiman, S. (2018). In situ transesterification of solid coconut waste in a packed bed reactor
with CaO/PVA catalyst. Waste Management 78: 929–937.
Talhaoui, N., Gómez-Caravaca, A.M., León, L. et al. (2014). Determination of phenolic compounds of
‘Sikitita’ olive leaves by HPLC-DAD-TOF-MS. comparison with its parents ‘Arbequina’ and ‘Picual’
olive leaves. LWT - Food Science and Technology 58: 28–34.
Talhaoui, N., Taamalli, A., Gómez-Caravaca, A.M. et al. (2015). Phenolic compounds in olive leaves: ana-
lytical determination, biotic and abiotic influence, and health benefits. Food Research International 77:
92–108.
Thushari, I. and Babel, S. (2018). Short communication: sustainable utilization of waste palm oil and
sulfonated carbon catalyst derived from coconut meal residue for biodiesel production. Bioresource Tech-
nology 248: 199–203.
Thushari, I., Babel, S., and Samart, C. (2019). Biodiesel production in an autoclave reactor using waste
palm oil and coconut coir husk derived catalyst. Renewable Energy 134: 125–134.
 Waste Biomass Suitable as Feedstock for Biofuels Production 47

Tian, X., Sui, S., Huang, J. et al. (2014). Neuroprotective effects of Arctium lappa L. roots against
glutamate-induced oxidative stress by inhibiting phosphorylation of p38, JNK and ERK 1/2 MAPKs in
PC12 cells. Environmental Toxicology and Pharmacology 38: 189–198.
Trevino-Cordero, H., Juárez-Aguilar, L.G., Mendoza-Castillo, D.I. et al. (2013). Synthesis and adsorption
properties of activated carbons from biomass of Prunus domestica and Jacaranda mimosifolia for the
removal of heavy metals and dyes from water. Industrial Crops and Products 42: 315–323.
Uddin, M.T., Rahman Md.A., Rukanuzzaman Md., and Islam Md.A. (2017). A potential low cost adsorbent
for the removal of cationic dyes from aqueous solutions. Applied Water Science 7: 2831–2842.
Union of Concerned Scientists (2012). The Promise of Biomass Clean Power and Fuel if Handled Right.
http://www.ucsusa.org/biomassresources (accessed 5 January 2019).
Vargas, A.M.M., Cazetta, A.L., Garcia, C.A. et al. (2011). Preparation and characterization of activated car-
bon from a new raw lignocellulosic material: flamboyant (Delonix regia) pods. Journal of Environmental
Management 92: 178–184.
Velazquez-Martı, B., Fernandez-Gonzale, E., Lopez-Cortes, I., and Salazar-Hernandez, D.M. (2011). Quan-
tification of the residual biomass obtained from pruning of trees in Mediterranean olive groves. Biomass
and bio energy 35: 3208–3217.
Vieira, S.S., Magriotis, Z.M., Santos, N.A.V. et al. (2012). Macauba palm (Acrocomia aculeata) cake from
biodiesel processing: an efficient and low cost substrate for the adsorption of dyes. Chemical Engineering
Journal 183: 152–161.
Xu, R., Ferrante, L., Briens, C., and Berruti, F. (2009). Flash pyrolysis of grape residues into biofuel in a
bubbling fluid bed. Journal of Analytical and Applied Pyrolysis 86: 58–65.
Yammine, S., Brianceau, S., Manteau, S. et al. (2018). Extraction and purification of high added
value compounds from by-products of the winemaking chain using alternative/nonconventional
processes/technologies. Critical Reviews in Food Science and Nutrition 58: 1375–1390. https://doi.org/
10.1080/10408398.2016.1259982.
Yang, Z., Koh, S.K., Ng, W.C. et al. (2016). Potential application of gasification to recycle food waste and
rehabilitate acidic soil from secondary forests on degraded land in Southeast Asia. Journal of Environ-
mental Management 172: 40–48.
Zabed, H., Sahu, J.N., Boyce, A.N., and Faruq, G. (2016). Fuel ethanol production from lignocellulosic
biomass: an overview on feedstocks and technological approaches. Renewable and Sustainable Energy
Reviews 66: 751–774.
Zhang, N., Hoadley, A., Patel, J. et al. (2017). Sustainable options for the utilization of solid residues from
wine production. Waste Management 60: 173–183.
 3
Multiscale Analysis for the
Exploitation of Bioresources: From
Reactor Design to Supply Chain
Analysis
Antonio Sánchez, Borja Hernández, and Mariano Martín
Department of Chemical Engineering, University of Salamanca, 37008, Salamanca, Spain

3.1 Introduction
Biomass as a resource has been around for centuries. Before 1800 the world ran on biomass.
The easy access and use of liquid fossil fuels changed the energy and transportation sector
(ExxonMobil 2013). However, in our attempt towards sustainable development, the use of
renewable resources and biomass in particular has gained support over the last decades.
The widespread use of biomass as a source of chemicals, fuels, and energy presents a new
paradigm. When biomass was the primary energy source, the use was mainly local. How-
ever, the global economy shows the advantages and challenges that biomass as a resource
possess, constituting an entire research field. The low density of biomass, the need to sub-
stitute a large amount of goods and the complex transformation processes from biomass to
fuels and power have determined the organization of this chapter. Lately, in a perspective
paper, Floudas et al. (2016) presented the need to evaluate systems at a multiscale level. This
need is inherent to the characteristics of the problems at hand. All the way from the molec-
ular level to the evaluation of the best use of the resources and the supply of the demand

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
50 Process Systems Engineering for Biofuels Development

involves size and temporal scales that must be analyzed. In this chapter we consider all
the stages, starting with the analysis of the individual transformations from the biomass
to make the most of its components either in the form of sugar or the elemental compo-
sition. Next, the development of optimal processes, where the low operating temperatures
result in the need to evaluate the integration of water and energy in detail. Subsequently, the
so-called process and product design problem is presented, where genetic design of crops
can play a role. Finally, at the macroscopic level, supply chain studies are presented. The
low density and even availability of biomass across the territories are the major variables.
Therefore, in this chapter we present an overview of the capabilities and possibilities that
process system engineering (PSE) has developed to address all these issues aiming at the
design of biorefineries and an entire biomass-based infrastructure.

3.2 Unit Level

The first stage for the process design is the modeling unit. The modeling of the process is
key for different operations: from process synthesis, to control, analysis of the fluctuation
in the raw materials or in the product specifications, optimization, etc. In this chapter, a
summary of the different modeling approaches for a unit is presented. The classification is
adapted and extended from Martín and Grossmann (2012a). These equations, that describe
the system behavior, can be written in an equations-based commercial software such as
GAMS©, gPROMS©, or MATLAB©. Some of these equations are already implemented
in the well-known modular commercial simulators such as ASPEN plus© or CHEMCAD©
(Martín 2014). Tian et al. (2018) provide a list of various software tools suitable for different
applications in the field of process intensification (PI) but extendable to chemical processes
in general.

3.2.1 Short Cut Methods

This modeling approach is the simpler to represent physical systems. It is based on mass
and energy balances. Despite being a simple representation of the problem, it is used in
a wide range of applications; for example, heat exchangers or splitters. One of the main
drawbacks of the short cut methods is the difficulty to represent a complex system with this
approach, therefore, the mass and energy balances are complemented with other equations
to represent the performance of the desired unit as shown later. In heat exchangers, a simple
model can be presented as (Eqs. (3.1) and ((3.2))):

ṅ in
i = ṅ i
out
∀i (3.1)
∑ ∑
Q̇ unit = ṅ out out
i Hi − ṅ in in
i Hi (3.2)
i i

where ṅ i is the inlet/outlet molar/mass flow in the heat exchanger for each component i and
Hi is the specific enthalpy per component. To calculate this value, it is necessary to take
into account the different contributions to the total enthalpy: enthalpy of formation, latent
heat, sensible heat or mixing and solution heat (Felder et al. 2016).
 Multiscale Analysis for the Exploitation of Bioresources 51

Another example is a splitter unit (Eqs. (3.3) and (3.4)).

ṅ ij = 𝜂ji • ṅ iin ∀i, j (3.3)

∑∑ ∑
Q̇ unit = ṅ ij Hji − ṅ iin Hin
i
(3.4)
j i i

where i is the set of components and j the number of output stream. 𝜂ji is the yield of each
component for the different output streams (Biegler et al. 1997).
In the biomass field, the use of mass and energy balances determine, for instance, the
performance of the combustion of the biomass. A general biomass combustion reaction is
as follows (Van Loo and Koppejan 2008; Zainal et al. 2001):
( )
1 − yO2
Ca Hb Oc Nd Se + n N2 + O2
yO2

→ mCO2 + nCO + oCf Hg + pH 2 O + qO2 + rN 2 + sNO + tNO2 + uN 2 O + vSO2

(3.5)

The mass and energy balances determine the final composition of the outlet gases and
the final temperature (Eqs. (3.6)–(3.12)).

C ∶ a = m + n + o•f (3.6)

H ∶ b = o • f + 2p (3.7)

O ∶ c + 2n = 2m + n + p + 2q + s + 2t + u + 2v (3.8)
( )
1 − yO2
N ∶ d+2 n = 2r + s + t + 2u (3.9)
yO2
S∶ e=v (3.10)
∑ ∑
fi hi − fi hi = 0 (3.11)
i∈IN i∈OUT
T
hi = h0f ,i + Cp,i (T)dT (3.12)
∫Tref

where fi is the flow for each component (i) and hi is the enthalpy per component. This
enthalpy is the sum of two terms: the enthalpy of formation (h0f ,i ) and the sensible enthalpy.

3.2.2 Mechanistic Models

These types of models use a first principles approach in more detailed compared with the
previous one. They are based on the underlying chemistry, physics, or biology that governs
the behavior of the system (Rasmuson et al. 2014). Several examples are discussed in the
following.
52 Process Systems Engineering for Biofuels Development

3.2.2.1 Equilibrium Models

Some reactions are limited by thermodynamic equilibrium. If the equilibrium conditions are
reached inside the reactor, the behavior of the unit can be modeled using such conditions.
This approach has been used to evaluate the performance of a biomass gasifier (Zainal
et al. 2001; Sharma and Sheth 2016; Aydin et al. 2017). Some of them combine the thermo-
dynamic modeling with some empirical correlations or parameters. The schematic general
reaction in gasification processes can be written as follows (Eq. (3.13)):

CH a Ob + wH 2 O + mO2 + 3.76mN 2 → x1 H2 + x2 CO + x3 CO2 + x4 H2 O + x5 CH 4

+ x6 N2 + x7 C (3.13)

In the outlet gases, three main equilibrium reactions take place (Sharma and Sheth 2016)
(Eqs .(3.14)–(3.19)).

PCH 4
C + 2H2 ⇌ CH 4 → K1 = (3.14)
P2H
2

7082.848 7.466 • 10−3 2.164 • 10−6 2

ln(K1 ) = − 6.567 ln(T) + T− T
T 2 6
0.701 • 10−5
+ + 32.541 (3.15)
2T 2
PCO2 PH2
CO + H2 O ⇌ CO2 + H2 → K2 = (3.16)
PCO PH2 O
5870.53 58200
ln(K2 ) = + 1.86 ln(T) − 2.7 • 10−4 T − − 18.007 (3.17)
T T2
P3H PCO
CH 4 + H2 O ⇌ CO + 3H2 → K3 = 2
(3.18)
PCH 4 PH2 O
−22784.85
ln(K3 ) = + 7.951 ln(T) − 4.354 • 10−3 T + 3.6067 • 10−7 T 2
T
4850
+ 2 − 24.899 (3.19)
T
Another example of an equilibrium model is hydrocarbon reforming. In general, the
reforming of hydrocarbons or biogas is employed when it is necessary to produce hydro-
gen, mainly, for synthesis purposes including methanol or ammonia production. In steam
methane reforming or autothermal reforming, two main equilibrium reactions are consid-
ered: the decomposition of methane and the water-gas shift reaction (Aasberg-Petersen
et al. 2013). However, over the last years, other kinds of reforming are attracting atten-
tion, for example, dry reforming (Hernández and Martín 2016), tri-reforming (Hernández
and Martin 2018), or super-dry reforming (Buelens et al. 2016). The target in these new
alternatives is to reduce the greenhouse gas emissions (mainly CO2 ) associated with the
methane reforming and to become more efficient. All these reactions can be modeled as
equilibrium systems.
 Multiscale Analysis for the Exploitation of Bioresources 53

3.2.2.2 Kinetic Models

In many cases equilibrium conditions are not achieved because a very large residence time
in the reaction is required. Kinetic models are widely used to describe the chemical reac-
tor operation. For example, Sánchez and Martín (2018a) used this approach to model a
multibed ammonia reactor as shown in Figure 3.1.
The mass balance and kinetic expression is given by (Eqs. (3.20) and (3.21)):
dX rA
= 0t (3.20)
dz FH
( 3 )𝛼 2 ( 2 )1−𝛼
⎡ aH aNH ⎤
r = 3 kreac ⎢Kp2 aN 2
− 3 ⎥ ΦΩ (3.21)
⎢ 2
a2NH a3H ⎥
⎣ 3 2 ⎦
And heat balances are presented, such as the heat balance to the reaction bed
(Eq. (3.22)), as:
[ ]
dT 1 2 dX dT dT
= |ΔHr |FH0 − ṁ bed1 Cp ext − ṁ bed1 Cp central (3.22)
dz ̇ p 3
mC 2 dz dz dz

Main Inlet Stream

Secondary Inlet
Stream
Catalytic Beds

Secondary Inlet
Stream

Heat Exchanger

Outlet Stream

Figure 3.1 Scheme of the multibed ammonia reactor.

54 Process Systems Engineering for Biofuels Development

Those corresponding to the heat balances to the annulus and central section (Eqs. (3.23)
and (3.24)) are:
dT central UL(T − Tcentral )
=− (3.23)
dz ṁ bed1 Cp
′
dT ext k L (T − Text )
= + isolation (3.24)
dz ṁ bed1 Cp 𝛿

And finally, the pressure drop (Eq. (3.25)) is given by:

(1 − 𝜀)2 𝜇vgas (1 − 𝜀) 𝜌vgas

2
dP
= −150 − 1.75 (3.25)
dz 𝜀 3 2
dp 𝜀3 dp

In the nomenclature at the end of the chapter, the meanings of the different variables
and parameters are given. Other examples can be found in modeling second-generation
bioethanol production (Martín 2016a; Martín et al. 2018) or biogas production from anaer-
obic digestion of residues (Taifouris and Martín 2018).
The kinetic models involve a set of differential and algebraic equations (DAE). For
the optimization purposes, this DAE system is transformed in a large nonlinear program
through discretization. One of the most used discretization methodologies is the orthogo-
nal collocation method (Biegler 2010). This approach has been applied in different reactor
systems (Flores-Tlacuahuac et al. 2008) to optimize the synthesis of biodiesel with immo-
bilized enzymes (Andrade et al. 2019) or to model a glycerol etherification reactor (Martín
and Grossmann 2014a).

3.2.2.3 Phase Equilibrium

Some of the most important operations in the chemical industry involve the presence of
different phases; for example, distillation columns or adsorption processes. Distillation
has multiple applications in biomass processes: to bioethanol dehydration (Martín and
Grossmann 2012b), in the production of biobutanol through acetone butanol fermentation
(Malmierca et al. 2017) or in dimethyl ether (DME) synthesis (Peral and Martín 2015; Haro
et al. 2013).
For a single equilibrium stage, a flash unit, the following equations (Eqs. (3.26)–(3.28))
determine the operation of the system (Biegler et al. 1997):

F =L+V (3.26)

Fzi = Lxi + Vyi (3.27)

yi = Ki xi (3.28)

where Ki is the equilibrium constant. One of the biggest challenges in this approach is to
select the thermodynamic system to calculate this equilibrium constant. In the ideal case, it
can be expressed as follows (Eq. (3.29)):
Pivap
Ki = (3.29)
Ptotal
 Multiscale Analysis for the Exploitation of Bioresources 55

This last equation is a combination of Dalton’s and Raoult’s laws. The vapor pressure
(Pivap ) can be obtained from the Antoine equations for each component.
To model a multicomponent distillation column, two kinds of models have been pro-
posed: rigorous and short-cut methods. The first ones are based on tray by tray calcula-
tions. In each tray, the compositions, flows, temperatures, etc. are computed using mass
and energy balances and equilibrium relationships (Viswanathan and Grossmann 1990).
Short-cut methods establish a simple approach to design distillation columns. One of the
most popular is the Fenske–Underwood–Gilliland method (Geankoplis 2006).
Other phase equilibrium systems have also been modeled in the literature; for example,
cooling water towers (Martín and Martín 2017), adsorption beds (Raghavan and Ruthven
1985; Agarwal et al. 2009; Sánchez and Martín 2018b), or membrane systems (Lima et al.
2012; Zarca et al. 2018).

3.2.2.4 Rigorous Models

The capabilities of computational resources during the last two decades have allowed the
use of more rigorous models such as computational fluid dynamics (CFD). In this approach,
the continuity, momentum and energy equations are defined by means of partial differential
equations that are solved for each of the finite volumes in which the system is discretized.
The complexity and time-consuming numerical methods involved in the computation of
the finite volumes has promoted the use of commercial tools that have implemented these
numerical methods such as ANSYS Fluent®, COMSOL Multiphysics®, or OpenFOAM®
(Martín 2014). Their use simplifies the application of CFD to complex systems like most of
the chemical and mechanical units. Regarding their application to biomass processing, the
use of CFD has mainly focused on studying the growth of algae cultures, the fermentation
of biomass and wastes or the processing of sequential products such as biodiesel or syngas.
In the area of algae growth, different systems have been studied such as panels, tubular
airlift photobiorectors or race-ways systems, where the unit composed by microalgaes has
been modeled by means of a CFD discrete phase model approach (Bitog et al. 2011; Ye
et al. 2018; Perner-Nochta and Posten 2007; Amini et al. 2018). The biochemical treatment
of biomass has also been studied with CFD techniques. Numerous papers have focused
on the study of anaerobic digestion for biomass or waste (Vesvikar and Al-Dahhan 2004;
Wu and Chen 2007) hydrolysis (Carvajal et al. 2012) or fermentation (Haringa et al. 2018;
Wright et al. 2018; Madhania et al. 2019), or the hydrodynamics of distillation columns
(May Vázquez et al. 2017). Finally, combustor modeling (Abagnale et al. 2016) as well as
the transport phenomena of the oil transesterification under supercritical conditions have
also been addressed (Mekala et al. 2019).
However, the modeling using CFD techniques is mainly limited to fluid and dilute mul-
tiphase systems. In order to model systems involving biomass in the form of grains or
particles, the discrete element method (DEM) is applied. DEM is a particle scale method
that allows modeling the particle mechanics with further resolution but with consumption
of higher computational resources and with limitations at large scales. Despite DEM being
a recent technology, it has been already applied to study operations involved in biomass
processing such as milling and blending (Zhu et al. 2007, 2008). The use of these DEM
models is not exclusively limited to systems composed by particles. The combination with
CFD has helped to model the dynamics of two-phase flow systems with high concentration
56 Process Systems Engineering for Biofuels Development

of particles; for example, for modeling a rotatory dryer (Scherer et al. 2016) of biomass or
fluidized beds such as the ones applied for pyrolysis or gasification (Xiong et al. 2017; Ku
et al. 2015).

3.2.3 Rules of Thumb

They provide a simple approach to model the units. They are based on operational data
from industry that, in general, are scarce. Therefore, they are limited to the common range
of operation. For chemical processes, some books collect this information for multiple units
as heat exchangers, distillation columns, reactors, etc. (Couper et al. 2005; Branan 2005). A
large percentage of these rules can be applied to biochemical processes. For more specific
and new processes, they are still under development. For example, carbon dioxide capture
and storage (CCS) technologies can be used within a thermal biorefinery to reduce the CO2
emissions from the biomass. One of these technologies is the Rectisol process based on
physical absorption. In the design of the absorption column, the rules of thumb state that to
achieve 97% of CO2 removal, 14–18 theoretical equilibrium stages are required (Sadhukhan
et al. 2014). Another example of rules of thumb is to calculate the heat and power necessary
in the monoethanolamine (MEA) process also for CO2 capture (GPSA 2004).

3.2.4 Dimensionless Analysis

Dimensionless analysis provides the tools to handle a lot of experimental data in a simple
way. It establishes what dimensionless groups determine the performance of the system.
Then, using experimental data, it is possible to create correlations between them (Munson
et al. 2009). The scale up/down issues are captured in these sorts of models. A key point
to obtain an accurate model is the selection of the variables involved. A large number of
applications in fluid mechanics and heat or mass transfer uses these kinds of models. For
example, in forced convention heat transfer inside tubes, with a fully developed turbulent
flow, the following correlation (Eq. (3.30)) holds (Dittus–Boelter correlation):
n
Nu = 0.023Re0.8 Pr (3.30)

where exponent n takes the value of 0.3 for heating of the fluid and 0.4 for cooling. Another
example is the oxygen mass transfer from growing bubbles that takes place in key appli-
cations in biochemical processes. Martín et al. (2007) proposed the following correlation
(Eq. (3.31)) between the Sherwood (Sh) number and the Reynolds (Re) and Schmidt (Sc)
numbers.
Sh = 10Re0.24 Sc0.32 (3.31)

3.2.5 Surrogate Models

In some cases, the models developed to simulate the behavior of some units can be too
complex to be used in other applications, for example, optimization or control. For instance,
a CFD model. In these situations, the development of surrogate models has been proposed as
an effective alternative. The surrogate technique creates simple models using the data from
 Multiscale Analysis for the Exploitation of Bioresources 57

the rigorous ones. A good accuracy between the model data and the surrogate is necessary.
Sampling to select the appropriate point is key to obtain a representative model. To do
this, the methodology from the design of experiments (DOE) is widely applied. Queipo
et al. (2005) defined the four stages to build a surrogate model from a rigorous one: DOE,
numerical simulations at sampling points for the previous step, construction of a surrogate
model, and model validation.
DOE uses different techniques to distribute the points uniformly through the space in
order to reduce the bias error. The full factorial designs, generally, are very computationally
expensive for a large majority of models. Therefore, fractional factorial designs are used.
Latin hypercube sampling is a stratified method where every portion has a representative
value. This method shows problems with the uniformity of the samples (Queipo et al. 2005).
The orthogonal array (OA) is a method where the columns (samples) are orthogonal to
each other (Arora 2017). This design depends on four variables: the number of samples, the
number of variables, the levels of the variables and the number of effects to be accounted for
in the model (Queipo et al. 2005). The Latin hypercube is a particular case of the OA. One
of the problems involved in this method is the possibility of creating duplicate points. This
is not desirable in surrogate models from rigorous simulations. A method that uses random
sampling is the Monte Carlo simulation. In this technique, with the random sampling and
the statistical inference tools, it is possible to extract information about the system (Lemieux
2009). Finally, the maxmin approach maximizes the minimum distance between two points
(Quirante et al. 2015).
To build the surrogate model using the data from the previous DOE, different techniques
have been proposed. One of them is the use of polynomial regression models. This method
models the relation between the variables using a polynomial (Montgomery et al. 2012).
For example, a second-order polynomial for regression models can be expressed as follows
(Eq. (3.32)):
∑n
∑
n
∑n
f (x) = 𝛽0 + 𝛽i xi + 𝛽ij xi xj (3.32)
i=1 i=1 j≤i

3.2.5.1 Kriging Models

Kriging models estimate the relationship between variables as a sum of two terms
(Eq. (3.33)): the first one is a linear model (for example a polynomial) and the second one
represents the fluctuations on the mean of the data set y(x) (Quirante et al. 2015).

y(x) = f (x) + Z(x) (3.33)

Z(x) is a stochastic Gaussian process with a mean value of 0 and a covariance given by
(Eq. (3.34)):
cov(Z(w), Z(x)) = 𝜎 2 R(w, x) (3.34)

where 𝜎 2 is the process variance and R(w, x)is a correlation function. One of the most
popular structures for this correlation takes the following form (Eq. (3.35)):
( n )
∑
R(w, x) = exp − 𝜃j (wj − xj ) Pj
(3.35)
j=1
58 Process Systems Engineering for Biofuels Development

i–1

1 1
x1 Oi

x2 w1 Oi
h f(x)
2 Σ 2
w2
xn
Oi
wn
i
n n

i+1

Previous layer Current layer Next layer

Figure 3.2 Scheme of a neural unit.

where 𝜃 j ≥ 0 and 0 ≤ Pj ≤ 2 are adjustable parameters (Sacks et al. 1989). The ordinary
Kriging model employs as f(x) a constant value with successful results over a widespread
range (Kleijnen 2009).
Another example of these methods to build a surrogate model is the radial basis function
(RBF). In this case, the model (Eq. (3.36)) is a linear combination of a set of radially sym-
metric functions (Queipo et al. 2005):
∑
n
f (x) = wi hi (x) (3.36)
i=1

where wi are the weights in the linear combination and hi are the RBFs. A typical radial
function is (Eq. (3.37)): ( )
(x − c)2
hi (x) = exp − (3.37)
𝛿2
In this equation, c is the center and 𝛿 is the radius of the function. Both are adjustable
parameters in the model. A generalization of the RBF model is the Kernel regression model.
The last method described in this summary is the artificial neural network (ANN). In this
methodology, the model tries to imitate the behavior of human neural networks in the brain.
A scheme of the basic structure of a neuron is shown in Figure 3.2 (Himmelblau 2000).
The input signal (xi ) from each previous neuron is weighted using different weight value
(wi ). After the summation, the signal is sent to a transfer function f(x). This function, typi-
cally, takes the following form (Eq. (3.38)):
1
f (x) = (3.38)
1 + eh
A collection of neuron nodes connected between them constitutes the neural network.
The neurons are organized in layers. To establish the value of the weight for each neuron
node, a training of the neural network is carried out using a collection of data (training
set).
 Multiscale Analysis for the Exploitation of Bioresources 59

Some applications of surrogate models in the biorefinery field are presented here. One
of the most interesting fields for the neural network is control systems. For example, Gad-
kar et al. (2005) studied the control in a fed batch yeast fermentation to produce ethanol.
The only measured variable is the oxygen concentration in the bioreactor. Quirante and
Caballero (2016) simulated, using Aspen HYSYS, a sour water plant and with these data a
kriging model is created. The kriging is used to optimize the plant. Henao and Maravelias
(2011) used the ANN to create surrogate models from a commercial modular simulator and
the integration of these kinds of models within a mixed integer nonlinear problem (MINLP)
superstructure formulation.
Some software packages have implemented these techniques to automate the process of
creating a surrogate model (Bhosekar and Ierapetritou 2018). ALAMO (Automatic Learn-
ing of Algebraic Models) software creates a surrogate model using integer optimization
techniques and an adaptive sampling based on error maximization sampling (EMS). The
tool is able to avoid overfitting or using unnecessary terms in the model (Wilson and Sahini-
dis 2017). This package has been developed using GAMS, where a mixed-integer quadratic
programming solver is necessary, preferably BARON, and MATLAB (Cozad et al. 2014).
A thorough overview of software tools for surrogate models is presented by Boheskar and
Ierapetritou (Bhosekar and Ierapetritou 2018).

3.2.6 Experimental Correlations

In some situations, an experimental data set is adjusted to a statistical model, for example,
the previous ones used to develop surrogate models. The results are based on the experi-
mental data and not only on theoretical models. However, it is necessary to have a good
data set, with all the variables involved in the phenomena to create a good model for the
unit. The range of applications for the correlation is limited by the range of the experimental
data. One of the most popular is the polynomial model. For instance, Hernández and Martín
(2017) developed a polynomial empirical model to predict the algae growth as a function
of the nutrient concentration (Eq. (3.39)).
( g )
growth =0.418528 • TotP + 0.52762 • TotN
m2 d
+ 0.225013 • TotN • TotP − 0.20754 • TotP2 − 0.03026 • TotN2 (3.39)

Warnes et al. (1998) applied RBF and ANNs to simulate the production of recombinant
protein using Escherichia coli fermentation. Liu et al. (2009) modeled and optimized the
production of hyaluronic acid. To model the system, a combination of RBF and ANN is
used. Besides, this approach was compared with the polynomial one for the same system.
Additional models for biomass pretreatment, Eq. (3.40) for the yield from biomass to cellu-
lose when using dilute acid pretreatment (Galán et al. 2019) or the recent work of Sánchez
et al. (2019) modeling gasifiers are examples of this too.

𝜂Cellulose = −0.00055171 + 0.00355819 ⋅ T_acid + 0.00067402 ⋅ conc_acid_mix

+ time_pret ⋅ 0.00100531 − enzyme_add ⋅ 0.0394809
− 0.0186704 ⋅ T_acid ⋅ conc_acid_mix + 0.00043556 ⋅ T_acid ⋅ time_pret
60 Process Systems Engineering for Biofuels Development

+ 0.0002265 ⋅ T_acid ⋅ enzyme_add − 0.0013224 ⋅ conc_acid_mix ⋅ time_pret

− 0.00083728 ⋅ time_pret ⋅ enzyme_add
+ 0.044353 ⋅ conc_acid_mix ⋅ enzyme_add + 0.000014412 ⋅ T_acid2 ; (3.40)

3.3 Process Synthesis

There are a number of approaches that can be followed for the design of a transformation
process. We can classify them into heuristic based, where rules and know-how provide the
criteria for decision making (Douglas 1988; Smith 2005), to mathematical optimization,
where the selection among alternative technologies is formulated as a mathematical prob-
lem that is solved for the optimal value of the objective function (Biegler et al. 1997). Both
approaches are not exclusive and hybrid ones can be the way to go, so that a screening of the
technologies is carried out based on experience and the final decision is taken by solving a
mathematical problem (Guerras and Martín 2019).

3.3.1 Heuristic Based

Two different approaches have been presented in the literature, the Douglas hierarchy (Dou-
glas 1988) and the onion model (Smith 2005). In both cases a procedure is presented to
structure a flowsheet following a series of rules. The Douglas hierarchy consists of five
steps. The method first evaluates the operation of the process, whether continuous or batch.
Next, it makes decisions on the input-output structure evaluating the feeds and the prod-
ucts. Subsequently, the recycle structure is targeted, to consider the reactor operation and its
conversion. Then the separation structure is addressed, where the purification of the reac-
tor exit stream and the recovery of the products is determined as a function of the state
of aggregation and the thermodynamics of the species involved. At the last stage heat, but
lately also water, is integrated to reduce the consumption. The onion scheme presented by
Smith (2005) has its starting point in the reactor, as the heart of the transformation process.
From that point on several layers are evaluated in the following order. First the separa-
tion and recycle system, based on the exit stream from the reactor and its conversion. Next
the heat recovery system, aiming at reducing the consumption of utilities. Subsequently,
the evaluation of the heating and cooling utilities. Finally, a treatment of water and other
effluents.
Following either approach a flowsheet is developed. In most of the stages to carry out the
same operation a number of technologies can be available. Screening of the alternatives can
be carried out based on industrial experience that sometimes is gathered in design books in
the form of rules of thumb (Couper et al. 2005; Branan 2005).
Apart from flowsheeting, energy and water integration have been addressed since the
1960s. From the heuristic point of view, there are a number of approaches that have been
widely accepted in industry due to the ease of application. Among them, we can consider the
development of rules of thumb to help select the matches between streams such as the ones
provided by Masso and Rudd (1969) and the ones by Ponton and Donaldson (1974). A step
forward was carried out when a more theoretical approach was used to address the problem.
The thermodynamic approach was proposed by Hohmann and Lockhart (1976) to establish
 Multiscale Analysis for the Exploitation of Bioresources 61

the minimum utility consumption. A few years later, the most important contribution in this
area was presented, the pinch technology proposed by Linnhoff and Hindmarsh (1983), that
allows the computing of the minimum heat and cooling needs. While these methods are
widely used, the highly combinatorial problem results in difficulties when applying them,
in spite of the large savings that have been provided to the gas and oil industry over the
years. The use of pinch technology has been extended to entire complexes within the total
site integration concept so that the energy is integrated within plants for a more efficient
operation (Klemes 2013; Liew et al. 2013).
The initial works on water usage reduction were performed in industry (Carnes et al.
1973; Skylov and Stenzel 1974; Hospondarec and Thompson 1974; Mishra et al. 1975;
Anderson 1977; Sane and Atkins 1977). Unlike the energy integration, where the first works
can be classified within this approach, the seminal paper on the optimization of water use
was in fact proposed within the mathematical programing approach. However, the difficul-
ties in solving the mathematical problem led to the use of conceptual design for many years
(Takama et al. 1980). Thus, in 1989, El-Halwagi and Manousiouthakis (1989) developed a
targeting graphical method to plot the cumulative exchanged mass versus the composition
for a set of contaminated streams. Later in 1994, Wang and Smith (1994) proposed the so
called “water pinch,” based on the same principles as the “temperature pinch.” It is also an
end-of-pipe approach, but most importantly it is basically useful for one contaminant only,
and the network configuration is built based on rules of thumb. When several contaminants
are involved, a network for each contaminant is designed and after that, a merging procedure
results in the complete network. A good review can be found in Belhateche (1995).
In an attempt to deal with both water and energy simultaneously within this approach,
temperature–concentration diagrams have been used (Martínez-Patiño et al. 2011).

3.3.2 Supestructure Optimization

3.3.2.1 General Problem Formulation
Alternatively to the heuristic design procedure or as a second stage once a prescreening of
the alternatives is performed, a mathematical optimization approach can be used to build
a flowsheet selecting among technologies and operating conditions (Guerras and Martín
2019). It follows from the modeling step presented in Section 3.2. The models of the differ-
ent units are linked to build a so called superstructure of alternatives that embeds options
for the processing stages. The solution of such a superstructure yields a flowsheet, see
Figure 3.3. There is no unique solution, it depends on the decision criterion, the objective
to be optimized, whether we are looking for minimum energy consumption, environmen-
tal impact, safety, heat and water consumption, or a combination of some of the objectives.

Rules of thumb
Thermodynamics
First principles
Experimental
Machine learning

Figure 3.3 Superstructure formulation.

62 Process Systems Engineering for Biofuels Development

Examples can be found in the biofuels production industry such as the case of second gener-
ation of bioethanol since multiple paths from the raw material can be followed. It is possible
to identify the thermal path (Martín and Grossmann 2011a,2011b) or the hydrolysis path
(Martín and Grossmann 2012b). Even within each path, there are a number of alternative
technologies for most of the transformation stages. For instance, in the thermochemical
path, different gasification technologies, direct or indirect, various reforming stages, steam
reforming, partial oxidation, autoreforming, a number of technologies to remove the CO2
and to adjust the H2 to CO ratio including water-gas shift, hydrogen recovery membranes
for the first stage and chemical absorption, i.e. ethanolamines, physical adsorption, pressure
swing adsorption, or membranes for the removal or CO2 , are available. Finally, syngas can
be used to produce ethanol either via catalytic synthesis or fermentation. Depending on the
synthesis stage, the purification can be carried out either following a distillation column
sequence or similar to first generation of bioethanol using a beer column and molecular
sieves.

3.3.2.2 Energy and Water Integration

The superstructure typically considers the raw material processing stages. It is also possi-
ble to formulate within the mathematical problem of the superstructure the optimization of
the energy (Martín and Grossmann 2012b) and/or the water and energy consumption (Bal-
iban et al. 2013). Alternatively, this can be carried out in a second stage optimizing a heat
exchanger network and a water treatment network (Martín and Grossmann 2011a,2011b).

3.3.2.2.1 Heat Exchanger Networks

The high share of the operating costs that utilities and, in particular, steam has led the effort
toward developing strategies and optimization formulation to address the design problem.
Following the works discussed in Section 3.3.1, mathematical optimization formulations
were developed and improved over time. In the first case, sequential approaches were
presented such as MAGNETS, developed by Floudas et al. (1986). In this case the heat
exchanger network (HEN) design problem is decomposed in three stages. The first step
involves the solution of the transshipment model, a linear programming (LP) problem, or
the extended transshipment (mixed integer linear programming [MILP]), by Papoulias and
Grossmann (1983), to determine the minimum consumption of utilities, cooling water, and
steam. With the utility consumption fixed, a MILP model is formulated to compute the min-
imum number of stream matches and the demands associated. Finally, in the third step the
network is defined by minimizing the area cost once the heat loads and matches are fixed
by solving a nonlinear programming model (Floudas et al. 1986) to determine the optimal
network configuration. Sequential approaches present limitations related to evaluating the
trade-offs between investment and operating costs leading to suboptimal networks.
In order to overcome these limitations, in 1990, Yee and Grossmann (1990) presented
SYNHEAT. Here, a superstructure optimization approach is used accounting for all costs
(piping layout, fixed and variable heat exchanger costs, and utility cost) but still assum-
ing isothermal stream mixing. Isothermal streams, such as phase change, were included in
the model by Ponce-Ortega et al. (2008) using disjunctions. For a detailed review of heat
exchanger methodologies see Furman and Sahinidis (2002).
 Multiscale Analysis for the Exploitation of Bioresources 63

Heat integration studies have been extended to account for entire chemical complexes.
The concept of total site integration appears as presented within the heuristic rules. The
idea is to integrate energy within processes to improve the systems efficiency (Nemet et al.
2015).

3.3.2.2.2 Heat Integrated Distillation Columns and Columns Sequencing

Two different examples can be identified. Energy can be integrated within the same sepa-
rating task, or along a distillation column sequence. The first example is also known as a
multi-effect column. In this particular structure, the feed is split and fed to columns operat-
ing at different pressures so that the boiler of the lower pressure column acts as condenser
for the higher pressure column. The variables are the operating pressures of the columns as
well as the fraction of the feed that goes to each one of them. Using multi-effect columns, we
cannot only reduce the energy consumption by nearly one half (Karuppiah et al. 2008), but
also the cooling needs are also reduced (Ahmetovic et al. 2010). Alternatively, the different
boiling and dew points of mixtures and species can be an advantage so that the energy to be
removed from condensers can be further used to save utilities by providing heat to reboilers.
The problem of distillation sequences has been addressed in the literature formulating MILP
problems to select the sequence (Andrecovic and Westerberg 1985; Eliceche and Sargent
1986). The optimal column sequence model can be extended to consider the integration of
energy within the distillation columns of that sequence (Floudas and Paules IV 1988).

3.3.2.2.3 Simultaneous Optimization and Heat Integration

In process flowsheeting, flowrates and temperatures are variables that can affect the energy
recovery. Therefore, heat integration should be taken into account together with the synthe-
sis problem. In particular, when the conversion in the reactor is low and/or it is governed by
chemical equilibrium, the operating conditions determine the separation stages, the energy
and cooling requirements. In the biofuel industry, biodiesel production is an example of
this particular case. Biodiesel production is based on the transesterification of oil with an
excess of alcohol. The conversion of this reaction does not only depend on the excess of
alcohol, but also on the operating pressure and temperature, the catalyst load and compo-
sition. Furthermore, the energy consumption in the process relies on the recovery of the
excess of alcohol. Thus, the simultaneous optimization and heat integration is the appropri-
ate approach for this example. To accomplish this, the mathematical problem formulated
must include the heat integration problem together with the process flowsheet model.
Duran and Grossmann (1986) proposed a procedure for solving nonlinear optimization
and synthesis problems of chemical processes simultaneously with the minimum utility
target for heat recovery networks. It is based on the use of a pinch point location method
that can be formulated for embedding the minimum utility target within the process opti-
mization. Since no temperature intervals are required in the proposed procedure, variable
flowrates and temperatures of the process streams can be handled, and thus process models
can be treated explicitly.
In order to have an optimization-based software for process design, Kravanja and Gross-
mann (1990) made use of several concepts developed within Grossmann’s group in order
to develop PROSYM, currently renamed as MYPSYN (Kravanja 2010). This software
is an implementation of the modeling and decomposition (M/D) strategy by Kocis and
64 Process Systems Engineering for Biofuels Development

Grossmann (1989) and the outer approximation and equality relaxation (OA/ER) algorithm
by Kocis and Grossmann (1987). The main characteristic is that it enables automated exe-
cution of simultaneous topology and parameter optimization of processes including simul-
taneous optimization and heat integration.

3.3.2.3 Water Networks

It took several years before the work by Takama et al. (1980) was revisited in 1998. The
basic idea holds: a superstructure of all possible re-use and recycle opportunities is formu-
lated to come up with the optimal water network (WN). The difficulty lies on the solvabil-
ity of the MINLP formulations. The mathematical complexity arises in the bilinear terms
resulting from the product between flows and concentrations as well as concave terms in
the objective function preventing obtaining the global optimum. Furthermore, it is worth
pointing out that most of the published articles do not consider all possible interconnec-
tions, multiple sources of water of different quality, pretreatment of the water, and mass
transfer and non-mass transfer water-using operations. In the literature, the cost of water
pumping is neglected as well as the investment cost of the pipes. Furthermore, the total
WN is decomposed into two parts (network with water-using operations and wastewater
treatment network) which are solved separately.
In order to obtain a solution, over the years some simplifications have been made includ-
ing linear relaxations to obtain lower bounds (Quesada and Grossmann 1995), the use of
linear models to initialize the problems (Doyle and Smith 1997), the use of a heuristic
search based on the successive solution of a lineal relaxation (Galan and Grossmann 1998),
bounding the flows (Alva-Argáez et al. 1998a,b), the definition of the necessary optimality
conditions to eliminate the nonlinearities (Savelski and Bagajewicz 2000, 2003). Two-stage
approaches have also been used (Gunaratnam et al. 2005; Li and Chang 2007; Castro et al.
2009) as well as the development of a spatial branch and contract algorithm (Karuppiah
and Grossmann 2006). In the same year, Alva-Argáez et al. (2007) proposed a method-
ology based on the water-pinch decomposition making use of the water-pinch insights to
define successive projections in the solution space and also considered the solution of a
sequence of MILP that replaces the original problem. A similar approach was used by
Faria and Bagajewicz (2010). Ahmetovic and Grossmann (2011) proposed a superstructure
that consists of multiple sources of water, water-using processes (involving mass trans-
fer or not), water treatment technologies and all possible connections in the network. The
model includes cost of piping, which represents a concave function, the cost of water pump-
ing (a linear function of the flowrate) so that it was possible to account for the trade-offs
between processing and piping. All this work involves isothermal WNs. An extension to
account for non-isothermal WNs is to be considered within the water and energy nexus
initiative.

3.3.2.4 Water and Energy Nexus

The strong link between energy and water consumption has been the focus of many works
lately (Water in the West 2013; Tan and Zhi 2016). The systematic approach has evolved
over time to account for the industrial implications (Varbanov 2014). Sequential and simul-
taneous approaches have been presented in the literature (Ahmetovic and Kravanja 2012).
 Multiscale Analysis for the Exploitation of Bioresources 65

The sequential approach has been used in different forms, for instance, it involves the WN
synthesis by minimizing freshwater consumption followed by the synthesis of HEN for
minimum utility cost or HEN cost. Alternatively, Bogataj and Bagajewicz (2008) presented
an approach for mixing and splitting of hot and cold streams inside the HEN and combined
this network with the WN. The WN is solved as a nonlinear programming model to clas-
sify the streams and provide an initial point. The application of this to biorefineries design
makes use of the water and energy nexus link. Grossmann and Martin (2010) presented a
two-stage procedure. A process was designed for minimum energy consumption including
Duran and Grossmann’s formulation and heat exchanger design. Next, a WN was devel-
oped. This approach cannot prove global optimality.
Due to the mathematical complexities of solving the entire problem, Yang and Gross-
mann (2013) developed a targeting procedure to simultaneously account for water and
energy integration. An LP targeting model is used to address the minimum water consump-
tion for the WN with only water-using process units based on the superstructure proposed
by Ahmetovic and Grossmann (2011). Thus, in a first step, the economics of the flowsheet
as well as the cost of HEN and WN targets subject to process constraints are simultaneously
optimized. In the first step the flowsheet is defined. The second step consists of determining
the detail HEN and WN structures and corresponding capital and utility costs using the
fixed heat capacity flowrate, inlet and outlet temperatures, and water-using process unit
flowrates. The targets are not used in the network synthesis problem so that the trade-off of
the capital costs of the HEN and WN are considered now. Mathematical approaches have
also been developed integrating a WN within superstructure optimization (Baliban et al.
2013). Dong et al. (2008) presented a simultaneous approach using a modified state-space
superstructure (Bagajewicz et al. 2002) in order to incorporate wastewater treatment and
direct heat exchange. It is also possible to use a two-step strategy. The first step provides
initialization as well as good upper bounds for freshwater and utilities consumption. In
the second step, the overall heat and water netwrok problem is solved simultaneously
(Ahmetovic et al. 2014).

3.3.3 Environmental Impact Metrics

The optimization of energy and water presents a number of trade-offs difficult to resolve.
Energy is in general expensive and its production contributes to the generation of CO2
and greenhouse gas emissions. On the other hand, freshwater is cheap, even though it also
shares a great impact with its extensive use. In the design of biomass-based processes as
in any other, addressing conflicting objectives has become paramount for the sustainable
development. A number of approaches can be used but all of them rely on quantifying the
environmental impact of alternative solutions. The use of life cycle assessments is becom-
ing a widely used tool to consider the environmental burden of a process or a bioproduct
(IRAM-ISO-14040 2006). The method can be divided in three steps. First, we set the start-
ing and final phases for the life cycle analysis (LCA). Next, we define the objective of the
analysis and identify the environmental metrics. Finally, we need to quantify the materials
and energy used in the process as well as the waste produced. With this information, it is
possible to compute the environmental impact that can be used as an objective function for
any design problem (Gebreslassie et al. 2013).
66 Process Systems Engineering for Biofuels Development

3.3.4 Safety Considerations

Process plants are places that must deal with hazardous species, and extreme operating
conditions to allow for the transformations. Safety in process design was established as
a systematic procedure in the work by Kletz (1984). Two different approaches have been
used over time. The traditional approach aims at mitigating the risk derived from the pro-
cess as it is as well as providing prevention measures. In order to reduce the frequency
and consequences of accidents, process industries have developed hazard identification and
analysis techniques (e.g. Failure, Modes, and Effects Analysis, FMEA; Fault Tree Analysis,
FTA; Event Tree Analysis, ETA; Cause−Consequence Analysis, CCA; Preliminary Haz-
ard Analysis, PHA; Human Reliability Analysis, HRA; and Hazard and Operability Study,
HAZOP). Based on these studies another approach is not only advantageous but also prof-
itable. Inherently safe design includes safety within the design process to take into account
the risks at an early stage. This approach consists of four pillars: minimization, substitution,
moderate, and simplify. The aim behind them is to reduce storage of hazardous materials,
replace substances, equipment, or operations with less hazardous ones, reduce hazards by
dilution, refrigeration, or process alternatives that operate at less-hazardous conditions, and
eliminate unnecessary complexity. While the application of the method has been carried out
along the design process using a heuristic approach, at three levels, the feasibility, at prelim-
inary and at basic engineering, where DOW index and HAZOP techniques are being used,
it is also possible to include safety issues within the process design mathematical problem.
Different papers have been published latterly on systematics methodologies for inherently
safe process synthesis. Ruiz-Femenia et al. (2017) used the DOW index and a pareto curve
optimization approach for the inherently safe design of processes. More detailed safety
indexes have been used to evaluate the optimal operating conditions of the different units
characterizing the risk of different accidents using a quantitative risk analysis. The risk
refers to the frequency and magnitude of damage that a person may suffer at a distance due
to the exposure to a physical variable as a result of an accident. To account for the various
objectives, a function including normalized objectives was developed (Martínez-Gomez
et al. 2017; Peña-Lamas et al. 2018).
The solution to multi-objective problems can be addressed following different strategies
including the 𝜀-constraint method, weighted averages and normalized objectives among
others. The 𝜀-constraint method is useful for a reduced number of objectives, working bet-
ter for up to three. Pareto curves or surfaces are developed that represent the geometrical
limit where an objective cannot be improved without worsening any of the others. Weighted
averages generate a single objective function out of the contradicting objectives. The chal-
lenge is providing the appropriate weights. In this way, efforts such as ecoprofit (Cucek
et al. 2012) or RePSIM (Martín 2016b) developed strategies in these lines. To avoid using
weights it is possible to normalize each of the objectives so that their contribution is the
same (Martínez-Gomez et al. 2017).

3.4 The Product Design Problem

3.4.1 Product Design: Engineering Biomass
Biomass is a mixture of compounds. For instance, lignocellulosic biomass consists mainly
of lignin, cellulose, and hemicelluloses; algae are comprised of lipids, carbohydrates, and
 Multiscale Analysis for the Exploitation of Bioresources 67

proteins as well as other species in smaller amounts. For a long time, the biomass as such
was used toward the production of biofuels making almost no distinction on its composi-
tion, apart from the different pretreatments required to make use of the sugars in the case
of lignocellulosic or the extraction of the lipids from seeds and algae. However, the con-
cept of biorefinery that resembles current refineries as multiproduct facilities and the need
for higher yields and better economics has led to the use of all the compounds within the
biomass. For instance, second-generation bioethanol production facilities show an interest-
ing feature. While the use of a thermal path yields a syngas, the biochemical path makes
use of the sugars. The thermal path shows the need to adjust the composition of the syngas
as a function of the product. Methanol aims at H2 to CO ratios over 2 (Martín and Gross-
mann 2017), FT-liquids require ratios around 1.7 (Martín and Grossmann 2011a), while
ethanol (Martín and Grossmann 2011b) or DME (Peral and Martín 2015) requires H2 to
CO ratios of 1. Thus, product design determines the operating conditions and the selection
of the technologies. Larger H2 to CO ratios select indirect gasification while ratios closer
to 1 suggest the use of direct gasification for ethanol production while indirect gasification
is chosen for the production of DME. However, the biochemical path relies on the usage of
hemicelluloses and celluloses. The two main products are xylose and glucose. The yields
to ethanol are larger from glucose, conversions up to 92% are reported, than from xylose,
with conversions around 80% (Aden and Foust 2009; Kazi et al. 2010). Furthermore, the
focus on bioethanol production has deviated the work from the fact that each kind of sugar
can be used for a different purpose. For instance, glucose is the raw material for the pro-
duction of dimethyl furfural while xylose can be used for the production of furfural (Martín
and Grossmann 2015). Other examples can be presented such as the production of i-butene
that has been experimentally validated from glucose alone (Martín and Grossmann 2014b).
Therefore, it is possible to produce renewable ethyl tert-butyl ether from lignocellulosic
raw materials where xyloses are used to produce ethanol and glucose for the production of
i-butene. The cellulose and hemicellulose ratio for the optimal integration can help select
the raw material or guide research into engineered species (Galán et al. 2019).
After a discussion on the use of the fraction of lignocellulosic biomass, we can switch
gears into algae. While for years the main aim has been the production of lipids, the use of
commercial alcohols was the best practice (Zhang et al. 2003). However, the aim toward
sustainable fuel production evaluated the possibility of using either ethanol (Severson et al.
2013) or the integration of renewable methanol within the algae biorefinery (Martín and
Grossmann 2013). Thus it is possible to evaluate the integrated production of biodiesel from
algae where apart from lipids, the composition of the carbohydrates becomes paramount as
the source of glucose. Therefore, within the process design, the optimal algae composition is
computed for the optimal integrated facility. It turns out that the optimal algae composition
results in 60% oil, 30% starch, and 10% protein. Further integration is possible where the
glycerol is used to produce ethanol (Martín and Grossmann 2014c), where the suggested
algae composition does not change, due to the high yield and low energy consumption
involved in the biodiesel production section versus the high energy intense ethanol dehy-
dration stage required for fuel or synthetic ethanol production. A step forward of integration
can lead to a facility that obtains i-butene and ethanol from glucose for the production of
biodiesel and glycerol ethers. In this case, a larger fraction of starch is required and almost
50% of each of lipids and starch are required for the facility to operate without the need for
external addition of ethanol or i-butene (De la Cruz et al. 2014).
68 Process Systems Engineering for Biofuels Development

3.4.2 Blending Problems

Coal or gasoline blending is a well-known topic aiming at selecting the proper mixture of
fossil fuels to limit the sulfur content and to optimize the combustion properties (Shih and
Frey 1995). So far, the blending studies in the literature for different feedstocks for coal
(Shih and Frey 1995), gasoline (Zhao and Wang 2009), refrigerants (Churi and Achenie
1997), detergents (Martín and Martínez 2013), and paints (Conte et al. 2011) have limited
the addition of process constraints on the composition of the feed to the mathematical for-
mulation of a mixing problem. However, the treatment stages must be considered in detail
to account for the actual removal yield and performance.
Lately, a more process design oriented blending problem has been presented so that detail
process models are coupled with raw material blending (Ameri et al. 2018). In particular, for
the case of biorefineries, Hernández et al. (2017) presented a methodology to evaluate the
mixture of different types of waste including manure, sludge, urban food waste, and urban
green waste for the production of syngas with the appropriate composition depending on its
further use such as the production of power, aiming at maximizing methane content in the
syngas, or the production of chemicals such as methanol, ethanol, and FT-liquids. Each one
of them requires a different H2 to CO ratio as presented above. The entire process from the
digester all the way to the synthesis including gas clean-up, reforming, additional syngas
composition adjustment and CO2 capture were included in the formulation. As a result of
considering the entire problem, interesting features were unveiled such as the allocation of
the feed of steam to increase the hydrogen content in the syngas as well as the selection of
the waste or waste mixture for each application. The selection depends on the credit that
is expected from the digestate so that either sludge or a mixture of 65% cattle slurry and
35% urban food waste is selected. Guerras and Martín (2019) followed a similar approach
to address the coal blending problem.

3.5 Supply Chain Level

3.5.1 Introduction
The upper level of operation and integration corresponds to network studies that show the
big picture of the operation of the system. The concept of enterprise wide optimization
(Grossmann 2005) is becoming widely used to evaluate the operation of an entire com-
pany. The main challenge in terms of biomass is the distribution of the resources across
regions. In the case of biomass-based networks, the aim has been to decide on the proper
technologies for transforming the biomass mainly into fuels and allocate them across the
territory depending on the availability of raw materials so as to meet the demand for gaso-
line and diesel substitutes. In the literature, five major components of the supply chain have
been identified as: biomass production system, biomass logistics system, biofuel production
system, biofuel distribution system, and biofuel end-use (Gold and Seuring 2011; Iakovou
et al. 2010). Although these components can be found in most supply chains, the particular
features of biomass and fuels results in the need for a different approach in terms of pro-
duction, transportation, or storage. Yue et al. (2014) presented the challenges that biomass
supply presents for these five components. In terms of feedstock supply and its logistics,
the main issue is related to its availability, that is seasonal and its yield is not constant over
 Multiscale Analysis for the Exploitation of Bioresources 69

time, the difficult storage due to its volume and decay under weather conditions and the
transportation, due to the sparse spatial distribution, that results in issues related to fleet
scheduling and vehicle routing. The production processes depend on the biomass type and
the final product. From terrestrial biomass, we can identify biochemical routes via hydrol-
ysis and fermentation, where the intermediates are sugars, thermal paths through syngas
that allows a wide range of products, or via pyrolysis producing bio-oil. From algae, it is
possible to obtain sugars, oils and higher added value products such as carotenoids.
Supply chain studies can be classified by the aim, and the raw materials and final prod-
ucts. In terms of aim, most of the work focuses on the design of the supply chain that
can include the selection of technologies. Another issue considered is the decentralized
production, the multi-objective optimization, including sustainability issues and consider-
ing uncertainties. By the raw material, early work focused on crops for the production of
ethanol and diesel. Second generation uses linocellulosic raw material to fuels, algae to bio-
fuels and finally the use of biomass to power, heat, food, etc. The area of study constitutes
another problem, related to the size of the area and the particular availability in time in and
yield to meet the demand. Although the contributions are numerous, we are going to focus
on some examples of interest.
For instance, some studies have been developed in the United States at state and national
level, focusing on lignocellulosic raw materials for ethanol production (Marvin et al. 2012),
considering multi-objective optimization (You et al. 2012; Akgul et al. 2012) and even
hybrid, fossil-renewable, feedstock including the production of power (Elia et al. 2012)
and evaluating different feedstocks (Elia et al. 2012; Osmani and Zhang 2013). The sea-
sonality in the availability of the biomass leads to the extension of the problem to mul-
tiperiod approaches. A particularly interesting work was developed to include not only
second-generation lignocellulosic biomass, but also oil (waste cooking oil and algae) to
meet the demand of fuels and food in Europe. Seasonality was also included in the problem
(Cucek et al. 2014). The large model, more than a million equations and 20 million vari-
ables, results in the need for proper solution procedures such as the development of model
reduction techniques. The nature of the raw materials poses different challenges. First, the
allocation of resources and the facilities (Ng et al. 2018; Ng and Maravelias 2017a,2017b).
Secondly, the high dependency on the weather suggests the addition of uncertainty in such
a problem. As a result, the mathematical problem becomes a challenge but is required due
to the nature of the resources, exogenous uncertainty, or the effect that the decrease in the
cost of technologies will have on its deployment as they become mature has been pre-
sented in Kim et al. (Kim et al. 2011), endogenous uncertainty. Furthermore, supply chains
are conditioned by political decisions, including legal and complex contractual constraints
(Gupta 2013; Parunak 1999) and risks (Giarola et al. 2013). We highlight the effect that
policies can have in the deployment of a renewable-based network, since they are respon-
sible for the actual structure and layout of the network. We have opposing objectives from
the various stakeholders involved, i.e. different industry sectors, society, etc. (Sampat et al.
2017). Those policies can be related to the rate of deployment, the need to support certain
regions, or the need for subsidies to meet a target. The implementation of a new system
is a dynamic problem. Agent-based computing has often been suggested as a promising
technique for problem domains that are distributed, complex and heterogeneous dealing
with opposed interests, i.e. social vs. economic ones (Sorda et al. 2013).
70 Process Systems Engineering for Biofuels Development

Finally, electricity has been addressed from the electrical engineering standpoint.
However, biomass can also be a source for power production (Martín and Grossmann
2018; Cucek et al. 2010), integrating or linking both networks.

3.5.2 Modeling Issues

Dong (2001) suggested five basic types of supply chain modeling. The basic one was intro-
duced in 1974 given by the network design that computed the flows of materials for a present
configuration of facilities, etc. The use of mathematical optimization is gaining attention
support as the solvers improve their capability of addressing larger problems. Another alter-
native is the use of stochastic programming that can be considered an extension of the
previous approach to account for the uncertainties. Heuristic methods can also be used to
generate supply chain alternatives that are appropriate for tactical and operational planning
decisions. Finally, simulation-based methods are used in particular to capture the dynamic
nature of the supply chains. Up to three production planning strategies typically guide the
supply chain: Make to Stock (MTS), where the production is carried out before there is an
order on the product; Assemble to Order (ATO), where the products are assembled after the
order is placed being useful in the case of high variability of options for the final products;
and finally, Make to Order (MTO), where the customer orders the product and the chain
reacts to meet the demand.
Supply chain modeling requires the analysis of two scales at least, that of the process
and the one corresponding to macroscopic level (Taifouris and Martín 2018). Process level
analysis is related to the yield and the economics of the technologies involved. Different
approaches can be found in the literature. If different technologies are to be evaluated within
the supply chain, it is highly desirable that the same level of detail for all the processes is
considered to avoid bias toward the processes developed under favorable conditions. In this
way, Cucek et al. (2010) developed a supply chain model where all the processes had been
optimized previously and their economics followed the same estimation procedure. The
information required for the supply chain includes the yield from raw material to prod-
uct and products as well as the needs for utilities and secondary raw materials such as the
alcohol for the production of biodiesel. Furthermore, production and the investment costs
of the processes are required. Ideally, scale up/down effects are needed, so as to evaluate
the effect of the production capacity on both costs. This additional information allows to
evaluate centralized versus distributed production. The methodology to obtain this kind of
surrogate model for the performance of the processes has been described in several works
(Martín 2016c; Martín and Grossmann 2018). A black box model is developed from the
optimization of the processes as an entire unit (Martín 2016c), or by pieces so that mul-
tiproduct facilities can be included (Martín and Grossmann 2018). Figure 3.4. shows the
procedure. Each of the pieces or the entire process is represented by an input-output model.
Piecewise linear approximations are included to capture the scale up issues.
Once the surrogate models for the processes are developed, the network analysis is
formulated. The synthesis is based on the superstructure optimization of a MILP or
MINLP. It consists of mass and energy balances, production and conversion constraints,
transportation, storage, pretreatment and investment cost, and economic objective
function – maximizing the economic profit. Graph networks (Lam et al. 2010; Sampat
 Multiscale Analysis for the Exploitation of Bioresources 71

FT

Bioethanol
Water
Syngas
Biomass
Hydrogen

W, Q Methanol

Prodi(process) = Ki·In(process)⩝i ∈ {Rawmaterials, Wind, Solar, Heat}

Feedj(process) = Qi·In(process)⩝j ∈ {Rawmaterials, Heat}
Power

Figure 3.4 Scheme for process surrogate modeling.

et al. 2017) or multiple layer/echelon models can be found in the literature (You and
Grossmann 2010; Cucek et al. 2010; Elia et al. 2011; Papapostolou et al. 2011; Marvin
et al. 2012, 2013).
In the case of graph networks, the interactions between products are modeled using a
hierarchical graph that maps products at each node using a transformation matrix and that
maps network nodes by using transportation links (arcs). This modeling approach allows
differentiating between in-network and out-of-network flows (sources and sinks). The graph
consists of a set of nodes N, links (arcs) F, products P, out-of-network sources S, and
out-of-network sinks D, see Figure 3.5. To each of the nodes, product balances are for-
mulated to build the supply chain framework.
A different approach consists of the multilayer model. The development of lay-
ers/echelons provides flexibility to the network to account for the different actors involved
(Marvin et al. 2012, 2013). However, recycles are difficult to be implemented so that a
product can be used at the same production level (Cucek et al. 2010). A particular example
is shown in Figure 3.6 and it is used for the explanation. It follows the four-layer nature of
the network’s superstructure, starting from the harvesting and supply (L1) layer, collection
and pre-processing (L2), main processing (L3), up to the use (L4) layer. It includes
intermediate storage at L2 and L3, and transportation links between and within the layers
(Cucek et al. 2010).

3.6 Multiscale Links and Considerations

The models developed for units and processes are typically too detailed for a broader
view type of analysis. Thus, most supply chain and unit commitment problems have been
developed without paying enough attention to the technologies involved. However, this
72 Process Systems Engineering for Biofuels Development

Fnin

Fn,in A Fn,in C

B
Sn n Dn

Fnout

Figure 3.5 Sketch of input and output flow sets into node n. Source: Sampat et al. 2017. Reproduced
with permission of Elsevier.

Δ Δ Δ Layer 4: End users j

Δ
Δ Δ
Δ Δ Δ
Δ Δ m,T,L3,L4
Δ Δ qn,t,j,pp,tp
Δ Δ Δ Δ# Layer 3: Biorefineries n
Δ Δ#
#
q mm,L2,L4
, j , pd, tp
m,T,L3,L3
qn,nn,t,tt,poutpin,tp
# #
#x
x
m,L2,L3 Layer 2: Collection and pre-
x qm,n,pm,tp processing centres m
x x m,T,L3,L2
x q n,m,t,tt,poutpin,tp
x x
m,buy,T,L3
x
•
x
• •
qn,t,pbuy,tp
• • • •
• • • •
q im,L1,L2
,m, pi, tp • • • •
• • • • Layer 1: Supply zones i
• • • • m,buy,T,L2
• • • • q m,t,pbuy,tp
• • • •

Legend: • Supply zone x Collection centre # Biorefinery Δ End user

Figure 3.6 Four-layer supply chain design. Source: Adapted from Lam et al. (2010).

approach although widely spread and used presents several drawbacks in the selection of
technologies for the use of resources avoiding issues related to the use of data from dif-
ferent sources and computed using different assumptions. Lately, the need for a multiscale
analysis that goes all the way from the molecular level to the supply chain analysis has been
proved to provide informed decisions, see Figure 3.7. Floudas et al. (2016) highlighted the
 Multiscale Analysis for the Exploitation of Bioresources 73

Supply chain level Process level

Unit level
Resources

Figure 3.7 Links between different scales.

need for multiscale analysis applied to a case study in the carbon capture and utilization
field. The idea consists of analyzing from the molecular level, to design the appropriate
adsorbent materials, to a second scale corresponding to individual process units, a third
scale at the level of the entire process design and, finally, the evaluation of the macroscopic
scale, the supply chain. The links between all the stages are paramount for informed deci-
sions at macroscopic level. Additional cases of studies have been presented lately related
to biorefineries. In particular, in the use and exploitation of waste. The yield, kinetics and
investment from different waste types range over a wide spectrum and the analysis of the
performance of the availability of the various resources and their effect on the production
of biogas and the size of the facilities are closely related. Thus, investment decisions are
directly associated to them. In Taifouris and Martin (2018), such a study was developed
for the selection of the use of four types of waste, lignocellulosic, mature, municipal solid
waste and sludge across the region of Castile and Leon, in the center of Spain. The analysis
of the kinetics of each one was performed for the design of the digester as a central piece
of a biogas production facility. The techno-economics of the facilities was linked to the
waste use so that the supply chain study was capable of selecting the use of waste per shire
across the territory for the best usage of the resources given a certain regional budget. If
50% of the regional budget for environmental infrastructure projects is used, the municipal
solid waste is the best alternative. Only if it is possible to use the entire regional budget are
lignocellulosic residues also used, due to their higher yield to biogas. Neither sludge nor
manure is chosen. However, even with the entire budget of 351 M€, less than 10% of the
total residues can be processed. Therefore, it would require a multiyear investment effort
to treat a larger amount of waste with the advantage that the waste available can provide
several times the natural gas consumed in the region and 5% of the national demand with
74 Process Systems Engineering for Biofuels Development

a total investment of 4.4 × 109 €. Only through a multiscale analysis of the problem all the
way from the biomass processing to the supply chain of the biomass and the final products
would it be possible to address the difficult problem of determining the best exploration of
the biomass resources due to their features.

Acknowledgment
The authors acknowledge the FPU, Spain grant (FPU16/06212) from MECD, Spain to
Mr. A. Sánchez and European Project IProPBio H2020-MSCA-RISE-2017 Project number
778168.

Nomenclature

At Cross-sectional area in the reactor (m2 )

ai Activity of each component (i)
Cp Heat capacity
dp Catalytic particle diameter
Fi0 Inlet molar flow of component i
Hi Specific enthalpy for each component (i)
K, Kp Equilibrium constant
Ki Gas–liquid phase equilibrium constant
kreac Kinetic constant (kmol/m3 h)
L, L’ Section for heat transfer
ṅ i Mass/molar flow for each component (i)
Nu Nusselt number
P Pressure
Pi Partial pressure of component i
Pivap Vapor pressure for each component (i)
Pr Prandtl number
Q̇ unit Heat involved in a unit
Re Reynolds number
r Reaction rate
Sh Sherwood number
Sc Schmidt number
T Temperature
U Global heat transfer coefficient
vgas Gas velocity
wi Weights for each data (i)
X Conversion
xi Molar fraction of component i
yi Molar fraction of component i
zi Molar fraction of component i
𝜂ji Yield in the splitter for different components (i) and different outlet streams (j)
𝛼 Kinetic parameter
 Multiscale Analysis for the Exploitation of Bioresources 75

𝜙 Effectiveness factor
Ω Catalyst activity
ΔHr Reaction heat
𝛿 Isolation thickness
𝜀 Catalytic porosity
𝜇 Viscosity
𝜌 Density
𝛽i Polynomial regression coefficient
𝜎2 Variance
𝜃 j , Pj Adjustable parameters

References
Aasberg-Petersen, K., Christensen, T.S., Stub Nielsen, C., and Dybkjaer, I. (2013). Recent developments
in autothermal reforming and pre-reforming for synthesis gas production in GTL applications. Fuel Pro-
cessing Technology 83: 253–261.
Abagnale, C., Cameretti, M.C., De Robbio, R., and Tuccillo, R. (2016). CFD study of a MGT combustor
supplied with syngas. Energy Procedia 101: 933–940.
Aden, A. and Foust, T. (2009). Technoeconomic analysis of the dilute sulfuric acid and enzymatic hydrolysis
process for the conversion of corn-stover to ethanol. Cellulose 16: 535–545.
Agarwal, A., Biegler, L.T., and Zitney, S.E. (2009). Simulation and optimization of pressure swing
adsoprtion systems using reduced-order Modeling. Industrial & Engineering Chemistry Research 48:
2327–2343.
Ahmetovic, E. and Grossmann, I.E. (2011). Global superstructure optimization for the design of integrated
process water networks. AIChE Journal 57 (2): 434–457.
Ahmetovic, E. and Kravanja, Z. (2012). Solution strategies for the synthesis of heat integrated process water
networks. Chemical Engineering Transactions 29: 1015–1020.
Ahmetovic, E., Martín, M., and Grossmann, I.E. (2010). Optimization of water consumption in process
industry: corn – based ethanol case study. Industrial and Engineering Chemistry Research 49 (17):
7972–7982.
Ahmetovic, E., Ibric, N., and Kravanja, Z. (2014). Optimal design for heat-integrated water-using and
wastewater treatment networks. Applied Energy 135: 791–808.
Akgul, O., Shah, N., and Papageorgiou, L.G. (2012). An optimisation framework for a hybrid first/second
generation bioethanol supply chain. Computers and Chemical Engineering 42: 101–114.
Alva-Argáez, A., Kokossis, A.C., and Smith, R. (1998a). Automated design of industrial water networks.
American Institute of Chemical Engineering Annual Meeting, Paper 13f, Miami, FL.
Alva-Argáez, A., Kokossis, A.C., and Smith, R. (1998b). Wastewater minimisation of industrial systems
using an integrated approach. Computers and Chemical Engineering 22 (Suppl): S741–S744.
Alva-Argáez, A., Kokossis, A.C., and Smith, R. (2007). The design of water-using systems in petroleum
refining using a water-pinch decomposition. Chemical Engineering Journal 128: 33–46.
Ameri, M., Moktari, H., and Mostafavi-Sani, M.E. (2018). Analyses and multi-objective optimization of
different fuels application for a large combined cycle power plant. Energy 156: 371–386.
Amini, H., Wang, L., Hashemisohi, A. et al. (2018). An integrated growth kinetics and computational fluid
dynamics model for the analysis of algal productivity in open raceway ponds. Computers and Electronics
in Agriculture 145 (2018): 363–372.
Anderson, D. (1977). Practical aspects of industrial reuse. Institution of Chemical Engineers Symposium
Series 52: 2–11.
76 Process Systems Engineering for Biofuels Development

Andrade, T.A., Martín, M., Errico, M., and Christensen, K.V. (2019). Biodiesel production catalyzed by
liquid and immobilized enzymes: optimization and economic analysis. Chemical Engineering Research
and Design 141: 1–14.
Andrecovic, M.J. and Westerberg, A.W. (1985). An MILP formulation for heat integrated distillation
sequence synthesis. AIChE Journal 31: 1461.
Arora, J.S. (2017). Introduction to Optimum Design, 4e. Elsevier.
Aydin, E.S., Yudel, O., and Sadikoglu, H. (2017). Development of a semi-empirical equilibrium model for
downdraft gasification systems. Energy 130: 86–98.
Bagajewicz, M., Rodera, H., and Savelski, M. (2002). Energy efficient water utilization systems in process
plants. Computers and Chemical Engineering 26: 59–79.
Baliban, R.C., Elia, J.A., and Floudas, C.A. (2013). Biomass to liquid transportation fuels (BTL)
systems: process synthesis and global optimization framework. Energy & Environmental Science 6:
267–287.
Belhateche, D.H. (1995). Choose appropriate wastewater treatment technologies. Chemical Engineering
Progress 91: 32.
Bhosekar, A. and Ierapetritou, M. (2018). Advances in surrogate based modelling, feasibility analysis and
optimization: A review. Computers & Chemical Engineering 108: 250–267.
Biegler, L.T. (2010). Nonlinear Programming. Concepts, Algorithms, and Applications to Chemical Pro-
cesses. Society for Industrial and Applied Mathematics.
Biegler, L.T., Grossmann, I.E., and Westerberg, A.W. (1997). Systematic Methods of Chemical Process
Design. Prentice Hall.
Bitog, J.P., Lee, I.B., Lee, C.G. et al. (2011). Application of computational fluid dynamics for model-
ing and designing photobioreactors for microalgae production: A review. Computers and Electronics in
Agriculture 76 (2011): 131–147.
Bogataj, M. and Bagajewicz, M.J. (2008). Synthesis of non-isothermal heat integrated water networks in
chemical processes. Computers and Chemical Engineering 32: 3130–3142.
Branan, C.R. (2005). Rules of Thumb for Chemical Engineers, 4e. Elsevier.
Buelens, L.C., Galvita, V.V., Poelman, H. et al. (2016). Super-dry reforming of methane intensifies CO2
utilization via Le Chatelier’s principle. Science 354: 449–452.
Carnes, B.A., Fordm D.L., and Brady, S.O. (1973). Treatment of refinery wastewaters for reuse. National
Conference on Complete Water Reuse, Washington, D.C.
Carvajal, D., Marchisio, D.L., Bensaid, S., and Fino, D. (2012). Enzymatic hydrolysis of lignocellulosic
biomasses via CFD and experiments. Industrial and Engineering Chemistry Research 51 (22):
7518–7525.
Castro, P., Teles, J.P., and Novais, A. (2009). Linear program-based algorithm for the optimal design of
wastewater treatment systems. Clean Technologies and Environmental Policy 11: 83–93.
Churi, N. and Achenie, L.E.K. (1997). The optimal design of refrigerant mixtures for a two-evaporator
refrigeration system. Computers and Chemical Engineering 21: S349–S354.
Conte, E., Gani, R., and Ng, K.M. (2011). Design of formulated products: A systematic methodology.
AIChE Journal 2011 (57): 2431–2449.
Couper, J.R., Penney, W.R., Fair, J.R., and Walas, S.M. (2005). Chemical Process Equipment: Selection
and Design, 2e. Gulf Professional Publishing/Elsevier.
Cozad, A., Sahinidis, N.V., and Miller, D.C. (2014). Learning surrogate models for simulation-based opti-
mization. AIChE Journal 60 (6): 2211–2227.
Cucek, L., Lam, H.L., Klemes, J.J. et al. (2010). Synthesis of regional networks for the supply of energy
and bioproducts. Clean Technologies and Environmental Policy 12: 635–645.
Cucek, L., Drobez, R., Pahor, B., and Kravanja, Z. (2012). Sustainable synthesis of biogas processes using
a novel concept of eco profit. Computers and Chemical Engineering 42: 87–100.
Cucek, L., Martín, M., Grossmann, I.E., and Kravanja, Z. (2014). Large-scale biorefinery supply net-
work – case study of the European Union. Computer Aided Chemical Engineering 33: 319–324.
 Multiscale Analysis for the Exploitation of Bioresources 77

De la Cruz, V., Hernández, S., Martín, M., and Grossmann, I.E. (2014). Integrated synthesis of biodiesel,
bioethanol, I-butene and glycerol ethers from algae. Industrial and Engineering Chemistry Research 53
(37): 14397–14407.
Dong, H.G., Lin, C.Y., and Chang, C.T. (2008). Simultaneous optimization approach for integrated
water-allocation and heat-exchange networks. Chemical Engineering Science 63: 3664–3678.
Dong, M. (2001). Process Modeling, Performance Analysis and Configuration Simulation in Integrated
Supply Chain Network Design. https://vtechworks.lib.vt.edu/handle/10919/28779 (accessed 25 January
2018).
Douglas, J. (1988). Conceptual Design of Chemical Processes. New York: McGraw-Hill.
Doyle, S.J. and Smith, R. (1997). Targeting water reuse with multiple contaminants. Process Safety and
Environmental Protection 75: 181–189.
Duran, M.A. and Grossmann, I.E. (1986). Simultaneous optimization and heat integration of chemical pro-
cesses. AIChE Journal 32: 123–138.
El-Halwagi, M.M. and Manousiouthakis, V. (1989). Synthesis of mass exchange networks. AIChE Journal
35 (8): 1233–1244.
Elia, J.A., Baliban, R.C., Xiao, X., and Floudas, C.A. (2011). Optimal energy supply network determina-
tion and life cycle analysis for hybrid coal, biomass, and natural gas to liquid (CBGTL) plants using
carbon-based hydrogen production. Computers and Chemical Engineering 35: 1399–1430.
Elia, J.A., Baliban, R.C., and Floudas, C.A. (2012). Nationwide energy supply chain analysis for hybrid
feedstock processes with significant CO2 emissions reduction. AIChE Journal 58 (7): 2142–2154.
Eliceche, A.M. and Sargent, R.W.H. (1986). Synthesis and design of distillation of sequences. IChemE
Symposium Series 61: 1–22.
ExxonMobil (2013). The Outlook for Energy: A View to 2040. https://www.clingendaelenergy.com/inc/
upload/files/2013_ExxonMobil_Outlook_for_Energy_.pdf (accessed March 2020).
Faria, D.C. and Bagajewicz, M.J. (2010). On the appropriate modeling of process plant water systems.
AIChE Journal 56: 668–689.
Felder, R.M., Rousseau, R.W., and Bullard, L.G. (2016). Elementary Principles of Chemical Processes, 4e.
Wiley.
Flores-Tlacuahuac, A., Moreno, S.T., and Biegler, L.T. (2008). Global optimization of highly nonlinear
dynamic systems. Industrial & Engineering Chemistry Research 47: 2643–2655.
Floudas, C.A. and Paules, G.E. IV (1988). A mixed-integer nonlinear programming formulation for the
synthesis of heat-integrated distillation sequences. Computers & Chemical Engineering 12 (6): 531–546.
Floudas, C.A., Ciric, A.R., and Grossmann, I.E. (1986). Automatic synthesis of optimum heat exchanger
network configurations. AIChE Journal 32: 276–290.
Floudas, C.A., Niziolek, A.M., Onel, O., and Mathews, L.R. (2016). Multi-scale systems engineering for
energy and the environment: challenges and opportunities. AIChE Journal 62 (3): 602–523.
Furman, K.C. and Sahinidis, N.V. (2002). A critical review and annotated bibliography for heat exchanger
network synthesis in the 20th century. Industrial & Engineering Chemistry Research 41 (10): 2335–2370.
Gadkar, K.G., Mehra, S., and Gomes, J. (2005). On-line adaptation of neural networks for bioprocess con-
trol. Computers & Chemical Engineering 29: 1047–1057.
Galan, B. and Grossmann, I.E. (1998). Optimal design of distributed wastewater treatment networks. Indus-
trial and Engineering Chemistry Research 37: 4036–4048.
Galán, G., Martín, M., and Grossmann, I.E. (2019). Integrated renewable production of ETBE from Switch-
grass. ACS Sustainable Chemistry & Engineering 7 (9): 8943–8953.
Gas Processors Suppliers Association, GPSA (2004). Engineering Data Book. Volumes I & II.
Geankoplis, C.J. (2006). Transport Processes and Unit Operations, 4e. Prentice Hall.
Gebreslassie, B.H., Slivinsky, M., Wang, B., and You, F. (2013). Life cycle optimization for sustainable
design and operations of hydrocarbon biorefinery via fast pyrolysis, hydrotreating and hydrocracking.
Computers and Chemical Engineering 50: 71–91.
78 Process Systems Engineering for Biofuels Development

Giarola, S., Bezzo, F., and Shah, N. (2013). A risk management approach to the economic and environmental
strategic design of ethanol supply chains. Biomass and Bioenergy 58: 31–51.
Gold, S. and Seuring, S. (2011). Supply chain and logistics issues of bio-energy production. Journal of
Cleaner Production 19: 32–42.
Grossmann, I.E. (2005). Enterprise-wide optimization: a new frontier in process systems engineering.
AIChE Journal 51 (7): 1846–1857.
Grossmann, I.E. and Martin, M. (2010). Energy and water optimization in biofuel plants. Chinese Journal
of Chemical Engineering 18 (6): 914–922.
Guerras, L.S. and Martín, M. (2019). Optimal gas treatment and coal blending for reduced emissions in
power plants: a case study in Northwest Spain. Energy 169: 739–749.
Gunaratnam, M., Alva-Aragez, A., Kokossis, A. et al. (2005). Automated design of total water systems.
Industrial and Engineering Chemistry Research 44: 588–599.
Gupta, V. (2013). Modeling and computational strategies for optimal oilfield development planning under
fiscal rules and endogenous uncertainties. Dissertations. Paper 231.
Haringa, C., Tang, W., Wang, G. et al. (2018). Computational fluid dynamics simulation of an industrial
P. chysogenum fermentation with a coupled 9-pool metabolic model: towards rational scale-down and
design optimization. Chemical Engineering Science 175: 12–24.
Haro, P., Ollero, P., Villanueva-Perales, A.L., and Gómez-Barea, A. (2013). Thermochemical biorefinery
based on dimethyl ether as intermediate: technoeconomic assessment. Applied Energy 102: 950–961.
Henao, C.A. and Maravelias, C.T. (2011). Surrogate-based superstructure optimization framework. AIChE
Journal 57 (5): 1216–1232.
Hernández, B. and Martín, M. (2016). Optimal process operation for biogas reforming to methanol:
effects of dry reforming and biogas composition. Industrial & Engineering Chemistry Research 55 (23):
6677–6685.
Hernández, B. and Martín, M. (2017). Optimal integrated plant for production of biodiesel from waste. ACS
Sustainable Chemistry & Engineering 5: 6756–6767.
Hernández, B. and Martin, M. (2018). Optimization for biogas to chemicals via tri-reforming. Analysis of
Fischer-Tropsch fuels from biogas. Energy Conversion and Management 174: 998–1013.
Hernández, B., León, E., and Martín, M. (2017). Bio-waste selection and blending for the optimal pro-
duction of power and fuels via anaerobic digestion. Chemical Engineering Research and Design 121:
163–172.
Himmelblau, D.M. (2000). Applications of artificial neural networks in chemical engineering. Korean Jour-
nal of Chemical Engineering 17 (4): 373–392.
Hohmann, E. and Lockhart, F. (1976). Optimum heat exchangers network synthesis. Proceedings of the
American Institute of Chemical Engineers. Atlantic City, NJ.
Hospondarec, R.W. and Thompson, S.J. (1974). Oil-steam system for water reuse. Proceedings of the Amer-
ican Institute of Chemical Engineering Workshop, Vol. 7. New York, NY: AIChE.
Iakovou, E., Karagiannidis, A., Vlachos, D. et al. (2010). Waste biomass-to-energy supply chain manage-
ment: a critical synthesis. Waste Management 30: 1860–1870.
IRAM-ISO-14040 (2006). Gestión Ambiental. Análisis del ciclo de vida: Principios y marco. Autores:
IRAM.
Karuppiah, R. and Grossmann, I.E. (2006). Global optimization for the synthesis of integrated water sys-
tems in chemical processes. Computers and Chemical Engineering 30: 650–673.
Karuppiah, R., Peschel, A., Grossmann, I.E. et al. (2008). Energy optimization of an ethanol plant. AIChE
Journal 54 (6): 1499–1525.
Kazi, F.K., Fortman, J.A., Anex, R.P. et al. (2010). Technoeconomic comparison of process technologies
for biochemical ethanol production from corn stover. Fuel 89: S20–S28.
Kim, J., Realff, M.J., and Lee, J.H. (2011). Optimal design and global sensitivity analysis of biomass
supply chain networks for biofuels under uncertainty. Computers & Chemical Engineering 35 (9):
1738–1751.
 Multiscale Analysis for the Exploitation of Bioresources 79

Kleijnen, J.P.C. (2009). Kriging metamodeling in simulation: a review. European Journal of Operational
Research 192: 707–716.
Klemes, J.J. (2013). Handbook of Process Integration (PI): Minimisation of Energy and Water Use, Waste
and Emissions. Cambridge, UK: Woodheat Publishing.
Kletz, T.A. (1984). Cheaper, Safer Plants, or Wealth and Safety at Work. Rugby: The Institution of Chemical
Engineers.
Kocis, G.R. and Grossmann, I.E. (1987). Relaxation strategy for the structural optimization of process
flowsheets. Industrial and Engineering Chemistry Research 26: 1869–1880.
Kocis, G.R. and Grossmann, I.E. (1989). A modelling and decomposition strategy for the MINLP opti-
mization of process flowsheets. Computers and Chemical Engineering 13: 797–819.
Kravanja, Z. (2010). Challenges in sustainable integrated process synthesis and the capabilities of an
MINLP process synthesizer MipSyn. Computers and Chemical Engineering 34: 1831–1848.
Kravanja, Z. and Grossmann, I.E. (1990). PROSYN—an MINLP process synthesizer. Computers and
Chemical Engineering 14: 1363–1378.
Ku, X., Li, T., and Lovas, T. (2015). CFD-DEM simulation of biomass gasification with steam in a fluidized
bed reactor. Chemical Engineering Science 122: 270–283.
Lam, H.L., Klemes, J.J., and Kravanja, Z. (2010). Model-size reduction techniques for large-scale biomass
production and supply networks. Energy 36 (8): 4599–4608.
Lemieux, C. (2009). Monte Carlo and Quasi-Monte Carlo Sampling. Springer.
Li, B.H. and Chang, C.T. (2007). A simple and efficient initialization strategy for optimizing water-using
network designs. Industrial and Engineering Chemistry Research 46: 8781–8786.
Liew, P.Y., Wan Alwi, S.R., Varbanov, P.S. et al. (2013). Centralized utility system planning for a total site
heat integration network. Computers and Chemical Engineering 57: 104–111.
Lima, F.V., Daoutidis, P., Tsapatsis, M., and Marano, J.J. (2012). Modeling and optimization of membrane
reactors for carbon capture in integrated gasification combined cycle units. Industrial & Engineering
Chemistry Research 51: 5480–5489.
Linnhoff, B. and Hindmarsh, E. (1983). The pinch design method for heat exchanger networks. Chemical
Engineering Science 38 (5): 745–763.
Liu, L., Sun, J., Xu, W. et al. (2009). Modeling and optimization of microbial hyaluronic acid production
by Streptococcus zooepidemicus using radial basis function neural network coupling quantum-behaved
particle swarm optimization algorithm. Biotechnology Progress 25 (6): 1819–1825.
Madhania, S., Muharam, Y., Winardi, S., and Puwanto, W.W. (2019). Mechanism of molasses–water mixing
behavior in bioethanol fermenter. Experiments and CFD modeling. Energy Reports 5: 454–461.
Malmierca, S., Díez-Antolínez, R., Paniagua, A.I., and Martín, M. (2017). Technoeconomic study of
biobutanol AB production. 2. Process design. Industrial & Engineering Chemistry Research 56:
1525–1533.
Martín, M. (2014). Introduction to Software for Chemical Engineers. CRC Press.
Martín, M. (2016a). Industrial Chemical Process Analysis and Design, 1e. Elsevier.
Martín, M. (2016b). RePSIM metric for design of sustainable renewable based fuel and power production
processes. Energy 114: 833–845.
Martín, M. (2016c). Methodology for solar and wind energy chemical storage facilities design under
uncertainty: methanol production from CO2 and hydrogen. Computers & Chemical Engineering 92:
43–54.
Martín, M. and Grossmann, I.E. (2011a). Process optimization of FT- diesel production from lignocellulosic
switchgrass. Industrial and Engineering Chemistry Research 50 (23): 13485–13499.
Martín, M. and Grossmann, I.E. (2011b). Energy optimization of bioethanol production via gasification of
switchgrass. AIChE Journal 57 (12): 3408–3428.
Martín, M. and Grossmann, I.E. (2012a). BIOpt: a library of models for optimization of biofuel production
processes. Computer Aided Chemical Engineering 30: 16–20.
80 Process Systems Engineering for Biofuels Development

Martín, M. and Grossmann, I.E. (2012b). Energy optimization of bioethanol production via hydrolysis of
switchgrass. AIChE Journal 58 (5): 1538–1549.
Martín, M. and Grossmann, I.E. (2013). ASI: toward the optimal integrated production of biodiesel with
internal recycling of methanol produced from glycerol. Environmental Progress & Sustainable Energy
32 (4): 791–801.
Martín, M. and Grossmann, I.E. (2014a). Simultaneous optimization and heat integration for the coproduc-
tion of diesel substitutes: biodiesel (FAME and FAEE) and glycerol ethers from algae oil. Industrial &
Engineering Chemistry Research 53: 11371–11383.
Martín, M. and Grossmann, I.E. (2014b). Optimization simultaneous production of ethanol and i-butene
from switchgrass. Journal of Biomass Bioenergy 61: 93–103.
Martín, M. and Grossmann, I.E. (2014c). Design of an optimal process for enhanced production of
bioethanol and biodiesel from algae oil via glycerol fermentation. Applied Energy 135: 108–114.
Martín, M. and Grossmann, I.E. (2015). Optimal production of furfural and DMF from algae and switch-
grass. Industrial & Engineering Chemistry Research 55 (12): 3192–3202.
Martín, M. and Grossmann, I.E. (2017). Optimal integration of a self sustained algae based facility with
solar and/or wind energy. Journal of Cleaner Production 145: 336–347.
Martín, M. and Grossmann, I.E. (2018). Optimal integration of renewable based processes for fuels and
power production: Spain case study. Applied Energy 213: 595–610.
Martín, M. and Martín, M. (2017). Cooling limitations in power plants: optimal multiperiod design of
natural draft cooling towers. Energy 135: 625–636.
Martín, M. and Martínez, A. (2013). Methodology for simultaneous process and product design in the
consumer products industry: the case study of the laundry business. Chemical Engineering Research
and Design 91: 795–809.
Martín, M., Montes, F.J., and Galán, M.A. (2007). Oxygen transfer from growing bubbles: effect of the
physical properties of the liquid. Chemical Engineering Journal 128: 21–32.
Martín, M., Sánchez, A., and Woodley, J.M. (2018). Fermentative Alcohol Production, Bioreactors for
Microbial Biomass and Energy Conversion, 319–357. Elsevier.
Martínez-Gomez, J., Peña-Lamas, J., Martín, M., and Ponce-Ortega, J.M. (2017). A multi-objective opti-
mization approach for the selection of working fluids of geothermal facilities: economic, environmental
and social aspects. Journal of Environmental Management 203: 962–972.
Martínez-Patiño, J., Picón-Núñez, M., Serra, L.M., and Verda, V. (2011). Design of water and energy net-
works using temperature–concentration diagrams. Energy 36: 3888–3896.
Marvin, W.A., Schmidt, L.D., and Benjaafar, S. (2012). Economic optimization of a lignocellulosic
biomass-to-ethanol supply chain. Chemical Engineering Science 67 (1): 68–79.
Marvin, W.A., Schmidt, L.D., and Daoutidis, P. (2013). Biorefinery location and technology selec-
tion through supply chain optimization. Industrial and Engineering Chemistry Research 52 (9):
3192–3208.
Masso, A.H. and Rudd, D.F. (1969). The synthesis of system designs. II. Heuristic structuring. AIChE
Journal 15 (1): 10–17.
May-Vázquez, M.M., Rodríguez Ángeles, M.A., Gómez-Castro, F.I., and Uribe-Ramírez, A.R. (2017).
Hydrodynamic feasibility of the production of biodiesel fuel in a high-pressure reactive distillation col-
umn. Chemical Engineering and Processing Process Intensification 112: 31–37.
Mekala, S., Fontanals, A., and Guardo, A. (2019). CFD study of the supercritical transesterification of
triolein: wall-to-fluid/particle-to fluid transport effects. Journal of Supercritical Fluids 143: 42–54.
Mishra, P.N., Fan, L.T., and Erickson, L.E. (1975). Application of mathematical optimization techniques
in computer aided design of wastewater treatment systems. American Institute of Chemical Engineers
Symposium Series Water (II) 71 (145): 136.
Montgomery, D.C., Peck, E.A., and Vining, G.G. (2012). Introduction to Linear Regression Analysis, 5e.
Wiley.
 Multiscale Analysis for the Exploitation of Bioresources 81

Munson, B.R., Young, D.F., Okiishi, T.H., and Huebsch, W.W. (2009). Fundamentals of Fluid Mechanics,
6e. Wiley.
Nemet, A., Klemes, J.J., and Kravanja, Z. (2015). Designing a total site for an entire lifetime under fluctu-
ating utility prices. Computers and Chemical Engineering 72: 159–182.
Ng, R.T.L. and Maravelias, C.T. (2017a). Economic and energetic analysis of biofuel supply chains. Applied
Energy 205: 1571–1582.
Ng, R.T.L. and Maravelias, C.T. (2017b). Design of biofuel supply chains with variable regional depot and
biorefinery locations. Renewable Energy 100: 90–102.
Ng, R.T.L., Kurniawan, D., Wang, H. et al. (2018). Integrated framework for designing spatially explicit
biofuel supply chains. Applied Energy 216: 116–131.
Osmani, A. and Zhang, J. (2013). Stochastic optimization of a multi-feedstock lignocellulosic-based
bioethanol supply chain under multiple uncertainties. Energy 59: 157–172.
Papapostolou, C., Kondili, E., and Kaldellis, J.K. (2011). Development and implementation of an optimi-
sation model for biofuels supply chain. Energy 36: 6019–6026.
Papoulias, S.A. and Grossmann, I.E. (1983). A structural optimization approach in process synthesis-II.
Heat recovery networks. Computers & Chemical Engineering 7: 707–721.
Parunak, H.V. (1999). Industrial and practical applications of DAI. In: Multiagent Systems (ed. G. Weiss),
377–421. The MIT Press.
Peña-Lamas, J., Martínez-Gomez, J., Martín, M., and Ponce-Ortega, J.M. (2018). Optimal production of
power from mid-temperature geothermal sources: scale and safety issues. Energy Conversion and Man-
agement 165: 172–182.
Peral, E. and Martín, M. (2015). Optimal production of dimethyl ether from switchgrass-based syngas via
direct synthesis. Industrial & Engineering Chemistry Research 54 (30): 7465–7475.
Perner-Nochta, I. and Posten, C. (2007). Simulations of light intensity variation in photobioreactors. Journal
of Biotechnology 131 (3): 276–285.
Ponce-Ortega, J.M., Jiménez-Gutiérrez, A., and Grossmann, I.E. (2008). Optimal synthesis of heat
exchanger networks involving isothermal process streams. Computers & Chemical Engineering 32 (8):
1918–1942.
Ponton, J.W. and Donaldson, R.A.B. (1974). A fast method for the synthesis of optimal heat exchanger
networks. Chemical Engineering Science 29: 2375–2377.
Queipo, N.V., Haftka, R.T., Shyy, W. et al. (2005). Surrogate-based analysis and optimization. Progress in
Aerospace Sciences 41: 1–28.
Quesada, I. and Grossmann, I.E. (1995). Global optimization of bilinear process networks with multicom-
ponent flows. Computers and Chemical Engineering 19: 1219–1242.
Quirante, N. and Caballero, J.A. (2016). Large scale optimization of a sour water stripping plant using
surrogate models. Computers & Chemical Engineering 92: 143–162.
Quirante, N., Javaloyes, J., and Caballero, J.A. (2015). Rigorous design of distillation columns using sur-
rogate models based on Kriging interpolation. AIChE Journal 61 (7): 2169–2187.
Raghavan, N.S. and Ruthven, D.M. (1985). Pressure swing adsorption. Part III: numerical simulation of a
kinetically controlled bulk gas separation. AIChE Journal 31 (12): 2017–2025.
Rasmuson, A., Andersson, B., Olsson, L., and Andersson, R. (2014). Mathematical Modeling in Chemical
Engineering. Cambridge University Press.
Ruiz-Femenia, R., Fernández-Torres, M.J., Salcedo-Díaz, R. et al. (2017). Systematic tools for the con-
ceptual design of inherently safer processes. Industrial and Engineering Chemistry Research 56 (25):
7301–7313.
Sacks, J., Welch, W.J., Mitchell, T.J., and Wynn, H.P. (1989). Design and analysis of computer experiments.
Statistical Science 4 (4): 409–423.
Sadhukhan, J., Ng, K.S., and Hernandez, E.M. (2014). Biorefineries and Chemical Processes: Design,
Integration and Sustainability Analysis. Wiley.
82 Process Systems Engineering for Biofuels Development

Sampat, A.S., Martín, E., Martín, M., and Zavala, V.M. (2017). Optimization formulations for multi-product
supply chain networks. Computers and Chemical Engineering 104: 296–310.
Sánchez, A. and Martín, M. (2018a). Optimal renewable production of ammonia from water and air. Journal
of Cleaner Production 178: 325–342.
Sánchez, A. and Martín, M. (2018b). Scale up and scale down issues of renewable ammonia plants: towards
modular design. Sustainable Production and Consumption 16: 176–192.
Sánchez, A., Martín, M., and Vega, P. (2019). Biomass based sustainable ammonia production: digestion
vs gasification ACS. ACS Sustainable Chemistry & Engineering https://doi.org/10.1021/acssuschemeng
.9b01158.
Sane, M. and Atkins, U.S. (1977). Industrial water management. Institution of Chemical Engineers Sym-
posium Series 52: 1–9.
Savelski, M. and Bagajewicz, M. (2003). On the necessary conditions of optimality of water utilization
systems in process plants with multiple contaminants. Chemical Engineering Science 58: 5349–5362.
Savelski, M.J. and Bagajewicz, M.J. (2000). On the optimality conditions of water utilization systems in
process plants with single contaminants. Chemical Engineering Science 55: 5035–5048.
Scherer, V., Mönnigmann, M., Berner, M.O., and Sudbrock, F. (2016). Coupled DEM-CFD simulation of
drying wood chips in a rotary drum – baffle design and model reduction. Fuel 184: 896–904.
Severson, K., Martín, M., and Grossmann, I.E. (2013). Optimal integration for biodiesel production using
bioethanol. AIChE Journal 59 (3): 834–844.
Sharma, S. and Sheth, P.N. (2016). Air-steam biomass gasification: experiments, modelling and simulation.
Energy Conversion and Management 110: 307–318.
Shih, J.S. and Frey, H.C. (1995). Coal blending optimization under uncertainty. European Journal of Oper-
ational Research 83 (3): 452–465.
Skylov, V., Stenzel, R.A. (1974). Reuse of wastewaters – possibilities and problems. Proceedings of the
American Institute of Chemical Engineering Workshop, Vol. 7. New York, NY: AIChE.
Smith, R. (2005). Chemical Process: Design and Integration. New York, NY: Wiley.
Sorda, G., Sunak, Y., and Madlener, R. (2013). An agent-based spatial simulation to evaluate the promotion
of electricity from agricultural biogas plants in Germany. Ecological Economics 89: 43–60.
Taifouris, M.R. and Martín, M. (2018). Multiscale scheme for the optimal use of residues for the production
of biogas across castile and Leon. Journal of Cleaner Production 185: 239–251.
Takama, N., Kuriyama, T., Shiroko, K., and Umeda, T. (1980). Optimal water allocation in a petroleum
refinery. Computers and Chemical Engineering 4: 251–258.
Tan, C. and Zhi, Q. (2016). The energy-water nexus: A literature review of the dependence of energy on
water. Energy Procedia 88: 277–284.
Tian, Y., Demirel, S.E., Hasan, M.M.F., and Pistikopoulos, E.N. (2018). An overview of process systems
engineering approaches for process intensification: state of the art. Chemical Engineering and Processing
Process Intensification 133: 160–210.
Van Loo, S. and Koppejan, J. (2008). The Handbook of Biomass Combustion & co-Firing. London: Earth-
scan.
Varbanov, P.S. (2014). Energy and water interactions: implications for industry. Current Opinion in Chem-
ical Engineering 5: 15–21.
Vesvikar, M.S. and Al-Dahhan, M. (2004). Flow pattern visualization in a mimic anaerobic digester using
CFD. Biotechnology and Bioengineering 89 (6): 719–732.
Viswanathan, J. and Grossmann, I.E. (1990). A combined penalty function and outer-approximation method
for MINLP optimization. Computers & Chemical Engineering 14 (7): 769–782.
Wang, Y.P. and Smith, R. (1994). Wastewater minimization. Chemical Engineering Science 49: 981–1006.
Warnes, M.R., Glassey, J., Montague, G.A., and Kara, B. (1998). Application of radial basis function and
feedforward artificial neural networks to the Escherichia coli fermentation process. Neurocomputing 20:
67–82.
 Multiscale Analysis for the Exploitation of Bioresources 83

Water in the West (2013). Water and Energy Nexus: A Literature Review. http://waterinthewest.stanford
.edu/sites/default/files/Water-Energy_Lit_Review_0.pdf (accessed March 2020).
Wilson, Z.T. and Sahinidis, N.V. (2017). The ALAMO approach to machine learning. Computers & Chem-
ical Engineering 106: 785–795.
Wright, M.R., Bach, C., Gernaey, K.V., and Krühne, U. (2018). Investigation of the effect of uncertain
growth kinetics on a CDF based model for the growth of S. cerevisae in an industrial bioreactor. Chemical
Engineering Research and Design 140: 12–22.
Wu, B. and Chen, S. (2007). CFD simulation of non-Newtonian fluid flow in anaerobic digesters. Biotech-
nology and Bioengineering 99 (3): 700–711.
Xiong, Q., Yang, Y., Xu, F. et al. (2017). Overview of computational fluid dynamics simulation of
reactor-scale biomass pyrolysis. ACS Sustainable Chemistry & Engineering 5: 2783–2798.
Yang, L. and Grossmann, I.E. (2013). Water targeting models for simultaneous flowsheet optimization.
Industrial and Engineering Chemistry Research 52: 3209–3224.
Ye, Q., Cheng, J., Yang, Z. et al. (2018). Improving microalgal growth by strengthening the flashing light
effect simulated with computational fluid dynamics in a panel bioreactor with horizontal baffles. RSC
Advances 2018 (8): 18828–18836.
Yee, T.F. and Grossmann, I.E. (1990). Simultaneous optimization models for heat integration. II. Heat
exchanger network synthesis. Computers and Chemical Engineering 14: 1165–1184.
You, F. and Grossmann, I.E. (2010). Integrated multi-echelon supply chain design with inventories under
uncertainty: MINLP models, computational strategies. AIChE Journal 56: 419–440.
You, F., Tao, L., Graziano, D.J., and Snyder, S.W. (2012). Optimal design of sustainable cellulosic biofuel
supply chains: multiobjective optimization coupled with life cycle assessment and input-output analysis.
AIChE Journal 58 (4): 1157–1180.
Yue, D., You, F., and Snyder, S.W. (2014). Biomass-to-bioenergy and biofuel supply chain optimization:
overview, key issues and challenges. Computers & Chemical Engineering 66: 36–56.
Zainal, Z.A., Ali, R., Lean, C.H., and Seetharamu, K.N. (2001). Prediction of performance of a downdraft
gasifier using equilibrium modeling for different biomass materials. Energy Conversion and Management
42: 1499–1515.
Zarca, G., Urtiaga, A., Biegler, L.T., and Ortiz, I. (2018). An optimization model for assessment of
membrane-based post-combustion gas upcycling into hydrogen or syngas. Journal of Membrane
Science 563: 83–92.
Zhang, Y., Dube, M.A., McLean, D.D., and Kates, M. (2003). Biodiesel production from waste cooking
oil: 1.Process design and technological assessment. Bioresource Technology 89: 1–16.
Zhao, X. and Wang, Y. (2009). Gasoline blending scheduling based on uncertainty. In: Proceedings of the
2009 International Conference on Computational Intelligence and Natural Computing, Wuhan, 84–87.
Zhu, H.P., Zhou, Z.Y., Yang, R.Y., and Yu, A.B. (2007). Discrete particle simulations of particulate systems:
theoretical developments. Chemical Engineering Science 62: 3378–3396.
Zhu, H.P., Zhou, Z.Y., Yang, R.Y., and Yu, A.B. (2008). Discrete particle simulation of particulate systems:
a review of major applications and findings. Chemical Engineering Science 63: 5728–5770.
 4
Challenges in the Modeling of
Thermodynamic Properties and
Phase Equilibrium Calculations
for Biofuels Process Design
Roumiana P. Stateva1 and Georgi St. Cholakov2
1 Institute of Chemical Engineering, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
2 Department of Organic Synthesis and Fuels, University of Chemical Technology and Metallurgy, 1756
Sofia, Bulgaria

4.1 Introduction
Global warming and environmental pollution from fossil fuels have become major issues.
Fossil fuel resources are limited and create problems globally, but are of even greater con-
cern for Europe, which relies heavily on their import from other countries.
The alternative is biofuels, which are renewable and less polluting. They are the key to
the EU directive requiring 10% of all transportation fuels to be renewable by 2020, and
achieving 25% in 2030 and beyond without compromising agricultural production. Similar
solutions are implemented globally and concern also other energy needs (electricity, heat,
etc.), so the problem of intensification of the production and use of biofuels is a major issue
of the future development of our civilization.
The successful achievement of the described targets presents chemical and biochemical
engineering challenges that need to be addressed recurrently and evaluated systematically

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
86 Process Systems Engineering for Biofuels Development

both from the point of view of production of biofuels and of the evaluation of their
environmental impact. The major goal of the required green engineering is to design
computationally, from the molecules up, economically lucrative and environmentally
benign processes.
In the context of process design, biofuels pose some different problems as compared
with their major rivals – fossil fuels. Fuels are commodities that are relatively cheap, but
marketed in huge quantities. They are continuously produced from nonpolar hydrocarbons
in refineries, some of which process up to 62 million tons of crude oil per year (Cholakov
2016). In contrast, present day biofuels are produced in small biorefineries that are predom-
inantly batch oriented, process different biomass feedstocks, ranging from sugar and starch,
lignocellulose, algae and lipids to energy crops. In order to be competitive, they must turn to
circular continuous production and full, but alternative valorization of the potential of mixed
feedstocks not only into biofuels, but also into high value-added functional products. These
features complemented by preprocessing and co-processing, including even fossil resources
and, where appropriate, municipal solid wastes, outline the main requirements for future
smart biorefineries. The latter need integrated computer-aided process and product design,
which can systematically produce synergetic economically feasible and environmentally
compatible alternative optimal solutions.
In view of the above, the development and application of a robust, reliable and versatile
tool (a thermodynamic modeling framework – TMF) based on advanced scientific knowl-
edge for modeling and simulation of biofuel-related systems is of key importance, since fast
screening, evaluation and optimization of process alternatives, can provide rapid answers to
the challenges to the production and use of biofuels, before investment decisions are made.
Data produced by the TMF can be fed into commercial simulators for development of
flowsheets of installations and their economic evaluation, sustainability and environmental
analysis, etc. and will thus have a major role in the process of simulation and design of
innovative biofuel production processes.
An integral part of the TMF is the library of thermophysical properties, which should
include compounds relevant both for the production of biofuels and the high added value
products of potential interest for the smart biorefinery. It has to offer and update also meth-
ods for prediction of properties for which there are no measured data.

4.2 Thermodynamic Modeling Framework: Elements, Structure,

and Organization
The proper design of biofuel processes requires the application of a next generation rigorous
TMF that overcomes the existing challenges, with the potential to improve the accuracy and
provide high predictive ability and reliable description of the thermodynamics and kinetics
of the processes involved (in biomass conversion routes, and in separation and purification
technologies) at a wide range of operating conditions. Thus, there is a clear-cut need of
reliable, robust and efficient methods capable of describing the thermodynamics of mixtures
relevant to biofuels, the development of which, however, is a very demanding task as these
mixtures can exhibit intricate phase behavior.
Building a TMF, particularly focused on biofuel production processes, is far from trivial
as there are considerable challenges and difficulties characteristic of the modeling of phase
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 87

equilibria of the complex systems involved that should be taken into account and resolved,
namely:
• The mixtures are highly asymmetric, non-ideal, and may contain polar molecules with
strong intra-, intermolecular and associative interactions. Hence, they can exhibit com-
plex phase behavior that immensely complicates the modeling task involved in the sim-
ulation and design of biofuels processing (Reynel-Ávila et al. 2019).
• In multi-phase processes, the prediction of the correct phase configuration of the systems
that corresponds to the minimum of the Gibbs energy, requires rigorous thermodynamic
stability analyses routines, the convergence of which, however, is further complicated
by the necessity to imbed within those methods for estimation of the parameters of the
specific thermodynamic model applied.
• Last, but not least, in cases of biofuel processes that are performed under supercriti-
cal conditions (e.g. biodiesel synthesis by transesterification in supercritical short-chain
alcohol without catalysts) it would be useful to understand the phase behavior of the
system throughout the reaction, in order to identify the conditions that provide the best
reaction outcomes. Yet, the proximity to the critical point, which is mathematically sin-
gular, is a challenge to any model applied.
The structure of the TMF is modular and hierarchical, and, as shown in Figure 4.1,
includes three elements: (i) a library of the available thermophysical properties of pure
compounds and methods for estimating the missing ones; (ii) thermodynamic models for
mixture properties; and (iii) model-independent routines, namely methods, algorithms, and
numerical techniques, for performing stability analysis and solving the phase and chemical
equilibrium relations.
The elements of the TMF are highly interconnected. The library with pure component
properties and the set of methods available for estimating the missing ones, though situated

TMF–Flow of Information

Thenmodynamic Modeling
Framework

Thermodynamic
Phase Identification Models for Mixture
Methods and Properties
Flash Routines

Library of thermophysical
Stability Analysis– parameters of pure
Routines and substances; Methods for
Numerical Methods estimating the missing
properties.

Figure 4.1 Paradigm of the TMF organization.

88 Process Systems Engineering for Biofuels Development

on the lowest level, are of vital importance as the objects of the biofuel industry are often
compounds for which there are either very limited or completely missing property data
because many of the properties required cannot be measured experimentally.
When the missing properties required for a given compound are identified those are
estimated and the information is sent to the element thermodynamic models for mixture
properties. The performance of a thermodynamic model chosen to correlate and predict the
phase equilibria of mixtures at different stages of a biodiesel production (Cotabarren et al.
2014) is a test for the quality and reliability of the thermophysical properties estimated and
in some cases a different method/methods can be chosen and tried. The iterative process
will continue until a satisfactory agreement between the available experimentally measured
and the calculated phase equilibria data is achieved.
The third element of the TMF is of high significance because it is focused on predicting
and modeling the phase behavior exhibited by the systems of interest to biofuel production.
It includes reliable and robust methods for thermodynamic stability analysis of complex
non-ideal multicomponent systems at normal, and/or supercritical conditions, and a set of
efficient flash routines for both non- and chemically reacting systems, which are interwoven
in an appropriate software medium.
The reliability of the modeling of the thermodynamics of a biofuel process and the robust-
ness of the results obtained are contingent on the performance of each of the three elements
of the TMF, on the quality of information exchanged among them and efficiency of inter-
action.
The data produced by the TMF is used as a basis for developing flowsheets of biofuel
processes, and is an essential component of the consequent economic evaluation, sustain-
ability, and environmental analysis of the latter that can save time and money, because it
will advise which alternative solutions are promising and are worth investing in (Fornari
and Stateva 2015).
For the simulation, design, and development of technologies for biofuels production the
prediction of the stable phase configuration and calculation of the components’ distribution
among the equilibrium phases of reacting and non-reacting systems of interest is of utmost
importance. Hence, first a concise summary of the thermodynamics of the phase equilibria
of biofuel mixtures will be given.

4.3 Thermodynamics of Biofuel Systems

The majority of biofuel related systems exhibit either vapor–liquid equilibria (VLE) or
liquid–liquid equilibria (LLE), as well as, in certain limited cases, vapor–liquid–liquid
(VLL) behavior. In what follows, we will briefly discuss the thermodynamics of VLE
and LLE.

4.3.1 Phase Equilibria

The necessary condition for a closed, heterogeneous system consisting of 𝜋 phases and Nc
species to be in a state of internal equilibrium in terms of intensive quantities temperature,
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 89

pressure, and chemical potential are:

T (1) = T (2) = … = T (𝜋)
P(1) = P(2) = … = P(𝜋)
𝜇1(1) = 𝜇1(2) = … = 𝜇1(𝜋)
𝜇2(1) = 𝜇2(2) = … = 𝜇2(𝜋)
⋮
𝜇N(1) = 𝜇N(2) = … = 𝜇N(𝜋) (4.1)
c c c

However, because the chemical potential does not have an immediate equivalent in the
physical world, it is necessary to express it in terms of some auxiliary function that relates
it to physically measurable quantities such as T, P, and composition.
Such function is supplied by the concept of fugacity, and the above necessary conditions
for equilibrium can be rewritten, for example for a two-phase, multicomponent system, as
follows:
fiI (T, P, x, V) = fiII (T, P, y, V) (4.2)
For a VL equilibrium, the necessary equilibrium relations are then expressed as:
fiL = fiV (4.3)
where fiL and fiV are the fugacities of component i, where i = 1, 2, … , Nc , in the liquid and
vapor phases, respectively.
There are two different methods for the description of the fugacity of the ith component
in the liquid and vapor phases, namely:
• A symmetric, also called phi-phi method, for which Eq. (4.3) will be expressed as:
fiL = xi 𝜑Li P
fiV = yi 𝜑V
i P (4.4)
where xi and yi are the mole fractions, and 𝜙Li and 𝜙V i
are the fugacity coefficients of the
ith component in the liquid and vapor phase, respectively.
• An asymmetric, also known as the gamma-phi method, where the fugacity of the ith
component in the liquid phase is expressed by the activity ai of the ith component via
the relation:
ai (T, P, x1 x2 , … , xN ) = fi (T, P, x1 x2 , … , xN )∕fi0 (T) (4.5)
Introducing further the activity coefficient 𝛾 i , defined as:
a
𝛾i = i (4.6)
xi
The VL equilibrium relation can be expressed as:
p
fiL = xi 𝛾i fi
fiV = yi 𝜑V
i P (4.7)
90 Process Systems Engineering for Biofuels Development

In a complete analogy, the necessary condition for LL equilibrium can be expressed as:
L1 L2
fi = fi (4.8)
where L1 and L2 pertain to liquid phase 1 in equilibrium with liquid phase 2, respectively.
Again, for the description of the fugacity of the ith component in the equilibrium L1 and
L2 phases either of the above two approaches can be applied.
To calculate the distribution of the components between the equilibrium phases, follow-
ing any of the two approaches, a suitable and reliable thermodynamic model via which the
fugacity and/or the activity coefficients can be calculated is required.
The element of the TMF “Thermodynamic models for mixture properties” comprises
one of the three main elements of the TMF, and, in what follows, we will briefly discuss
two groups of models commonly used to model the phase equilibria of biofuel pertaining
systems, and some of the challenges associated with their application.

4.3.2 Thermodynamic Models

Robust thermodynamic models that can describe reliably the complex systems involved
in biofuel related processes are an essential element of the TMF. The perfect thermody-
namic model would be the one that employs easily measured physical properties to predict
phase equilibria at all conditions and is, in addition, theoretically based. Nevertheless, till
present no current model satisfies completely those requirements, as discussed by Fornari
and Stateva (2015).
In what follows, we will present in a concise manner two groups of models often used
for the design, optimization and debottlenecking of biofuel processes.

4.3.2.1 Equation of State

Equations of state (EoSs), in combination with different mixing rules, are currently often
the choice to model the phase behavior of biofuels related systems since, in general (and in
particular cubic EoSs) they are exceedingly simple.
The general form for of a cubic EoS may be written as:
RT a(T)
P= − (4.9)
Vm − b (Vm + c1 b)(Vm + c2 b)
where Vm is the volume of the system, and a(T) and b are the attraction and repulsive terms,
respectively.
All of the most widely used modified cubic EoSs conform to this general type and are
characterized by integer values of the parameters c1 and c2 .
An extensively used cubic EoS is the Soave–Redlich–Kwong (SRK):
RT a(T)
P= − (4.10)
v − b v(v + b)
where the attraction and repulsive terms for pure components are related to:
R2 Tc2
a = 0.42747 𝛼(Tr ) (4.11)
Pc
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 91

RT c
b = 0.08664 (4.12)
Pc
where 𝛼(Tr ) is expressed in terms of the acentric factor 𝜔 as:
1∕2
𝛼(Tr ) = [1 + (0.480 + 1.574𝜔 − 0.176𝜔2 )(1 − Tr )] (4.13)

To represent mixtures, a mixing rule, for example the one fluid van der Waals, is introduced:
∑∑
amix = xi xj aij (4.14)
i j
∑∑
bmix = xi xj bij (4.15)
i j

where amix and bmix are, respectively, the mixture energy and co-volume parameter of
the EoS.
Usually, a geometric mean rule is applied to determine the cross-energy parameter,

aij = (aii ajj )0.5 (1 − kij ) (4.16)

and either an arithmetic mean (conventional linear rule)

∑
Nc
bmix = xi bii (4.17)
i=1

or a quadratic rule for the cross-co-volume parameter are used

( )
bii + bjj
bij = (1 − lij ) (4.18)
2
When lij is set to zero, the mixing rule is usually referred to as the one-binary
interaction-parameter-per-pair version of the van der Waals mixing rule or 1PVDW.
If Eq. (4.10) is rewritten in terms of the compressibility factor Z = PV
RT
, then the SRK
EoS is:
Z 3 − Z 2 + (A − B − B2 )Z − AB = 0 (4.19)

where
aP
A= (4.20a)
(RT)2
bP
B= (4.20b)
RT
The fugacity coefficient can be calculated according to:
( ∑N )
( )
bi A 2 j=1 xj aij bi B
ln 𝜙i = (Z − 1) − ln(Z − B) − − ln 1 + (4.21)
b B a b Z

Hence, the critical temperature and pressure, and the acentric factor of each of the pure
components in the system at issue must be known in order to calculate its fugacity coeffi-
cient.
92 Process Systems Engineering for Biofuels Development

Within the framework of the symmetric phi-phi method, EoSs can be applied to model
both VLE and LLE at wide ranges of temperatures and pressures, including the supercriti-
cal region. However, taking into consideration that cubic EoSs were originally developed to
model hydrocarbon systems, in order to improve their performance and extend their appli-
cation to strongly non-ideal systems, two paths usually can be pursued – improvement of
mixing rules and improvement of the existing EoS and/or development of new models.
With regard to the first option, an attractive alternative to e.g. 1PVDW or 2PVDW mixing
rules is to integrate into the EoSs activity coefficient models. The resulting thermody-
namic models demonstrate enhanced performance since they integrate the advantages of
successful cubic EoSs and models like e.g. UNIFAC, NRTL, etc. Some of those, and in par-
ticular the group contribution (GC) EoS and the GC with association EoS were proven to
have excellent predictive capabilities and to very successfully represent the complex phase
behavior of mixtures containing natural products (Fornari and Stateva 2015) and biofuels
(Soria et al. 2011). The interested readers can find further information in Fornari and Stat-
eva (2015). Still, it should be noted that the necessity to have information about the pure
compounds physical properties is not alleviated.
With regard to development of novel efficient thermodynamic methods, a real advance
in the modeling of polar and highly non-ideal systems came with the advancement of more
rigorous explicit association models like the statistical associating fluid theory (SAFT)
equation, which was adopted by both the academic and industrial communities. The
approach is very successful in modeling the phase equilibria of complex systems, see for
example Pereira et al. (2016), Kanda et al. (2018), Rodriguez and Beckman (2019), and
Abala et al. (2019), to name just a few.
The performance of SAFT and perturbed-chain statistical associating fluid theory
(PC-SAFT) equations for mixtures containing strongly polar or hydrogen-bonding com-
pounds is superior to that of cubic EoSs. However, in comparison with the latter, the amount
of information required about the characteristic parameters of the pure compounds by the
SAFT-type equations is increased and the computations are considerably more demanding.
More importantly, as the number of adjustable parameters increases, the parameter
estimation problem becomes more complicated due to parameter identifiability issues.
In what follows we will focus briefly on the challenges inherent to the application of
cubic EoSs to model the phase behavior of biofuel related systems.

4.3.2.2 Determining the Model’s Parameters

The binary interaction parameters kij and lij are of the same order of magnitude and their
numerical values are in the range of −1 to +1. As discussed by Fornari et al. (2010), a
negative value of kij indicates that specific chemical interactions, such as hydrogen bonding,
are present in the mixture, however, there are no conclusive indications how to interpret a
negative value for lij .
The values of the interaction parameters are found from the best fit to available exper-
imental data and their determination is performed by solving a nonlinear optimization
problem, which may be challenging, and far from trivial.
On the one hand, if local optimization methods are used, the values of the parameters
obtained depend on the searching interval and on the initial values of the interaction param-
eters used to start the iterative procedure. Thus, those values correspond to a local rather
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 93

than to the global minimum of the objective function and represent just one of the possible
sets. If, on the other hand, a global optimization method is applied then the parameters’
values located will be the ones corresponding with the global minimum, but the numerical
effort involved will increase considerably with increasing the dimension of the problem.
An additional problem in adjusting the model parameters to match experimental data is
that because for some systems the experimental measurements are challenging there is a
large scatter of the data, and the solver fails to converge to some experimental points that
are either too close to the critical point and/or phase boundary or seem not to follow the
general trend. Hence, it might be decided to ignore such “difficult” points in the course of
parameter determination, which might influence in a profound manner the ability of the
model to predict correctly the phase behavior (Fornari et al. 2010).
Furthermore, it is often considered safe and acceptable, in the cases when the binary inter-
action parameters of a pair of components are not available as a result of insufficient data,
to set those to zero. This, however, can lead to inaccurate phase equilibrium predictions,
particularly for strongly non-ideal systems.

4.3.2.3 Organization of the Equilibrium Calculations When Applying an EoS

Higher level algorithms for simulation and design, e.g. the flash routines, based on the
phi-phi approach have a hierarchical structure of organization. On the lower level is
the EoS routine, which has an extremely important role – to provide to the higher level
algorithm valid values for the density (volume) of the system under consideration. The
application of the phi-phi model in predicting and calculating the phase behavior of
multicomponent systems is often preferable. The reason behind that is that it requires few
parameters to describe fluid behavior adequately over a wide range of temperatures and
pressures. Furthermore, a consistent prediction of mixture properties in the sub- and/or
supercritical region is usually achieved.
Some of the challenges associated with the prediction and calculation of phase equilibria
applying an EoS as the thermodynamic model, are a result of computational pitfalls inherent
to the nature of the EoS, while others are a result of a poor organization of the calculations
in the higher level algorithms.
For example:
• A thermodynamically stable phase equilibria configuration satisfies the equi-fugacity (a
necessary) condition – Eq. (4.2) – and corresponds with the global minimum of the Gibbs
energy (sufficient condition). The practice of searching for all roots to the equi-fugacity
condition and testing for phase stability appears not to be a widespread practice, as
pointed out by Fornari et al. (2010). Yet, this is of immense importance in the design
of any biofuel process, particularly in the cases of possible complex phase behavior,
when, at certain values of temperature, pressure, and kij , and lij multiple solutions to the
equi-fugacity condition can exist, and the incorrect rather than the correct phase config-
uration can be accepted as the solution. Moreover, because phase equilibrium equations
are often difficult to converge in the critical region, the quality of initial estimates used
is crucial. Taking into consideration that inappropriate estimates can lead to the trivial
solution, the availability in a TMF of a robust stability analysis routine that will provide
a good set of initial estimates and will guarantee steady convergence of the flash routines
is of huge importance.
94 Process Systems Engineering for Biofuels Development

• In the phi-phi model the EoS is an integral part of the higher level simulation algorithms
like the LV, LL, LLV, etc. flash routines. Its fundamental role is to supply a value for the
mixture density (or volume) of each phase at a given pressure, temperature, and com-
position. The EoS model interacts with the higher level algorithms and provides liquid
and vapor density values, which are then used by the latter to evaluate key thermody-
namic properties like the fugacity coefficients. However, during the iteration process,
often the EoS is “fed-back” by higher level algorithm conditions, which are either far
from the physical solution or where one phase physically does not exist. Obviously, under
such infeasible conditions, and particularly in the cases of phi-phi models, where both
phases are described by one and the same thermodynamic model, the latter may return a
liquid-like density for the vapor phase or vice versa.
If the infeasible specifications problem is not taken care of, the higher level algorithms
will either yield unreasonable results, or converge to the trivial solution or not converge
at all.
To take care of infeasible specifications different extrapolating methods that should pro-
vide values for pseudoproperties (e.g. pseudofugacity coefficients) can be adopted. The
methods should guarantee that the pseudoproperties are continuous with the true values,
mimic the physical behavior and follow the trends that are characteristic for each true prop-
erty and each phase. Lastly, but most importantly, the extrapolating methods should improve
the convergence characteristics of the higher level algorithms – see for example Watson and
Barton (2017). In view of this, an analysis of the influence of the properties’ extrapolation,
during the course of iterative calculations, on the convergence pattern of higher level simu-
lation algorithms, is necessary as discussed in detail by Watson and Barton (2017), Stateva
and Tsvetkov (1992), and Stateva et al. (1990).

4.3.2.4 Missing Data on Thermophysical Properties of the Pure Components

Constituents of a Mixture
The application of an EoS to predict and model the phase behavior of mixtures requires
knowledge of the pure component properties, e.g. their critical temperature and pressure.
For some of the pure compounds of interest to biofuels production, there are no data avail-
able for the properties required as they have either not been measured, or it is not possible
to measure them experimentally. Hence, they have to be estimated by different methods, as
will be discussed in Section 4.5.1. These hypothetical values might have a profound effect
on the quality of the phase equlibria predictions and calculations as will be discussed in
more detail in Section 4.7.

4.3.2.5 Activity Coefficient Models

An appropriate EoS provides a thermodynamically consistent route to the fugacity of com-
ponents in both the vapor and liquid phases and thus offers a very convenient basis for phase
equilibrium calculations. Yet, in cases of complex systems with strong specific interactions
they might fail to give good results for the liquid phase. In those cases, better results can
be obtained when the fugacity of the components in the liquid phase are estimated from an
activity coefficient model.
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 95

The best activity coefficient models are semi-theoretical and allow the calculation of
activity coefficients for both binary and multicomponent systems over a range of tempera-
tures requiring only experimental input from the binary constituents at one temperature.
To determine the parameters of any activity coefficient model usually one of the following
approaches is used.
Minimization of the difference between existing experimental VLE data on binary mix-
tures and model predictions, usually at a single temperature, for all possible binary pairs
is the first one. Then, the model can be used to calculate activity coefficients in a multi-
component mixture over a range of temperature and pressure. Hence, though the model
permits extrapolation with respect to temperature or pressure, it is just a correlation for
binary systems. For multicomponent systems, however, it can be considered as truly pre-
dictive. Typical examples are the Wilson, T-K-Wilson, NRTL, and UNIQUAC equations.
The alternative approach is to estimate the parameters applying a GC method, which does
not require experimental data. The GC principle postulates that molecules can be broken
down into functional groups that can be treated as being independent of the rest of the
molecule, the functional group parameters can be determined by regression against a very
large data base of, typically, experimental VLE results. The UNIFAC equation is an example
and it is considered by some authors as completely predictive.
We will very briefly discuss two representatives of activity coefficient models used often
to model the phase behavior of biofuel related systems.

4.3.2.6 Non-Random Two-Liquid Model

The non-random two-liquid (NRTL) model was developed in an attempt to overcome the
inadequacy of the Wilson equation to describe LLE. It contains three parameters per binary
interaction and is based on similar local composition treatment. The three parameters are
independent of temperature – the first two signify energy parameters while the third is
purely empirical.
Provided the NRTL model is applied to model the LLE of a biofuel related system, then
rewriting Eq. (4.8) in terms of the activity coefficient of component i, the equi-fugacity
relation will be:
(xi 𝛾i )LI = (xi 𝛾i )LII (4.22)
The activity coefficient is given by:
∑
Nc
⎛ ∑Nc
⎞
𝜏ki xk Gki ⎜ 𝜏k xkj Gkj ⎟
∑ xj Gij
Nc
ln 𝛾i =
k=1
+ ⎜𝜏ij − k=1 ⎟ (4.23)
∑
Nc
j=1 ∑
Nc ⎜ ∑Nc ⎟
xk Gki xk Gkj ⎜ xk Gkj ⎟
k=1 k=1 ⎝ k=1 ⎠
where
Gij = exp(−aij 𝜏ij ) (4.24)

𝜏ij = gij ∕RT (4.25)

and
𝜏ij ≠ 𝜏ji , 𝜏ii = 𝜏jj = 0 (4.26)
96 Process Systems Engineering for Biofuels Development

In the above, gij are energy parameters that characterize interaction between molecules i
and j, 𝛼 ij is the non-randomness factor in the mixture, 𝜏 ij are dimensionless interaction
parameters, and R is the universal gas constant.

4.3.2.7 The Universal Quasichemical Model

The universal quasichemical (UNIQUAC) equations were developed and introduced by
Abrams and Prausnitz (1975), who applied a semi-theoretical approach to the mixture prob-
lem which includes a local composition model. The pioneering concept was based on the
recognition that the non-ideality of liquid mixtures has contributions not only from specific
interactions but also from the differences in the size and shape of the molecules. Thus the
Gibbs free energy of a mixture is correlated by the sum of two separate terms:

GEm ∕RT = (ΔGcm + ΔGrm )∕RT (4.27)

which include a contribution, known as the configurational term, owing to the differences
in size and shape; and another one, known as the residual term, owing to the energetic
interactions between the molecules.
Subsequently, the logarithm of the activity coefficient for each component also has a
configurational and residual contribution.
The configurational term of the activity coefficient is given by:
( ) ( )
𝜙i ∑
n
𝜙i z 𝜃i
ln 𝛾i = ln
c
+ ln + li − xl (4.28)
xi 2 xi xi j=1 j j

and the residual by:

⎡ ( n ) ⎤
⎢ ∑ ∑n 𝜃
′
𝜏 ij
⎥
ln 𝛾ir = qi ⎢1 − ln ⎥
′ ′ j
𝜃k 𝜏ki − (4.29)
⎢ j=1 ∑ 𝜃 ′ 𝜏
n ⎥
⎢ k=1
k kj ⎥
⎣ k=1 ⎦
where ( )
uij
𝜏ij = exp − (4.30)
RT
z
li = (ri − qi ) − (ri − 1) (4.31)
2
In the above ri is determined by the volume and qi by the surface area of the molecule
both of which, in theory, may be determined from crystallographic measurements on the
pure solid. Yet, both parameters are typically obtained by a GC method. Furthermore, there
are tables which contain information allowing the determination of ri and qi for a very large
selection of molecules.
Z is a co-ordination number to which the value 10 is normally, but not necessarily,
assigned. Lastly, the model contains two adjustable binary energy parameters u12 and u21
that participate in the residual part of the excess Gibbs free energy through the quantities 𝜏 ij .
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 97

The pure substance parameters are given by:

xi ri
𝜙i = (4.32)
∑
n
xj rj
j=1
xi qi
𝜃i = (4.33)
∑n
xj qj
j=1
′
′ xi qi
𝜃i = (4.34)
∑
n
′
xj qj
j=1

where 𝜙i and 𝜃 i are the area fractional and the molar weighted segment components for
the ith molecule in the total system, respectively; ri and qi are calculated from the group
surface area and volume contributions according to:
∑
ri = 𝜈k Rk (4.35)
∑
k
qi = 𝜈k Qk (4.36)
k

where the dimensionless parameters Rk and Qk are a measure of the van der Waals molecular
volume and surface area and are available in tabulated forms.
The UNIQUAC model is one of the most successful two-parameter models for LL pre-
dictions and calculations and requires pure component and binary parameters to apply it to
multicomponent mixtures.
The challenges associated with the application of activity coefficient models can be con-
cisely summarized as follows:
• Analogously to EoSs, the binary interaction parameters 𝜏 ij and 𝜏 ji are considered as
adjustable parameters and the goal is to determine values for them that will provide the
best fit to the experimental data available.
The procedure most often employed is based on some type of least squares or maxi-
mum likelihood criterion and requires the solution of a nonlinear optimization problem.
As discussed previously, the objective function in such nonlinear parameter estimation
problems is often non-convex and thus has several local minima. Furthermore, the meth-
ods typically used to solve such problems are local rather than global solvers and hence
do not provide any guarantee that the global optimum has been found. Thus, in most
of the cases, the sets of the binary interaction parameters located represent just one of
the possible sets of parameters. The latter in some cases might lead to a physically not
correct prediction of the phase behavior, e.g. prediction of an erroneous phase spit.
Along those lines, an additional problem is the necessity to predict and model, e.g.
the LL phase behavior of a system for which just experimental VL data are available
only. From a physical point of view these interaction parameters should be equally useful
for the LLE. However, because of the complexity of the parameter determination, the
extrapolation is not always successful.
98 Process Systems Engineering for Biofuels Development

• For some complex molecules the ri and qi parameters are not known and those should be
determined. Usually, these structural parameters can be estimated using semi-theoretical
methods like that of Bondi. However, a potential pitfall here is associated with the correct
determination of the respective groups building the given molecule, for instance, cafestol
and kahweol diterpenoids food additives, found in spent coffee grounds.
In addition, the ri and qi parameters must be known before the optimum values of the
binary interaction parameters are determined, and hence any insufficiency that exists in
the selection of these structural parameters will inevitably affect the group interaction
parameters determination.
• Some complex molecules contain functional groups for which parameters are
unavailable. And even if all group definitions exist, a subset of the relevant parameters
is unavailable. Hence, those must be estimated from limited experimental data and most
probably their values will be valid only for the particular systems discussed.
• For GC approaches like UNIFAC the interaction parameters between groups are deter-
mined by a fit of the model to experimental data. If, for example, the UNIFAC model
is considered, for the determination of its parameters more than 60 different functional
groups are defined simultaneously in an extensive optimization involving minimization
of the total deviation between the UNIFAC model and huge amounts of experimental
data. Because the optimization problem is extremely complex and with high dimension,
parameter values are predominantly correlation values rather than describing the actual
physics in each particular interaction, which limits the possibility to extrapolate the use
into different situations (Fornari and Stateva 2015).
The third element of the TMF is the library of thermophysical properties, the creation
of which starts with searching into the available sources of data for biofuels process
design.

4.4 Sources of Data for Biofuels Process Design

Measured data provide the best estimation of the processes being modeled, but are also
essential for the development of property prediction methods, which cannot be developed
without measured data. They are more precise than predicted data because the latter always
have the uncertainties of the experimental data, from which they are obtained, in addition
to those due to the prediction.
The primary sources of experimental data are their respective original publications.
However, process design needs measured data of high quality in terms of availability of
measurement methods at time of publication, uncertainties, access to original papers, etc.
Hence, the data for chemical and biochemical engineering process design are collected,
critically evaluated, updated, processed, and filled into relevant databases by well recog-
nized organizations, often affiliated with government agencies and/or with the respective
professional societies.
Published experimental data by definition should be the same in all chemical engineering
databases, but different data experts usually recommend their own estimations. Moreover,
when there are no measured data, data experts might suggest estimated values.
Many owners of databases, identify, finance and direct measurements of data of particular
interest to the engineering community. A typical example is the Design Institute for Physical
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 99

Properties (DIPPR) of the American Institute of Chemical Engineering, which with the help
of sponsors, presently runs four projects:
DIPPR Project 801 has the task to evaluate, organize, update, and expand the existing
DIPPR database with thermodynamic and physical property data. As stated by DIPPR, it
includes 2332 industrially important compounds; values for 49 thermophysical properties
for compounds in the database, including raw data and expert recommended values for
34 constant properties (normal boiling point, critical properties, acentric factor, autoigni-
tion temperature, temperature and percent upper and lower flammability limits, etc.) and
equation coefficients for 15 temperature dependent properties (liquid and solid density, heat
capacity of ideal gas, of liquid and of solid, surface tension, viscosity of liquid and of vapor,
etc.). When experimental chemical data are not available, the project recommends predicted
values as well as references, notes, and quality codes for all data points.
DIPPR Project 805 aims to obtain experimentally physical properties of industrially
important compounds with emphasis on phase equilibria and other data for binary and
multicomponent mixtures; presently available are measured data for 315 systems.
DIPPR Project 851 is devoted specifically to measuring data on critical properties of pure
compounds. It works also on developing measurement methods for unstable compounds.
DIPPR Project ESP monitors 57 physical properties relevant to environment and safety
for more than 1080 chemicals; data for about 39% of them are from Project 801.
In order to cover costs, most of the database developers charge for obtaining particu-
lar data, for subscription or licenses. A useful free database (https://www.chemeo.com) is
maintained by Céondo GmbH in Germany. It offers experimental and/or calculated, but
clearly identified property values for pure compounds. The Russian “Chemister” database,
available also in English, offers on the internet free experimental data, collected from pub-
lished sources and data predicted by the method of Joback and Reid (Joback and Reid 1987).
The National Institute of Standards and Technology (NIST) of the U.S. Department of
Commerce is an exemplary combination of free and paid databases. It presently offers 75
free Standard Reference Databases (SRD), 19 of which contain critically evaluated data that
can be used in chemical engineering process design. The very popular Chemistry Webbook,
provides free data for properties of pure compounds, complemented by paid data from
the well-known Texas Research Center (TRC) commercial database. Information about all
SRD can be found from the sites of NIST. The principles of their critical evaluation are
described in one of their latest publications (Kaiser and Hanish 2018). The latest NIST
Reference Fluid Thermodynamic and Transport Properties Database (REFPROP) has 147
pure fluids, 5 pseudo-pure fluids and mixtures with up to 20 components.
The Dortmund Data Bank Software and Separation Technology (DDBST) is considered
the largest database of thermophysical properties of pure components and their mixtures,
especially on phase equilibria. It also includes own measurements and has a powerful soft-
ware package for data handling, correlation, property estimation and process synthesis.
DDBST is a close partner of DECHEMA and employs experts from Germany, Austria, and
other countries who develop highly regarded original correlations and reviews, relevant
to biofuel process design (Wallek et al. 2013, 2018). Its UNIFAC Consortium is respon-
sible for revising and extending GC methods – Modified UNIFAC (Dortmund), Original
UNIFAC, Predictive Soave–Redlich–Kwong, and Volume-translated Peng–Robinson.
The development of Ge models, especially the original UNIFAC, is one of the highly
regarded contributions also of CAPEC (the Computer-Aided Process Engineering Center,
100 Process Systems Engineering for Biofuels Development

presently part of the Prosys Research Center) of the Technical University of Denmark
(Mustaffa et al. 2014). However, it is even better known globally for achievements in the
theory of computer-aided process and product engineering, and their practical realization
in widely recognized group/bond contribution methods (Damaceno et al. 2018a; Kalakul
et al. 2018). It maintains one of the best updated databases, containing only critically
evaluated experimental values. A significant amount of its recent activities is targeted on
property prediction for lipids (Cunico et al. 2013), which is also relevant to biodiesel and
drop-in biofuels.
Experimental and predicted data can be also obtained from monographs, text books,
handbooks, and other reference sources. However, this cannot be recommended, since their
data are not regularly updated and even probably the best recognized of them (Poling et al.
2001) will leave teen age this year.
Some of the databases, available originally only as printed handbooks, are presently
offered online in an interactive electronic form. A typical example is the widely used “Yaws’
Critical Property Data for Chemical Engineers and Chemists” (Yaws 2012) database. Its
online version, last updated in 2017, offers values for thermophysical, thermodynamic, and
other properties of more than 5000 inorganic and 35 000 organic substances, complemented
by properties of materials and other engineering databases. The used data sources and/or
methods for their estimation are identified, but there is no clear distinction between exper-
imental and estimated data, unless the calculations have been done by the authors.
Databases contain three types of property data – data for constant properties, data for
temperature dependent properties of pure compounds, and data for phase equilibrium prop-
erties of mixtures. Some of the more general challenges, relevant also to biofuels design
data needs can be summarized as in the following.
Experimental property data are much more expensive than predicted (often by several
orders of magnitude) and in some cases impossible to obtain, at least at present. Typical
examples involve compounds which polymerize or degrade at measurement conditions,
compounds not in the relevant state at measurement conditions and compounds that cannot
be analyzed because of safety reasons (e.g. toxic, explosive, etc.). The challenges in the
measurement of critical properties can be illustrated with published data for DIPPR Project
851, which, between 1985 and 2004, studied the properties of 168 compounds for Tc , 134
for Pc , 60 for Vc , and 19 for dc , respectively. After 2004, Project 851 added data for 64
compounds, 40 of them new (Wilson et al. 2018).
The above problem has led to lack of measured data and abundance of data, predicted
by different methods. This postulates a need for estimation of the quality of the available
measured and predicted data. This is a challenge of serious importance which might prede-
termine the success of any modeling. There are specific methods, proposed by developers
of predictive methods for addressing it, for instance by testing the consistency of VLE data
by the QVLE factor and taking it into account in the correlations (Mustaffa et al. 2014).
Biofuel design engineers typically use computer simulators with built-in databases and
look only for missing data. In such cases it is important to check if measured data are clearly
identified. Some databases, when there are more than one published experimental data, sug-
gest an average (e.g. of 89 normal boiling point measurements for methyl ethyl ketone).
Averaging, however, might not be appropriate, because a significant number of measure-
ments have been done at the beginning of the twentieth century and even at the end of the
nineteenth century (Cholakov et al. 1999). Serious differences in values, recommended by
two of the recognized databases, have been also reported (Brauner et al. 2005).
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 101

It is also a good practice to compare the recommended values with those for compounds
with similar chemical structures and eventually check the original publications. For some
properties there are simple options for testing consistency. For instance, for predicted
critical properties it has been demonstrated (e.g. by Coelho et al. 2018) that the Tc /Pc
ratio should be close to the Tc /Pc ratio calculated from the UNIFAC van der Waals surface
area as recommended by Žbogar et al. (2006). The critical temperature is always higher
than the normal boiling temperature. In homologous series the relationships between
property values and number of CH2 groups are asymptotic and approach a limiting
value.
The above challenges are the same in the process design for biofuels. There are also
specific challenges, some of which are:
• Biofuels include very large groups of intermediates and products, which differ in feed-
stocks, transformation processes, structures, and properties, end use, etc. For instance,
products of pyrolysis and gasification can be used directly for energy or, after further
treatment, as energy carriers and/or as feedstocks for synthesis of alcohols and ethers,
hydrocarbons, basic chemicals, etc. The so called XtL or resource X (e.g. biomass)
to liquid technology can be done with different raw biomaterials (DECHEMA – VDI
2018), non-edible glycerides may be transesterified to biodiesel additives with differ-
ent low molecular mass alcohols or hydrogenated to hydrocarbon “drop-in” fuels and
so on.
• In order to be competitive and realize circular chemical and biochemical transformation
of adequate amounts of feedstocks with specified property ranges, advanced biore-
fineries should co-process mixed feeds, eventually including also fossil raw materials,
and look for synergy of production processes (DECHEMA – VDI 2018). For instance,
production of biodiesel should not only extract glycerides from mixed solid waste
biomasses, e.g. from spent coffee grounds (Cholakov et al. 2013), industrial grape
seeds (Coelho et al. 2018), wine marc, etc., but also obtain high added value functional
additives – polyphenols, sterols, tocopherols, caffeine, diterpenes (Georgieva et al.
2018), etc. Biofuel process design in this case should also include the utilization of the
depleted solid matrixes, for instance, into active carbon that might be used as adsorbent
for the waste waters and/or as a transesterification catalyst carrier.
• The above two challenges outline the need for experimental or predicted data for the
properties of numerous structurally different pure compounds and mixtures. The latter
might be very broadly classified into two groups for biofuels process design – relatively
lower molecular mass compounds like gases (methane, ethane, CO and CO2 , H2 ), alkyl
alcohols (methanol, ethanol, butanol), alkyl ethers (methyl tert-butyl ether [MTBE], ethyl
tert-butyl ether [ETBE], dimethyl and diethyl ethers), etc. and higher molecular mass
compounds like fatty acids, mono-, di- and triacylglycerides, methyl, ethyl and butyl
esters of fatty acids, polyphenols, tocopherols, diterpenes, etc.
• The immediate challenges being addressed, as evidenced from DIPPR Project 851, are
the critical properties of the compounds in the second group. Critical properties are
among the primary constant properties. They are employed in all cubic EoSs, in the
calculation of the acentric factor, in the extrapolation to higher temperatures of liquid
densities, heats of vaporization, surface tension, liquid viscosities and other important
physical parameters, used in biofuels process design. Table 4.1 illustrates the availability
of critical data for relevant high molecular mass compounds.
102 Process Systems Engineering for Biofuels Development

Table 4.1 Availability of Tc and Pc data needed for biodiesel process design.

Compounds Tc Pc
Saturated fatty acids (FA) C8:0 –C22:0 (without C13:0 ) Yes No
Unsaturated FA C8 –C22 No No
Saturated methyl esters of FA C8 –C22 (only up to C18:0 , without C13:0 ) Yes Yes
Unsaturated methyl esters of FA (C18:1 , C18:2 , C18:3 , and C22:1 ) Yes Yes
Saturated ethyl esters of FA C8 –C22 (without C13 and up to C16 only) Yes Yes
Unsaturated ethyl esters of FA C8 –C22 No No
Butyl esters of FA (any) No No
Monoglycerides No No
Diglycerides No No
Saturated triglycerides of FA (only C8 , C10 , C12 , and C14 ) Yes Yes

The table includes the latest measurements on biodiesel relevant compounds (Nikitin
and Popov 2015, 2016a,2016b, Bogatishcheva et al. 2017). Such compounds are unsta-
ble at high temperatures, and are analyzed with techniques especially designed for such
measurements.
Cunico et al. (2013) have shown that the CAPEC database contains limited amounts of
experimental data on lipids, not only for Tc and Pc , but also for six other primary proper-
ties and properties for phase equilibrium systems. For standard Gibbs energy of formation,
standard enthalpy of formation, and standard enthalpy of fusion the data are for less than
10 pure lipid compounds, and for normal melting and normal boiling points for around 40
compounds. Equilibrium data are also not abundant – for mono- and diglycerides, respec-
tively, there are data only for 2 and 4 solid–liquid equilibria binary mixtures. It should be
noted, however, that recently the measured data for biofuel process design were comple-
mented with comparatively significant amounts of new contributions (Belting et al. 2012,
2013; Damaceno and Ceriani 2017; Damaceno et al. 2018b; Garmus et al. 2019). A seri-
ous increase of data for eight constant properties of pure compounds was reported for the
CAPEC database (Cunico et al. 2013, 2014).
When all measured data needed for biofuels process design are not available, the rest of
the data have to be predicted.

4.5 Methods for Predicting Data for Biofuels Process Design

The methods for predicting data for process design may be classified by different criteria.
According to the type of properties predicted there are methods for predicting constant prop-
erties of pure compounds, methods for predicting temperature dependent properties of pure
compounds, and methods for predicting properties of mixtures. In the first group, methods
devoted to the most frequently used primary properties prevail; in the second group, liquid
density, viscosity, surface tension, heat capacity and liquid vapor pressure are among those
most used for biofuel process design. The third group typically employs phase equilibria
data and Ge -based models.
According to the type of relationship targeted, prediction methods typically use either
properties to estimate properties or chemical structures to predict properties. The first group
is probably the oldest (Watson and Nelson 1933). Such methods are still widely used in
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 103

crude oil processing design for middle and heavier cuts, the composition of which are
presented with physical pseudo-components (fractions, boiling in a narrow range) and/or
chemical lumps (Cholakov 2011). Typically, they employ easily measured properties to pre-
dict other properties, which are more difficult to measure with the needed precision (Riazi
and Daubert 1987). In a modern interpretation (Varamesh et al. 2017; Hemmati-Sarapardeh
et al. 2018) these methods can be used with and for pure component properties.
The most used methods that predict properties from chemical structure are the GC meth-
ods and the quantitative structure–property relationships (QSPRs). Molecular dynamics
simulation methods can in principle also predict thermophysical properties of compounds
and materials at the nano scale (Cagin et al. 1999). However, they are still very expensive
and used mostly in computer-aided drug design and molecular biology (Hollingsworth and
Dror 2018; Liu et al. 2019). They are not discussed here.

4.5.1 Group Contribution Methods for Biofuels Process Design

GC methods are the most widely used methods, which have been and are still serving the
chemical engineering community for process and product design. Suitable GC correlations
have been developed for prediction of all three types of properties, discussed above. They
are offered in chemical process simulators and used in biofuel design exemplary modules,
e.g. of reactive distillation for MTBE or transesterification for biodiesel.
The relationships for the prediction of constant, temperature dependent and mixture prop-
erties of biofuel design relevant compounds have been comprehensively presented by their
authors and recently reviewed (Cunico et al. 2013, 2014; Ceriani et al. 2013; He et al. 2018).
In this text we shall use only the prediction of constant properties in order to outline the
principles of GC methods.
The most popular correlations for constant properties are presented, illustrated with
examples and compared in Poling et al. (2001), which is a recommended reference source,
complemented with respective tabulated GC values, and checked for errors. Among the
methods reviewed are those of Joback and Reid, Wilson and Jasperson, Marrero and
Pardillo, and Constantinou and Gani. Previous editions of this reference describe the older
methods of Lydersen and of Ambrose.
We shall outline briefly the principles of GC methods with the method of Marrero and
Gani as described by its authors (Marrero and Gani 2001) and later reviewed for lipid
compounds by Cunico et al. 2013. The method is semi-empirical and uses three levels of
sophistication of chemical structures:
∑ ∑ ∑
f (X) = Ni Ci + 𝜔 Mj Dj + z Ok Ek (4.37)
i j k

where Ci , Dj , and Ek are the contributions of the first-, second-, and third-order groups of
type i, j, and k, respectively. Ni , Dj , and Ok are the occurrences of the respective groups in a
compound and f(X) is a function of the respectively targeted property X. They may contain
property-specific adjustable parameters. f(X) functions have been proposed by Marrero and
Gani for nine properties.
The groups for each of the orders, which offer increased sophistication of the presenta-
tion of chemical structures, have been selected according to chemical engineering theory
104 Process Systems Engineering for Biofuels Development

criteria, formulated by the authors (Marrero and Gani 2001). The types of f(X), asymptotic
or linear, have been chosen to reflect the behavior of the respective property for homologous
series of relevant classes of pure compounds as a function of CH2 group number.
The first-order groups are the largest compilation and can represent a wide variety of
aromatic and aliphatic compounds. They cannot overlap, do not capture proximity effects,
and do not distinguish isomers. The second-order groups distinguish isomers and can
describe cyclic, polar and non-polar polyfunctional compounds. They are designed to
represent specific structural elements, do not cover the entire compound and can overlap.
The third-order groups are the lowest number and are devoted to complex heterocyclic and
acyclic molecules, insufficiently described on the second level.
The database for the development of the method used the available at the time exper-
imental data, after analysis of their reliability. The parameters of the correlations were
determined in three steps by minimization of the sum of squares of the differences between
experimental and estimated values of the respective property by the Levenberg–Marquardt
optimization algorithm. In the first step 𝜔 and z were set to zero. The values of Ci and the
universal constants were determined with the available data. In the next step 𝜔 was set to
1, z was set to zero and the regression of the same experimental data was performed, using
the values of Ci and the universal constants from the first step. Thus, the values of Dj were
estimated. In the third step, the values of the parameters established in the previous steps
and the same experimental data were used to estimate Ek , by setting both 𝜔 and z to unity.
The algorithm described was employed to calculate the values of (Cunico et al. 2013)
220 first-order, 130 second-order and 74 third-order groups. The total of the experimental
values used, for instance for the critical properties, was (Marrero and Gani 2001): 783 for
Tc , 775 for Pc , and 762 for Vc , respectively. These rather low ratios of number of data points
to number of parameters determined (the differences are the degrees of freedom) are typical
for GC methods and limit the possibility of using cross-validation techniques. The models
were reported with values determined with all available data (Cunico et al. 2013), but a test
with training and validation sets had been performed and estimated positively with the sum
of the mean square residuals (Marrero and Gani 2001). It should be pointed out also that
the group selection and the used step algorithm capture very closely the tendencies in the
experimental data and the structure–property relationships.
For overcoming the inherent problems with interactions between groups and dis-
tinguishing among isomers, different other options have been proposed as well. Some
examples include definition of group interaction contributions (Marrero-Morejon and
Pardillo-Fontdevila 1999), identification of the specific positions of groups, elements and
chemical bonds, etc. (Li et al. 2017). The method of Marrero and Gani and the older
method of Constantinou and Gani (1994), which this year will mark 25 years of service but
still preferred by many, are among the most used GC methods, chosen for biofuel process
design (Garcia et al. 2010; Arvelos et al. 2014).
The method of Marrero and Gani has been modified to offer an original solution of
another of the inherent problems of GCs – missing groups, by introduction of atom con-
nectivity topological indexes (Hukkerikar et al. 2012; Cunico et al. 2013). The modified
connectivity index (CI) method has the form
∑ ∑ ∑
f (X) = Ni Ci + f (X ∗ ) + Mj Dj + Ok Ek (4.38)
i j k
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 105

f(X* ) is calculated from

[ ]
∑
f (X ∗ ) = ( nm f (Xm ) + d (4.39)
m
∑ v v
f (Xm ) = am,i Am,i + b( ℵ0 )m + 2c( ℵ1 )m + d (4.40)
i

where f(X), the left-hand side of (4.38), has the same functional form and the values of the
universal constants for a particular property X, are the same as in the Marrero and Gani
method (4.37); am,i is the contribution of atom of type i, that appears Am,i times in the
molecular structure m, The atom zeroth and first-order CIs, denoted, respectively, as (v ℵ0 )
and (v ℵ1 ), are the Kier et al. (1975; Murray et al. 1975) CIs. Their calculation is described
comprehensively and illustrated with examples in Gani et al. 2005. The number of times a
missing (m) group appears in the molecule is given by nm ; b and c are adjustable param-
eters, d is a universal parameter. The values of the unknown parameters in Eq. (4.40) are
determined by regression of the available experimental data for the targeted property (Gani
et al. 2005). The CI method, with parameter values revised by regression after addition of
new data and analysis of uncertainties, has been proposed by Hukkerikar et al. (2012) with
the name GC+ .
In the same context, it can be noted that the popular method of Joback and Reid (1987)
is also having its Renaissance (Nishiumi 2016).
The text above outlines the principles of development and use of the GC methods for
prediction of biofuels property values with a focus on permanent properties. Its references
present theoretical considerations and applications in the prediction of temperature depen-
dent properties and properties of mixtures.
The methods for the mathematical statistics evaluation, cross-validation, and uncertain-
ties of GC-based models are similar to those used for QSPR, so they will be discussed
later.

4.5.2 Quantitative Structure–Property Relationships for Biofuels Process Design

In the previous section we have discussed how GC methods can overcome the problem with
missing groups. QSPRs offer other options for doing that and a somewhat different philos-
ophy for process and product design. They stem from the successful application of quan-
titative structure–activity relationships (QSARs) in drug design for properties of bioactive
molecules. Today QSPRs are an integral part of cheminformatics with diverse applications,
which include not only properties needed in organic syntheses (Gasteiger 2016), but are
increasingly targeting also properties employed in process and product design (Katritzky
et al. 2010; Abudour et al. 2014; Zhou et al. 2017). To outline the principles of QSPR meth-
ods for biofuels process design we shall again focus on the prediction of constant properties
of pure compounds.
QSPRs for individual molecules are built by relating two databases. The “structure”
database contains numerical descriptors of the structure of molecules important to the
studies. Specialized software programs are employed for drawing the structures and
eventually for preparing molecular models with minimized energy, which are then picked
106 Process Systems Engineering for Biofuels Development

up by programs for evaluation of the descriptors. These programs calculate descriptors

with group/bond contribution, topological, simulated molecular mechanics and quantum-
chemical methods and presently provide values for several thousands of them. Descriptors
pair-wise correlated with coefficients, typically higher than ±0.75, are identified by the
software and only one of a pair is left in the database (Shacham et al. 2010).
The corresponding “property” database lists experimental values of the targeted property
for the molecules in the “structure” database, critically evaluated, as discussed previously.
As with GC methods the database may contain all available data, but in contrast also a
smaller structurally related set of structures might be used, as would be discussed later. It
was suggested (Wakeham et al. 2002) to consider as pseudoexperimental, data for homol-
ogous series, predicted by asymptotic correlations with significantly higher precision than
with QSPRs, if clearly identified as such.
Numerical techniques used to build QSPR models (Katritzky et al. 2010), involve
multilinear regression (MLR), principal component regression, partial least square, etc. as
formally linear approaches. Nonlinear techniques like artificial neural networks, genetic
algorithms, support vector machines, etc. are increasingly popular. Further elaboration
of QSPRs for individual molecules depends on the sophistication of the presentation of
chemical structures and the availability of experimental data. In this context, it might be
pointed out that many of the calculated descriptors are related nonlinearly to the chemical
structures and can reflect nonlinear interactions. Moreover, the relationship between
chemical structures and targeted property is property-specific, i.e. for different properties
the different structural elements and interactions have different impact.
The minimization of the sum of squares of the differences between experimental and esti-
mated values of the respective property is realized usually stepwise, employing the obtained
coefficients of multiple correlation and the average standard deviations for choosing an
adequate model, when the effect of increasing the number of independent variables upon
the values of the chosen criteria becomes insignificant. The sufficient number of the thus
selected descriptors is determined by factors like the variation of chemical structures in
the database, achieving average standard deviation within the experimental precision of the
property measurement method, degrees of freedom, etc.
The importance of the variation of the chemical structures is often explained by the obser-
vation that properties of structures which prevail in the database are predicted with lower
uncertainties than those of the rest. The relationship probably is not so straightforward, since
smaller molecules might act as building blocks of larger molecules. The similarity of molec-
ular structures is in the basis of the quantitative structure–structure property relationships
(QSSPRs), which are going to be discussed later.
It might be noted also that the precision of any model cannot be better than that of the
property measurement method. In MLR having at least 20 experimental data per calculated
model parameter might be considered appropriate to avoid chance correlations.
It has been suggested that the descriptors, which provide the best fit to the experimental
data for the chosen mathematical model, represent the “significant common features” of the
targeted molecular structures which are most relevant for the targeted property (Wakeham
et al. 2002). The number of the descriptors selected for the final QSPR depends on the
number of groups of chemical structures, presented significantly in the database and their
structural similarity, the total number of the used experimental data, etc. and is rarely more
than 10.
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 107

The models, developed for property prediction by GC or QSPR, as any models, have to
be evaluated and validated. This is done by methods of mathematical statistics. The most
used estimates are presented below.
Average absolute deviation, AAD:
√
√ n
√∑ ∣ (yi − y′ ∣
AAD = √ i
(4.41)
i
n

Relative absolute deviation, AAD (%):

√
√ n∣
100 √
√∑ (yi − yi ∣
′

AAD(%) = (4.42)
n i
yi

Standard deviation, 𝜎: √
√ n
√∑ (yi − y′ )2
𝜎= √ i
(4.43)
i=1
n

where n is the total number of compounds, used for development of the model, yi is the mea-
′
sured property value of the ith compound, and yi is its predicted value. These estimations
are usually included in the statistical programs, some of which offer also calculation of the
squared and the adjusted squared correlation coefficients, R2 and eventually Student’s and
Fisher’ criteria, compared with their tabulated values for 95% confidence.
As discussed above, due to the lack of data, the validation of GC models in earlier times
was done with single training and validation sets. The compounds in the first set, called
further “internal set,” were used for the development of the model, and the compounds
in the second set (called “external”) for validation. It is presumed that the compounds in
the internal set are predicted with lower uncertainties, than those in the external set. The
members of the external set are an arbitrary number and randomly selected.
More recently, with the option included in statistical software, the “leave one out (LOO)”
or leave some arbitrary number of compounds out, cross-validation procedures become
increasingly popular. They evaluate a respective number of internal (training) and external
(validation sets), so that the prediction ability of the model is tested for all available exper-
imental data by moving them from one set to the other. The estimation of the following
parameters is described and recommended, for instance by Golbraikh and Tropsha (2002),
Tropsha et al. (2003), and Peraza et al. (2014):
The standard deviations of the external cross-validation sets are estimated by
√
√ ne
√∑ (yi − y′ )2
𝜎ext = √ i
(4.44)
i=1
ne

where 𝜎 ext is the standard deviation of the external cross-validation set, yi is the measured
′
value of the property, yi is its predicted value, and ne is the number of the samples in the
external set.
Equation (4.44) is used also for estimation of 𝜎 int , the standard deviation of the respective
internal cross-validation sets, when ne is replaced with nint , the number of the samples in
108 Process Systems Engineering for Biofuels Development

the internal set, yi is the measured value of the property for the ith member of the internal
′
cross-validation set, and yi is its predicted value. For the standard deviation of the internal
(training) sets 𝜎 tr the respective replacements in the equation are the number of compounds
′
in the training set ntr , and the measured (yi ) and predicted (yi ) values of the ith member of
the training set.
For estimation of the residual standard deviation of training sets 𝜎 trres the form of
Eq. (4.44) is similar to that for 𝜎 tr , but the denominator is (ntr – k), where k is the number
of coefficients in the model, including also the free parameter.
The cross-validated coefficients of the training sets are estimated by:

∑
ntr
′
(yi − yi )2
i=1
Q2 Tr = 1 − (4.45)
∑
ntr
(yi − yav )2
i=1

where yav is the average value of yi for the respective training set, the rest of the symbols
have the same meaning as in Eq. (4.44). Q2 Tr characterizes the interpolation ability within
a training set.
It has been shown, that for cross-validation of the extrapolation ability of the QSPR, it is
correct to use Q2 Ext, the cross-validated coefficients of the external sets. They are calculated
by an equation similar to Eq. (4.45), in which all symbols are replaced with those for the
respective external sets.
The above outlined problems have motivated the creation of increasing numbers of new
options for the development of GC and QSPR methods. One of these options is offered by
the QSSPRs (QS2PR).
QSSPR methods stem from the ideas of the stepwise regression using orthogonalized
variables (SROV) algorithm of Shacham and Brauner (2003). The first described QSSPR
method (Shacham et al. 2004) starts with regression of the vector of structural descriptors of
a target compound with unknown properties against a database with vectors of descriptors
of compounds with measured properties to obtain a linear relationship between the vector
of the target and those of structurally similar compounds with known properties. It has
been demonstrated that the coefficients of the model obtained represent numerically the
molecular similarity of the target and the predictive compounds. The coefficients have the
same values for the structural and the property correlations of the target and the predictive
compounds. This fact was used to calculate the unknown values of 31 properties of 18
compounds from the known values of the properties of their structurally related predictive
compounds.
Moreover, the method proposed can estimate the uncertainties of the unknown property
values, predicted for the targets, directly from the known experimental uncertainties of the
predictive compounds and can be used for consistency analysis of property data (Brauner
et al. 2005).
The established principles for obtaining structure–structure correlations were employed
in the development of targeted (T) QSSPR methods – TQSPR (Brauner et al. 2006) and
TQSPR for homologous series (Cholakov et al. 2008). In the TQSPR methods structural
correlation with the vector of descriptors of the target identifies within the whole prop-
erty database a small number (10–50) of predictive compounds with which the QSPR is
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 109

developed. For homologous series the TQSPR algorithm searches for a descriptor vector
collinear with the studied property and develops a linear correlation, which predicts within
experimental precision from only several measured data. Definition and identification of
the importance of dominant descriptors (Shacham et al. 2007a) led to refinement of the
TQSPR method (Kahrs et al. 2008) and the development of the references series method
(Shacham et al. 2013a) and the TQSPR1 method (Shacham et al. 2013b).

4.6 Challenges for the Biofuels Process Design Methods

The future places production of biofuels in smart biorefineries with circular transformation
of biomass. Smart biorefineries are systems with highly integrated process and product
design, using biomass as a source of diverse raw materials that are transformed sustainably
into an optimized variety of intermediates and products (chemicals, bioenergy/biofuels,
new materials, etc.). German experience has defined five main types of advanced biore-
fineries, classified by raw material platforms: No. 1, sugar and starch; No. 2, vegetable
and algal lipids; No. 3, lignocellulosic; No. 4, synthesis gas; and No. 5, biogas (Biore-
fineries Roadmap Working Group 2012). All of them are expected to use mixed feeds,
co-processing, alternative chemical and biochemical transformation paths, etc., in order to
find synergy that would achieve the desired performance and competitiveness of the smart
biorefinery (DECHEMA – VDI 2018).
In the smart biorefinery, biofuel process design is part of the integrated computer-aided
process and product design. It starts with pre-processing of mixed feeds, which have to
be transformed not only into fuels complying with respective specifications, but also into
many other products, thus ensuring high valorization of all raw materials and intermediates,
minimal overall energy consumption and environmental pollution, and sustainable use of
outside resources like water and land, etc. The results of biorefinery design would be sys-
tematically evaluated alternatives rather than a single optimal solution (Dimian et al. 2014,
Kalakul et al. 2018).
The above context outlines challenges to biofuel process design as part of TMF, some of
which are summarized below, using biodiesel for illustration:
• Biofuel design would need data not only for the processes involved directly in its produc-
tion, e.g. for biodiesel the properties of acylglycerols, fatty acids, low molecular mass
alcohols and their respective esters, etc. The production would be based on mixed feed
stocks – non-edible vegetable oils and fats, cooking oils, algal lipids, oils extracted from
liquid and solid wastes like fish oils, spent coffee grounds, grape marc, etc.
So, property data would be needed also for compounds and mixtures contained as
impurities and/or as valuable compounds in the feedstocks. Examples include polymer-
ized impurities in cooking oils, valuable polyphenols, diterpenoids, tocopherols, etc. in
extracted oils, omega 3 fatty acids in waste fish oils and so on. Thus, the biodiesel process
design has also to evaluate the feasibility, valorization effects, effects on circular produc-
tion and environment, etc. of alternative processes for pre-refining, as well as production
of antioxidants, food additives, etc. Moreover, alternative processes for transformation of
some of the acylglycerols into fatty acids for syntheses of biosurfactants, biolubricants,
etc., and production of active carbon from wastes like oil-depleted solids, might also be
considered.
110 Process Systems Engineering for Biofuels Development

• Many of the compounds of the potential high added-value products have sophisticated
chemical structures and are even more unstable than the biodiesel related compounds, so
their properties would be more difficult for measurement and prediction, especially by
GC methods. Typical examples can be the 𝜔-3 eicosapentadienoic acid (present in fish
waste and algal oils) and chlorogenic acid (representative of polyphenols).
• It has been demonstrated that GC methods can generate new groups and design molecules
with preset properties (Hada et al. 2014; Ng et al. 2015). However, the contribution of
new groups, still has to be estimated with the already existing experimental data. Molecu-
lar design solves the reverse problem, so its efficency depends strongly on the correlation
method used to solve the forward problem. The sophistication of chemical structure in
GC already leads to loss of their most important advantage – easy selection of the cor-
rect groups and interpretation of structure–property relationships. Typical examples are
second-order groups like CH3 —CHm =CHn , where m can be 0 or 1, and n equal 0 or
2, which codes six groups (Poling et al. 2001). Moreover, different methods use spe-
cific groups and philosophies, some of which are even more difficult to understand (e.g.
Dalmazzone et al. 2006).
QSPR methods do not have problems with presentation of molecular structures as long
as they can be drawn. Most structures needed for biofuel process design can be found on the
Internet, often together with *.mol files, containing molecular drawings, that can be picked
up by the usual software for further processing (i.e. energy minimization and/or calculation
of descriptors).
The commonly used software accounts for proximity effects, but cannot distinguish some
structural details like cis and trans isomers, if needed. The use of minimized energy molec-
ular models introduces additional computational uncertainties (Hechinger et al. 2012), due
to differences in the definition of force fields and possibilities to find global minima for
more complex structures. Due to the need of specialized software and training of chemical
engineers, QSPRs are far less popular than GC methods. Most descriptors are developed
with the task to describe molecular features important for QSAR studies and the behavior of
biologically active molecules. The result is that the meaning of the property-specific “sig-
nificant common features,” represented by the descriptors, selected for a particular QSPR
often remains hidden.
QSSPR-based methods clearly relate the molecular structures of the target and predictive
compounds. With the improvement of structure description and algorithms for identifica-
tion of structure–structure relationships, it would be even easier to find compounds with
measured data within the available database, than with traditional QSPR methods. TQSPR
methods for homologous series and the eventual use of reference series allow for the gener-
ation of “pseudoexperimental” data from a very small number of series members and thus
allow for filling-in gaps in databases.
The objects of industrial interest, including the feedstocks, intermediates, biofuels, and
fuel bioadditive products are typically mixtures. Most process and product design methods
for mixtures employ experimental property data, data for pure components and experimen-
tal phase equilibria data, for their binary and/or ternary mixtures. Mixture composition
can be represented with one or several surrogate molecules. This semi-empirical approach,
based on sound theory and experience has been and is still widely used in biofuel process
design. For functional properties, needed for product design, GC methods still dominate,
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 111

though they are typically less sophisticated and efficient. QSPRs are being increasingly
employed, but in both cases mixtures are modeled either with surrogate molecules or with
the properties of homologous series, present in the real products.
Some of the properties needed for biofuels product design are needed also for their pro-
cess design, e.g. density, viscosity, vapor pressure, boiling points. Other properties like
cetane numbers, lubricity, and cold filter plugging point for biodiesel, are only functional
properties for product design. In both cases, more data are available for lower molecular
mass pure bioadditive compounds: biohydrogen, biomethane, and bioethane, C1 –C4 bioal-
cohols and their ethers, dimethyl bioether, etc.
The measured data for biofuel relevant functional properties of pure compounds are con-
siderably less than those for thermodynamic properties. However, there are huge amounts
of experimental data on real biofuel mixtures, namely biogases, gasoline, and diesel fuel
bioadditives and hydrocarbon fuels, containing them, that are not used for fuel design. This
outlines an important opportunity to at least double presently available data.
Presentation of real fuels with surrogate molecules, used in process and/or product
design, is arbitrary and not always straightforward. Biodiesels, for instance, are produced
already from mixed feedstocks containing several original oils and fats, including also
wastes like cooking oil. The proportions are selected on the basis of specifying input
properties of the feed (e.g. for acid number) and balance the output properties of the
product within specification (e.g. for cetane number and cold filter plugging point). The
result is a mixture in which the fatty moieties of the acylglycerols are almost equally
distributed between several compounds. The use of several surrogates, especially if their
data are not experimental, increases both the uncertainties and computation time (Coelho
et al. 2018). It may happen also that for some reasons (e.g. availability of data) one and the
same biofuel mixture should be presented differently for process and for product design,
which would not be convenient for integration of process and product design.
The above presented text for this challenge outlines the need for an alternative method
for presentation of the composition of mixtures. It should be compatibe with GC and QSPR
methods, represent all their components, use the experimental data for real fuels, and allow
for the same presentation of mixtures for process and product design. In recent years the
theory and practice of product design have seen new valuable contributions (Liu et al. 2019),
together with a specialized software package. Thus, the development of the described new
method might be a timely effort.
In contrast to designed experiments, in which the response of the dependent variable is
tested according to a particular plan, the available data for structure–property correlations
use experimental data for molecules, randomly selected, depending on what has been of
interest for a particular research. The result is that molecules better presented in the database
are described with lower uncertainties, so the response space has to be scanned and tested.
We have presented some of the formulas that can estimate interpolation and extrapola-
tion abilities of models. There are also algorithms for determination of their applicability
domains and outliers (Shacham et al. 2007b; Zhou et al. 2017).
The number of available experimental data for developing a particular correlation are,
at best, several thousands and in many cases considerably less. The number of compounds
of interest of the present chemical industry are several hundreds of thousands, while drug
design is creating libraries of tens of millions of compounds that might be biologically
active (Lyu et al. 2019). Therefore, today’s efforts and methods for process and product
112 Process Systems Engineering for Biofuels Development

design concern rather small sets. In this context, it seems to be a better strategy to develop
models not targeted on all available data, but focused on smaller structurally related groups
of compounds with clearly defined applicability and uncertainties.
In the light of the challenges outlined, presently the hopes for a significant breakthrough
in the availability of properties for biofuel process design are related to the development
of quantum chemical methods. They might not solve Schroedinger’s equation ab initio, but
could provide estimations of property values satisfying the needs of chemical engineering
and smart biorefineries.

4.7 Influence of Uncertainties in Thermophysical Properties of Pure

Compounds on the Phase Behavior of Biofuel Systems
The correct estimation of properties plays an important role in the design of any process.
Even small uncertainties produced with the methods employed can dramatically mislead
the designer. Caution must be taken when choosing the estimation method that should be
employed to estimate the required property, because it might lead to bad, or sometimes
impossible correlations, and hence misrepresent completely the design. Unfortunately, no
ready recipe of how to choose the best and most adequate model for properties estimation is
available. Table 4.2 presents experimental and calculated data for four lower molecular mass
members of the homologous series of saturated triglycerides, reported by Bogatishcheva
et al. (2017).
Table 4.3 compares the Tc /Pc ratios of the experimental and estimated critical data from
Table 4.2, as recommended in Zbogar et al. 2006.
In the interpretation of the above data, it should be noted that the compounds are mem-
bers of a series, so their asymptotic relationship is usually well represented in databases.
However, the experimental data for them were not available at the time of development of
either of the methods. In the first case, it is expected for such data to be predicted better,
but in the second case worse.
The two tables show that the MCG method predicts Tc with close to experimental pre-
cision, but fails to do that for Pc . The latter is reflected in the Tc /Pc ratios predicted. On
the whole, however, this method performs better than its later developments and the older
Joback and Reid (1987) method.
Recent comparisons of CG predictions and experimental data for compounds relevant
to biodiesel process design are made in the respective publications by Nikitin and
Table 4.2 Critical properties of triglycerides (modified from Bogatishcheva et al. 2017; the data for the
Joback and Reid method are calculated by the authors).

Triglycerides of Critical temperature, Tc (K) Critical pressure, Pc (bar)

Exp. CG MG Hukk-sw Exp. CG MG Hukk-sw JR
C8:0 836.0 851.2 887.3 859.8 8.3 7.4 10.3 7.6 7.0
C10:0 864.0 882.8 924.1 890.3 7.5 5.9 9.1 5.8 5.3
C12:0 899.0 909.7 955.8 916.4 6.5 4.9 8.4 4.6 4.0
C14:0 925.0 933.1 983.7 939.3 6.1 4.2 7.9 3.8 3.3

CG, method of Constantinou and Gani; MG, method of Marrero and Gani; Hukk-sw, method of Hukkerikar et al. 2012,
with step regression; and JR, method of Joback and Reid. Fatty acid moieties are presented with the common code, i.e.
C8:0 is the triglyceride of octanoic acid. Experimental (Exp.) uncertainties: 0.015Tc , K; 0.04Pc , bar.
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 113

Table 4.3 Experimental and predicted Tc /Pc ratios of the triglycerides.

Tc/Pc ratios
Triglycerides of Exp. CG MG Hukk-sw Zbogar et al. (2006)
C8:0 100.7 114.9 86.5 112.5 124.3
C10:0 115.2 149.9 101.3 152.2 170.8
C12:0 138.3 186.4 114.1 197.9 224.9
C14:0 151.8 223.8 125.0 247.2 284.4

Abbreviations are as in Table 4.2.

Popov (2015, 2016a,2016b) and Wallek et al. (2013, 2018). The latter compares GC and
corresponding state models for prediction of 8 properties of the homologous series of fatty
acids and their esters. The investigation proves yet again that there are methods, which
could be recommended for a given property, but there is no universal method.
Uncertainties in predicted thermophysical properties are upper bounds, resulting from
prediction errors on top of measurement errors. They are estimated by averaged quantities
like average absolute and relative deviation, standard deviation, etc., which do not reveal
the distribution of uncertainties of the predicted values. Consequently, the error distribution
is often estimated also for different groups of structurally related compounds. The latter is
laid in the foundations of the QSSPR methods (Shacham et al. 2013b), which allow also for
estimation of the experimental errors of the targets from those of the predictive compounds.
Chemical engineers have long recognized that if they have a fair estimation of the uncer-
tainties they can eventually find suitable applications for measured and/or predicted values.
The application of EoS-type models for prediction and calculation of the phase behavior
in process design requires knowledge of the pure component parameters, e.g. their critical
temperature and pressure. The uncertainties in their values can critically influence the qual-
ity of the calculation and prediction as Sovova et al. (2001) demonstrated. In the example of
modeling phase behavior of the triolein + CO2 binary system it was shown that the uncer-
tainties in triolein critical properties could have a dramatic effect on the calculations and
extent of the VL region predicted. For example, for some estimations of the triolein param-
eters, the system was incorrectly predicted to be homogeneous at very low pressure values.
For systems with a supercritical solvent, Pfohl et al. (1998) showed that over-estimation
of Pc and Tc of pure components might predict wrong phase splitting in the region close to
the solvent’s critical point in mixtures.
On the other hand, when estimating binary interaction parameters from the best fit to the
available experimental data, the kij parameters might, to some extent, incorporate the errors
involved.
Wakeham et al. (2001) have studied the propagation of prediction errors in the normal
boiling temperature (Tb ), Tc , and Pc , through to the final design of a distillation column,
which separates binary and ternary mixtures of model hydrocarbons. This investigation is
relevant to the production of Fischer–Tropsch biofuels and jet biofuels, obtained by hydro-
genation of bioesters. The values of the targeted properties were estimated by several well
recognized methods, still offered by process simulators. They demonstrated that even small
uncertainties can dramatically mislead the designer. Additional errors (e.g. from selection
of the models and quality of equilibrium data) involved in the practical design can further
aggravate wrong conclusions.
114 Process Systems Engineering for Biofuels Development

However, it should be emphasized that the detailed analysis of the influence of the ther-
mophysical properties uncertainties on the phase behavior predictions and calculations of
systems of interest for the biofuels industry (or any other complex system for that matter) is
not only a very demanding but also an extremely challenging task, which requires expertise,
sophisticated understanding, and the need to explore alternative designs.
The very substantial deviations between the pure component parameters values estimated
by various authors with different methods is just one example of the challenges encoun-
tered. For this reason, effective methods to estimate the solute properties should be available
within the TMF. Then, as suggested by Hajipour and Satyro (2011), Monte Carlo techniques
should be used to evaluate error propagation from uncertainties in property values and its
effects on process simulation and design results.
In any case, the experience and common sense should also be a practical guide to any
modeler (The Self Portrait Mona Lisa, 1973 by Salvador Dali, https://www.dalipaintings
.com/self-portrait-mona-lisa.jsp#prettyPhoto[image2]/0).

4.8 Conclusions
Phase equilibrium has been measured for some but not for all possible biofuel related sys-
tems. However, to model, design, optimize, and debottleneck biofuel related processes it
is extremely important to understand the physics and thermodynamics of those processes.
A significant limitation is that there is either lack of phase equilibrium data or the data
available are often very contradictory.
In addition, for many complex compounds of interest to biofuels production there is no
information on their thermophysical properties required for the modeling and, regrettably,
the methods for estimating those are insufficiently advanced. Hence, the uncertainties in
the values of the pure component parameters can seriously influence the quality of the final
results, sometimes in a dramatic fashion.
It is a misconception that process simulators are generally “user-friendly” and always
give an applicable solution. Hence, the user is likely to adhere to that solution without
further deliberation and/or searching expert advice. Yet, as demonstrated, in some cases
inaccurate and false results can be obtained, and the user, if not cautioned, will accept those
as the true solution to the simulation or design task being solved.
In our view, the two avenues to be pursued for the future are: development of novel robust
thermodynamic models and new advanced methods for the estimation of thermophysical
properties. The challenges for these avenues were discussed in the text.

Acknowledgment
The authors acknowledge the funding received from the European Union’s Horizon 2020
research and innovation programme under the Marie Sklodowska-Curie grant agreement
No. 778168.

Exercises
The main aim of the exercises proposed is to gradually introduce the reader to the main steps
in the selection and critical analysis of missing property data of exemplary components of
biofuels, as well as comparison of alternative methods for their estimation.
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 115

1. On the Internet you may find useful information for the process and product design of
biofuels. Please verify that by obtaining the molecular and structural formulas of some or
all of the suggested components of gasoline biofuels and additives: methane, methanol,
ethanol, MTBE, ETBE, tert-amyl methyl ether (TAME, 2-methoxy-2-methylbutane),
and tert-amyl ethyl ether (TAEE, 2-ethoxy-2-methylbutane).
2. Please obtain the molecular and structural formulas of some or all components of the
suggested diesel biofuels and additives: dimethyl ether, methyl ethyl ketone, butanol,
methyl oleate, ethyl oleate, propyl oleate, butyl oleate, methyl linoleate, ethyl linoleate,
methyl linolenate, methyl palmitate, ethyl palmitate, methyl stearate, butyl stearate,
glyceryl trioleate (triolein), glyceryl trilinoleate (trilinolein), glyceryl tripalmitate
(palmitin), and glyceryl tristearate (stearin).
3. Get acquainted with the sources of property data, presented in the chapter, and find
the available values for the normal boiling temperatures, critical temperatures and pres-
sures of the compounds selected in the first two exercises. Try to identify experimental
values and the methods of estimation of those, which are not experimental. Prepare a
table (preferably in Excel) with the found values, classifying them also as experimental,
estimated, and unknown.
4. Get acquainted with the methods for property estimation, presented in the chapter and
select two different group contribution methods (e.g. the method of Joback and Reid and
the method of Constantinou and Gani). Using the structural formulas found in Exercise
1, present the chemical structures of your selected compounds with the groups, required
by the Joback and Reid and the Constantinou and Gani methods.
5. Calculate the normal boiling temperatures, and critical temperatures and pressures of
your compounds by the selected methods. Copy them in your table. Using the formulas
given in the chapter, calculate and copy in to the table, the average absolute, the relative
average absolute, the maximal and the standard deviations of the predictions of your
chosen estimation methods from the experimental values. Formulate conclusions about
the suitability of the compared methods for your particular data and design projects.

References
Abala, I., M’hamdi Alaoui, F.E., Chhiti, Y. et al. (2019). Experimental density and PC-SAFT modeling of
biofuel mixtures (DBE + 1-heptanol) at temperatures from (298.15 to 393.15) K and at pressures up to
140 MPa. The Journal of Chemical Thermodynamics 131: 269–285.
Abrams, D.S. and Prausnitz, J.M. (1975). Statistical thermodynamics of liquid-mixtures: A new expression
for excess Gibbs energy of partly or completely miscible systems. AIChE Journal 21: 116–128.
Abudour, A.M., Mohammad, S.A., Robinson, R.L. Jr., and Gasem, K.A.M. (2014). Generalized
binary interaction parameters for the Peng–Robinson equation of state. Fluid Phase Equilibria 383:
156–173.
Arvelos, S., Rade, L.L., Watanabe, E.O. et al. (2014). Evaluation of different contribution methods over
the performance of Peng–Robinson and CPA equation of state in the correlation of VLE of triglycerides,
fatty esters and glycerol + CO2 and alcohol. Fluid Phase Equilibria 362: 136–146.
Belting, P.C., Rarey, J., Gmehling, J. et al. (2012). Activity coefficient at infinite dilution measurements for
organic solutes (polar and non-polar) in fatty compounds – Part I: Saturated fatty acids. The Journal of
Chemical Thermodynamics 55: 42–49.
Belting, P.C., Rarey, J., Gmehling, J. et al. (2013). Activity coefficient at infinite dilution measurements
for organic solutes (polar and non-polar) in fatty compounds – Part II: C18 fatty acids. The Journal of
Chemical Thermodynamics 60: 142–149.
116 Process Systems Engineering for Biofuels Development

Biorefineries Roadmap Working Group (2012). Biorefineries Roadmap. https://www.bmbf.de/pub/Road

map_Biorefineries_eng.pdf (accessed 3 March 2019).
Bogatishcheva, N.S., Faizullin, M.E., Popov, A.P., and Nikitin, E.D. (2017). Critical properties, heat capac-
ities and thermal diffusivities of four saturated triglycerides. The Journal of Chemical Thermodynamics
113: 308–314.
Brauner, N., Shacham, M., Cholakov, G.St., and Stateva, R.P. (2005). Property prediction by similarity of
molecular structures—practical application and consistency analysis. Chemical Engineering Science 60:
5458–5471.
Brauner, N., Stateva, R.P., Cholakov, G.S., and Shacham, M. (2006). Structurally targeted QSPR method
for property prediction. Industrial Engineering Chemistry Research 45: 8430–8437.
Cagin, T., Che, J., Qi, Y. et al. (1999). Computational materials chemistry at the nanoscale. Journal of
Nanoparticle Research 1: 51–69.
Ceriani, R., Gani, R., Liu, Y.A., and Y.A. (2013). Prediction of vapor pressure and heats of vaporization of
edible oil/fat compounds by group contribution. Fluid Phase Equilibria 337: 53–59.
Cholakov, G., Toteva, V., Nikolov, R. et al. (2013). Extracts from coffee by-products as potential raw mate-
rials for fuel additives and carbon adsorbents. Journal of Chemical Technology and Metallurgy 48 (5):
497–504.
Cholakov, G.S. (2016). Air quality and the petroleum industry. In: Comprehensive Analytical Chemistry,
The Quality of Air, vol. 73 (eds. M. de la Guardia and S. Armenta), 563–587. Elsevier.
Cholakov, G.St, Wakeham, W.A., and Stateva, R.P. (1999). Estimation of normal boiling points of hydro-
carbons from descriptors of molecular structure. Fluid Phase Equilibria 163: 21–42.
Cholakov, G.St. (2011). Towards computer aided design of fuels and lubricants. Journal of the University
of Chemical Technology and Metallurgy 46 (3): 217–236.
Cholakov, G.S., Stateva, R.P., Brauner, N., and Shacham, M. (2008). Estimation of properties of homol-
ogous series with targeted quantitative structure-property relationships. Journal of Chemical and Engi-
neering Data 53: 2510–2520.
Coelho, J.A.P., Filipe, R.M., Robalo, M.P., and Stateva, R.P. (2018). Recovering value from organic waste
materials: supercritical fluid extraction of oil from industrial grape seeds. Journal of Supercritical Fluids
141: 68–77.
Constantinou, L. and Gani, R. (1994). New group contribution method for estimating properties of pure
compounds. AIChE Journal 40 (10): 1697–1710.
Cotabarren, N., Hegel, P., and Pereda, S. (2014). Thermodynamic model for process design, simulation and
optimization in the production of biodiesel. Fluid Phase Equilibria 362: 108–112.
Cunico, L.P., Hukkerikar, A.S., Ceriani, A. et al. (2013). Molecular structure-based methods of property
prediction in application to lipids: a review and refinement. Fluid Phase Equilibria 357: 2–18.
Cunico, L.P., Ceriani, A., Sarup, B. et al. (2014). Data, analysis and modeling of physical properties for
process design of systems involving lipids. Fluid Phase Equilibria 362: 318–327.
Dalmazzone, D., Salmon, A., and Sofiane, G.S. (2006). A second order group contribution method for the
prediction of critical temperatures and enthalpies of vaporization of organic compounds. Fluid Phase
Equilibria 242: 29–42.
Damaceno, D.S. and Ceriani, R. (2017). Vapor-liquid equilibria of monoacylglycerol+ monoacylglycerol
or alcohol or fatty acid at subatmospheric pressures. Fluid Phase Equilibria 452: 135–142.
Damaceno, D.S., Perederic, O.A., Ceriani, R. et al. (2018a). Improvement of predictive tools for
vapor-liquid equilibrium based on group contribution methods applied to lipid technology. Fluid Phase
Equilibria 470 (18): 249–258.
Damaceno, D.S., Jesus, E.P., and Ceriani, R. (2018b). Experimental data and prediction of normal boiling
points of partial acylglycerols. Fuel 232: 470–475.
DECHEMA-VDI (2018). Advanced alternative liquid fuels: For climate protection in the global raw mate-
rials change. Position paper of the Process Net working group “Alternative Liquid and Gaseous Fuel”.
https://processnet.org/en/Publications/Topics+of+Strategic+Relevance.html (accessed February 2019).
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 117

Dimian, A.C., Bildea, C.S., and Kiss, A.A. (2014). Integrated Design and Simulation of Chemical Pro-
cesses, 2e. Elsevier B.V.
Fornari, T. and Stateva, R.P. (2015). Thermophysical properties of pure substances in the context of sus-
tainable high pressure food processes modelling. In: High Pressure Fluid Technology for Green Food
Process (eds. T. Fornari and R.P. Stateva). Springer.
Fornari, T., Luna, P., and Stateva, R.P. (2010). The vdW EoS hundred years later, yet younger than before.
Application to the phase equilibria modeling of food-type systems for a green technology. The Journal
of Supercritical Fluids 55: 579–593.
Gani, R., Harper, P.M., and Nostrup, M. (2005). Automatic creation of missing groups through connectiv-
ity index for pure-component property prediction. Industrial and Engineering Chemistry Research 44:
7262–7269.
García, M., Gonzalo, A., Sánchez, J.L. et al. (2010). Prediction of normalized biodiesel properties by sim-
ulation of multiple feedstock blends. Bioresource Technology 101: 4431–4439.
Garmus, T.T., Giani, N.A.O., Filho, W.A.R. et al. (2019). Solubility of oleic acid, triacylglycerol and their
mixtures in supercritical carbon dioxide and thermodynamic modeling of phase equilibrium. The Journal
of Supercritical Fluids 143: 275–285.
Gasteiger, J. (2016). Chemoinformatics: achievements and challenges, a personal view. Molecules 21 (2):
151.
Georgieva, S.S., Coelho, J.A.P., Campos, F.C. et al. (2018). Green extraction of high added value substances
from spent coffee grounds: preliminary results. Journal of Chemical Technology and Metallurgy 53:
640–646.
Golbraikh, A. and Tropsha, A. (2002). Beware of q2! Journal of Molecular Graphics and Modelling 20:
269–276.
Hada, S., Solvason, C.C., and Eden, M.R. (2014). Characterization-based molecular design of bio-fuel
additives using chemometric and property clustering techniques. Frontiers in Energy Research 2, Article
20: 1–12.
Hajipour, S. and Satyro, M.A. (2011). Uncertainty analysis applied to thermodynamic models and process
design – 1. Pure components. Fluid Phase Equilibria 307: 78–94.
He, M., Wang, Ch., Chen, J., and Liu, X. (2018). Prediction of the critical properties of mixtures based on
group contribution theory. Journal of Molecular Liquids https://doi.org/10.1016/j.molliq.2018.08.048.
Hechinger, M., Leonhard, K., and Marquardt, W. (2012). What is wrong with quantitative structure-property
relations models based on three-dimensional descriptors? Journal of Chemical Information and Modeling
52: 1984–1993.
Hemmati-Sarapardeh, A., Ameli, F., Varamesh, A. et al. (2018). Toward generalized models for estimating
molecular weights and acentric factors of pure chemical compounds. International Journal of Hydrogen
Energy 43: 2699–2717.
Hollingsworth, S.A. and Dror, R.O. (2018). Molecular dynamics simulation for all. Neuron 99:
1129–1143.
Hukkerikar, A.S., Sarup, B., Kate, A.T. et al. (2012). Group-contribution+ (GC+) based estimation of
properties of pure components: improved property estimation and uncertainty analysis. Fluid Phase
Equilibria 321: 25–43.
Joback, K.G. and Reid, R.C. (1987). Estimation of pure component properties from group contributions.
Chemical Engineering Communications 57: 233–243.
Kahrs, O., Brauner, N., Cholakov, G.S. et al. (2008). Analysis and refinement of the targeted QSPR method.
Computers and Chemical Engineering 32: 1397–1410.
Kaiser, D.L. and Hanisch, R.J. (2018). Technical review of NIST’s free standard reference data products.
NIST Special Publication 1223, 1–15. National Institute of Standards and Technology, U.S. Department
of Commerce.
Kalakul, S., Zhang, L., Fang, Z. et al. (2018). Computer aided chemical product design–ProCAPD and
tailor-made blended products. Computers and Chemical Engineering 116: 37–55.
118 Process Systems Engineering for Biofuels Development

Kanda, L.R.S., dos Santos, K.C., Santos, J.T.F. et al. (2018). Vapor-liquid and liquid-liquid equilibrium
modeling of systems involving ethanol, water, and ethyl valerate (valeric acid) using the PC-SAFT
equation of state. Fluid Phase Equilibria 474: 92–99.
Katritzky, A., Kuanar, M., Slavov, S., and Hall, C. (2010). Quantitative correlation of physical and chemical
properties with chemical structure: Utility for prediction. Chemical Reviews 110: 5714–5789.
Kier, L.B., Hall, L.H., Murray, W.J., and Randic, M. (1975). Molecular connectivity I: relationship to non-
specific local anesthesia. Journal of Pharmaceutical Sciences 64: 1971–1974.
Li, Z., Zuo, L., Wu, W., and Chen, L. (2017). The new method for correlation and prediction of ther-
mophysical properties of fluids. Critical temperature. Journal of Chemical and Engineering Data 62:
3723–3731.
Liu, Q., Zhang, Z., Liu, L. et al. (2019). OptCAMD: an optimization-based framework and tool for
molecular and mixture product design. Computers and Chemical Engineering https://doi.org/10.1016/j
.compchemeng.2019.01.006.
Lyu, J., Wang, S., Balius, T.E. et al. (2019). Ultra-large library docking for discovering new chemotypes.
Nature 566: 224–246.
Marrero, J. and Gani, R. (2001). Group-contribution based estimation of pure component properties. Fluid
Phase Equilibria 183–184: 183–208.
Marrero-Morejon, J. and Pardillo-Fontdevila, E. (1999). Estimation of pure compound properties using
group-interaction contributions. AIChE Journal 45: 615–621.
Murray, W.J., Hall, L.H., and Kier, L.B. (1975). Molecular connectivity II: relationship to water solubility
and boiling point. Journal of Pharmaceutical Sciences 64: 1974–1977.
Mustaffa, A.A., Gani, R., and Kontogeorgis, G.M. (2014). Development and analysis of the original
UNIFAC-CI model for prediction of vapor–liquid and solid–liquid equilibria. Fluid Phase Equilibria
366: 24–44.
Ng, L.Y., Chong, F.H., and Chemmangattuvalappil, N.G. (2015). Challenges and opportunities in
computer-aided molecular design. Computers and Chemical Engineering 81: 115–129.
Nikitin, E.D. and Popov, A.P. (2015). Vapor–liquid critical properties of components of biodiesel. 1. Methyl
esters of n-alkanoic acids. Fuel 153: 634–639.
Nikitin, E.D. and Popov, A.P. (2016a). Vapor–liquid critical properties of components of biodiesel. 2. Ethyl
esters of n-alkanoic acids. Fuel 166: 502–508.
Nikitin, E.D. and Popov, A.P. (2016b). Vapor–liquid critical properties of components of biodiesel. 3.
Methyl esters of linoleic, linolenic, and erucic acids. Fuel 176: 130–134.
Nishiumi, H. (2016). Thermodynamic property prediction for high molecular weight molecules based on
their constituent family. Fluid Phase Equilibria 420: 1–6.
Peraza, A.R., Rojas, L., Bolivar, C., and Ruette, F. (2014). A QSPR strategy to select models with extrap-
olative ability – critical temperatures of linear and aromatic hydrocarbons. Journal of Computational
Methods in Science and Engineering 14: 155–167.
Pereira, C.G., Ferrando, N., Lugo, R. et al. (2016). Predictive evaluation of phase equilibria in biofuel
systems using molecular thermodynamic models. The Journal of Supercritical Fluids 118: 64–78.
Pfohl, O., Giese, T., Dohrn, R., and Brunner, G. (1998). Comparison of 12 equations of state with respect to
gas-extraction processes: reproduction of pure-component properties when enforcing the correct critical
temperature and pressure. Industrial and Engineering Chemistry Research 37: 2957–2965.
Poling, B.E., Prausnitz, J.M., and O’Connell, J.P. (2001). The Properties of Gases and Liquids, 5e. New
York: The McGraw-Hill Companies, Inc.
Reynel-Ávila, H.E., Bonilla-Petriciolet, A., and Tapia-Picazo, J.C. (2019). An artificial neural
network-based NRTL model for simulating liquid-liquid equilibria of systems present in biofuels
production. Fluid Phase Equilibria 483: 153–164.
Riazi, M.R. and Daubert, T.E. (1987). Characterization parameters for petroleum fractions. Industrial Engi-
neering Chemistry Research 26: 755–759.
 Challenges in the Modeling of Thermodynamic Properties and Phase Equilibrium Calculations 119

Rodriguez, G. and Beckman, E.J. (2019). Modelling phase behavior of biodiesel related systems with CO2
using a polar version of PC-SAFT. Fluid Phase Equilibria 485: 32–43.
Shacham, M. and Brauner, N. (2003). The SROV program for data analysis and regression model identifi-
cation. Computers and Chemical Engineering 27: 701–714.
Shacham, M., Brauner, N., Cholakov, G.S., and Stateva, R.P. (2004). Property prediction by correlations
based on similarity of molecular structures. AIChE Journal 50 (10): 2481–2492.
Shacham, M., Kahrs, O., Cholakov, G.St. et al. (2007a). The role of the dominant descriptor in targeted
quantitative structure property relationships. Chemical Engineering Science 62 (22): 6222–6233.
Shacham, M., Brauner, N., Cholakov, G.S., and Stateva, R.P. (2007b). Identifying applicability domains for
quantitative structure–property relationships. In: Computer Aided Chemical Engineering, vol. 24 (eds.
V. Plesu and P.S. Agachi), 327–333. Elsevier.
Shacham, M., Brauner, N., and Paster, I. (2013a). Property prediction and consistency analysis by a refer-
ence series method. AIChE Journal 59 (2): 420–428.
Shacham, M., Elly, M., Paster, I., and Brauner, N. (2013b). Evaluation and refinement of the property
prediction stage of the targeted QSPR method for 15 constant properties and 80 groups of compounds.
Chemical Engineering Science 97: 186–197.
Shacham, M., Cholakov, G.St., Stateva, R.P., and Brauner, N. (2010). Quantitative structure–property rela-
tionships for prediction of phase equilibrium related properties. Industrial and Engineering Chemistry
Research 49: 900–912.
Soria, T.M., Andreatta, A.E., Pereda, S., and Bottini, S.B. (2011). Thermodynamic modeling of phase equi-
libria in biorefineries. Fluid Phase Equilibria 302 (1–2): 1–9.
Sovova, H., Stateva, R.P., and Galushko, A.A. (2001). Essential oils from seeds. Solubility of limonene in
supercritical CO2 and how it is affected by fatty oil. Journal of Supercritcial Fluids 20: 113–129.
Stateva, R.P. and Tsvetkov, St. (1992). Influence of thermophysical properties extrapolation, in the course of
iterative calculations, on the convergence pattern of a simulation algorithm – a general analysis. Journal
of Chemical Engineering Japan 25: 327–333.
Stateva, R.P., Tsvetkov, St., and Boyadjiev, C. (1990). An approach to organizing the EOS model in process
simulators. AICHE Journal 36: 132–136.
Tropsha, A., Gramatica, P., and Gombar, V.K. (2003). The importance of being earnest: validation is the
absolute essential for successful application and interpretation of QSAR models. QSAR and Combina-
torial Science 22: 69–77.
Varamesh, A., Hemmati-Sarapardeh, A., Moraveji, M.K., and Mohammadi, H. (2017). Generalized models
for predicting the critical properties of pure chemical compounds. Journal of Molecular Liquids 240:
777–793.
Wakeham, W.A., Cholakov, G. St., and Stateva, R.P. (2001). Consequences of property errors on the design
of distillation columns. Fluid Phase Equilibria 185: 1–12.
Wakeham, W.A., Cholakov, G.St., and Stateva, R.P. (2002). Liquid density and critical properties of hydro-
carbons estimated from molecular structure. Journal of Chemical and Engineering Data 47: 559–570.
Wallek, T., Rarey, J., Metzger, J.O., and Gmehling, J. (2013). Estimation of pure-component properties
of biodiesel-related components: fatty acid methyl esters, fatty acids, and triglycerides. Industrial and
Engineering Chemistry Research 52: 16966–16978.
Wallek, T., Knobelreiter, K., and Rarey, J. (2018). Estimation of pure-component properties of
biodiesel-related components: fatty acid ethyl esters. Industrial and Engineering Chemistry Research
57: 3382–3396.
Watson, H.A.J. and Barton, P.I. (2017). Reliable flash calculations: Part 3. A non-smooth approach to den-
sity extrapolation and pseudoproperty evaluation. Industrial and Engineering Chemistry Research 56
(50): 14832–14847.
Watson, K. and Nelson, E. (1933). Improved methods for approximating critical and thermal properties of
petroleum fractions. Industrial and Engineering Chemistry 25: 880–887.
120 Process Systems Engineering for Biofuels Development

Wilson, L.C., Jasperson, L.V., VonNiederhausern, D. et al. (2018). DIPPR project 851 − thirty years of
vapor−liquid critical point measurements and experimental technique development. Journal of Chemical
and Engineering Data 63: 3408–3417.
Yaws, C.L. (2012). Critical Property Data for Chemical Engineers and Chemists. Knovel. ISBN:
978-1-61344-932-5.
Žbogar, A., Lopes, F.V.D.S., and Kontogeorgius, G.M. (2006). Approach suitable for screening estimation
methods for critical properties of heavy compounds. Industrial and Engineering Chemistry Research 45:
476–480.
Zhou, L., Wang, B., Jiang, J. et al. (2017). Predicting the gas-liquid critical temperature of binary mix-
tures based on the quantitative structure property relationship. Chemometrics and Intelligent Laboratory
Systems 167: 190–195.
 5
Up-grading of Waste Oil: A Key
Step in the Future of Biofuel
Production
Luigi di Bitonto and Carlo Pastore
Istituto di Ricerca Sulle Acque (IRSA), Consiglio Nazionale delle Ricerche (CNR), 70132 Bari, Italy

5.1 Introduction
Biodiesel is a mixture of fatty acid methyl esters (FAMEs), which has very similar physi-
cal and chemical properties to petroleum-derived fuel (Borugadda and Goud 2012; Yusuf
et al. 2011). It is industrially produced by transesterification of glycerides (animal and/or
vegetal) with methanol (MeOH) under homogeneous alkaline conditions, by using sodium
or potassium hydroxide, carbonates, or alkoxides (Figure 5.1; Baskar and Aiswarya 2016;
Tubino et al. 2016; Bozbas 2008).
Despite the fact that the use of a co-solvent would facilitate the formation of a single
phase by improving the reaction rate (tetrahydrofuran, dimethyl sulfoxide, acetone, cyclic
esters, and n-hexane have been thoroughly investigated [Alhassan et al. 2014; Thanh et al.
2013; Liu et al. 2012]), in an industrial context, homogeneous basic catalysis is adopted in
the presence of MeOH only, for the cost-effectiveness and the relevant good performances.
However, such application can only be adopted on refined oils or highly pure fat: free
fatty acids (FFAs) content must not exceed 0.1–0.5 wt% and moisture has to be absent
(Eze et al. 2018; Verma and Sharma 2016). For this reason, the industrial production of
biodiesel was initially conducted to specifically convert dedicated and refined feedstocks
(namely, first-generation biodiesel). However, such an approach was soon inappropriate:
for an economic reason, since the overall production cost was strongly influenced (>80%)
by the price of the starting vegetable oil (Gebremariam and Marchetti 2018; You et al. 2008)
to the point that it compromised the entire profitability of the industrial production chain
with respect to fossil fuel, as well as for an ethical concern, due to the land for fuel use.

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
122 Process Systems Engineering for Biofuels Development

. cat.

FAMEs

Triglycerides Glycerol

Figure 5.1 General equation to obtain biodiesel from triglycerides.

In any case, production of FAMEs still remains one of the main important European
targets in terms of renewables in liquid transport fuels for the future (Ajanovic 2013). In
fact, a value of 10% of renewables was fixed by the European Parliament as a final objective
to be achieved by 2020. In addition, new forthcoming rules have been introduced which also
consider the sustainability criteria. More specifically, in order to limit the environmental
impact correlated with the use of “first-generation” feedstock, and to better regulate the
indirect land use change due to the cultivation of crop-for-fuel, a cap for this specific fuel
was fixed to 7% (meaning 7 Mton/yr).
On the other hand, the use of alternative waste-oily sources and renewable cleaner energy
sources (renewable electricity for rails and cars) (Directive 2009/28/EC 2009) are going to
be privileged by introducing the “double counting” rule/concept. Such a decision actually
“forces” 1.5 Mton of less valuable feedstocks to be used to produce biofuel for promoting
a cleaner mobility.
Waste cooking oils (WCO), waste animal grease (WAG) and non-edible oils were defini-
tively recognized as the most important alternative feedstocks (“second-generation feed-
stocks”) for a sustainable biodiesel production (Ashraful et al. 2014; Atabani et al. 2013).
In Italy, the total annual consumption of vegetable oils is around 1.4 Mton, and it was
estimated that over one forth (0.26 Mton) ended up as waste (CONOE 2018): 94 000 ton
(36%) from restaurants, bars, hotels, etc., and 166 000 ton from domestic users. Only a very
limited amount (72 000 ton in 2018) was effectively and properly collected and recycled,
while over 60% of waste oil was dispersed into the environment, most presumably into the
water collection system.
Besides WCO, in Italy, over 0.25 Mton WAG, which should be disposed of through incin-
eration (cat1 and cat2, as defined by European Regulation No. 1069/2009), is collected
annually and could be potentially used for producing liquid biofuels. These considerations
of consumption and effectiveness of collection, can be easily extended to most of the rich
and developed American, Asiatic, and European countries. In any case, most uncollected
oils can be potentially recovered as brown grease, trap grease, sewer grease, and sewage
scum at different points of the water collection systems depending on the specific country
considered.
For example, in European countries, where grease traps are hardly installed, most of the
WCO is supposed to be potentially collected in wastewater treatment plants (WWTPs),
as sewage scum or sewer grease, even if it is not ever collected and separated (di Bitonto
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 123

et al. 2016). In the US or in some developed Asiatic countries, where grease traps are more
widely installed, a huge amount of yellow grease is effectively collected and even found to
be convertible into biodiesel (Montefrio et al. 2010).
All these second-generation feedstocks normally contain high levels of FFAs (1–90 wt%)
and could be efficiently converted into biodiesel by adopting appropriate strategies. In fact,
the presence of FFAs, water, or other contaminants, makes the direct application of the
conventional base-catalyzed process more complicated. Specifically, the addition of a base
to oil which contains FFAs, leads to the formation of soaps and water (Figure 5.2).
This reaction not only leads to the neutralization of the catalyst, but also complicates
the industrial operations in recovering both the products (FAMEs and glycerol), since the
resulting soaps and water may cause emulsification, with an increase in separation costs.
In general, two main strategies have been considered for the conversion of waste oils into
biodiesel (Figure 5.3):
• Physicochemical pretreatments of waste oils for the removal of FFAs.
• Direct treatment and conversion of FFAs into esters (FAMEs or glycerides).
Evaporation, steam injection, filtration, and dewatering (Figure 5.3, route a) represent
the most common physicochemical operations adopted to remove impurities from waste
oils (water, FFAs, etc.) with the aim of obtaining refined glycerides to be transesterified
under the conventional route. Alternatively (Figure 5.3, route b), FFAs can be directly con-
verted in situ in the relevant FAMEs. To this end, homogeneous mineral acids represent

RCOOH + MOH RCOO⊝M⊕ + H2O

M = Na, K, etc.

Figure 5.2 Reaction for saponification of FFAs.

Waste Oil
WCO, Animal grease, Brown grease, Low-quality oils
FFAs, water, glycerides

(a) (b)

FFAS or Soaps Refined Oils Pretreated Oils

Biodiesel

Figure 5.3 Rationale of conversion of waste oils into biodiesel.

124 Process Systems Engineering for Biofuels Development

the conventional choice. Recently, several further catalysts have been investigated and pro-
posed: heterogeneous acids, ionic liquids (ILs), and enzymes.
In this chapter, a critical overview of these different approaches has been assessed, by
reporting the respective advantages and drawbacks.

5.2 Physicochemical Pretreatments of Waste Oils: Removal

of Contaminants
Different types of pretreatments have been studied with the aim of reducing the presence
of contaminants and for improving the quality of an oil to be converted into biodiesel:
steam injection, neutralization, vacuum evaporation, and vacuum filtration represent the
most simple procedures investigated (Kulkarni and Dalai 2006).
When not considerable differences were determined between waste oils (as could hap-
pen for some WCO) and virgin oils (Knothe et al. 1997), with FFA content less than 3 mg
KOH/g, heating and removal by filtration of solid particles are often sufficient as pretreat-
ments for the subsequent conventional transesterification reaction. In fact, heating WCO by
steam to 65 ∘ C, and allowing sedimentation, reduction of the water (from 1.4 to 0.4%) and
FFA contents (from 6.3 to 4.3%) were observed, which allowed higher yield of FAMEs to
be obtained (from 67.5 to 83.5%) (Supple et al. 2002).
In some other cases, water and suspended solids were efficiently removed by mixing oily
feedstock with magnesium sulfate (Felizardo et al. 2006), or 10% silica gel (28–200 mesh)
(Issariyakul et al. 2007), or calcium chloride (Predojević 2008), and using vacuum filtration.
However, considering that waste oils often partially decompose and/or deteriorate, the
reduction of the final methyl ester yield and the co-formation of undesired products were
verified during biodiesel production.
So, in most of cases, pretreatment methods were focused on reducing FFAs (up to 3 mg
KOH/g), water (up to 0.1 wt%) and polymers, before the conventional transesterification
process. Typically, FFAs can be removed by neutralization and separated as soaps (caus-
tic washing), while activated charcoal was found to efficiently remove polymers through
adsorption (Cvengroš and Cvengrošová 2004). In this way, industrial production can be
maintained without the great problems of separations and emulsification.
For lower quality feedstocks, such as yellow grease (FFA content up to 15 wt%) or brown
grease (FFA content up to 99%), the preliminary treatment can result in a steam stripping
(Anderson 2014), in order to obtain FFAs, water and light contaminants on one side and
glycerides ready to be converted conventionally on the other side.
Although the removal of impurities such as water, FFAs, and polymers prior to transester-
ification can improve the quality of waste oil for producing FAMEs, pretreatments tend to
increase biodiesel production costs, and specific and detailed economic analysis needs to be
conducted. For this purpose, very limited studies were carried out to estimate and quantify
the additional costs correlated to the pretreatment steps. However, a significant improve-
ment of the economic viability of the process from waste oils was determined compared
with virgin oils (Yaakob et al. 2013). In some cases, reduction of about 45% of direct pro-
duction costs, even including the additional costs due to pretreatment (Zhang et al. 2003a),
was achieved.
In any case, removal of FFAs as soaps, prevent a huge amount of starting feedstock from
being converted into the desired FAMEs, by lowering the final proficiency of the whole
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 125

production chain. This approach could be feasible only on those oily feedstocks which
contain low FFA content.

5.3 Direct Treatment and Conversion of FFAs into Methyl Esters

As an alternative to the removal of FFAs from waste oils, a more profitable strategy is the
direct conversion of FFAs in the relevant methyl esters in situ (Figure 5.3, route b). In this
case, through the use of a two-step reaction, acid oil can be more efficiently converted into
biodiesel: in the first step an acid catalyst is used to promote the direct esterification of
FFAs, whereas alkaline catalysts are adopted in a second stage for the transesterification of
glycerides (Deng et al. 2010; Canakci 2007).
The most challenging part of this process is the first conversion of FFAs in FAMEs:
homogeneous and heterogeneous Brønsted and Lewis acids, ILs and enzymes represent
possible alternatives.

5.3.1 Homogeneous Catalysis: Brønsted and Lewis Acids

Industrially, sulfuric acid (H2 SO4 ) is the most used catalyst for promoting direct esteri-
fication of FFAs. It allows the reaction to be effectively carried out at low temperatures
(60–120 ∘ C), atmospheric pressure and in a short time (two to four hours).
The direct esterification suffers from several thermodynamic and kinetic constraints,
since the co-formation of water inhibits the reaction, requiring a double pretreatment
step (Canakci and van Gerpen 2001) or the use of a large excess of MeOH (up to 40:1
MeOH:FFAs) (Chai et al. 2014). In principle, the use of MeOH in this step allows FFAs to
be converted into FAMEs, by obtaining a final mixture of FAMEs and glycerides, which
can eventually be reacted under alkaline condition to produce biodiesel.
However, a series of drawbacks connected to the use of H2 SO4 is also included: (i) the
acid is partially soluble in the resulting esterified oils/fats; (ii) its recovery is only partial;
(iii) its recycle and/or reuse in a new pretreatment cycle is quite complicated; and (iv) the
use of costly materials (resistant to the acid treatment) for reactors is compulsory. In detail,
the partial dissolution of H2 SO4 in the resulting oily mixture needs costly downstream
operations. Washing procedures with solvents (MeOH [Kawahara and Ono 1977] or glyc-
erol [Brunner et al. 2000]) or chemical neutralization with bases (potassium methoxide or
sodium hydroxide, sodium hydroxide in MeOH [Garcia 2015]) are necessary.
Besides large consumption of the alkaline catalyst, a concomitant production of a new
salt, which needs to be disposed of at the end, characterizes such a pretreatment. Such draw-
backs negatively affect the economic balance sheet of the overall production of biodiesel
from waste grease, pushing forward the study and the optimization of alternative solutions.
For example, the use of sulfamic acid as a catalyst demonstrated elevated potential for an
environmentally friendly means of biodiesel production. In fact, it results in a safe chemical,
cheap and sufficiently active in promoting the direct esterification: about 90% of the starting
acidity was converted in the relevant ethyl esters under mild conditions. Finally, it was
even found potentially recoverable and recyclable. However, its separation from the reacted
mixture needs unpracticable operations for an industrial scale: (i) addition of hexane in
order to induce precipitation of most of the catalyst; and (ii) washing of the organic phase
126 Process Systems Engineering for Biofuels Development

was necessary to remove any traces of sulfamic acids from the pretreated organic mixture
FAMEs/glycerides (de Oliveira et al. 2016).
Besides conventional mineral Brønsted acids, Lewis acids have also been found active in
promoting vegetable oil transesterification. Several metal acetates and stearates (calcium,
barium, plumb, zinc, cobalt, and nickel) were found to be capable of producing esters from
low-quality feedstock, even in the presence of high FFA contents. In fact, they do not pro-
duce soaps, but if used in large excess, they precipitate as salts. In addition, they need the use
of high temperature (200–250 ∘ C) and high pressure (400–600 psi). Stearates were found
more active than the respective acetates, due to the relevant higher solubility in the oils,
showing better performance than Brønsted acids with lower concentration of catalysts and
lower oil/alcohol molar ratio (Di Serio et al. 2005).
Glycerolysis may be adapted to reduce FFA content in acidic and low-cost feedstocks for
biodiesel production (Figure 5.4): reduction of FFA content from 50 to 5% after three hours
of reaction at 200 ∘ C, even without the addition of any catalyst, was observed (Felizardo
et al. 2011). Zinc salts were tested as efficient catalysts. Once a mixture of monoacyl-
glycerol (MAG), diacylglycerol (DAG), and triacylglycerol (TAG) was obtained, the con-
ventional transesterification could be proficiently applied. Such an approach was really
promising, specially if carried out by directly using crude glycerin coming from a con-
ventional biodiesel plant. Produced in substantial excess as a by-product from the alkaline
transesterification process to make biodiesel, the direct utilization of crude glycerin with
minimum processing can significantly benefit the biodiesel industry.
Low quality grease (30% FFAs) was efficiently converted into MAG, DAG, and TAG
using crude glycerin, which could be obtained from a conventional biodiesel plant, evap-
orating methanol from the glyceric phase. Working under a 1:1 M ratio of crude glycerin
and FFAs at 230 ∘ C and for 150 minutes, the wt% FFAs was reduced to below 1 wt%. Such
a biodiesel production, based on glycerolysis-treated oil, resulted in a less energy intensive
route (0.251 MJ/kg biodiesel produced) than the conventional “esterification and transes-
terification” route (0.534 MJ/kg biodiesel produced) (Tu et al. 2017).

First Step Waste Cooking

Oils Glycerol
MeOH

FFAs + Glycerides

FAMEs
+ Glycerides
Glycerides Animal Grease

Biodiesel
Second Step

Figure 5.4 Scheme of two-step up-grading of waste oil and grease to biodiesel
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 127

Actually, the evaluation of the overall sustainability of this process needs to take into
account the effective recoverability of glycerol at the end of the whole process (glycerol-
ysis and transesterification). The presence of different contaminants in raw grease could
promote and induce side reactions of polymerization or decomposition of glycerol, even
into toxic compounds, such as acrolein. With respect to the environmental issue (check of
the emissions and treatment of toxic gaseous compounds), an economic constraint could be
associated in particular for brown grease, which has low glycerides content. In fact, degra-
dation of glycerin could be a critical aspect, since a partial refilling of fresh glycerol could
be always necessary, by compromising the economic profitability.

5.3.2 Heterogeneous Catalysis

In order to take advantage of an easy recoverability, several heterogeneous acid catalysts
were widely investigated. In fact, the use of “solid” acids allows an easy separation of the
catalyst from the final products at the end of the chemical pretreatment, which makes the
overall downstream simpler.
Strong acid cation exchange resins (Lewatit K2621, K2620, Amberlyst 15, 20BD, 35,
Dowex M 31, Tulsion T-42, Indion 130 [Sani et al. 2014; Park et al. 2010]) represent one
of the most investigated species. Largely available for several different industrial purposes,
they were reported as adequately active in promoting the direct esterification, under mild
conditions. Acceptable reduction of acidity, below 1 wt%, was obtained even starting from
pure acids, after two to six hours of reaction at 50–80 ∘ C and working with an oil:MeOH
molar ratio of 1:6. Specific surface analysis, number of acidic sites and the size of average
pore diameters play key roles in determining better performances (Özbay et al. 2008).
Another class of heterogeneous catalysts are H-form zeolites: microporous crystalline
solids with well-defined structures containing Si, Al, and O in their framework and cations.
Already used in oil refining, petrochemistry, and in the production of fine chemicals, they
can be prepared with strong Brønsted acid sites (also Lewis acid ones) with a quite good
resistance to high reaction temperatures. However, diffusion of the reactant molecules to
the active sites can be a limitation due to their small pore size (lower than 2 nm). In the
production of biodiesel, H-ZSM-5, H-MOR, H-BETA, and H-USY are found to show poor
esterification or transesterification catalytic activity, as a consequence of internal diffusion
limitations of the bulky reactant molecules (FFAs and glycerides) into the micropores of
zeolites (Kiss et al. 2006; Peters et al. 2006).
In addition several mixed metal oxides were found to be active in promoting direct ester-
ification and, in some cases, the transesterification also, in a one-pot conversion of acid oils
(Dibenedetto et al. 2014). In particular, bifunctional oxides, namely basic and acid ones, can
promote the two main reactions in a single step. High temperatures and oil:MeOH molar
ratio are however necessary.
It is worth specifying that a surface modification of some simple metal oxides (such
as ZrO2 , Ta2 O5 , Nb2 O5 , TiO2 ) through sulfuric acid, yield the relevant sulfated metal
oxides (SO4 2− /ZrO2 , SO4 2− /Ta2 O5 , SO4 2− /Nb2 O5 , and SO4 2− /TiO2 ), resulting in new
typical solid superacids. Since they possess both Brønsted and Lewis acid sites, they can
promote under relatively mild conditions direct esterification and transesterification of
FFAs and glycerides, respectively. Transesterification of tripalmitine (TP) with MeOH
128 Process Systems Engineering for Biofuels Development

was investigated under a N2 atmosphere at 65 ∘ C for 24 hours. TP (0.55 mmol) was

dispersed into 49 mmol of MeOH, and 5 wt% catalyst (referred to the overall reaction
mixture). A yield of 25% MP was obtained, but the use of some appropriately modified
catalysts (SO4 2− /ZrO2 –SiO2 (Et) hybrid material) produced more than 75% under the
same conditions (W. Li et al. 2010, 2012).
The use of some magnetic nano-sized solid acid catalysts has been recently proposed.
Basically, to a magnetic iron-based core may be grafted different functionalized arms (car-
bonaceous or also glycidylic), ending with sulfonic acid groups. In this way, to the acidity
of sulfonic groups, the easy recoverability of magnetic iron oxide was coupled, by guar-
anteeing good activity and recoverability. Direct esterification of FFAs (16 wt%) in grease
with MeOH using this kind of catalyst (4 wt%) gave 96% conversion to FAMEs within
two hours. At the end of a reactive run, the catalyst was easily separated under a magnetic
field and showed no loss of productivity during 10 cycles (Tan and Li 2012).
Direct production of biodiesel from Jatropha oil with high acid value (17.2 mg KOH/g)
was also positively tested at 200 ∘ C. In addition to a good (90.5% at 200 ∘ C) biodiesel yield,
there was also a very efficient recovery of the catalyst (Zhang et al. 2015).
Finally, very simple FeSO4 represents an effective example of heterogeneous catalysts,
even if it was demonstrated that the direct esterification reaction was promoted by the solu-
ble part of this salt and catalysis did not take place on the solid surface (Wang et al. 2007).
However, heterogeneous catalysis in general was quite expensive, often less active than
conventional mineral acids, and reusable for only limited cycles. The surface of catalysts
was subjected to passivation and/or deactivation phenomena (especially when very con-
taminated oils were treated) and the relevant reactivation requires costly procedures not
compatible with industrial constraints. In fact, washing operations with pure alcohol and/or
thermal treatment (up to calcination) were proposed as recovery procedures.

5.3.3 Enzymatic Biodiesel Production

5.3.3.1 The Use of Lipases for Biodiesel Production
A wide range of lipases and other esterases have been tested as catalysts for biodiesel pro-
duction (Jothiramalingam and Wang 2009; Fjerbaek et al. 2009). The alkaline catalysts are
able to convert glycerides as well as FFAs into biodiesel in the same mild conditions without
the formation of soaps. In addition, they offer several additional benefits:
• The utilization of low-quality and non-edible oils without a negative impact on the envi-
ronment (Tan et al. 2010).
• Fewer process steps with the reduction of energy consumption and wastewater volumes.
• Improved phase separation to obtain a higher quality of glycerol (Kumari et al. 2007;
Meher et al. 2006; Fukuda et al. 2001).
However, they present some drawbacks that limit the use on an industrial scale:
• Low reaction rate (Zhang et al. 2003b).
• Their costs (Andrade et al. 2019; Haas et al. 2006; Jaeger and Eggert 2002).
• Loss of activity.
Lipases are widely employed to catalyze hydrolysis, alcoholysis, esterification, and
transesterification of carboxylic esters.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 129

Several literature reports describe biodiesel synthesis with a range of enzymes, using
different feedstocks and alcohols under different conditions (Mata et al. 2017; Mulalee et al.
2016; Yan et al. 2012). The conversion of oils and fats into biodiesel is a complex reaction
in which several parallel reactions take place. The conversion of glycerides occurs through
the sequential transesterification reaction of TAG to DAG and subsequently to MAG and
finally to glycerol, by using MeOH, ethanol, propanol, and butanol as nucleophiles. At the
same time, FFAs can be esterified into the corresponding alkyl esters.
For several lipases, the esterification reaction is significantly faster with respect to the
transesterification process, since the smaller nucleophilic part of the former fits better in
the active site of the enzyme (Nordblad et al. 2016). The Candida antarctica lipase B
(CALB) has been extensively investigated in biodiesel applications for the high activity
and selectivity shown in the conversion of substrates rich in FFAs such as spent oils and
animal fats (Syrén and Hult 2010; Deng et al. 2005; Shimada et al. 2002; Watanabe et al.
2001, 2000). However, several papers describe the benefits of other lipases, Candida
rugosa lipase (CRL; Kaieda et al. 1999), Candida sp. 99–125 (C sp.; Lu et al. 2010; Nie
et al. 2006), Rhizomucor miehei lipase (RML; Nelson et al. 1996), Eversa Transform
(Andrade et al. 2017), Burkholderia cepacia (BCL) and Thermomyces lanuginosa lipase
(TLL; Li et al. 2006) that can be used alone or combined to catalyze both the processes (see
Table 5.1).
Many of these enzymes are commercially available, also in immobilized form, in order
to facilitate the relevant recovery and re-use. CALB, which is commercially known as
Novozym 435, has been used as a catalyst for the conversion of different vegetable oils
and WCO to obtain yields of FAMEs over 90–95% (Mulalee et al. 2016; Royon et al. 2007;
Du et al. 2004).
Candida sp. 99–125 lipase immobilized on textile membrane can catalyze the conversion
of several waste oils, lard, and various vegetable oils with yield higher than 90% (Lu
et al. 2007; Lee et al. 2002). Generally, enzymes show a higher yield and a longer lifetime
in substrates rich in FFAs than in substrates rich in triglycerides, but the presence of
phospholipids in the crude oil could inhibit the activity of lipases (Lai et al. 2005; Wei
et al. 2004). Therefore, preliminary treatments of degumming and filtration of starting oils
are often required before the process (Watanabe et al. 2002). Optimal temperature is in the
range between 50 ∘ C and 70 ∘ C, in relation to the nature of the enzyme and the thermal
stability of the carrier used. Generally, a stoichiometric amount of alcohol is required
for the complete conversion of glycerides and FFAs into the corresponding alkyl esters.
However, considering that an excess of alcohol causes inactivation of the enzyme, and
that the smaller the alcohol, the more important the inhibition effect, the stepwise addition
of MeOH during the process was the best strategy. Since the solubility of MeOH in the
alkyl esters is greater than in the oil, with the gradual addition of alcohol it is possible
to limit the enzyme deactivation (Shimada et al. 1999). This strategy was also effective
for other lipases such as C sp. (Lu et al. 2007), Rhizopus oryzae (Chen et al. 2006),
and Pseudomonas fluorescens (Soumanou and Bornscheuer 2003). Another approach to
resolve the lipase inactivation caused by MeOH, is the use of organic solvents. Royon
et al. (2007) have studied the enzymatic methanolysis of cotton seed oil in the presence
of solvents. Lipase inhibition was eliminated by using t-butanol as reaction medium, by
obtaining a noticeable increase of reaction rate and ester yields (97%).
Other solvents such as 1,4-dioxane (Iso et al. 2001) and ILs (Ha et al. 2007) might solve
the problem of lipase inactivation caused by MeOH, but other problems like environmental
130 Process Systems Engineering for Biofuels Development

Table 5.1 Enzymatic biodiesel production using various lipases.

Enzyme Substrate Alcohol Reaction conditions Yield References

CALB Sunflower oil MeOH 3:1 M alcohol 93.2% Deng et al.
10 wt% cat. based to oil (2005)
40 ∘ C, 24 h
Pretreated tallow MeOH 3:1 M alcohol 74% Lee et al.
(15% FFAs) 10 wt% cat. based to oil (2002)
30 ∘ C, 72 h
Restaurant grease EtOH 3:1 M alcohol >96% Wu et al.
(8.5% FFAs) 5 wt% cat. based to oil (1999)
30 ∘ C, 24 h
Waste oil MeOH 3:1 M alcohol 90% Watanabe
(2.5% FFAs) 1 wt% cat. based to oil et.al.
30 ∘ C, 24 h (2001)
Rice bran oil MeOH 3.5:1 M alcohol >98% Lai et al.
(85% FFAs) 5 wt% cat. based to oil (2005)
30 ∘ C, 7 h
CRL Soybean oil MeOH 3:1 M alcohol 90% Kaieda et al.
10 wt% cat. based to oil (1999)
45 ∘ C, 90 h
C sp. Soybean oil MeOH 3:1 M alcohol 83.2% Lu et al.
10 wt% cat. based to oil (2010)
40 ∘ C, 12 h
Waste oil MeOH 3:1 M alcohol 92% Nie et al.
(46.7% FFAs) 5 wt% cat. based to oil (2006)
40 ∘ C, 90 h
RO and CALB Soybean oil MeOH 4.5:1 M alcohol >99% Lee et al.
30 wt% cat. based to oil (2006)
45 ∘ C, 21 h
TLL and Waste oil MeOH 4:1 M alcohol 95% Li et al.
CALB (70% FFAs) 3 wt% cat. based to oil (2006)
35 ∘ C, 12 h
Eversa Castor oil MeOH 6:1 M alcohol 94% Andrade et al.
Transform 5 wt% cat. based to oil (2017)
35 ∘ C, 8 h

RO, Rhizopus oryzae.

issues (toxicity, emissions) and economic sustainability (due to recovery and losses of sol-
vents) are introduced and need to be taken into consideration.

5.3.4 ILs Biodiesel Production

5.3.4.1 ILs and Supported ILs Catalysts in Biodiesel Production
ILs are organic salts generally composed of an anion and an organic cation. With respect to
the inorganic salts, these are liquid at low temperature (even room temperature) and show
unique and advantageous properties (Greaves and Drummond 2008):
• Low vapor pressure.
• Good chemical and thermal stability.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 131

• Ability to dissolve a wide range of inorganic and organic compounds.

• Easy separation by obtaining products with high purity.
ILs can be used in various fields such as enzymatic catalysis (Zhao et al. 2002),
organic synthesis (Forbes et al. 2006), Diels–Alder cycloadditions (Janus et al. 2006),
Friedel–Crafts alkylations and acylations (Baleizão et al. 2004), spectroscopy and electro-
chemistry (Liu et al. 2010), polymerizations (Lu et al. 2009), nanomaterials (Tunckol et al.
2012), and extraction and separation processes (Nakashima et al. 2005).
In recent years, ILs have received much attention for their potential applications for
biodiesel production in replacing the conventional homogeneous and heterogeneous
catalysts. In fact, they are very effective in the direct esterification and transesterification
processes, due to the presence of Brønsted acid sites. In addition, they have a poor solubility
in the biodiesel phase, which facilitates the operations of separation at the end of the process
and reuse for several cycles of reaction (De Diego et al. 2011). N-triethylammonium sul-
fate ([Et3 NH][HSO4 ]), N-methyl-2-pyrrolidonium methyl sulfonate ([NMP][CH3 SO3 ]),
and 1-butyl-3-methylimidazolium hydrogen sulfate ([BMIM][CH3 SO3 ]) have been
positively tested as effective promoters in the conversion of non-edible oils and ani-
mal fats, as well as several –SO3 H functionalized strong Brønsted acidic ILs, e.g.
1-(4-sulfonic acid) butyl-pyridinium hydrogen sulfate ([BSPy][HSO4 ]), 1-(4-sulfonic
acid) butyl-pyridinium trifluoro methane sulfonate ([BSPy][CF3 SO3 ]) and 1-(3-sulfonic
acid) propyl-3-methylimidazolium hydrogen sulfate ([C3 SO3 H–mim][HSO4 ]) (Figure 5.5)
(Muhammad et al. 2015; Ullah et al. 2015; Su and Guo 2014, Liu et al. 2012; K.X. Li et al.
2010).
Zhang et al. (2009) have investigated the use of ILs containing imidazolium and pyrroli-
donium groups in biodiesel production, especially starting from pure FFAs or their mix-
tures. [NMP][CH3 SO3 ] showed the best catalytic behavior in the esterification of oleic acid
with ethanol, with a yield of 95.3%, into ethyl esters after eight hours at 70 ∘ C (molar ratio
of oleic acid:ethanol:IL = 1:12:0.2). In addition, the catalyst can be recovered at the end of
the process and reused up to eight times, maintaining a conversion of above 90%.

–
HSO4 – –
CH3SO3 HSO4
H O
+
N N + N
+
N H

[Et3NH][HSO4]
[NMP][CH3SO4] [BMIM][CH3SO3]

– – –
HSO4 SO3H CF3SO3 HSO4
SO3H

+ N + N
N +
N
SO3H

[BSPy][HSO4] [BSPy][CF3SO3] [C3SO3H–min][HSO4]

Figure 5.5 Chemical structure of –SO3 H functionalized strong Brønsted acidic ILs.
132 Process Systems Engineering for Biofuels Development

Han et al. (2009) reported biodiesel preparation from WCO using novel Brønsted acidic
ILs with an alkane sulfonic group (R-SO3 H) as catalysts, with yields of over 93.5% after
four hours at 170 ∘ C (molar ratio of oil:MeOH:IL = 1:12:0.06).
ILs containing pyridine rings were studied by K.X. Li et al. (2010, 2013) in the ester-
ification and transesterification of Jatropha oil, which presents a high content of FFAs.
A FAMEs yield of 95.1% was obtained after six hours at 100 ∘ C under a molar ratio of
oil:MeOH:IL = 1:18:0.15. The product was easily recovered from the reaction medium,
and [BSPy][CF3 SO3 ] in particular maintained its catalytic activity for seven reuses.
Olkiewicz et al. (2016) studied biodiesel production from sludge lipids catalyzed by dif-
ferent acidic imidazolium and long chain ammonium liquids with an alkane sulfonic acid
group and different anions. The FAMEs yield reached 90% at 100 ∘ C after five hours, with
a molar ratio of MeOH to saponifiable lipids of 10:1 and 7 wt% of catalyst with respect to
lipids.
However, in some cases ILs present some disadvantages such as partial catalyst loss
during the process, limited solubility with organic compounds (especially polar molecules),
and high viscosity that limits their application on an industrial scale.
To overcome these problems, the immobilization of ILs (on porous silica, polystyrene-
based polymers, polydivinylbenzene, and inorganic material supports) has been adopted
(Claus et al. 2018; Pârvulescu and Hardacre 2007; Mehnert 2005).
Karimi and Vafaeezadeh (2012) developed a supported 1-methyl-3-octylimidazolium
hydrogen sulfate on SBA-15-functionalized propylsulfonic acid ([MOIm]-HSO4 @SBA-
15-Pr-SO3 H) that shows a high activity in the direct esterification of different FFAs with
ethanol at room temperature (yields = 91–87%, molar ratio of FFAs:ethanol:catalyst =
1:0.6:0.1). The catalyst can be reused for four cycles without significant decrease of activity.
Liang (2013) reported the use of a novel liquid polymer based on the copolymerization
of IL acid oligomers with divinylbenzene (PDVB-[SO3 H(CH2 )3 VPy]HSO4 ), able to cat-
alyze the conversion of WCO with high FFAs content, obtaining a FAMEs yield of 99.1%
after 12 hours at 70 ∘ C (molar ratio of oil:MeOH = 1:15, 1 wt% of catalyst with respect
to oil). The catalyst was easily recovered by filtration and recycled six times without loss
of activity. Zhang et al. (2012) investigated the efficiency of Brønsted acidic IL supported
onto Fe-incorporated SBA-15 (Fe-SBA-15) for esterification of oleic acid with short-chain
alcohols. Using a molar ratio of oleic acid to MeOH = 1:6 and 5 wt% of catalyst, a yield
of 87.7% was obtained after three hours at 90 ∘ C. The reusability of IL/Fe-SBA-15 catalyst
was also evaluated with a yield of 80.8% after six reaction cycles.
He et al. (2013) synthetized and tested long-chain Brønsted acid ILs, 3-(N,N-dimethyl
dodecylammonium) propane sulfonic acid p-toluene sulfonate ([DDPA][Tos]), for synthe-
sis of biodiesel from FFAs. FAMEs were obtained in good yield (92.5–96.5%) at 60 ∘ C and
three hours with molar ratio of FFAs:alcohol:catalyst = 1:1.5:0.1. The catalysts could be
recycled up to nine times.

5.3.4.2 Combined Use of Lewis Acids with ILs in the Conversion of Non-edible Oils
Bifunctional catalysts with Brønsted and Lewis acid sites can improve the catalytic activity
in biodiesel production from non-edible oils due to their synergistic effect. The Brønsted
acid functionality promotes the direct esterification of FFAs, while the Lewis acid site facil-
itates the transesterification of glycerides (Di Serio et al. 2005).
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 133

A preliminary study of the effect of Lewis acids in the transesterification of soybean oil
with MeOH was conducted by Liang et al. (2009). Choosing triethylammonium chloride
([Et3 NH]Cl) as the model ammonium salt, different anhydrous metal chlorides were added
and the catalytic activity in the transesterification process was evaluated. [Et3 NH]Cl–AlCl3
showed the highest activity with a yield of 98.5% at 70 ∘ C after nine hours of reaction (molar
ratio of oil:MeOH:catalyst = 1:12:0.7). Conversely, Lewis acids such as Mg2+ and Zn2+
showed lower yields (32.1 and 64.2%, respectively), due to the characteristics of the metal
centers.
Liu et al. (2013) tested Brønsted–Lewis acidic ILs [HO3 S-(CH2 )3 -NEt3 ]Cl-xFeCl3
(molar fraction of FeCl3 , x = 0.67) in the conversion of waste oil with high level of FFAs
(acid value = 34.6 mg KOH/g). A yield of over 95% was obtained at 120 ∘ C after four hours
of reaction, due to the synergetic effects of the Brønsted and Lewis acid catalytic sites.
Li et al. (2014) prepared various Brønsted–Lewis catalysts using Brønsted acidic
IL 1-sulfobutyl-3-methylimidazolium hydrogen sulfate ([BSO3 HMIM]HSO4 ) with a
series of metal sulfates that were applied in the Camptotheca acuminata seed oil (acid
value = 23.7 mg KOH/g). [BSO3 HMIM]HSO4 -Fe2 (SO4 )3 showed the best catalytic
activity with respect to the other metal sulfates with a FAMEs yield of 95.7% (molar
ratio of oil:MeOH = 1:5, 5 wt% of catalyst with respect to oil). The order of activity was
as follows: Fe3+ > Cu2+ > Zn2+ > Mg2+ > Ca2+ in line with the respective capability to
accept electrons.
Finally, Casiello et al. (2019) proposed a very cheap binary catalytic system ZnO/TBAI
(where TBAI = tetrabutylammonium iodide) for the simultaneous transesterification and
esterification of vegetable oils, animal fats, WCO, and municipal sewage scum. A maximum
yield of 96% was achieved at 65 ∘ C after seven hours of reaction. Also, in this case, the
catalytic system can be separated and reused for five cycles without significant decrease of
activity.

5.3.4.3 Drawbacks Related to the Use of ILs

As already mentioned for enzymatic processes, IL systems suffer from economic viabil-
ity. In fact, as already stated above, ILs not only need high methanol:oil ratio, which bring
high costs in recovering final products, but they are really expensive in themselves. In addi-
tion, their effectiveness may be negatively influenced by the presence of water and/or other
contaminants contained in waste oils, compromising the reuse, and needing a proper final
treatment.
Different from the enzymes, they are very toxic compounds and have a problematic dis-
posal in terms of the environmental impact. In fact, at the end of their use they need to be
oxidized to more innocuous products.
All these aspects made the use of ILs for the production of biodiesel inconvenient and a
long way from being industrially applied.

5.3.5 Use of Metal Hydrated Salts

Most recently, new sustainable alternatives were found in using hydrated chlorides and
nitrates (Pastore and di Bitonto 2017), due to their cheapness and environmental safety.
134 Process Systems Engineering for Biofuels Development

Even commonly used as coadjuvants in wastewater treatment in sedimentation processes,

they were sufficiently active in promoting FFAs conversion into FAMEs under very mild
conditions. Real WCOs, having an acidity of 8.04 mg KOH/g, were reacted with MeOH
in the presence of 7 mol% of different hydrated chlorides and nitrates. The residual acidity
was found to be linearly correlated with the pH of the corresponding aqueous solutions
prepared by using their reaction concentration. Only some cases did not follow this general
trend.

6 6
H2O CuCI H2O
MeOH MeOH Zn(NO3)2·6H2O
5 CaCI2·2H2O 5
MgCI2·6H2O Ni(NO3)2·6H2O
4 CoCI2·2H2O 4 Cd(NO3)2·4H2O
Cu(NO3)2·2.5H2O
AlCI3·6H2O MnCI2·4H2O
pH

pH
3 3
Al(NO3)3·9H2O
2 SnCI2·2H2O
2 Fe(NO3)3·9H2O
FeCI3·6H2O
AlCI3·6H2O SnCI2·2H2O
1 1 Fe(NO3)3·9H2O
FeCI3·6H2O Al(NO3)3·9H2O
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5
Residual acidity (wt%) Residual acidity (wt%)
(a) (b)

Figure 5.6 Correlation between acidity (pH) and efficacy in promoting direct esterification of FFAs in
WCO expressed as residual acidity of (a) hydrated metal chlorides and (b) hydrated metal nitrates.

On the other hand, once these particular metal salts were dissolved into MeOH, the cor-
relation of the reactivity versus acidity was found better addressed (see the specific points
in Figure 5.6). Such a behavior can be attributed to the formation of new super-acid species
in an alcoholic environment, in particular for cases of aluminum, iron, and tin dissolved
in MeOH. After the formation of partially solvated species, obtained from the substitution
of coordinated water with MeOH, a rearrangement into dimeric (oligomeric) chains with
bridging water molecules was observed through electrospray ionization-mass spectrometry
experiments. The acidity of such hydrogen was clearly improved with respect to that of the
original hexa-aquo complex (Figure 5.7; Pastore et al. 2015a).
This provided evidence that, despite aluminum chlorides being typically identified as
Lewis acids, when dissolved in MeOH they act as Brønsted acids.
This point was also further experimentally demonstrated by studying in detail the reaction
of direct esterification of FFAs into FAMEs through the use of AlCl3 ⋅6H2 O (Eq. (5.1)))
(Pastore et al. 2014).
AlCl3 • 6H2 O
kFAMEs (5.1)
RCOOH + MeOH ROOMe + H2O
kFFAs

Kinetic experiments were carried out with the aim of evaluating the effect of the temper-
ature (40, 55, and 70 ∘ C) on the conversion of FFAs in FAMEs. The reduction of acidity
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 135

H H
O
(H2O)n–1(MeOH)6–nAI3 AI3 (H2O)n–1(MeOH)6–n 6 CI
O
H H
(b)
MeOH

–H2O
H
3 H
O
2 3CI (H2O)n–1(MeOH)6–nAI3 AI3 (H2O)n–1(MeOH)6–n 6 CI
O

H
H

(a) (c)

Figure 5.7 Mechanism of formation of alumino-aquo-MeOH super-acid species: (a) structure of starting
AlCl3 ⋅6H2 O; (b) dimeric chains with bridging water molecules, obtained from the partial substitution of
coordinated water with MeOH; and (c) super-acid species generated in an alcoholic environment.

and the concomitant formation of FAMEs were evaluated as a function of time with a
MeOH:FFAs ratio of 10:1.
The direct esterification was modeled as a second-order reversible reaction and consider-
ing the disappearance of the reagents. The rate of the reaction was studied by using Eq. (5.2):
dCFFAs m
ν=− = FFAs (kFFAs CFFAs CMeOH − kFAMEs CFAMEs CH2 O ) (5.2)
dt Mmix
nFFAs,t
CFFAs = (5.3)
Mmix
nMeOH,t nMeOH,t0 − nMeOH,react nMeOH,t0 − (nMeOH,t0 − nFFAs,t )
CFFAs = = = (5.4)
Mmix Mmix Mmix
nFAMEs,t nFFAs,t0 − nFFAs,t
CH2 O = CFAMEs = = (5.5)
Mmix Mmix

Under conditions defined by Eqs. (5.3–5.5), Eq. (5.2) became a nonlinear differential
equation in which the reaction time was the independent variable, while kFFAs and kFAMEs
were the variables to be determined.
Figure 5.8a shows the experimental points at 40, 55, and 70 ∘ C and the relevant fitting
curves were calculated through mathematical modeling, while in Figure 5.8b the lnKeq ,
ln(kFFAs ), and the ln(kFAMEs ) were plotted versus 1/T.
The expected linear trends (with an R2 higher than 0.998) were confirmed and the acti-
vation energies for the production of FAMEs and the back-reaction of hydrolysis were
appropriately calculated (43.9 and 24.7 kJ/mol) and found to be in good agreement with
values reported for direct esterification catalyzed by homogeneous strong acids, namely
136 Process Systems Engineering for Biofuels Development

1.5

Concentration (mmolFFAs/g solution)

40°C
55°C
70°C
Fitting 40°C
1
Fitting 55°C
Fitting 70°C

0.5

0
0 1 2 3 4 5 6 7 8
Time (h)
(a)

–1
In(kFFSs or FAMEs)

–2
In kFFAs
In kFAMEs
–3
In Keq
–4

–5

–6

2.90 2.95 3.00 3.05 3.10 3.15 3.20

1/T (K–1) × 10–3
(b)

Figure 5.8 (a) FFAs kinetic profiles obtained at 40, 55, and 70 ∘ C by reacting MeOH:FFAs:AlCl3 ⋅6H2 O
10:1:0.02. (b) Van’t Hoff plot and representations of ln(kFFAs ) and ln(kFAMEs ) with respect to 1/T.

sulfuric acid. The experimental evidence not only shows that AlCl3 ⋅6H2 O is as active as
sulfuric acid, but also suggests that it works as a Brønsted acid.
Such an activity in promoting direct esterification was confirmed with real samples of
WCO and WAG (di Bitonto and Pastore 2019).
WCO with a starting acidity of 8.04 mg KOH/g was brought to 0.77 mg KOH/g after
four hours at 70 ∘ C. AlCl3 ⋅6H2 O works in the homogeneous phase, since it is completely
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 137

soluble in MeOH, but at the end of the pretreatment, it induced a very advantageous
separation. In the upper MeOH phase, most of the catalyst was dissolved, together with
most of the residual FFAs, water, and contaminants salts, present in the initial feedstocks.
However, in the bottom organic layer, most of FAMEs and glycerides were solubilized,
resulting in a mixture ready to be directly transesterified using conventional alkaline
catalysis, without any further treatments. Such a separation, not only allows the catalyst
to be efficiently recoverable and re-usable for new pretreatment cycles, but also a series
of inconvenient pretreatments and downstream processes became unnecessary (such
as washing or neutralization of the residual acid catalyst onto the pretreated oil before
conventional alkaline transesterification, and preparation of starting feedstocks, etc.).
This behavior was confirmed in the pretreatment of WAG, where its conversion into
biodiesel was known to be hampered by certain limitations that have to be overcome.
The huge presence of proteins, phosphoacylglycerols, water, FFAs, pathogens etc., usu-
ally requires some preliminary operations to be conducted on raw WAG, in order to obtain
a more suitable substrate. For example, protein and phosphoacylglycerols (so-called gums)
need to be removed through a degumming process before biodiesel production. Also, the
huge presence of water and salts require some complicated and appropriate operations for
the relevant removal (Banković-Ilić et al. 2014). AlCl3 ⋅6H2 O was also positively tested
for the pretreatment of WAG. In detail, when a real sample of WAG was pretreated with
AlCl3 ⋅6H2 O, not only were 84% of the initial FFAs converted to FAMEs, but also the sub-
stantial removal of Na (62.7%), Ca (54.1%), Mg (83.5%), and P (90.2%) was specifically
revealed in the pretreated oily phase. At the end, the residual ash content was reduced to
one fourth of its initial value (437–612 ppm versus 1718 ppm).
If from one side this one-pot removal of water and salts from a waste oily feedstock asso-
ciated with the efficient conversion of FFAs to FAMEs can be seen as an advantage, some
specific drawbacks need to be considered. In particular, direct reuse of the methanolic phase
(Pastore et al. 2015a) could become more complicated by the increasing presence of water
and further contaminants. In fact, when the methanolic phase after a cycle of pretreatment
of WCO was recovered and directly reused for a second and a third reaction cycle with
fresh WCO, the residual acidities of the resulting oily phases after treatment at 70 ∘ C and
four hours were 1.51 and 4.35 mg KOH/g, respectively. These values were higher than the
0.77 mg KOH/g which was obtained after the first cycle under the same conditions. In order
to obtain residual acidities within 1 mg KOH/g, longer reaction times (8–12 hours) were
required, evidencing a kinetic inhibition. The effect of water on this reduction in reaction
rate of the direct esterification was specifically investigated. Considering the biphasic nature
of the reactive system in the pretreatment, where a methanolic phase was well separated by
an oily layer, the distribution of FFAs among these two phases was monitored in the pres-
ence of different amounts of water. The higher the water content in the reactive system, the
lower the amount of FFAs dissolvable in the methanolic layer. For instance, reactive mix-
tures containing 20 000 ppm of water, which can be achieved for the development of water
from the direct esterification only, the maximum concentration of dissolved FFAs in the
MeOH phase resulted in a halved value (21 mmol/l) with respect to that in anhydrous media
(43 mmol/l). The almost complete dissolution of the catalyst in MeOH showed that the
direct esterification reaction should occur in this specific part of the reactive system. For that
reason, the slowing down of the migration of FFAs from the waste oil/grease to MeOH due
to the presence of water, may have a negative influence on the overall kinetics of the process.
138 Process Systems Engineering for Biofuels Development

In any case, dewatering of the methanolic phase recovered from direct esterification was
positively achieved through the use of molecular sieves (3 Å) and silica.
In fact, after the dewatering treatment with these drying agents, the water content in the
regenerated methanolic phase dropped to less than 100 ppm. When recycling and reuse was
carried out by using this regenerated methanolic phase, the residual acidities of the resulting
treated oils were always lower than 1 mg KOH/g after four hours of treatment.
AlCl3 ⋅6H2 O has good capability in promoting direct esterification. It is completely solu-
ble in the MeOH layer during and after reaction. It induces the efficient separation of an oily
phase prompt to be directly reacted through alkaline transesterification. All these properties
allowed a new configuration of biodiesel plant to be designed (Figure 5.9).
Figure 5.9 shows the two-step process to obtain biodiesel from waste grease (WCO,
WAG, etc.) through the use of AlCl3 ⋅6H2 O as a catalyst in the pretreatment. It consists of
the following operations:
i. Waste raw grease having FFAs content of 1–90 wt% can be directly used as a feedstock
without any further preliminary treatment.
ii. AlCl3 ⋅6H2 O is solubilized in MeOH and added to the feedstock (i) in a batch reactor
(R1).
iii. Direct esterification can be operated at 70 ∘ C, for 2–16 hours.
iv. Two distinguished phases can be separated using a static separator (S1): a methanolic
layer (up) and an oily phase (down).
v. The oily phase can be directly transesterified with MeOH under alkaline catalysis, for
example using sodium hydroxide 0.5 wt% at 70 ∘ C for one hour in a new reactor (R2).
vi. A biphasic system is obtained in this case: FAMEs can be easily recovered in the upper
phase from glycerol in S2.

MeOH

Direct esterification (pretreatment)

dw
S1 R1
AICI3·6H2O
in MeOH
Raw grease

oil R2 S2
MeOH

DC
MeOH/base
Biodiesel
Transesterification

Glycerol

Figure 5.9 Scheme of the process to obtain biodiesel from WCO and WAG using AlCl3 ⋅6H2 O as a cat-
alyst in the pretreatment.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 139

vii. Unreacted MeOH can be distilled, recovered, and eventually recycled back to the reac-
tion in R2, by residuing crude biodiesel.
viii. Crude biodiesel can be washed through water and then dried through evaporation.
ix. The methanolic phase separated from step (iv) in which the AlCl3 ⋅6H2 O was mostly
dissolved can be treated with drying agents (dw) and reused for new cycles of direct
esterification in R1 with fresh raw grease.
The present scheme of production results in a more simplified process with respect to
the conventional two-step configuration which uses H2 SO4 and NaOH as catalysts. In fact,
pretreated oils can be directly transesterified under alkaline conditions and without any fur-
ther washing, and the catalyst can be efficiently recovered into the methanolic phase, which
can be recycled back for pretreating fresh raw grease in a new reactive cycle. Furthermore,
brown grease, sewage scum or substrates poor in glycerides can be easily converted into
biodiesel through the use of a simple direct esterification, without adopting the transesteri-
fication step (Pastore et al. 2015b). In this context, new sources of grease, namely sewage
sludge, could further implement the actual scenario of possible second-generation feed-
stock (Pastore et al. 2013). In particular, lipids that are in urban primary sludge could be
also conveniently converted into biodiesel through the abovementioned process scheme.
In any case, the use of waste oils, in which “inert and heavy” compounds are also present,
strongly needs a further final purification step, likewise distillation of FAMEs as pure prod-
ucts, in order to satisfy EN14214 specifications (FAMEs over 96.5 wt%).
The robustness of the catalyst was clearly demonstrated for different typology of feed-
stocks having very different starting compositions.

5.4 Future Trends of the Pretreatments of Waste Oils

In the future, improvements will be necessary to consistently decrease the overall economic
and environmental impact of the current production of biodiesel. Regarding the reaction
conditions, several innovative solutions have been already proposed. Enzymatic catalysts,
ILs, and hydrated salts resulted in very efficient promoters for the direct production of
biodiesel, or for the pretreatment of the starting waste oil. However, currently they are not
ready for industrial application because of economic limits.
Actually, these limits are not directly related to the reactive step, but more principally to
the costs of lipid feedstocks and the downstream purification of the final biodiesel. For this
reason, improvement of the present state of the art in biodiesel production can be achieved
more by developing new side technologies (i.e. recovering lipids from innovative feed-
stocks, purification processes, water separation from alcohol, etc.) than by improving the
reactive step.
The identification of new feedstocks represents a really challenging objective for the
future of this field. For example, in the European context, satisfaction of the economic con-
straints could be more easily achievable by using second-generation feedstock, for which
the double-counting rule could be applied. Besides the well-known WAG and WCO, sewage
grease, trap grease and grease recoverable from municipal waste (di Bitonto et al. 2018)
could represent profitable alternatives which satisfy the viability criteria.
The use of these biomasses can have a great economic potential: according to the lit-
erature data (Pastore and di Bitonto 2017), over 5 Mton/yr of biodiesel could be obtained
140 Process Systems Engineering for Biofuels Development

annually in Europe from sewage sludge only. This massive amount would represent half of
the present continental demand.
For these specific lipids, aluminum hydrated salts were very effective in promoting the
direct esterification reaction with methanol (di Bitonto et al. 2016; Pastore et al. 2014,
2015a). AlCl3 ⋅6H2 O not only satisfied the criterion of economic viability, but also guar-
anteed a high robustness in reactivity of these waste greases, which are very contaminated
by salts, water, polymers, etc.
Finally, even improvement of technologies for conducting side operations (separation of
water from methanol and/or purification of products in general) through integration and
intensification of processes may significantly influence the overall economic scenario of
biodiesel production. Development of these processes may make proficient some technolo-
gies that are already available but not usable currently due to some drawbacks and limits
related to side processes.

5.5 Conclusions
In the present chapter, a critical overview of the current biodiesel production scenario was
reported.
The industrial production is nowadays based on the application of simple homogeneous
acid and alkaline catalysts. However, they present several drawbacks, especially in terms
of bad recoverability and generation of waste.
Heterogeneous catalysts have been proposed as alternatives: in some cases they resulted
in very active compounds, but very rarely can they be considered economically viable. In
addition, they suffer from the problem of contamination of the active surface and further
costs need to be considered for their regeneration.
Lipase-catalyzed esterification and transesterification can be carried out in the presence
of relatively high water content, at low temperature, and convert more feedstock to biodiesel
in a single step. Nevertheless, it is costly in particular due to the expense of the enzymes
(especially for immobilization) and for the long reaction time to have good yields.
Conversion of waste oils/grease mediated by ILs resulted in easier separations of the final
products, due to formation of two phases, which reduces process costs. In addition, prop-
erties of catalysts can be modulated during the relevant preparation: high catalytic activity,
excellent stability and easy separability and reusability are the most important qualities.
On the other hand, they often require relatively more alcohol for producing effective yields,
which increases the overall production cost for the corresponding increase in the separation
costs of the final pure product. They are too expensive in themselves, especially if the final
treatment and disposal of spent residues have been taken into account.
Cheap and safe hydrated salts efficiently convert FFAs into methyl esters. They mediated
the direct esterification of FFAs into the methanolic phase. Hydrated metal chlorides and
nitrates are solubilized in MeOH (homogeneous catalysis), and for the same reason, they
can be easily reused several times, by simply recycling the methanolic phase after a reaction
cycle. They resulted in very effective and robust catalysts, to the point to be efficiently
applied for pretreating WCO, WAG, and sewage grease to obtain biofuel. Considering that
the final use of exhausted aluminum- and iron-based hydrated salts may be as coagulant in
WWTPs, they actually result in a “zero-waste” process.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 141

Acknowledgment
This research was partially supported by FESR “PON Ricerca e Innovazione 2014–2020.
Progetto: Energie per l’Ambiente TARANTO—Cod. ARS01_00637”.

Abbreviations

BCL Burkholderia cepacia

CFAMEs Concentration of FAMEs as mmol/gsolution
CFFAs Concentration of FFAs as mmol/gsolution
CH2O Concentration of H2 O as mmol/gsolution
CMeOH Concentration of MeOH as mmol/gsolution
C sp. Candida sp. 99–125
CRL Candida rugosa lipase
DAG Diacylglycerol
DC Distillation column
dw Dehydration of methanolic phase
FAMEs Fatty acid methyl esters
FFAs Free fatty acids
ILs Ionic liquids
Keq Equilibrium constant
kFFAs Kinetic constants referring to the direct esterification reaction
kFAMEs Kinetic constants of the hydrolysis
MAG Monoacylglycerol
mcat Amount of catalyst used in a batch reaction
Mmix Total mass of the solution
MP Methyl palmitate
MeOH Methanol
nFFAs,t Number of moles of FFAs at time t
nMeOH,t Number of moles of MeOH at time t
nFAMEs,t Number of moles of FAMEs at time t
nFFAs,t0 Number of moles of FFAs at time t = 0
nMeOH,t0 Number of moles of MeOH at time t = 0
nMeOH ,react Number of moles of MeOH reacted at time t
R1 Batch reactor for direct esterification
R2 Batch reactor for transesterification
RML Rhizomucor miehei lipase
RO Rhizopus oryzae
TAG Triacylglycerol
TLL Thermomyces lanuginosa lipase
TP Tripalmitine
𝜈 Rate of reaction
WAG Waste animal grease
WCO Waste cooking oil
WWTP Wastewater treatment plant
142 Process Systems Engineering for Biofuels Development

References
Ajanovic, A. (2013). Renewable fuels–a comparative assessment from economic, energetic and ecological
point-of-view up to 2050 in EU-countries. Renewable Energy 60: 733–738.
Alhassan, Y., Kumar, N., Bugaje, I.M. et al. (2014). Co-solvents transesterification of cotton seed oil into
biodiesel: effects of reaction conditions on quality of fatty acids methyl esters. Energy Conversion and
Management 84: 640–648.
Anderson, E. (2014). Glycerolysis for lowering free fatty acid levels. Render 43: 34–35.
Andrade, T.A., Errico, M., and Christensen, K.V. (2017). Influence of the reaction conditions on the enzyme
catalyzed transesterification of castor oil: a possible step in biodiesel production. Bioresource Technology
243: 366–374.
Andrade, T.A., Martín, M., Errico, M., and Christensen, K.V. (2019). Biodiesel production catalyzed by
liquid and immobilized enzymes: optimization and economic analysis. Chemical Engineering Research
and Design 141: 1–14.
Ashraful, A.M., Masjuki, H.H., Kalam, M.A. et al. (2014). Production and comparison of fuel properties,
engine performance, and emission characteristics of biodiesel from various non-edible oils: a review.
Energy Conversion and Management 80: 202–228.
Atabani, A.E., Silitonga, A.S., Ong, H.C. et al. (2013). Non-edible vegetable oils: a critical evaluation of
oil extraction, fatty acid compositions, biodiesel production, characteristics, engine performance and
emissions production. Renewable & Sustainable Energy Reviews 18: 211–245.
Baleizão, C., Pires, N., Gigante, B., and Curto, M.J.M. (2004). Friedel–Crafts reactions in ionic liquids: the
counter-ion effect on the dealkylation and acylation of methyl dehydroabietate. Tetrahedron Letters 45:
4375–4377.
Banković-Ilić, I.B., Stojković, I.J., Stamenković, O.S. et al. (2014). Waste animal fats as feedstocks for
biodiesel production. Renewable & Sustainable Energy Reviews 32: 238–254.
Baskar, G. and Aiswarya, R. (2016). Trends in catalytic production of biodiesel from various feedstocks.
Renewable & Sustainable Energy Reviews 57: 496–504.
di Bitonto, L. and Pastore, C. (2019). Metal hydrated-salts as efficient and reusable catalysts for pre-treating
waste cooking oils and animal fats for an effective production of biodiesel. Renewable Energy 143:
1193–1200.
di Bitonto, L., Lopez, A., Mascolo, G. et al. (2016). Efficient solvent-less separation of lipids from municipal
wet sewage scum and their sustainable conversion into biodiesel. Renewable Energy 90: 55–61.
di Bitonto, L., Antonopoulou, G., Braguglia, C. et al. (2018). Lewis-Brønsted acid catalysed ethanolysis of
the organic fraction of municipal solid waste for efficient production of biofuels. Bioresource Technology
266: 297–305.
Borugadda, V.B. and Goud, V.V. (2012). Biodiesel production from renewable feedstocks: status and oppor-
tunities. Renewable & Sustainable Energy Reviews 16: 4763–4784.
Bozbas, K. (2008). Biodiesel as an alternative motor fuel: production and policies in the European Union.
Renewable & Sustainable Energy Reviews 12: 542–552.
Brunner, K., Frische, R., and Ricker, R. (2000). Method for the production of fatty acid esters. US Patent
7109363 B2.
Canakci, M. (2007). The potential of restaurant waste lipids as biodiesel feedstocks. Bioresource Technology
98: 183–190.
Canakci, M. and Van Gerpen, J. (2001). Biodiesel production from oils and fats with high free fatty acids.
Transactions of the ASAE 44: 1429–1436.
Casiello, M., Catucci, L., Fracassi, F. et al. (2019). ZnO/Ionic liquid catalyzed biodiesel production from
renewable and waste lipids as feedstocks. Catalysts 9: 71–84.
Chai, M., Tu, Q., Lu, M., and Yang, Y.J. (2014). Esterification pretreatment of free fatty acid in biodiesel
production, from laboratory to industry. Fuel Processing Technology 125: 106–113.
Chen, G., Ying, M., and Li, W. (2006). Enzymatic conversion of waste cooking oils into alternative
fuel-biodiesel. Applied Biochemistry and Biotechnology 132: 911–921.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 143

Claus, J., Sommer, F.O., and Kragl, U. (2018). Ionic liquids in biotechnology and beyond. Solid State Ionics
314: 119–128.
CONOE (2018). CONOE: il sistema. http://www.conoe.it/wp-content/uploads/2018/11/ANNUAL-
REPORT-2018.pdf (accessed 31 March 2020).
Cvengroš, J. and Cvengrošová, Z. (2004). Used frying oils and fats and their utilization in the production
of methyl esters of higher fatty acids. Biomass and Bioenergy 27: 173–181.
De Diego, T., Manjón, A., Lozano, P. et al. (2011). An efficient activity ionic liquid-enzyme system for
biodiesel production. Green Chemistry 13: 444–451.
Deng, L., Xu, X.B., Haraldsson, G.G. et al. (2005). Enzymatic production of alkyl esters through alco-
holysis: a critical evaluation of lipases and alcohols. Journal of the American Oil Chemists’ Society 82:
341–347.
Deng, X., Fang, Z., and Liu, Y.H. (2010). Ultrasonic transesterification of Jatropha curcas L. oil to biodiesel
by a two-step process. Energy Conversion and Management 51: 2802–2807.
Di Serio, M., Tesser, R., Dimiccoli, M. et al. (2005). Synthesis of biodiesel via homogeneous Lewis acid
catalyst. Journal of Molecular Catalysis A: Chemical 239: 111–115.
Dibenedetto, A., Angelini, A., Colucci, A. et al. (2014). Tunable mixed oxides: efficient agents for the
simultaneous transesterification of lipids and esterification of free fatty acids from bio-oils for the effec-
tive production of fames. International Journal of Renewable Energy and Biofuels 204112: 1–18.
Directive 2009/28/EC (2009). Directive 2009/28/EC of the European Parliament and of the Council of 23
April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently
repealing Directives 2001/77/EC and 2003/30/EC.
Du, W., Xu, Y., Liu, D., and Zeng, J. (2004). Comparative study on lipase-catalyzed transformation of
soybean oil for biodiesel production with different acyl acceptors. Journal of Molecular Catalysis B:
Enzymatic 30: 125–129.
European Regulation No. 1069/2009 (2009). European Regulation No. 1069/2009 of the European Parlia-
ment and of the Council of 21 October 2009 laying down health rules as regards animal by-products and
derived products not intended for human consumption and repealing Regulation (EC) No. 1774/2002
(Animal by-products regulation).
Eze, V.C., Phan, A.N., and Harvey, A.P. (2018). Intensified one-step biodiesel production from high water
and free fatty acid waste cooking oils. Fuel 220: 567–574.
Felizardo, P., Correia, M.J.N., Raposo, I. et al. (2006). Production of biodiesel from waste frying oils. Waste
Management 26: 487–494.
Felizardo, P., Machado, J., Vergueiro, D. et al. (2011). Study on the glycerolysis reaction of high free fatty
acid oils for use as biodiesel feedstock. Fuel Processing Technology 92: 1225–1229.
Fjerbaek, L., Christensen, K.V., and Norddahl, B. (2009). A review of the current state of biodiesel produc-
tion using enzymatic transesterification. Biotechnology and Bioengineering 102: 1298–1315.
Forbes, D.C., Law, A.M., and Morrison, D.W. (2006). The Knoevenagel reaction: analysis and recycling
of the ionic liquid medium. Tetrahedron Letters 47: 1699–1703.
Fukuda, H., Kondo, A., and Noda, H. (2001). Biodiesel fuel production by transesterification of oils. Journal
of Bioscience and Bioengineering 92: 405–416.
Garcia, R. (2015). Improved method for producing biodiesel from natural and recycled vegetable oils. WO
2015 162307.
Gebremariam, S.N. and Marchetti, J.M. (2018). Techno-economic feasibility of producing biodiesel from
acidic oil using sulfuric acid and calcium oxide as catalysts. Energy Conversion and Management 171:
1712–1720.
Greaves, T.L. and Drummond, C.J. (2008). Protic ionic liquids: properties and applications. Chemical
Reviews 108: 206–237.
Ha, S.H., Lan, M.N., Lee, S.H. et al. (2007). Lipase-catalyzed biodiesel production from soybean oil in
ionic liquids. Enzyme and Microbial Technology 41: 480–483.
144 Process Systems Engineering for Biofuels Development

Haas, M.J., McAloon, A.J., Yee, W.C., and Foglia, T.A. (2006). A process model to estimate biodiesel
production costs. Bioresource Technology 97: 671–678.
Han, M., Yi, W., Wu, Q. et al. (2009). Preparation of biodiesel from waste oils catalyzed by a Brønsted
acidic ionic liquid. Bioresource Technology 100: 2308–2310.
He, L., Qin, S., Chang, T. et al. (2013). Biodiesel synthesis from the esterification of free fatty acids and alco-
hol catalyzed by long-chain Brønsted acid ionic liquid. Catalysis Science & Technology 3: 1102–1107.
Iso, M., Chen, B., Eguchi, M. et al. (2001). Production of biodiesel fuel from triglycerides and alcohol using
immobilized lipase. Journal of Molecular Catalysis B: Enzymatic 16: 53–58.
Issariyakul, T., Kulkarni, M.G., Dalai, A.K., and Bakhshi, N.N. (2007). Production of biodiesel from waste
fryer grease using mixed MeOH/ethanol system. Fuel Processing Technology 88: 429–436.
Jaeger, K.E. and Eggert, T. (2002). Lipases for biotechnology. Current Opinion in Biotechnology 13:
390–397.
̇
Janus, E., Goc-Maciejewska, I., Łozyński, M., and Pernak, J. (2006). Diels–Alder reaction in protic ionic
liquids. Tetrahedron Letters 47: 4079–4083.
Jothiramalingam, R. and Wang, K.M. (2009). Review of recent developments in solid acid, base, and
enzyme catalysts (heterogeneous) for biodiesel production via transesterification. Industrial & Engi-
neering Chemistry Research 48: 6162–6172.
Kaieda, M., Samukawa, T., Matsumoto, T. et al. (1999). Biodiesel fuel production from plant oil catalyzed
by Rhizopus oryzae lipase in a water-containing system without an organic solvent. Journal of Bioscience
and Bioengineering 88: 627–631.
Karimi, B. and Vafaeezadeh, M. (2012). SBA-15-functionalized sulfonic acid confined acidic ionic liquid:
a powerful and water-tolerant catalyst for solvent-free esterifications. Chemical Communications 48:
3327–3329.
Kawahara, Y. and Ono, T. (1977). Process for producing lower alcohol esters of fatty acids. US 4164506
A.
Kiss, A.A., Dimian, A.C., and Rothenberg, G. (2006). Solid acid catalysts for biodiesel production-towards
sustainable energy. Advanced Synthesis & Catalysis 348: 75–81.
Knothe, G., Dunn, R.O., and Bagby, M.O. (1997). Biodiesel: the use of vegetable oils and their derivatives
as alternative diesel fuels. In: Fuels and Chemicals from Biomass, ACS Symposium Series, vol. 666 (eds.
B.C. Saha and J. Woodward), 172–208. American Chemical Society.
Kulkarni, M.G. and Dalai, A.K. (2006). Waste cooking oil an economical source for biodiesel: a review.
Industrial & Engineering Chemistry Research 45: 2901–2913.
Kumari, V., Shah, S., and Gupta, M.N. (2007). Preparation of biodiesel by lipase catalyzed transesterifica-
tion of high free fatty acid containing oil from Madhuca indica. Energy Fuels 21: 368–372.
Lai, C.C., Zullaikah, S., Vali, S.R., and Ju, Y.H. (2005). Lipase-catalyzed production of biodiesel from rice
bran oil. Journal of Chemical Technology & Biotechnology 80: 331–337.
Lee, D.H., Kim, J.M., Shin, H.Y. et al. (2006). Biodiesel production using a mixture of immobilized Rhi-
zopus oryzae and Candida rugosa lipases. Biotechnology and Bioprocess Engineering 11: 522–525.
Lee, K.-T., Foglia, T., and Chang, K.-S. (2002). Production of alkyl ester as biodiesel from fractionated lard
and restaurant grease. Journal of the American Oil Chemists’ Society 79: 191–195.
Li, J., Peng, X., Luo, M. et al. (2014). Biodiesel production from Camptotheca acuminata seed oil catalyzed
by novel Brönsted–Lewis acidic ionic liquid. Applied Energy 115: 438–444.
Li, K.X., Chen, L., Yan, Z.C., and Wang, H.L. (2010). Application of pyridinium ionic liquid as a recyclable
catalyst for acid-catalyzed transesterification of Jatropha oil. Catalysis Letters 139: 151–156.
Li, K.X., Chen, L., Yan, Z.C., and Wang, H.L. (2013). Synthesis of biodiesel from Jatropha oil using
pyridinium ionic liquid as a catalyst. Energy Sources, Part A: Recovery, Utilization, and Environmental
Effects 35: 1150–1160.
Li, L., Du, W., Liu, D. et al. (2006). Lipase-catalyzed transesterification of rapeseed oils for biodiesel
production with a novel organic solvent as the reaction medium. Journal of Molecular Catalysis B:
Enzymatic 43: 58–62.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 145

Li, W., Jiang, Z., Ma, F. et al. (2010). Design of mesoporous SO4 2− /ZrO2 –SiO2 (Et) hybrid material as
an efficient and reusable heterogeneous acid catalyst for biodiesel production. Green Chemistry 12:
2135–2138.
Li, W., Ma, F., Su, F. et al. (2012). Pore morphology control of mesostructured SO4 2− /ZrO2 -based hybrid
catalysts functionalized by alkyl-bridged organosilica moieties for biodiesel production from non-edible
oil. ChemCatChem 4: 1798–1807.
Liang, X. (2013). Novel efficient procedure for biodiesel synthesis from waste oils using solid acidic ionic
liquid polymer as the catalyst. Industrial & Engineering Chemistry Research 52: 6894–6900.
Liang, X., Gong, G., Wu, H., and Yang, J. (2009). Highly efficient procedure for the synthesis of biodiesel
from soybean oil using chloroaluminate ionic liquid as catalyst. Fuel 88: 613–616.
Liu, C.Z., Wang, F., Stiles, A.R., and Guo, C. (2012). Ionic liquids for biofuel production: opportunities
and challenges. Applied Energy 92: 406–414.
Liu, H., Liu, Y., and Li, J. (2010). Ionic liquids in surface electrochemistry. Physical Chemistry Chemical
Physics 12: 1685–1697.
Liu, S., Wang, Z., Li, K. et al. (2013). Brønsted-Lewis acidic ionic liquid for the “one-pot” synthesis of
biodiesel from waste oil. Journal of Renewable and Sustainable Energy 5: 023111–023116.
Lu, J., Nie, K., Xie, F. et al. (2007). Enzymatic synthesis of fatty acid methyl esters from lard with immo-
bilized Candida sp. 99–125. Process Biochemistry 42: 1367–1370.
Lu, J., Yan, F., and Texter, J. (2009). Advanced applications of ionic liquids in polymer science. Progress
in Polymer Science 34: 431–448.
Lu, J., Deng, L., Zhao, R. et al. (2010). Pretreatment of immobilized Candida sp. 99–125 lipase to improve
its MeOH tolerance for biodiesel production. Journal of Molecular Catalysis B: Enzymatic 62: 15–18.
Mata, T.M., Andrade, S., Correia, D. et al. (2017). Acidity reduction of mammalian fat by enzymatic ester-
ification. Energy Procedia 136: 290–295.
Meher, L.C., Sagar, D.V., and Naik, S.N. (2006). Technical aspects of biodiesel production by transesteri-
fication: a review. Renewable & Sustainable Energy Reviews 10: 248–268.
Mehnert, C.P. (2005). Supported ionic liquid catalysis. Chemistry–A European Journal 11: 50–56.
Montefrio, M.J., Xinwen, T., and Obbard, J.P. (2010). Recovery and pretreatment of fats, oil and grease
from grease interceptors for biodiesel production. Applied Energy 87: 3155–3161.
Muhammad, N., Elsheikh, Y.A., Mutalib, M.I.A. et al. (2015). An overview of the role of ionic liquids in
biodiesel reactions. Journal of Industrial and Engineering Chemistry 21: 1–10.
Mulalee, S., Srisuwan, P., and Phisalaphong, M. (2016). Influences of operating conditions on biocatalytic
activity and reusability of Novozym 435 for esterification of free fatty acids with short-chain alcohols: a
case study of palm fatty acid distillate. Biochemical Engineering Journal 105: 52–61.
Nakashima, K., Kubota, F., Maruyama, T., and Goto, M. (2005). Feasibility of ionic liquids as alterna-
tive separation media for industrial solvent extraction processes. Industrial & Engineering Chemistry
Research 44: 4368–4372.
Nelson, L.A., Foglia, T.A., and Marmer, W.M. (1996). Lipase-catalyzed production of biodiesel. Journal
of the American Oil Chemists’ Society 73: 1191–1195.
Nie, K., Xie, F., Wang, F., and Tan, T. (2006). Lipase catalyzed MeOHysis to produce biodiesel: optimiza-
tion of the biodiesel production. Journal of Molecular Catalysis B: Enzymatic 43: 142–147.
Nordblad, M., Pedersen, A.K., Rancke-Madsen, A., and Woodley, J.M. (2016). Enzymatic pretreatment of
low-grade oils for biodiesel production. Biotechnology and Bioengineering 113: 754–760.
de Oliveira, P.M., Farias, L.M., Morón-Villarreyes, J.A., and Montes D’Oca, M.G. (2016). Eco-friendly
pretreatment of oil with high free fatty acid content using a sulfamic acid/ethanol system. Journal of the
American Oil Chemists’ Society 93: 1393–1397.
Olkiewicz, M., Plechkova, N.V., Earle, M.J. et al. (2016). Biodiesel production from sewage sludge lipids
catalysed by Brønsted acidic ionic liquids. Applied Catalysis B: Environmental 181: 738–746.
Özbay, N., Oktar, N., and Tapan, N.A. (2008). Esterification of free fatty acids in waste cooking oils (WCO):
role of ion-exchange resins. Fuel 87: 1789–1798.
146 Process Systems Engineering for Biofuels Development

Park, J.Y., Kim, D.K., and Lee, J.S. (2010). Esterification of free fatty acids using water-tolerable Amberlyst
as a heterogeneous catalyst. Bioresource Technology 101: S62–S65.
Pârvulescu, V.I. and Hardacre, C. (2007). Catalysis in ionic liquids. Chemical Reviews 107: 2615–2665.
Pastore, C. and di Bitonto, L. (2017). Procedimento di esterificazione diretta in fase omogenea di acidi
organici e procedimento per la preparazione di biodiesel. 102017000038638.
Pastore, C., Lopez, A., Lotito, V., and Mascolo, G. (2013). Biodiesel from dewatered wastewater sludge: a
two-step process for a more advantageous production. Chemosphere 92: 667–673.
Pastore, C., Lopez, A., and Mascolo, G. (2014). Efficient conversion of brown grease produced by municipal
wastewater treatment plant into biofuel using aluminium chloride hexahydrate under very mild condi-
tions. Bioresource Technology 155: 91–97.
Pastore, C., Barca, E., Del Moro, G. et al. (2015a). Recoverable and reusable aluminium solvated species
used as a homogeneous catalyst for biodiesel production from brown grease. Applied Catalysis A: Gen-
eral 501: 48–55.
Pastore, C., Pagano, M., Lopez, A. et al. (2015b). Fat, oil and grease waste from municipal wastewater:
characterization, activation and sustainable conversion into biofuel. Water Science & Technology 71:
1151–1157.
Peters, T.A., Benes, N.E., Holmen, A., and Keurentjes, J.T. (2006). Comparison of commercial solid acid
catalysts for the esterification of acetic acid with butanol. Applied Catalysis A: General 297: 182–188.
Predojević, Z.J. (2008). The production of biodiesel from waste frying oils: a comparison of different purifi-
cation steps. Fuel 87: 3522–3528.
Royon, D., Daz, M., Ellenrieder, G., and Locatelli, S. (2007). Enzymatic production of biodiesel from
cotton seed oil using t-butanol as a solvent. Bioresource Technology 98: 648–653.
Sani, Y.M., Wan Daud, W.M.A., and Aziz, A.R.A. (2014). Activity of solid acid catalysts for biodiesel
production: a critical review. Applied Catalysis A: General 470: 140–161.
Shimada, Y., Watanabe, Y., Samukawa, T. et al. (1999). Conversion of vegetable oil to biodiesel using
immobilized Candida antarctica lipase. Journal of the American Oil Chemists’ Society 76: 789–793.
Shimada, Y., Watanabe, Y., Sugihara, A., and Tominaga, Y. (2002). Enzymatic alcoholysis for biodiesel fuel
production and application of the reaction to oil processing. Journal of Molecular Catalysis B: Enzymatic
17: 133–142.
Soumanou, M.M. and Bornscheuer, U.T. (2003). Improvement in lipase-catalyzed synthesis of fatty acid
methyl esters from sunflower oil. Enzyme and Microbial Technology 33: 97–103.
Su, F. and Guo, Y. (2014). Advancements in solid acid catalysts for biodiesel production. Green Chemistry
16: 2934–2957.
Supple, B., Holward-Hildige, R., Gonzalez-Gomez, E., and Leahy, J.J. (2002). The effect of steam treating
waste cooking oil on the yield of methylester. Journal of the American Oil Chemists’ Society 79: 175–178.
Syrén, P. and Hult, K. (2010). Substrate conformations set the rate of enzymatic acrylation by lipases.
Chembiochem 11: 802–810.
Tan, G. and Li, Z. (2012). Highly active, stable, and recyclable magnetic nano-size solid acid catalysts:
efficient esterification of free fatty acid in grease to produce biodiesel. Green Chemistry 14: 3077–3086.
Tan, T., Lu, J., Nie, K. et al. (2010). Biodiesel production with immobilized lipase: a review. Bioresource
Technology 97: 671–678.
Thanh, L.T., Okitsu, K., Sadanaga, Y. et al. (2013). A new co-solvent method for the green production of
biodiesel fuel – Optimization and practical application. Fuel 103: 742–748.
Tu, Q., Lu, M., and Knothe, G. (2017). Glycerolysis with crude glycerin as an alternative pretreatment for
biodiesel production from grease trap waste: Parametric study and energy analysis. Journal of Cleaner
Production 162: 504–511.
Tubino, M., Junior, J.G.R., and Bauerfeldt, G.F. (2016). Biodiesel synthesis: A study of the triglyceride
MeOHysis reaction with alkaline catalysts. Catalysis Communications 75: 6–12.
Tunckol, M., Durand, J., and Serp, P. (2012). Carbon nanomaterial–ionic liquid hybrids. Carbon 50:
4303–4334.
 Up-grading of Waste Oil: A Key Step in the Future of Biofuel Production 147

Ullah, Z., Bustam, M.A., and Man, Z. (2015). Biodiesel production from waste cooking oil by acidic ionic
liquid as a catalyst. Renewable Energy 77: 521–526.
Verma, P. and Sharma, M.P. (2016). Review of process parameters for biodiesel production from different
feedstocks. Renewable & Sustainable Energy Reviews 62: 1063–1071.
Wang, Y., Ou, S., Liu, P., and Zhang, Z. (2007). Preparation of biodiesel from waste cooking oil via two-step
catalyzed process. Energy Conversion and Management 48: 184–188.
Watanabe, Y., Shimada, Y., Sugihara, A. et al. (2000). Continuous production of biodiesel fuel from veg-
etable oil using immobilized Candida antarctica lipase. Journal of the American Oil Chemists’ Society
77: 355–360.
Watanabe, Y., Shimada, Y., Sugihara, A., and Tominaga, Y. (2001). Enzymatic conversion of waste edible
oil to biodiesel fuel in a fixed-bed bioreactor. Journal of the American Oil Chemists’ Society 78: 703–707.
Watanabe, Y., Shimada, Y., Sugihara, A., and Tominaga, Y. (2002). Conversion of degummed soybean oil to
biodiesel fuel with immobilized Candida antarctica lipase. Journal of Molecular Catalysis B: Enzymatic
17: 151–155.
Wei, D., Xu, Y.Y., Jing, Z., and Liu, D.H. (2004). Novozym 435-catalysed transesterification of crude soya
bean oils for biodiesel production in a solvent free medium. Biotechnology and Applied Biochemistry
40: 187–190.
Wu, W.H., Foglia, T.A., Marmer, W.N., and Phillips, J.G. (1999). Optimizing production of ethyl esters
of grease using 95% ethanol by response surface methodology. Journal of the American Oil Chemists’
Society 76: 517–521.
Yaakob, Z., Mohammad, M., Alherbawi, M. et al. (2013). Overview of the production of biodiesel from
waste cooking oil. Renewable & Sustainable Energy Reviews 18: 184–193.
Yan, J., Li, A., Xu, Y. et al. (2012). Efficient production of biodiesel from waste grease: one-pot esterification
and transesterification with tandem lipases. Bioresource Technology 123: 332–337.
You, Y.D., Shie, J.L., Cheng, C.Y. et al. (2008). Economic cost analysis of biodiesel production: case in
soybean oil. Energy Fuel 22: 182–189.
Yusuf, N.N.A.N., Kamarudin, S.K., and Yaakub, Z. (2011). Overview on the current trends in biodiesel
production. Energy Conversion and Management 52: 2741–2751.
Zhang, F., Fang, Z., and Wang, Y.T. (2015). Biodiesel production direct from high acid value oil with a
novel magnetic carbonaceous acid. Applied Energy 155: 637–647.
Zhang, L., Xian, M., He, Y. et al. (2009). A Brønsted acidic ionic liquid as an efficient and environmentally
benign catalyst for biodiesel synthesis from free fatty acids and alcohols. Bioresource Technology 100:
4368–4373.
Zhang, L., Cui, Y., Zhang, C. et al. (2012). Biodiesel production by esterification of oleic acid over
brønsted acidic ionic liquid supported onto Fe-incorporated SBA-15. Industrial & Engineering
Chemistry Research 51: 16590–16596.
Zhang, Y., Dube, M.A., McLean, D.D., and Kates, M. (2003a). Biodiesel production from waste cooking
oil: 2. Economic assessment and sensitivity analysis. Bioresource Technology 90: 229–240.
Zhang, Y., Dube, M.A., Mclean, D.D., and Kates, M. (2003b). Biodiesel production from waste cooking
oil. 1. Process design and technological assessment. Bioresource Technology 89: 1–16.
Zhao, D., Wu, M., Kou, Y., and Min, E. (2002). Ionic liquids: applications in catalysis. Catalysis Today 74:
157–189.
 6
Production of Biojet Fuel from
Waste Raw Materials: A Review
Ana Laura Moreno-Gómez1 , Claudia Gutiérrez-Antonio1 , Fernando
Israel Gómez-Castro2 , and Salvador Hernández2
1 Facultad de Química, Universidad Autónoma de Querétaro, Querétaro, 76010 Querétaro, México
2 Departamento de Ingeniería Química, Universidad de Guanajuato, Guanajuato, 36050 Guanajuato,
México

6.1 Introduction
In the transport sector, aviation has the highest growth. The forecasts of the International
Air Transport Association, IATA, indicate that in the next two decades aviation will have
a compound annual growth rate of 3.5%, which implies that the number of passengers
could double to 8200 million in 2037 (IATA 2018a,b). As consequence, there will be an
increase of both fuel usage and carbon dioxide emissions. Therefore, in order to guarantee
the sustainable development of aviation sector, the Four-Pillar strategy was proposed (IATA
2009); this strategy considers technological improvements in engines and aircraft struc-
tures, operational improvements through online optimization of flight paths, market-based
measures, and development of alternative fuels (Gutiérrez-Antonio et al. 2016a). From these
alternatives, the IATA and the International Civil Aviation Organization, ICAO, agree that
renewable aviation fuel is the one that contributes most to the sustainable development of
the aviation sector.
Fossil aviation fuel consists of hydrocarbons in the boiling point range from C8 to C16,
including paraffinic, naphthenic and aromatic compounds. On the other hand, renewable

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
150 Process Systems Engineering for Biofuels Development

aviation fuel contains paraffinic and naphthenic compounds, but it may or may not contain
aromatic compounds, depending on the processing pathway. The absence of aromatic com-
pounds could cause leaks in the tank fuels, since aromatic compounds interact with the
elastomers in the tank seals. Due to this, renewable aviation fuel must be used blended with
fossil jet fuel, with a maximum composition of renewable fuel of 50%, according to the
standard ASTM D7566 (ASTM 2018).
Renewable aviation fuel is also known as biokerosene, biojet fuel or synthetic paraffinic
kerosene; and it can be generated from triglyceride, lignocellulosic, sugar and starchy
feedstock. The feedstock is transformed in biojet fuel through different pathways, which
include hydroprocessing of triglyceride feedstock, thermochemical processing of biomass,
and alcohol-to-jet (Gutiérrez-Antonio et al. 2016a). At present, five pathways have
been certified under ASTM D7566 standard: gasification followed by Fischer–Tropsch
synthesis, hydroprocessing of esters and fatty acids, hydroprocessing of fermented sugars,
Fischer–Tropsch synthesis with aromatics, and alcohol-to-jet (IATA 2018a,b). However,
another seven pathways are in the evaluation stage to reach certification (CAAFI 2019).
The certified conversion routes allow the production of renewable aviation fuel that meets
or even exceeds the quality standards. Despite these pathways being different, three possible
scenarios are observed: raw material of high cost with low processing costs, raw material of
low cost with high processing costs, and both raw material and processing with intermediate
costs. The reduction in the processing costs can be reached through several strategies, such
as process intensification or energy integration; several works have reported the application
of these tools to the production processes of biojet fuel (Gutiérrez-Antonio et al. 2016a,b,
2018a,b). On the other hand, the raw material costs depend on several factors, such as the
variety of the crops, the automation of the cultivation task, along with the transportation
distances; due to this an optimal supply chain must be established, and some works are
reported on this topic (Reimer and Zheng 2017; Domínguez-García et al. 2017a,b; Leila
et al. 2018). Another alternative is the use of waste raw materials, which are abundant and
have low cost; some of them even constitute a pollution problem, due to the high volumes
in which they are generated. In this way, a problem can be transformed into an energetic
solution. Therefore, the focus of this work is the review of the use of waste raw materi-
als to produce biojet fuel. As shown in Figure 6.1, there are three main waste feedstocks:
(i) triglyceride-containing materials, (ii) lignocellulosic, and (iii) sugar and starchy. Any of
this feedstock can be transformed into biojet fuel through different processing pathways.
Moreover, the main challenges found in these production processes are analyzed, and future
research trends are discussed.
The chapter is organized as follows: the conversion of waste triglycerides through
hydroprocessing is discussed in Section 6.2. The transformation of lignocellulosic
residues, through different technologies, to produce renewable aviation fuel is presented
in Section 6.3. The production of biojet fuel from sugar and starchy residues is shown in
Section 6.4. Finally, the main challenges and future trends in the use of residues to produce
aviation biofuel are discussed in Section 6.5.

6.2 Waste Triglyceride Feedstock

Waste triglyceride feedstock includes waste cooking oil, animal fats, oil extracted from
agroindustrial residues, and bio-oil obtained from pyrolysis. Waste cooking oil and animal
 Production of Biojet Fuel from Waste Raw Materials: A Review 151

Catalytic
Lipids
Hydrothermolysis

Bio -
Processing
Waste sugar Alcohol
and starchy
feedstock H
Saccharification Sugar Thermo - y
Processsing d
Waste r
lignocellulosic o
feedstock Bio-oil p
Pyrolysis Biojet
r
fuel
Fischer- o
Tropsch c
Gasification Syngas e
s
Thermo- s
Processing i
Waste n
triglyceride g
feedstock

Figure 6.1 Pathways to process waste feedstock to produce renewable aviation fuel.

fats are usually a by-product of food industries and/or residues from residential and
commercial sectors; they represent a contamination problem due to the high volume in
which they are generated, and usually there is a non-appropriated disposal of such residues.
For instance, in the United States the availability of 10 million tons of waste cooking oil
has been reported (Gui et al. 2008). In Mexico, the information is scattered; there are
some reports that indicate nearly 8300 and 60 000 l/yr of waste cooking oil in the Puerto
Interior Industrial Park, in Guanajuato (IEEG 2016), and Emiliano Zapata, in Tabasco
(Gasca González 2017), respectively. In Europe, the potential of waste cooking oil was
estimated as 3.55 million tons in 2015, with 1.748 million tons generated in the domestic
sector (EUBIA 2015).
Other interesting waste material is the spent coffee ground, which represents a good
source of triglycerides to produce biojet fuel. According to McNutt and He (2019), in 2016
the worldwide production of coffee was around 9.3 billion kilograms, and it is possible to
recover between 10% and 20% of oils from spent coffee grounds by conventional extraction
methods.
Therefore, there is a great amount of these polluting residues to produce biojet fuel
through hydroprocessing technology. The first production process to obtain biojet fuel
through hydroprocessing was proposed by UOP Honeywell (Vera-Morales and Schäfer
2009), and it consist of two consecutive reactors and a distillation train (Figure 6.2).
In the first reactor, the triglyceride feedstock is converted, at high pressure and tempera-
ture, to long lineal chain hydrocarbons; these hydrocarbons are cracked and isomerized in
the second reactor, which also operates at high pressure and temperature. In both reac-
tors, hydrogen is used as reactant along with solid catalysts. The reactor outlet stream
contains renewable hydrocarbons that include light gases, naphtha, biojet fuel and green
diesel; these products are purified through a distillation sequence. The UOP Honeywell’s
technology was proposed as an extension of a green diesel production process, in which
a selective cracking stage was added (Regalbuto 2010; Verma et al. 2015); therefore, the
152 Process Systems Engineering for Biofuels Development

CO2
H2O light gases
cracking/
pretreatment deoxygenation isomerization
naphtha
Waste Physical
Chemical
triglyceride Thermal
feedstock biojet fuel
Biological

green diesel

Water H2 H2 Distillation
Reactants

Figure 6.2 Hydrotreating process to produce renewable aviation fuel from waste triglyceride feedstock.

maximum selectivity to biojet fuel is 36%, with total conversion to hydrocarbons of 70%
(Regalbuto 2010; Verma et al. 2015). Another hydrotreating process is BioSynfining™, in
which renewable aviation fuel is produced from fatty acids and triglycerides. The fatty acid
chains are converted to n-paraffins through deoxygenation, and the long-chain paraffins are
hydrocracked to short-chain paraffins. The hydrocracked products have boiling points in
the ranges of kerosene and naphtha. The BioSynfining process was successfully used to
produce about 600 gal of biojet fuel (Liu et al. 2013), and it is very similar to that proposed
by UOP Honeywell.
The hydrotreating process can employ diverse catalysts, raw materials and different pro-
cessing conditions (Vásquez et al. 2017). In the hydrotreating process different catalysts
have been utilized, either nickel-based or bifunctionals, as well as different types of reac-
tors such as the fixed bed reactor. In most of the studies, a large amount of hydrogen is
required to perform the conversion to biojet fuel. Moreover, almost all the studies report
the use of vegetable oils, both edible and inedible, such as soybean, Jatropha curcas, castor
and microalgae oil (Gutiérrez-Antonio et al. 2017). However, the use of waste oils repre-
sents a good alternative to reduce the price of biojet fuel, and, at the same time, to solve
a contamination problem due to the accumulation of these materials. Next, we present a
review of the works where waste triglyceride feedstock is treated to generate biojet fuel.
The paper published by Tian et al. (2008) reported the production of different biofuels
through the cracking of animal fats. They obtained a high yield for liquefied petroleum gas
(45%) and olefins (47%), but a low yield for naphtha (approximately 4.17%). In the reactor,
CoRh was used as catalyst, and the hydrogen requirement was not reported. Additionally,
aromatic compounds in the range C7–C10 were obtained.
Other works (Bezergianni et al. 2009; Bezergianni and Kalogianni 2009) point out that
cracking is a prominent technology for biojet fuel production. These works compare the
yield of products and the quality of hydrocracking at three different temperatures, consid-
ering non-used oil and waste cooking oil as raw materials. The evaluation indicates that, for
both raw materials, a high diesel production is observed, while the yield for kerosene and
gasoline is low. They also noted that as temperature increases, the selectivity toward diesel
also increased for both raw materials; however, at 390 ∘ C the selectivity toward kerosene
was favored (22.24%).
 Production of Biojet Fuel from Waste Raw Materials: A Review 153

Later, Bezergianni et al. (2012) reported the conversion of waste cooking oil toward
biojet fuel using three catalysts: a hydrotreating catalyst, a mild-hydrocracking catalyst and
a severe hydrocracking catalyst. They used conditions of high temperature (330–390 ∘ C)
and pressure (8.27–13.79 MPa). As products, hydrocarbons in the range of gasoline and
diesel were obtained. The highest yield was 80% for diesel at 370 ∘ C and 8.27 MPa, with a
hydrotreating catalyst.
Shi et al. (2014) showed a new path for converting bio-oils, prepared from cornstalks liq-
uefaction, to diesel and hydrocarbons in the boiling point range of biojet fuel. The reaction
was carried out using Ni/ZrO2 as catalyst in the presence of supercritical cyclohexane at
573 K and 5 MPa of hydrogen. They obtained a hydrocarbon yield of 81.6%, with 90% of
hydrocarbons in the range of diesel and biojet fuel, and 7% in the range of gasoline.
Mosisa et al. (2018) studied the cracking of waste cooking oil for the production of liquid
fuels in a semi-batch reactor with a nitrogen atmosphere; they used zirconia oxide (ZrO2 )
as catalyst. It is important to mention that the raw material was cracked without the need
of a pretreatment using the mentioned catalyst. A yield of 83% for organic liquid products
was obtained. An important aspect of this process is that it does not require high energy
consumption, and the catalyst is easily regenerated and recycled, being environmentally
friendly.
Next, we present the review of the works where waste triglyceride feedstock is deoxy-
genated and cracked to generate biojet fuel.
Charusiri et al. (2006) studied the conversion of waste vegetable oils into liquid fuels
over sulfated zirconia, HZSM-5 and HZSM-5 hybrid catalysts; their experiments were per-
formed in a batch micro-reactor at temperature range of 380–430 ∘ C, the hydrogen initial
pressure was in the range of 10–20 bar, and the reaction time was 45–90 minutes. Most of
the obtained products were liquids (gasoline, kerosene), gases and a small amount of solids.
Waste cooking oil, which contains a high acidity value (28.7 mg KOH/g oil), is converted
through the hydrocracking process over a ruthenium catalyst to obtain diesel as product (Liu
et al. 2012); the advantage of obtaining diesel is that it can be cracked and then converted
to biojet fuel. The temperature, hydrogen pressure, retention time and H2 /oil ratio were
350 ∘ C, 2 MPa, 15.2 h, and 400 ml/ml, respectively. Free fatty acids and triglycerides present
in oil were deoxygenated at the same time to form hydrocarbons; the liquid hydrocarbons
had a yield of 98.9%, with octadecane, heptadecane, hexadecane, and pentadecane as the
predominant ones.
I.H. Choi et al. (2015) performed a study to obtain biojet fuel through deoxygenation,
isomerization, and cracking in a single stage. In order to carry out such reaction 300–420 m3
of H2 per m3 of waste oil were required. With this proposal, they achieved simplification of
the process, and simultaneously they decreased the consumption of hydrogen and energy.
In such study, they prepared a Pd catalyst supported in beta-zeolite. As raw materials, waste
soybean oil and palm fatty acid distillate were used.
Moreover, the work proposed by Hanafi et al. (2016) reported the production of hydro-
carbons in the boiling point range of jet fuel, as well as naphtha and light gases; this work
is important because fatty acids are deoxygenated and cracked in a single reactor, which
allows a reduction in the investment costs. Furthermore, waste chicken fats, obtained from
a company, are employed as raw material. In their work, the physical characteristics of the
raw material and activation energy are provided; the yield toward kerosene/diesel fraction
is 53%. The operating conditions for temperature, pressure, retention time and H2 feed are
400 ∘ C, 6 MPa, 4 h, and 450 v/v H2 /oil, respectively.
154 Process Systems Engineering for Biofuels Development

The study reported by Zhang et al. (2017) showed the production of fuel for airplanes
in a one-stage process; they used animal fat as raw material and Pt/SAPO-11 as catalyst in
a micro-reactor that operated at high pressure. According to their experimental results, the
optimum operation conditions to produce a bigger amount of hydrocarbons in the C8–C16
range are 4 MPa, 400 ∘ C, and 1000 ml H2 /ml oil; under such conditions there is a 96.6%
conversion and the selectivity toward C8–C16 hydrocarbons is 50.25% with selectivity to
their corresponding isomers around 35.68%. It should be noticed that the retention time
of the micro-reactor is 1.2 hours. In that study, they mentioned that isomerization gradually
increases with the rise of temperature, but when the temperature reaches 380 ∘ C the isomer-
ization of the hydrocarbons decreases gradually. The increase in the H2 /oil ratio is beneficial
for the hydrocarbon conversion and for the isomerization of alkanes; however, when the
H2 /oil ratio is too large, the hydrocarbon selectivity decreases, and thus the isomerization
percentage reduces.
In the following paragraphs, we present a review of the works where waste triglyceride
feedstock is deoxygenated to generate biojet fuel.
Zhang et al. (2014) carried out a kinetic study of the hydrodeoxygenation of waste
cooking oil with a CoMoS catalyst. Their results show that the hydrodecarbonyla-
tion/decarboxylation (HDC) are the predominant reaction paths for oxygen elimination;
the catalyst activity decreases as the amount of sulfur in the catalyst also reduces.
Moreover, the hydrodecarbonylation of fatty acids controls the HDC path; while through
the hydrodeoxygenation path fatty acids are transformed to aldehydes/alcohols, and
subsequently to C18 hydrocarbons as final product. The difference between the C18/C17
ratio with supported and unsupported catalyst shows that the acid Lewis sites are related to
the selectivity for the hydrodeoxygenation path, thus giving as a result a product of high
quality. The experiment was performed in a batch reactor using 0.6 g of catalyst and 120 g
of waste oil, with catalyst to oil ratio of 1:200 w:w.
Three types of zeolites (Meso-Y-SAPO-34, y HY) mixed with nickel were used to con-
vert waste cooking oil to aviation fuel (Li et al. 2015). The mesoporous-Y zeolite exhibited
53% selectivity and 13.4% of selectivity to aromatic compounds in the liquid products; this
zeolite showed a 40.5% yield at 400 ∘ C, but the yield for aromatic compounds decreased
2.1%. The experimental results show that the deoxygenated raw materials tend toward hep-
tadecane and pentadecane, through a decarbonylation path during the three first hours. In
this experiment, the raw material was previously dried in order to eliminate the water con-
tent (Li et al. 2015).
The direct conversion of waste oil to biojet fuel was researched by Zhang et al. (2018);
in this process a zeolite support with a core-shell structure USY-AL-SBA-15 and NiMo
as catalyst were used. The use of this support and catalyst contributed significantly to
improve the selectivity toward biojet fuel from 9.3% over NiMo/USY to 35.7% over
NiMo/USY-AL-SBA-15, with a high isomerization ratio (iso-n/n-paraffin = 2.7) and
18.7% of aromatic compounds. The authors mention that through a single path, using
either NiMo or NiW catalysts supported on ZSM-5, it is possible to obtain high yields to
biojet fuel, 40–45%, with an excellent isomer/alkane ratio in the range 2–6.
An interesting work proposes the one-step hydroprocessing of bio-oil generated in
the hydrothermal liquefaction of microalgae cultivated in wastewater (Ranganathan and
Savithri 2019). In this process, renewable hydrogen is also obtained from the biogas
generated from the anaerobic digestion of the sludge from the wastewater treatment.
 Production of Biojet Fuel from Waste Raw Materials: A Review 155

However, renewable hydrogen can also be obtained from the electrolysis of water (IEA
2017). Fu et al. (2015) proposed the direct conversion of microalgae lipids in water to
renewable aviation fuel. The catalyst used was Pt/C, and the optimal conditions were
360 ∘ C with a reaction time of 45 minutes; in these conditions, a total conversion was
observed with a selectivity of 90% to heptadecane. There are several works in the literature
where the hydroisomerization of model compounds is reported, mainly n-hexadecane and
n-dodecane. These works are important since in most cases the hydroprocessing consists
of two stages; in the first one the hydrodeoxygenation is performed, while the hydroiso-
merization and hydrocracking are carried out after that. These works are described next.
The first work in this category was presented by Zhang et al. (2000). They reported
the activity, selectivity and long-term stability of platinum-promoted tungstate-modified
zirconia (Pt/WO3 /ZrO2 ), under mild conditions; in the hydroisomerization reaction,
n-hexadecane was used as a model compound. A trickle bed reactor was used for the
experiments. The best results were a conversion of n-hexadecane of 79.1 wt%, while for
iso-hexadecane a selectivity of 89.9 wt% and a yield of 71.1 wt% i-C16 were reported.
Later, Zhang et al. (2001) studied the effect of activity and selectivity of tungstated zirconia
(8% w) on the isomerization of n-hexadecane. The study was carried out in a trickle bed
continuous reactor. The results showed that temperatures between 300 ∘ C and 400 ∘ C, for
three hours, were slightly beneficial for achieving high yields of iso-hexadecane. Gomes
et al. (2017) studied the performance of bifunctional Pt/alumina-beta zeolite catalysts for
the hydroisomerization of n-C16. A biphasic micro-reactor of plug flow was used to study
the effect of nC16 isomerization on the pour point of the products. They found that as
the pour point decreases at constant rate, the formation of cracked products was small;
products are essentially composed of mono- and disubstituted C16 isomers, while 50%
of n-C16 was converted. Moreover, the hydroisomerization of n-hexadecane was studied
in order to evaluate the activity of Pt/AlSBA-15 catalysts (Jaroszewska et al. 2017);
they determined that the catalyst AlSBA-15 using aluminum isopropoxide showed better
isomerization selectivity than AlSBA-15 using aluminum sulfate.
De Lucas et al. (2005) studied the performance of palladium and platinum beta
zeolite-based catalysts with or without binder in the hydroisomerization of n-octane, along
with the influence of Si/Al ratio. As result, they found that the catalytic activity of beta
zeolite catalysts decreased when the Si/Al ratio increased, in samples with or without
binder. In addition, the isomer selectivity rose from 54.3 to 67.8% in samples without
binder when the Si/Al ratio increased.
Wang et al. (2008) investigated the hydroisomerization of n-dodecane over Pt supported
in ZSM-22 unmodified and ZSM-22 modified with two treatments. The ZSM-22 unmodi-
fied showed low activity in the hydroisomerization of n-dodecane. The catalyst treated with
NH4 + ion and (NH4 )2 SiF6 showed a high selectivity for iso-dodecane (88.0%) and good
conversion (87.5%) at 300 ∘ C. The experiments were carried out in a fixed bed flow reactor
with an internal diameter of 12 mm.
Other studies were reported for the hydroisomerization of n-butane (Adeeva et al. 1998),
n-hexane and n-octane (Amanza et al. 1999), n-hexane, n-octane, and n-hexatriacontane
(Calemma et al. 2000), and n-hexane, n-heptane, and n-octane (Dhar et al. 2017).
To summarize, Table 6.1 presents the reported articles where waste triglyceride feedstock
is used to produce biojet fuel.
Table 6.1 Summary of the use of waste triglyceride feedstock to obtain biojet fuel.
Hydrogen
Raw material Catalyst Temperature(∘ C) Pressure requirements Reactor type Yield (%) References
Animal CORH 500 (first 1 atm Not TSRFCC∗ LPG: 47; liquid total: Tian et al. (2008)
Fats stage); mentioned 77.6
LTB-2 520 (second
stage)
Non used oil DMDS∗∗ 350 13 789.5 1068 nm3 /m3 Fixed bed Waste oil: 20.04; Bezergianni et al.
TBA∗∗∗ 370 kPa H2 /oil fresh (2009),
Waste 390 oil: 22.24 Bezergianni
cooking oil and Kalogianni
(2009)
Waste HDT 330–390 8.27–13.79 MPa 3000 l at NTP Fixed bed Gas oil: Bezergianni et al.
cooking oil MID-HDC 15; (2012)
HDC diesel:
79
Bio-oils Ni/ZrO2 573 5 MPa Not Nantong Hydrocarbon: 81.6; Shi et al. (2014)
mentioned Huaxing diesel-biojet fuel: 90;
Petroleum gasoline: 7
Waste Zirconia oxide 400–500 Not mentioned Not mentioned Semi-batch Liquid organic product: Mosisa et al.
cooking oil (ZrO2 ) 83 (2018)
Waste Sulfated Zirconia 380–430 10–20 bar Not mentioned Batch Gas oil: 6.5; Charusiri et al.
vegetable oil (HZSM-5); micro-reactor gasoline: 26.57; (2006)
Hybrid catalyst kerosene: 10.65;
(HZSM-5) light gases: 23.62;
residues: 12.88
Waste Ru supported on 350 2 MPa 400 ml H2 /ml oil Fixed-bed Liquid hydrocarbons: Liu et al. (2012)
cooking oil (Al13-Mont)† 98.9
Non-edible Pd/ beta-zeolite 270 15 bar 300–420 m3 H2 /m3 Not mentioned Biojet fuel: 40 I.H. Choi et al.
oil oil (2015)
Waste NiW/SiO2 -Al2 O3 400 60 atm 450 v/v, H2 /oil Fixed-bed Total conversion: 94; Hanafi et al.
Chicken selectivity to (2016)
kerosene: 40
Animal Pt/SAPO-11 400 4 MPa 1000 ml H2 /ml oil Micro-reactor Conversion: 96.6; Zhang et al.
Fat selectivity C8-C16: (2017)
50.25; isomers: 35.68
Waste CoMoS 375 88.4 atm Not mentioned Batch Naphtha: 10; diesel: 81 Zhang et al.
cooking oil (2014)
Waste Meso-Y 400 3 MPa 350 ml/min Batch reactor Aromatics: 13.4; jet Li et al. (2015)
cooking SAPO-34 fuel: 40.5
oil HY
Waste USY-AL-SBA-15 380 30 atm 250 ml H2 /ml oil Fixed bed flow Jet fuel: 39.7; aromatic Zhang et al.
oil NiMo fraction: 18.7 (2018)
n-Hexadecane Pt/WO3 /ZrO2 218 160 psig H2 /n-C16 mole Trickle bed Selectivity i-C16: 89.9; Zhang et al.
ratio = 2 total conversion: 79.1 (2000)
n-hexadecane Tungstated 300–400 500 psig; Not mentioned Trickle bed Iso-hexadecane: 87 Zhang et al.
zirconia (8% w) pressure drop continuous (2001)
73 psig
n-Hexadecane Bifunctional 260–320 50–100 bar 500 NTP l/l Micro-reactor of Conversion of n-C16: 50 Gomes et al.
Pt/alumina-beta H2 /reactant plug flow (2017)
zeolite
n-Hexadecane AlSBA-15 320–360 5 MPa H2 :CH = 350 Pressure fixed Conversion to iso-C16: Jaroszewska
NTP m3 /m3 bed 61 et al. (2017)
microreactor
Table 6.1 (continued)
Hydrogen
Raw material Catalyst Temperature(∘ C) Pressure requirements Reactor type Yield (%) References
n-dodecane Pt/ZSM-22 480 6.0 MPa H2 /n-C12 = 600:1 Not mentioned Unmodified ZSM-22 Wang et al.
unmodified showed low activity. (2008)
Pt/ZSM-22 Modified ZSM-22: 88.0
modified
n-octane Platinum 548 2.0 MPa 24 H2 /n-C6 Not mentioned n-Hexane: 52; n-octane: Amanza et al.
n-hexane 27 (1999)
n-hexadecane 0.3% platinum/ 345–380 2–13.1 MPa Not mentioned Stirred Iso-hexadecane: 58–62; Calemma et al.
n-octacosane amorphous microautoclave iso-octacosane: (2000)
n-hexatriacontane silica–alumina 49–46;
(MSA/E) iso-hexatriacontane:
39–32
n-hexane Pt doped on 140 20 bar Not mentioned Batch At 180 ∘ C Dhar et al.
n-heptane gamma alumina 160 n-hexane: 85; (2017)
n-octane 180 n-heptane: 68;
n-octane: 40
∗ Two-stage riser fluid catalytic cracking.
∗∗ DMDS, dimethyl disulfide.
∗∗∗ TBA,tert-butylamine.
† Aluminum-polyoxocation-pillared montmorillonite.
 Production of Biojet Fuel from Waste Raw Materials: A Review 159

6.3 Waste Lignocellulosic Feedstock

Waste lignocellulosic feedstock includes wood waste, agricultural residues, textile residues,
solid urban waste, among others. Lignocellulosic materials such as agricultural wastes
are attractive feedstock for biojet fuel production, since they are abundant and renew-
able. In Mexico during 2011, 52% of the solid urban waste included waste food and waste
organic materials (SEDESOL 2012); also, 75.73 millions of tons of agroindustrial residues
were generated in Mexico in 2006 (Saval 2012). On the other hand, Brazil generates large
amounts of agricultural residues from sugarcane cultivation; this residue could be used as
raw material to produce biojet fuel (Nicodème et al. 2018). This type of feedstock is trans-
formed through different types of processing technologies, such as gasification followed by
Fischer–-Tropsch synthesis, pyrolysis followed by hydroprocessing, or alcohol production
plus oligomerization. A review of the works where alcohol is generated as intermediate for
the subsequent production of biojet fuel is presented next.
Waste lignocellulosic feedstock can be converted to ethanol; however, a pretreatment
(hydrolysis) is needed in order to extract the sugar contained. Sun and Cheng (2002) did
a review of the hydrolysis of lignocellulosic materials for ethanol production. The studied
materials include agricultural residues and wastes such as nutshells, grasses, paper, and
newspaper. In this context, waste papers from chemical pulps are an interesting and new
raw material, which have been little explored to obtain bioethanol, and later biojet fuel.
Once bioethanol is produced, it can be processed to obtain biojet fuel (Figure 6.3).
In this context, a review for the production of biojet fuel from agricultural residues
through the production of alcohol as intermediate is not available in the literature. However,
Sarkar et al. (2012) did a comprehensive review of bioethanol production from agricultural
waste. The review includes different processes and methods to increase the concentra-
tion of fermentable sugars. In addition, information about the conversion of glucose and
xylose to ethanol through fermentation technologies is discussed. Other works reported
the production of bioethanol from kitchen waste (Tang et al. 2008), grape and sugar beet
pomaces (Rodríguez et al. 2010), lignocellulosic agro-waste (Mutreja et al. 2011), date
wastes (Acourene and Ammouche 2012), non-marketable dates (Louhichi et al. 2013),
sugarcane bagasse (Lin et al. 2013), oil palm fronds (Ofori-boateng and Lee 2014), and
a mixture of waste fruits juice (Mansouri et al. 2016). An important aspect to the selection

light gases
pretreatment hydrolysis fermentation oligomerization hydrogenation
naphtha
Physical
Waste Chemical
lignocellulosic Thermal biojet fuel
Biological
feedstock

green diesel

Water H2 Distillation
Reactants

Figure 6.3 Alcohol production plus oligomerization process to produce renewable aviation fuel from
waste lignocellulosic feedstock.
160 Process Systems Engineering for Biofuels Development

of the waste to produce bioethanol is the lignin content. The best lignocellulosic feedstocks
to produce bioethanol are those with reduced amount of lignin, since it is hydrophobic in
nature and is tightly bound to cellulose and hemicellulose (Sarkar et al. 2012).
After the generation of the alcohol, it is possible to obtain biojet fuel through thermopro-
cessing, as shown in Figure 6.1. The thermoprocessing includes several stages: dehydra-
tion, oligomerization, hydrogenation, and fractionation. In the dehydration stage, the water
molecule is removed from ethanol using a catalyst and heat. According to Sakthivel (2018),
the thermal decomposition of ethanol takes place in a temperature range of 400–450 ∘ C and
11 bar, and the catalyst is alumina or transition metal oxides. The oligomerization process
is the conversion of short chain into linear 𝛼-oleofins (long chain); this step needs a catalyst
such as chromium diphosphine and zeolites, in the case where acidic zeolites are used the
temperature range is 100–300 ∘ C at high pressure.
Brooks et al. (2016) showed a summary of biojet fuel production processes from alco-
hol, considering different intermediates. The best processes were those where butane and
carbonyl are intermediates, both with conversions of 70–90%; in comparison with a direct
alcohol to jet process with 30–70% of ethanol converted. Recently, in a review article about
biojet fuel production processes, a summary of oligomerization processes for biojet fuel
production was reported (Gutiérrez-Antonio et al. 2017). In that work the authors reported
that the catalyst and conditions depend on the monomer; however, these researches do not
conclude with the production of biojet fuel. In this context, the first research was realized by
Harvey and Quintana (2010); the raw material employed was 2-ethyl-1-hexene, the catalyst
was montmorillonite K-10 and sulfated zirconia, and the yield was 90% with a mixture of
diesel and jet fuel.
The oligomerization of propene, on solid phosphoric acid as catalyst, was studied by
Sakuneka et al. (2008). The operating conditions were 3.8 MPa and 160–240 ∘ C; also, the
alkylation of benzene and toluene with propene was analyzed. The results show that it is
possible to realize both reactions in the same catalyst, producing a synthetic jet fuel that
meets Jet-A1 specifications.
Olcay et al. (2013) studied the conversion of C5 sugars derived from lignocellulosic
feedstock to produce hydrocarbons. The yield was 55%, and the hydrocarbons included
gasoline, jet fuel, diesel fuel, and fuel oil. The operation conditions were 80–140 ∘ C,
5.5–8.27 MPa and Ru/Al2 O3 catalyst for the hydrocycloaddition stage, and for the
hydrodeoxygenation stage NaOH and Pt/SiO2 –Al2 O3 were employed.
In all the proposed studies the obtained products need additional processing, at least
one additional distillation stage. Next, a review on the production of biojet fuel through
pyrolysis and hydroprocessing of waste lignocellulosic feedstock is presented.
In the pyrolysis, the waste lignocellulosic feedstock is heated in a special process to
produce an oily product, bio-oil, which subsequently is refined to obtain biojet fuel (Air
Transport Action Group 2011), as shown in Figure 6.4. The pyrolysis is the thermal cracking
of biomass in the absence of oxygen (Jenkins et al. 2016); the product yield and distribution
depends on the operating conditions, such as temperature, pressure, and residence time.
Jenkins et al. (2016) presented an extensive review of different pyrolysis technologies,
operating conditions and obtained yields of the pyrolysis products; the conversion technol-
ogy was classified as slow (300–700 ∘ C; 5–500 mm), fast (400–800 ∘ C; <3 mm) and flash
(800–1000 ∘ C; <0.2 mm) pyrolysis according to the operating conditions. In fast pyrolysis,
the amount of bio-oil increased while biochar decreased. In addition, there were different
 Production of Biojet Fuel from Waste Raw Materials: A Review 161

CO2
H2O light gases
cracking/
pretreatment pyrolysis deoxygenation isomerization
naphtha
Waste Physical
lignocellulosic Chemical
feedstock Thermal biojet fuel
Biological

green diesel

Water H2 H2 H2 Distillation
Reactants

Figure 6.4 Pyrolysis followed by hydroprocessing process to produce renewable aviation fuel from waste
lignocellulosic feedstock.

type of reactors, whose selection depended on the feed size and desired purity of the bio-oil
(Thangalazhy-gopakumar and Adhikari 2016).
There are different waste raw materials that can be processed through pyrolysis. After the
pyrolysis stage, hydroprocessing must be performed in order to obtain biojet fuel. In addi-
tion, pyrolysis has been studied for power generation (Chiaramonti et al. 2007). Table 6.2
contains some of the reported waste raw materials studied for bio-oil production, which can
be further converted to biojet fuel; however, the conversion to biojet fuel is not included in
those works.
Another technology for the conversion of lignocellulosic materials into biofuels is the
Fischer–Tropsch synthesis, which was developed to convert synthesis gas, containing
hydrogen and carbon monoxide, to hydrocarbon products (Steynberg 2004). German
researchers Franz Fischer and Hans Tropsch developed this technology, which is an
indirect liquefaction process where raw material is the synthesis gas produced from
biomass. The composition of the oil product from Fischer–Tropsch synthesis depends on
the raw material and the operations conditions. In order to produce aviation fuels, three
stages are needed: conversion of biomass to synthesis gas, conversion of synthesis gas
to oil, and oil refining to aviation fuels (Figure 6.5; De Klerk 2016). According to the
applied temperature, the Fischer–Tropsch process can be divided into low-temperature
Fischer–Tropsch (LTFT) and high temperature Fischer–Tropsch (HTFT), where the
temperature ranges are 200–240 ∘ C and 300–350 ∘ C, respectively (Liu et al. 2013).
There are different works related to the conversion of waste lignocellulosic biomass to
renewable aviation fuel through Fischer–Tropsch synthesis. Schablitzky et al. (2011) used
biowaxes for the production of hydrocarbon fuels using Fischer–Tropsch synthesis. They
obtained as products naphtha, kerosene and diesel, with a selectivity of 12.7% to kerosene.
According to Liu et al. (2013), the Atomic and Alternative Energy Commission (CEA)
announced the construction of a pilot plant to produce diesel, kerosene and naphtha using
wood/straw/green waste. In 2014, the United States Department of Agriculture (USDA)
granted a loan guarantee to Fulcrum Sierra Biofuels, LLC, to build a biorefinery for bio-
jet fuel production from municipal solid waste using Fischer–Tropsch synthesis (USDA
2014).
Table 6.2 Summary of the use of waste lignocellulosic feedstock in the pyrolysis process, as an initial step to obtain biojet fuel.
Raw material Temperature (∘ C) Catalyst Products Yield Description References
Wood 350 Not mentioned Liquid products Not mentioned More than 99.925% of Helsen et al.
waste Gas heavy metals and (1998)
Heavy metals minerals are
Wood waste Coal captured.
impregnated The raw material was
with CCA dried in a conventional
oven with air
circulation, at a
temperature of 120 ∘ C
for 90 min
Peanut shell 500 Not mentioned Char 65% pine wood; None Wang et al.
Bio-oil 54% peanut shell; (2007)
Maize stalk Gas 45% maize stalk

Pine wood
Waste 350–700 Not mentioned Char Waste wood Slow pyrolysis Phan et al.
wood Liquid (% w/w): (2008)
Gas char (33.2%), Composition of each
Card-board liquid (44.7%), product:
gas (33.1%) Liquids: water,
Textile Cardboard: char (33.9%), oxygenated
residues liquid (32.0%), compounds, heavy oil
gas (42.3%) Gases: CO, CO2 , CH4 ,
Textile: H2
char (31.6%), Liquid had a high
liquid (47.0%), calorific value of
gas (45.8%) 10–12 MJ/kg
Waste rice husk 500 Meso-MFI Acids With Meso-MFI good The sample particles Jeon et al. (2012)
Hydrocarbons selectivity to aromatics were 8–10 mm long,
Pt-Meso-MFI Oxygenates 2.0–2.5 mm wide,
Phenolics 0.1–0.15 mm thick
Non-catalytic Anhydro sugars
Aromatics
Gas
Jatropha waste 500 HZSM-5 Acid The highest yield was 95% to The raw material was Vichaphund
Alcohol aromatic and aliphatic dried at 60 ∘ C for et al. (2014)
Aldehyde hydrocarbons 24 h, with particle
Aromatic HC size equal or less than
Ether 125 μm
HC
Ketone
Nitrogen compounds
Phenol
Sugar
Waste pepper stems 550 HZSM-5 (23) and Gas 15% by aromatics with Fast pyrolysis Park et al. (2015)
SiO2 /Al2 O3 Acids HZSM-5(23) The raw material has
Oxygenates hard texture
HZSM-5 (280) with Phenolics Large lignin content
SiO2 /Al2 O3 Aliphatics
Aromatics
Without catalyst PAHs
Nitrogen compounds
Camelina straw 500 HZSM-5 zeolite Char 55 wt% toward bio-oil in The catalyst was Hernando et al.
Coke sample partially de-ashed pelletized with a (2017)
Gas particle size of
H2 O 180–250 μm; the raw
Bio-oil material was reduced
to 0.5–1 mm and
dried at 105 ∘ C for
48 h
Waste pepper 500 Waste FCC Char 58% Oil; The reported yield is for Yoo et al. (2018)
Oil 10% char; the relation
HY zeolite Gas 35% gas FCC:acetone 1:10
164 Process Systems Engineering for Biofuels Development

CO
CO2
H 2O light gases
Fischer-Tropsch cracking/
gasification synthesis hydrotreating isomerization
naphtha
Waste
lignocellulosic
feedstock biojet fuel

H2O
green diesel

H2 H2 Distillation

Figure 6.5 Gasification followed by Fischer–Tropsch synthesis to produce renewable aviation fuel from
waste lignocellulosic feedstock.

Yamamoto et al. (2016) produced biojet fuel from woody biomass. The first step was the
pulverization of the biomass, followed by its gasification; thus, syngas consisting mainly
of H2 and CO was generated. The syngas was cooled and cleaned, before being supplied
to the synthesis process; the gas was pressurized to normal conditions in order to acceler-
ate the Fischer–Tropsch reaction. The CO2 was recovered through decarbonation after the
pressurization. Hydrogen was fed to the hydrocarbon generated by the synthesis, and then
the isomerization process generated isoparaffins; the products such as light oil, jet fuel, and
wax were recovered through distillation. The process converted 37% of the biomass into
biofuels, including biojet fuel, light oil, and wax.
Kim et al. (2016) studied the production and evaluation of diesel using the
Fischer–Tropsch process at pilot scale. Wood pellets were used as raw material. The
produced fuel reached the automotive fuel standard. Diesel could be hydrocracked and
hydroisomerized later to generate biojet fuel; however, this additional processing was not
included in the study.
In the study reported by Snehesh et al. (2017), 1000 kg/h of casuarina wood chips were
gasified in order to enter into a Fischer–Tropsch unit. The maximum conversion was 73%.
In addition, through an economic analysis, they found that a 50% increase in the biomass
cost led to an 18% increase in the cost of the fuel.

6.4 Waste Sugar and Starchy Feedstock

Waste sugar and starchy feedstock include mainly the residues from the food industry,
which produces large amounts of these materials each year. Big amounts of fruit waste are
produced daily, from agricultural processes worldwide; many of these residues are often
dumped into landfills or the ocean (I.S. Choi et al. 2015). Fruit wastes have high levels of
sugar, including sucrose, glucose, and fructose. Therefore, these residues can be used as
raw material to produce biojet fuel, generating an alcohol as an intermediate component.
Russ and Meyer-Pittroff (2004) studied the use of waste from the food production and
processing industries. In particular, waste sugar feedstock can be obtained from sugar
 Production of Biojet Fuel from Waste Raw Materials: A Review 165

light gases
pretreatment fermentation oligomerization hydrogenation
naphtha
Physical
Waste Chemical
sugar and Thermal biojet fuel
Biological
starchy
feedstock

green diesel

Water H2 Distillation
Reactants

Figure 6.6 Alcohol generation plus oligomerization process to produce renewable aviation fuel from
waste sugar and starchy feedstock.

production, preparation and processing of fruit, vegetables, grain, edible oil, cocoa, coffee
and tobacco, waste from the production of baked goods, along with sweets, waste bakeries,
confectioners and candy producers. These residues can be converted to bioethanol through
a fermentation process; this bioethanol can be used as raw material to produce biojet fuel
(Figure 6.6).
Mahro and Timm (2007) explored the possibilities of using residues from food pro-
cessing to generate bioenergy, biomaterial production, chemical feedstock or animal feed.
They reported that from some residues, such as whey, molasses and pomace, sugars can be
extracted; these sugars can be converted into alcohol through a fermentation process, and
later converted to biojet fuel. An important aspect is that these types of residues contains
large amounts of water; thus, they have to be used as local resources, since they degrade
easily.
Another interesting proposal considered the use of coffee pulp as raw material for a
biorefinery, where valuable bioproducts and biofuels are obtained (Hughes et al. 2014).
With regard to biofuels, bioethanol was produced using Kluyveromyces marxianus yeast.
Moreover, I.S. Choi et al. (2015) proposed a pathway to produce bioethanol through the
combination of fruits and waste fruits. A high conversion, 90%, was obtained after 48 hours.

6.5 Main Challenges and Future Trends

Waste feedstock is an interesting renewable raw material for the production of biojet fuel. It
has as its main advantage its high-volume, periodical generation, which guarantees the sup-
ply for the production process. Moreover, in most of the cases these residues are not used
to generate value-added products. Additionally, they have potential to be harmful for the
environment and are produced in a big volume, thus representing a contamination problem;
therefore, its use as raw material for the production of renewable aviation fuel helps to solve
the issue of their disposal while a new energetic source emerges. The cost of waste feed-
stock is another advantage. Now, these raw materials can be obtained free or with reduced
costs, mainly due to its transportation; however, its price could increase with time if the
commercialization of the value-added products is successful.
166 Process Systems Engineering for Biofuels Development

The main disadvantage of waste feedstock is that usually it is scattered in different cul-
tivation sites or processing industries. Thus, the transportation costs play an important role
in the profitability of the production processes. This shows the importance of the devel-
opment of regional or local supply chains to produce renewable aviation fuel, in order to
decrease the transportation costs, especially of lignocellulosic materials with low densities.
Thus, the optimization of the supply chain for the production of biojet fuel is mandatory.
Another disadvantage is the heterogeneity of the composition of the waste feedstock, as a
function of the period where it is generated as the main product from where the residue
is obtained. This issue could be addressed through pretreatments, where the raw material
composition can be standardized; nevertheless, this means additional processing and the
need of a specially designed control structure, which leads to an increase in the production
costs.
An interesting challenge is the flexible design of the production processes in order to
manage mixtures of waste feedstock, which helps to guarantee the supply of raw materials
to the process. In order to reach this flexible design, more studies are needed of the con-
version of mixtures of waste feedstock at different proportions and operating conditions.
On this topic, only one study is reported (I.S. Choi et al. 2015). Another challenge is the
development of more robust and flexible pretreatment operations, which helps to reduce
the processing costs. Moreover, energy integration and process intensification are powerful
tools to decrease operation and capital costs in the processing of waste feedstock to produce
renewable aviation fuel. In addition, the production of biojet fuel in a biorefinery scheme
from waste feedstock has not been reported, and it represents an important challenge from
the design point of view.
According to the review of the literature, in the case of waste triglyceride feedstock most
reported works are focused on the use of waste cooking oil, and some of them on the use
of animal fats as raw materials. However, to the best knowledge of the authors, there are no
reported works about the use of waste oil from the industrial sectors, where different kind
of machines are employed. Moreover, research in the use of different animal fats must be
intensified. In addition, the production of biojet fuel with oil extracted from spent coffee
grounds remains an area of opportunity, following reports on the production of biodiesel
with such raw materials (McNutt and He 2019).
Regarding the waste lignocellulosic feedstock, most of the works are focused on the use
of agriculture and forest residues through different pathways to produce biojet fuel. As
suggested by Sun and Cheng (2002), the use of paper and newspaper residues represent an
interesting feedstock to produce renewable aviation fuel, especially because of their homo-
geneous composition. Other residues such as used tires or plastics have not been analyzed
for their potential use to produce renewable aviation fuel.
On the other hand, few works have reported the use of waste sugar and starchy feed-
stock. The existing works are focused on the use of fruit residues to produce bioethanol.
No reported works for the production of biojet fuel from this waste feedstock were found.
For all type of waste feedstock, the use of mixtures has not been addressed in the
literature. From the design point of view the processing of mixtures is a great challenge,
especially for the heterogeneity of the composition of the feedstock; however, it will help
to guarantee the constant supply of raw material for the processing centers. Moreover, the
application of energy integration and process intensification remains an area of opportunity
in the processing of all types of waste feedstock (Gutiérrez-Antonio and Hernández 2019).
 Production of Biojet Fuel from Waste Raw Materials: A Review 167

The use of waste feedstock for the production of renewable aviation fuel will help to reach
the objectives established for the aviation sector related to its sustainable development. In
addition, it will incorporate the circular economy in other productive chains, which will
help to produce biojet fuel with the required quality and at competitive prices.

6.6 Conclusions
A review on the use of waste feedstock for the production of renewable aviation fuel
has been presented. The waste feedstock can be classified according to its nature in
triglyceride-containing materials, lignocellulosic, sugar and starchy. Waste cooking
oil, animal fats, along with agriculture and forest residues have been explored for the
production of biojet fuel; in the case of food processing residues only the production of
bioethanol has been tackled, but this biofuel can be further processed to produce biojet
fuel. The use of mixtures of waste feedstock for the production of renewable aviation
fuel remains an area of opportunity, as well as the application of energy integration and
process intensification strategies for the decrease of operating and capital costs. Therefore,
research efforts must be focused on the processing of waste feedstock to produce biojet
fuel to contribute to the sustainable development of the aviation sector.

Acknowledgments
Financial support provided by CONACyT, grants 239765 and 279753, for the develop-
ment of this project is gratefully acknowledged. Moreover, Ana Laura Gómez-Moreno is
supported by a scholarship from CONACYT-SENER.

References
Acourene, S. and Ammouche, A. (2012). Optimization of ethanol, citric acid, and a -amylase production
from date wastes by strains of Saccharomyces cerevisiae, Aspergillus Niger, and Candida guilliermondii.
Journal of Industrial Microbiology and Biotechnology 39: 759–766.
Adeeva, V., Liu, H., Xu, B., and Sachtler, W.M.H. (1998). Alkane isomerization over sulfated zirconia and
other solid acids. Topics in Catalysis 6 (1–4): 61–76.
Air Transport Action Group (2011). Beginner’s Guide to Aviation Biofuels. https://aviationbenefits.org/
media/166152/beginners-guide-to-saf_web.pdf (accessed 26 February 2019).
Almanza, O.L., Narbeshuber, T., d’Araujo, P. et al. (1999). On the influence of the mordenite acidity in
the hydroconversion of linear alkanes over Pt-mordenite catalysts. Applied Catalysis A: General 178 (1):
39–47.
ASTM D7566-18a (2018). Standard specification for aviation turbine fuel containing synthesized hydro-
carbons. West Conshohocken, PA: ASTM International.
Bezergianni, S. and Kalogianni, A. (2009). Hydrocracking of used cooking oil for biofuels production.
Bioresource Technology 100 (17): 3927–3932.
Bezergianni, S., Voutetakis, S., and Kalogianni, A. (2009). Catalytic hydrocracking of fresh and used cook-
ing oil. Industrial and Engineering Chemistry Research 48 (18): 8402–8406.
Bezergianni, S., Kalogianni, A., and Dimitriadis, A. (2012). Catalyst evaluation for waste cooking oil
hydroprocessing. Fuel 93: 638–641.
Brooks, K.P., Snowden-Swan, L.J., Jones, S.B. et al. (2016). Low-carbon aviation fuel through the alcohol
to jet pathway. In: Biofuels for Aviation (ed. C. Chuck), 109–150. Elsevier Inc.
168 Process Systems Engineering for Biofuels Development

Calemma, V., Peratello, S., and Perego, C. (2000). Hydroisomerization and hydrocracking of long chain
n-alkanes on Pt/amorphous SiO2 -Al2 O3 catalyst. Applied Catalysis A: General 190 (1–2): 207–218.
Charusiri, W., Yongchareon, W., and Vitidsant, T. (2006). Conversion of used vegetable oils to liquid fuels
and chemicals over HZSM-5, sulfated zirconia and hybrid catalysts. Korean Journal of Chemical Engi-
neering 23 (3): 349–355.
Chiaramonti, D., Oasmaa, A., and Solantausta, Y. (2007). Power generation using fast pyrolysis liquids
from biomass. Renewable and Sustainable Energy Reviews 11 (6): 1056–1086.
Choi, I.H., Hwang, K.R., Han, J.S. et al. (2015). The direct production of jet-fuel from non-edible oil in a
single-step process. Fuel 158: 98–104.
Choi, I.S., Lee, Y.G., Khanal, S.K. et al. (2015). A low-energy, cost-effective approach to fruit and citrus
peel waste processing for bioethanol production. Applied Energy 140: 65–74.
Commercial Aviation Alternative Fuels Initiative (2019). Fuel Qualification. http://caafi.org/focus_areas/
fuel_qualification.html#approved (accessed 10 February 2019).
De Klerk, A. (2016). Aviation turbine fuels through the Fischer–Tropsch process. In: Biofuels for Aviation
(ed. C. Chuck), 241–260. Elsevier Inc.
De Lucas, A., Ramos, M.J., Dorado, F. et al. (2005). Influence of the Si/Al ratio in the hydroisomerization
of n-octane over platinum and palladium beta zeolite-based catalysts with or without binder. Applied
Catalysis A: General 289 (2): 205–213.
Dhar, A., Vekariya, R.L., and Sharma, P. (2017). Kinetics and mechanistic study of n-alkane hydroisomer-
ization reaction on Pt-doped 𝛾-alumina catalyst. Petroleum 3 (4): 489–495.
Domínguez-García, S., Gutiérrez-Antonio, C., De Lira-Flores, J.A. et al. (2017a). Strategic planning for the
supply chain of aviation biofuel with consideration of hydrogen production. Industrial and Engineering
Chemistry Research 56 (46): 13812–13830.
Domínguez-García, S., Gutiérrez-Antonio, C., De Lira-Flores, J.A., and Ponce-Ortega, J.M. (2017b). Opti-
mal planning for the supply chain of biofuels for aviation in Mexico. Clean Technologies and Environ-
mental Policy 19 (5): 1387–1402.
European Biomass Industry Association. (2015). Transformation of Used Cooking Oil into Biodiesel: From
Waste to Resource. UCO to Biodiesel 2030. https://www.eubren.com/UCO_to_Biodiesel_2030_01.pdf
(accessed 3 June 2019).
Fu, J., Yang, C., Wu, J. et al. (2015). Direct production of aviation fuels from microalgae lipids in water.
Fuel 139: 678–683.
Gasca González, R. (2017). Diseño de un proceso de bajo costo para la producción de biodiésel a partir
de aceites de re-uso. Guanajuato: Universidad de Guanajuato.
Gomes, L.C., de Oliveira Rosas, D., Chistone, R.C. et al. (2017). Hydroisomerization of n-hexadecane
using Pt/alumina-beta zeolite catalysts for producing renewable diesel with low pour point. Fuel 209:
521–528.
Gui, M.M., Lee, K.T., and Bhatia, S. (2008). Feasibility of edible oil vs. non-edible oil vs. waste edible oil
as biodiesel feedstock. Energy 33 (11): 1646–1653.
Gutiérrez-Antonio, C. and Hernández, S. (2019). Process intensification applied to waste-to-energy pro-
duction. In: Waste-to-Energy, 43–56. Nova Science Publishers.
Gutiérrez-Antonio, C., Romero-Izquierdo, A.G., Gómez-Castro, F.I. et al. (2016a). Simultaneous energy
integration and intensification of the hydrotreating process to produce biojet fuel from Jatropha curcas.
Chemical Engineering and Processing: Process Intensification 110: 134–145.
Gutiérrez-Antonio, C., Romero-Izquierdo, A.G., Gómez-Castro, F.I. et al. (2016b). Energy integration of
a hydrotreatment process for sustainable biojet fuel production. Industrial and Engineering Chemistry
Research 55 (29): 8165–8175.
Gutiérrez-Antonio, C., Gómez-Castro, F.I., de Lira-Flores, J.A., and Hernández, S. (2017). A review on the
production processes of renewable jet fuel. Renewable and Sustainable Energy Reviews 79: 709–729.
Gutiérrez-Antonio, C., Gómez-De la Cruz, A., Romero-Izquierdo, A.G. et al. (2018a). Modeling, simula-
tion and intensification of hydroprocessing of micro-algae oil to produce renewable aviation fuel. Clean
Technologies and Environmental Policy 20 (7): 1589–1598.
 Production of Biojet Fuel from Waste Raw Materials: A Review 169

Gutiérrez-Antonio, C., Soria-Ornelas, M.L., Gómez-Castro, F.I., and Hernández, S. (2018b). Intensification
of the hydrotreating process to produce renewable aviation fuel through reactive distillation. Chemical
Engineering and Processing: Process Intensification 124: 122–130.
Hanafi, S.A., Elmelawy, M.S., Shalaby, N.H. et al. (2016). Hydrocracking of waste chicken fat as a cost
effective feedstock for renewable fuel production: a kinetic study. Egyptian Journal of Petroleum 25 (4):
531–537.
Harvey, B.G. and Quintana, R.L. (2010). Synthesis of renewable jet and diesel fuels from 2-ethyl-1-hexene.
Energy and Environmental Science 3: 352–357.
Helsen, L., Van Den Bulck, E., and Hery, J.S. (1998). Total recycling of CCA treated wood waste by
low-temperature pyrolysis. Waste Management 18 (6–8): 571–578.
Hernando, H., Fermoso, J., Moreno, I. et al. (2017). Thermochemical valorization of camelina straw waste
via fast pyrolysis. Biomass Conversion and Biorefinery 7 (3): 277–287.
Hughes, S.R., López-Núñez, J.C., Jones, M.A. et al. (2014). Sustainable conversion of coffee and other
crop wastes to biofuels and bioproducts using coupled biochemical and thermochemical processes in a
multi-stage biorefinery concept. Applied Microbiology and Biotechnology 98 (20): 8413–8431.
Instituto de Ecología del Estado de Guanajuato (2016). Padrón de Prestadores de Servicios de RME. https://
smaot.guanajuato.gob.mx/sitio/papsrme (accessed 21 February 2019).
International Air Transport Association (2009). A Global Approach to Reducing Aviation Emissions-First
Stop: Carbon Neutral Growth from 2020. www.iata.org (accessed 7 February 2019).
International Air Transport Association (2018a). IATA Fact sheet. https://www.iata.org/pressroom/facts_
figures/fact_sheets/Documents/fact-sheet-alternative-fuels.pdf (accessed 10 February 2019).
International Air Transport Association (2018b). IATA forecast predicts 8.2 billion air travelers in
2037.Press release no. 62. https://www.iata.org/pressroom/pr/Pages/2018-10-24-02.aspx (accessed 7
February 2019).
International Energy Agency (2017). Global Trends and Outlook for Hydrogen https://ieahydrogen.org/
pdfs/Global-Outlook-and-Trends-for-Hydrogen_Dec2017_WEB.aspx (accessed 03 June 2019).
Jaroszewska, K., Masalska, A., Czycz, D., and Grzechowiak, J. (2017). Activity of shaped Pt/AlSBA-15
catalysts in n-hexadecane hydroisomerization. Fuel Processing Technology 167: 1–10.
Jenkins, R.W., Sutton, A.D., and Robichaud, D.J. (2016). Pyrolysis of biomass for aviation fuel. In: Biofuels
for Aviation (ed. C. Chuck), 191–216. Elsevier Inc.
Jeon, M.J., Kim, S.S., Jeon, J.K. et al. (2012). Catalytic pyrolysis of waste rice husk over mesoporous
materials. Nanoscale Research Letters 7: 1–5.
Kim, Y.D., Yang, C.W., Kim, B.J. et al. (2016). Fischer– Tropsch diesel production and evaluation as alter-
native automotive fuel in pilot-scale integrated biomass-to-liquid process. Applied Energy 180: 301–312.
Leila, M., Whalen, J., and Bergthorson, J. (2018). Strategic spatial and temporal design of renewable diesel
and biojet fuel supply chains: case study of California, USA. Energy 156: 181–195.
Li, T., Cheng, J., Huang, R. et al. (2015). Conversion of waste cooking oil to jet biofuel with nickel-based
mesoporous zeolite Y catalyst. Bioresource Technology 197: 289–294.
Lin, Y., Lee, W., Duan, K., and Lin, Y. (2013). Ethanol production by simultaneous saccharification and
fermentation in rotary drum reactor using thermotolerant Kluyveromyces marxianus. Applied Energy
105: 389–394.
Liu, G., Yan, B., and Chen, G. (2013). Technical review on jet fuel production. Renewable and Sustainable
Energy Reviews 25: 59–70.
Liu, Y., Sotelo-Boyás, R., Murata, K. et al. (2012). Production of bio-hydrogenated diesel by hydrotreatment
of high-acid-value waste cooking oil over ruthenium catalyst supported on Al-polyoxocation-pillared
montmorillonite. Catalysts 2 (4): 171–190.
Louhichi, B., Belgaib, J., Benamor, H., and Nejib, H. (2013). Production of bio-ethanol from three varieties
of dates. Renewable Energy 51: 170–174.
Mahro, B. and Timm, M. (2007). Potential of biowaste from the food industry as a biomass resource.
Engineering in Life Sciences 7 (5): 457–468.
170 Process Systems Engineering for Biofuels Development

Mansouri, A., Rihani, R., Laoufi, A.N., and Özkan, M. (2016). Production of bioethanol from a mixture of
agricultural feedstocks: biofuels characterization. Fuel 185: 612–621.
McNutt, J. and He, Q.S. (2019). Spent coffee grounds: a review on current utilization. Journal of Industrial
and Engineering Chemistry 71: 78–88.
Mosisa, F., Shemsedin, A., Bhalerao, M.S., and Goud, V.V. (2018). Catalytic cracking of waste cooking oil
for biofuel production using zirconium oxide catalyst. Industrial Crops and Products 118: 282–289.
Mutreja, R., Das, D., Goyal, D., and Goyal, A. (2011). Bioconversion of agricultural waste to ethanol by
SSF using recombinant cellulase from clostridium thermocellum. Enzyme Research 2011 (1): 1–6.
Nicodème, T., Berchem, T., Jacquet, N., and Richel, A. (2018). Thermochemical conversion of sugar indus-
try by-products to biofuels. Renewable and Sustainable Energy Reviews 88: 151–159.
Ofori-boateng, C. and Lee, K.T. (2014). Ultrasonic-assisted simultaneous saccharification and fermentation
of pretreated oil palm fronds for sustainable bioethanol production. Fuel 119: 285–291.
Olcay, H., Subrahmanyam, A.V., Xing, R. et al. (2013). Production of renewable petroleum refinery diesel
and jet fuel feedstocks from hemicellulose sugar streams. Energy and Environmental Science 6: 205–216.
Park, Y., Lang, M., Ho, S., and Hoon, S. (2015). Catalytic fast pyrolysis of waste pepper stems over
HZSM-5. Renewable Energy 79: 20–27.
Phan, A.N., Ryu, C., Sharifi, V.N., and Swithenbank, J. (2008). Characterisation of slow pyrolysis products
from segregated wastes for energy production. Journal of Analytical and Applied Pyrolysis 81: 65–71.
Ranganathan, P. and Savithri, S. (2019). Techno-economic analysis of microalgae-based liquid fuels pro-
duction from wastewater via hydrothermal liquefaction and hydroprocessing. Bioresource Technology
284: 256–265.
Regalbuto, J. (2010). An NSF perspective on next generation hydrocarbon biorefineries. Computers and
Chemical Engineering 34 (9): 1393–1396.
Reimer, J.J. and Zheng, X. (2017). Economic analysis of an aviation bioenergy supply chain. Renewable
and Sustainable Energy Reviews 77: 945–954.
Rodríguez, L.A., Toro, M.E., Vazquez, F. et al. (2010). Bioethanol production from grape and sugar beet
pomaces by solid-state fermentation. International Journal of Hydrogen Energy 35 (11): 5914–5917.
Russ, W. and Meyer-Pittroff, R. (2004). Utilizing waste products from the food production and processing
industries. Critical Reviews in Food Science and Nutrition 44 (1): 57–62.
Sakthivel S. (2018). Jet Pathway: Production of low-carbon Jet fuel from alcohol.
Sakuneka, T.M., De Klerk, A., Nel, R.J.J., and Pienaar, A.D. (2008). Synthetic jet fuel production by
combined propene oligomerization and aromatic alkylation over solid phosphoric acid. Industrial and
Engineering Chemistry Research 47 (6): 1828–1834.
Sarkar, N., Ghosh, S.K., Bannerjee, S., and Aikat, K. (2012). Bioethanol production from agricultural
wastes: an overview. Renewable Energy 37 (1): 19–27.
Saval, S. (2012). Aprovechamiento de residuos agroindustriales: pasado, presente y futuro. BioTecnología
16 (2): 14–46.
Schablitzky, H.W., Lichtscheidl, J., Hutter, K. et al. (2011). Hydroprocessing of Fischer – Tropsch biowaxes
to second-generation biofuels. Biomass Conversion and Biorefinery 1 (1): 29–37.
SEDESOL (2012). Residuos, México. https://apps1.semarnat.gob.mx:445/dgeia/informe_12/pdf/Cap7_
residuos.pdf (accesed 26 February 2019).
Shi, W., Gao, Y., Song, S., and Zhao, Y. (2014). One-pot conversion of bio-oil to diesel- and jet-fuel-range
hydrocarbons in supercritical cyclohexane. Industrial and Engineering Chemistry Research 53 (28):
11557–11565.
Snehesh, A.S., Mukunda, H.S., Mahapatra, S., and Dasappa, S. (2017). Fischer-Tropsch route for the con-
version of biomass to liquid fuels – technical and economic analysis. Energy 130: 182–191.
Steynberg, A.P. (2004). Introduction to Fischer-Tropsch technology. Studies in Surface Science and Catal-
ysis 152: 1–63.
Sun, Y. and Cheng, J. (2002). Hydrolysis of lignocellulosic materials for ethanol production: a review.
Bioresource Technology 83: 1–11.
 Production of Biojet Fuel from Waste Raw Materials: A Review 171

Tang, Y.Q., Koike, Y., Liu, K. et al. (2008). Ethanol production from kitchen waste using the flocculating
yeast Saccharomyces cerevisiae strain KF-7. Biomass and Bioenergy 32 (11): 1037–1045.
Thangalazhy-gopakumar, S. and Adhikari, S. (2016). Fast pyrolysis of agricultural wastes for bio-fuel and
bio-char. In: Recycling of Solid Waste for Biofuels and Bio-Chemicals (eds. O. Parthiba Karthikeyan, K.
Heimann and S. Senthilkannan Muthu), 301–332. Springer.
Tian, H., Li, C., Yang, C., and Shan, H. (2008). Alternative processing technology for converting vegetable
oils and animal fats to clean fuels and light olefins. Chinese Journal of Chemical Engineering 16 (3):
394–400.
USDA (2014). USDA announces loan guarantee to help innovative company turn waste into renewable jet
fuel. http://www.usda.gov/wps/portal/usda/usdahome?Contentidonly=true&contentid=2014/09/0195
.xml (accessed 18 February 2019).
Vásquez, M.C., Silva, E.E., and Castillo, E.F. (2017). Hydrotreatment of vegetable oils: a review of the
technologies and its developments for jet biofuel production. Biomass and Bioenergy 105: 197–206.
Vera-Morales M., and Schäfer A. (2009). Fuel-cycle assessment of alternative aviation fuels.OMEGA Alter-
native Fuels Report (Draft), February, 1–39.
Verma, D., Rana, B.S., Kumar, R. et al. (2015). Diesel and aviation kerosene with desired aromatics from
hydroprocessing of jatropha oil over hydrogenation catalysts supported on hierarchical mesoporous
SAPO-11. Applied Catalysis A: General 490 (1): 108–116.
Vichaphund, S., Aht-Ong, D., Sricharoenchaikul, V., and Atong, D. (2014). Effect of synthesis time on
physical properties and catalytic activities of synthesized HZSM-5 on the fast pyrolysis of Jatropha waste.
Research on Chemical Intermediates 40 (7): 2395–2406.
Wang, G., Liu, Q., Su, W. et al. (2008). Hydroisomerization activity and selectivity of n-dodecane over
modified Pt/ZSM-22 catalysts. Applied Catalysis A: General 335 (1): 20–27.
Wang, H., Chen, H.P., Yang, H.P. et al. (2007). Fast pyrolysis of agricultural wastes in a fluidized bed
reactor. In: Proceedings of the 20th International Conference on Fluidized Bed Combustion, 719–725.
Springer.
Yamamoto, T., Tsubaki, N., Shinoda, K., and Hishida, M. (2016). Development of jet fuel production system
from woody biomass. In: Proceedings of the National Symposium on Power and Energy Systems. E223.
Kitakyushu, Japan: Japan Society of Mechanical Engineers.
Yoo, M.L., Yong, Y.P., and Park, H. (2018). Fast pyrolysis of waste pepper stem over waste FCC. Research
on Chemical Intermediates 44 (6): 3773–3786.
Zhang, H., Lin, H., Wang, W. et al. (2014). Hydroprocessing of waste cooking oil over a dispersed nano
catalyst: kinetics study and temperature effect. Applied Catalysis B: Environmental 150–151: 238–248.
Zhang, S., Zhang, Y., Tierney, J.W., and Wender, I. (2000). Hydroisomerization of normal hexadecane
with platinum-promoted tungstate-modified zirconia catalysts. Applied Catalysis A: General 193 (1–2):
155–171.
Zhang, S., Zhang, Y., Tierney, J.W., and Wender, I. (2001). Anion-modified zirconia: effect of metal promo-
tion and hydrogen reduction on hydroisomerization of n-hexadecane and Fischer-Tropsch waxes. Fuel
Processing Technology 69 (1): 59–71.
Zhang, X., Chen, B., Li, X.Y. et al. (2017). One-step preparation of biological aviation kerosene by catalytic
hydrogenation of waste lard over Pt/SAPO-11. IOP Conference Series: Earth and Environmental Science
93 (1): 1–9.
Zhang, Z., Wang, Q., Chen, H., and Zhang, X. (2018). Hydroconversion of waste cooking oil into bio-jet
fuel over a hierarchical NiMo/USY@Al-SBA-15 zeolite. Chemical Engineering and Technology 41 (3):
590–597.
 7
Computer-Aided Design for
Genetic Modulation to Improve
Biofuel Production
Feng-Sheng Wang and Wu-Hsiung Wu
Department of Chemical Engineering, National Chung Cheng University, Chiya 62102, Taiwan

7.1 Introduction
Bioprocess engineering is a branch of biotechnology that is responsible for translating the
discoveries of science into practical products, processes, or systems that can serve the needs
of society. The improvement in production processes to achieve commercially viable pro-
duction levels is a prerequisite for any bioprocesses. Improvements in the product yield,
rate of production, and final product concentration are common goals in achieving more
efficient and cost-effective bioprocesses. These improvements can be achieved by two main
approaches: process development and genetic modulation. Process improvements as shown
in Figure 7.1 involve the adjustment of the environment of the organisms and the opti-
mization of parallel and downstream processes to achieve the best possible performance.
Process improvements integrate bioreactor, recovery, separation, purification, and utility in
order to evaluate process economics and its performance to achieve maximum productiv-
ity and minimal cost. Optimization is generally applied to bioprocesses toward achieving
those goals. For example, Singh and Rangaiah applied multi-objective optimization (mini-
mizing greenhouse gas emissions and cost of manufacture) to solve a bioethanol recovery
and dehydration process (Singh and Rangaiah 2017). Differential evolution is applied to
solve optimal control of a fermentation process for xylitol production (Koop et al. 2017).

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
174 Process Systems Engineering for Biofuels Development

Strain improvement Process development

Raw Downstream
Cell Bioreactor
Materials Processes

Intracellular Extracellular Extracellular

effectors effectors effectors
(chemical) (physical) Product

Figure 7.1 Two approaches for improving the production rate of bioprocesses.

Hybrid differential evolution has been applied to design multi-stage integrated extractive
fermentation with cell recycling for ethanol production and to determine optimal dilution
rate, fed sugar concentration, and bleed ratio (Chen and Wang 2010). To continuously
produce ethanol, a two-tank fermenter with a cell recycling system is constructed and exper-
imented on to prohibit the growth of contaminant bacteria, and further stabilize the system,
resulting in high productivity (Wang et al. 2013, 2014). However, ethanol inhibition remains
a problem for the system. Computer-aided process/solvent design is introduced to identify
a biocompatible solvent for an extractive fermentation and separation processes to address
the issue of ethanol inhibiting cell growth (Cheng and Wang 2010).
The second approach to improving productivity is a molecular biological method that
uses a biotechnological strategy for strain improvement, as shown in Figure 7.1. Biotech-
nology is the application of biological science and engineering to the use of living organisms
to enhance bioproducts or to perform functions that can benefit the product condition.
Improving the productivity of commercially viable microbial strains is an important field
in biotechnology, especially since wild strains isolated from nature usually produce a low
level of products. To produce economically viable biofuel from microbial cells, it is gener-
ally necessary to modify the metabolic pathway since microorganisms are typically evolved
for maximizing growth in their natural environment. Random mutagenesis and screening
are traditional and time-consuming approaches to improving productive microbial strains
(Ng et al. 2012). The rapid progress in molecular biology and the development of tools for
directed genetic modifications, high throughput measurements, and genome sequencing
enable the construction of mathematical models that provide a new paradigm for the ratio-
nal design of strains (Orth et al. 2010; Chung et al. 2012; Maranas and Zomorrodi 2016).
The use of these computer-aided strain designs, often referred to as metabolic engineering,
for improvement of cell factories is not a novel concept (Stephanopoulos et al. 1998), but
recent synergies with tools developed in systems biology have enabled the production of
a variety of products through biotechnology with significantly reduced time and resources
required for commercialization (Otero and Nielsen 2010).
Mathematical models can be used to represent biological models that range from
global views of cellular systems to detailed descriptions involving different levels of
cellular organizations, including genes, proteins, the metabolism, and signaling pathways.
Metabolic network models may be classified into two coarse categories: static and dynamic
approaches. The first category consists of constraint-based models (CBMs) or stoichiomet-
ric models, which use steady-state approaches. These models use the metabolic network,
which contains the biochemical reactions involved in the cell metabolism. The metabolism
can be compared with a chemical engine that drives the living process. CBMs enforce
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 175

cellular limitations on biological networks such as physicochemical constraints, spatial or

topological constraints, environmental constraints, and gene regulatory constraints (Orth
et al. 2010), and have two advantages: they are algebraic, which renders genome-scale
networks possible, and they are relevant, as many systems operate close to a steady
state. The second category is dynamic modeling, which has the potential to capture the
complex behavior of microbes more accurately. Kinetic models are, at least in principle,
especially capable of simulating time course data and permit a variety of dynamical
analyses of metabolic pathways. However, dynamic models are more difficult to analyze,
and require much greater data support than CBMs. Furthermore, kinetic data available for
the simulation of networks are quite scarce, limiting the number and size of systems in
different species that can be studied through this approach. This chapter focuses on the
development of a systems biology platform (SBP) for simulation of biological models using
constraint-based approaches. The platform can be applied on the design of growth-coupled
production strains that can give a hint to molecular biologists on performing experiments
to obtain strains for industrial production through adaptive evolution.

7.2 Method
Constraint-based modeling uses physicochemical constraints such as mass balance, energy
balance, and flux limitations to describe the potential behavior of an organism (Orth et al.
2010; Bordbar et al. 2014; Rau and Zeidan 2018). Under a given environmental condition,
the organism will reach a steady state that satisfies the physicochemical constraints. As the
constraints on a cellular system are never completely known, multiple steady-state solutions
are possible. In typical metabolic networks, the number of fluxes is greater than the number
of metabolites, because the same metabolite is usually involved in more than one reaction
and one compartment of a cell. To identify a physiologically meaningful steady state, an
optimization is performed to find the optimal flux of a specified objective function with
respect to the constraints identified.

7.2.1 Flux Balance Analysis

Flux balance analysis (FBA) is a constraint-based modeling approach in which the
stoichiometry of the underlying biochemical network constrains the solution. The stoi-
chiometric matrix of a typical metabolic system is underdetermined and infinitely many
solutions are possible. FBA formulates the metabolic network as a linear programing prob-
lem as shown in Eq. (7.1), where the solution of the underdetermined system is a member
of the solution space, and optimizes an objective function of choice, such as maximal
growth.
⎧max z = cT v
⎪ v
⎪subject to
⎪Nv = 𝟎
⎨vLB ≤ v ≤ vUB (7.1)
⎪ rev rev rev
⎪𝟎 ≤ virrev ≤ virrev
UB

⎪v ∈ Rn
⎩
where the matrix N is an m × n stoichiometry matrix with m metabolites and n reactions and
c is the vector representing the linear objective function. The decision variables v represent
176 Process Systems Engineering for Biofuels Development

fluxes, and vectors vUB specify the upper bounds for reversible or irreversible reactions,
respectively. The lower bounds vLB for the reversible reactions are negative values. FBA
assumes that metabolic networks will reach a steady state constrained by the stoichiome-
try. Though the stoichiometric constraints lead to an underdetermined system, a bounded
solution space of all feasible fluxes can be identified. The most widely used objective func-
tion in FBA of metabolic networks is the maximization of the cell growth rate built upon
the assumption that the cell is striving to maximally allocate all available resources toward
growth. FBA can be applied to predict flux capability if the cell is mutated. For example,
a gene is deleted so that the corresponding fluxes are set to zero in Eq. (7.1), i.e. vj = 0,
j∈ΩKO . In this situation, the problem is referred to as mutant FBA.

7.2.2 Flux Variability Analysis

Generally, linear optimization problems arising in FBA involve many optimal flux distri-
butions leading to the same maximum cell growth rate. Flux variability analysis (FVA) is a
constraint-based modeling variant of FBA. It addresses the well-known situation in linear
programming (LP) in which a problem has infinitely many solutions, because the optimal
solution is not one of the vertices of the solution simplex. Therefore, FVA is applied to deter-
mine the maximum and minimum values of all the fluxes that will satisfy the constraints
and allow for the same optimal objective value. It is expressed as follows:

⎧max ∕ min vi , i = 1, … , n
⎪ v v
⎪subject to
⎪ Nv = 𝟎
⎪
rev ≤ vrev ≤ vrev
⎨ vLB UB (7.2)
⎪
⎪ 𝟎 ≤ virrev ≤ virrev
UB

⎪ cT v ≥ z∗
⎪
⎩ v ∈ Rn
where z* is a maximum solution to Eq. (7.1). Generally, we need to solve 2n optimization
problems.

7.2.3 Minimization of Metabolic Adjustment

Flux distributions in a cellular metabolic network alter when a gene is knocked out in vivo.
However, on a global level, dramatic shifts in metabolism toward optimality are not imme-
diate. FBA predicts the long-term evolved state of a knockout cell but does not forecast the
immediate outcome of genetic manipulations. The minimization of metabolic adjustment
(MOMA) (Segrè et al. 2002) has the same stoichiometric constraints as FBA, but relaxes the
optimal growth flux for mutants and seeks an approximate solution for a suboptimal growth
flux state, which is nearest in flux distribution to the unperturbed state. MOMA adopts a
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 177

quadratic optimization function, resulting in a quadratic programming (QP) problem as

follows:
⎧ ∑m
⎪min (v − wi )2 = min (v − w)T (v − w)
v i=1 i v
⎪
⎪subject to
⎪
⎨ Nv = 𝟎 (7.3)
⎪ vLB ≤ v ≤ vUB
⎪ rev rev rev
⎪ 𝟎 ≤ virrev ≤ virrev
UB

⎪ v = 0, j ∈ Ω
⎩ j KO

An important feature of MOMA is that the wild-type flux distribution used, w, need not
be obtained by performing an FBA. Instead, an experimentally determined flux distribution
could serve better. Thus, objective functions for optimization, which at times may not reflect
the physiological situation very accurately, can be circumvented using MOMA. MOMA
also does not assume optimality of growth or any other metabolic function.

7.2.4 Regulatory On-Off Minimization

Another variant of FBA called regulatory on–off minimization (ROOM) problems (Shlomi
et al. 2005), similar to MOMA, attempts to minimize the number of significant flux changes
from the wild-type flux distribution. The ROOM formulation requires the solution of a
mixed integer linear programming (MILP) problem expressed as follows:

⎧ ∑m
⎪min yi
⎪ v , y i=1
⎪subject to
⎪
⎪ Nv = 𝟎
⎪ LB
⎨ vrev ≤ vrev ≤ vrev
UB
(7.4)
⎪ 𝟎 ≤ virrev ≤ vUB
⎪ irrev
⎪ vj = 0, j ∈ ΩKO
⎪ v − y (vUB − wU ) ≤ wU , v − y (vLB − wL ) ≥ wL , y ∈ {0, 1}
⎪ iU i i i i i i i i i i
⎪ wi = wi + 𝛿|wi | + 𝜀 = wi + Δwi , wLi = wi − 𝛿|wi | − 𝜀 = wi − Δwi
⎩
where yi = 1 for a significant flux change in vi and yi = 0 otherwise, and wi U and wi L are
thresholds determining significance of the flux change, with 𝛿 and 𝜀 specifying the relative
and absolute ranges of tolerance, respectively. wi and ΩKO are as in MOMA.

7.2.5 Optimal Strain Design Problem

A manual strategy, such as mutated FBA, MOMA, and ROOM, is a one-by-one to select a
gene to find a mutant strain that maximizes a desired production rate. However, selecting
178 Process Systems Engineering for Biofuels Development

suitable genes from a large-scale metabolic network is not a straightforward task. From
the computational standpoint, we can apply a bi-level optimization problem (BLOP) as
shown in Algorithm 7.1 to identify candidate genes. Several questions should be con-
sidered when using BLOP. What is the minimum set of genes in a given microorganism
that should be knocked out in order to maximize the synthesis flux of the desired end
products, to simultaneously minimize that of by-products, and to obtain a growth-coupled
strain?

Algorithm 7.1 Bi-level Optimization Problem for Determining Optimal Production Strains

1. Set the bioengineering and cellular objectives.

2. Set the maximum number of knockout genes.
3. Generate mass balance equations.
4. Generate enzyme capacity constraint equations.
5. Generate thermodynamics equations.
6. Delete reactions regulated by knockout genes.
7. Solve the inner optimization problem to maximize the cellular objective subject to steps
3–6.
8. Generate constraint equation to limit the number of knockout genes.
9. Solve the outer optimization problem to maximize the bioengineering objective subject
to steps 7 and 8.

Based on the BLOP framework shown in Algorithm 7.1, several methods, such as
OptKnock (Burgard et al. 2003), OptStrain (Pharkya et al. 2004), OptReg (Pharkya and
Maranas 2006), OptForce (Ranganathan et al. 2010), OptORF (Kim and Reed 2010),
EMILiO (Yang et al. 2011), and ReacKnock (Xu et al. 2013), are developed to identify
strains for maximizing the production rate. These methods transform the BLOP into a
single-level MILP problem by applying duality theory. However, such a duality transforma-
tion can increase computational time exponentially when the problem dimension increases.
Although the OptKnock algorithm requires a long CPU time (up to one week) to predict
a five-reaction knockout design using the Escherichia coli iAF1260 model, it was the first
constraint-based method used to predict strain designs for various substrates and products
(Feist et al. 2010). To avoid non-uniquely growth-coupled strains, RobustKnock (Tepper
and Shlomi 2009) extended the OptKnock (Burgard et al. 2003) method to yield guaranteed
production rates by accounting for the presence of competing pathways in the network
model iJR904 (Tepper and Shlomi 2009). Evolutionary algorithms have been applied
to identify the modulated genes of strain design problems with a user-defined objective
function in which complicated nonlinear objective functions can be used (Patil et al. 2005;
Rocha et al. 2008; Lun et al. 2009; Angione et al. 2012; Bautista et al. 2013; Rashid
et al. 2013).
The abovementioned methods may not be guaranteed to obtain a growth-coupled produc-
tion strain. Wang and Wu have applied FVA to the desired product as a constraint in the inner
optimization problem of Algorithm 7.1 to ensure that growth-coupled strains are obtained
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 179

(Wang and Wu 2015). The strain design problem then becomes a triple-level mixed-integer
linear optimization problem.

⎧max f = vminFVA
⎪ z 1 bioeng
⎪max f2 = vmaxFBA
⎪ z cellular
∑
⎪min f3 = z
⎪ z z∈ΩKO
⎪subject to
⎪
⎪ vminFVA = vmaxFVA
⎪ bioeng bioeng
⎨ ⎧max v (7.5)
⎪ ⎪ v cellular
⎪ ⎪subject to
⎪⎪
⎪ ⎪ Nv = 𝟎
⎪ ⎨ vLB ≤ v ≤ vUB , z ∉ Ω
⎪ ⎪ rev rev rev KO
⎪ ⎪ 𝟎 ≤ virrev ≤ vUB , z ∉ Ω
⎪⎪ irrev KO

⎪⎪ vj = 0, j ∈ ΩKO
⎩⎩

where the growth-coupled constraint, vminFVA

bioeng
= vmaxFVA
bioeng
, is used to solve the knocked-out
FVA for the desired product only, not all fluxes in the metabolic network. This triple level
optimization problem is a mixed-integer optimization problem. Classical algorithms for
solving BLOPs have applied the duality theory to convert the inner level optimization
problem as constraints in the outer-level problem. However, the duality transformation is
difficult for applying multi-level optimization problems, such as Eq. (7.5). Wang (2016) has
discussed in detail how to apply the nested hybrid differential evolution (NHDE) algorithm
to solve mixed-integer optimization problems.

7.3 Computer-Aided Strain Design Tool

A genome-scale metabolic network is complex, including a huge number of species,
reactions, gene regulation, and reaction interactions, and is hard for the human mind to
grasp. To understand the function of metabolism and control mechanisms of prokary-
otic/eukaryotic organisms, it is necessary to provide a platform for simulation and
analysis of genome-scale integrated metabolic networks. The second issue is to create an
efficient platform for systematic analysis of complex integrated metabolic networks using
computational tools, including constraint-based modeling (FBA, FVA, robust analysis,
and flux envelope). For this reason, the SBP program (Appendix 7.A) was developed in
the Windows system and is easy to use for end users who are not computer programming
specialists.
For communication and exchange between different analytical tools, genome-scale
metabolic network models are stored in XML-based systems biology makeup language
(SBML) format, a representation suitable for machine reading but difficult for human
180 Process Systems Engineering for Biofuels Development

reading. Many common tools support SBML. For example, SBML2LATEX (Dräger
et al. 2009), a Java-based program, retrieves the information of genome-scale metabolic
network models in SBML format and transforms it into a LaTeX document. Via the
transformation to LATEX, the output of SBML2LATEX supports scientific writing and is
easy for humans to read. When the size of the SBML input files increases, SBML2LATEX
takes a long time (many days) to generate the output and the number of output pages
(several hundred pages) is huge. Its performance limits its applications in genome-scale
mathematical models. The JSMBL project has developed a Java library for the creation of
Java web applications and the development of plugins for other applications (Dräger et al.
2011). JSBML parses a mathematical formula in SBML files and converts it into a syntax
tree that can be stored in memory using a data structure and is easily traced and displayed
for graphical visualization. Although the document generated by SBML2LATEX is easy
reading for humans, it is difficult to use in an analysis. To summarize the information in
an SBML file for error checking of mathematical models, SBP provides an SBML parser
that efficiently transfers model information into different worksheets in an Excel file and
ultimately, a model file in General Algebraic Modeling System (GAMS) format. This
transformation takes a few seconds to convert a genome-scale metabolic network model
into a GAMS model for further optimization analysis.
Simulation is one of the major components of systems biology studies of genome-scale
metabolic networks. Many simulation tools based on FBA have been developed and pub-
lished (Becker et al. 2007; Klamt et al. 2007; Rocha et al. 2010; Hyduke et al. 2011; Liao
et al. 2012). Objective optimization, FVA, robust analysis, gene essentiality/deletion analy-
sis, and flux envelop analysis are the basic functions implemented by most simulation tools.
These basic functions are implemented by solving LP problems, and their optimal solu-
tions can be found by LP problem solvers. Many advanced and complex analyses, such as
optimal gene knockout problems, are required in systems biology studies of genome-scale
metabolic networks. To implement these analyses, it may be necessary to solve MILP, non-
linear, or mixed integer nonlinear problems. The optimal solutions for these problems are
not easily found by different solvers. The GAMS, a commercial tool, consists of a language
compiler and a set of stable and high-performance solvers. Using GAMS, researchers can
focus on the applications of large-scale models.
A platform development for optimal gene intervention strategies is a key step in design-
ing microbial strains with enhanced capabilities. Such approaches can be implemented as
a toolbox of computer-aided design for genetic modulation. The SBP platform can trans-
form the mathematical models represented by SBML into GAMS codes and call different
GAMS solvers to solve these models coded in GAMS language. More accurate decisions
may then be made by comparing the results obtained from different solvers. Even the most
complete metabolic network reconstructions are not perfect and may contain missing infor-
mation. There is still an enormous need to improve the existing metabolic networks and
generate new metabolic networks for microorganisms. The model-building algorithm can
generate mathematical models for new metabolic networks automatically and efficiently.
With the assistance of this platform, users can focus on the simulation and analysis of
global metabolic networks and keep pace with the latest reconstructed metabolic networks.
In addition to the basic simulation functions (FBA, FVA, robust analysis, and flux envelope)
implemented by most simulation tools, the SBP platform implements extra simulation func-
tions, including metabolic adjustment and gene essentiality analysis, under the assumptions
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 181

of MOMA (Segrè et al. 2002), ROOM (Shlomi et al. 2005), and fuzzy equal metabolic
adjustment (FEMA) (Hsu and Wang 2013).

7.4 Examples
To illustrate the applications of the triple-level mixed-integer linear optimization prob-
lem and SBP, they were applied to the design of growth-coupled production strains for
a small-scale metabolic network of E. coli core text model and a genome-scale metabolic
model of E. coli iAF1260. Both models can be accessed from the biological database, BiGG
Models (Angione et al. 2012; King et al. 2016; Reed 2019).

7.4.1 E. coli Core Model

The core model consists of 72 metabolites and 95 reactions, which is a small-scale network,
used for education. The maximization of biomass growth rate is considered the cellular
objective. The bioengineering objective is to design growth-coupled strain to yield maxi-
mum ethanol production rates using glucose and anaerobic conditions.
Optimal growth-coupled strains for ethanol production with one-hit or two-hit enzyme
knockouts are shown in Table 7.1. For knockout gene tdcD, the optimal biomass growth rate
and ethanol production rate achieved were 0.474 h−1 and 31.443 mmol/(h gDW), respec-
tively. Other optimal results obtained by alternative one-hit genes were nearly identical to
those of tdcD. Table 7.1 shows that the ethanol production rate (32.768) was slightly higher
than the optimal result obtained from the one-hit knocked out, but the biomass growth rate
was lower if two-hit mutants were applied.

Table 7.1 One-hit and two-hit knockout enzymes and their encoding genes, and the regulated reactions
for the small-scale metabolic network solved by the NHDE algorithm.
minFVA
vbioeng maxFBA
vcellular
(mmol/ (mmol/
KO Gene Regulated reaction (h gDW)) (h gDW))
tdcD Acetate + ATP ⇐⇒ Acetyl-phosphate + ADP 31.443 0.474
One-hit enzyme

pta Acetyl-CoA + Phosphate ⇐⇒ Acetyl-phosphate 31.443 0.474

+ Coenzyme-A
ftsA Formate_(e) + H_(e) → Formate + H 31.448 0.451
Formate → Formate_(e)

pta Acetyl-CoA + Phosphate ⇐⇒ Acetyl-phosphate 32.768 0.470

+ Coenzyme-A
Two-hit enzymes

talA Glyceraldehyde-3-phosphate + Sedoheptulose-7-

phosphate ⇐⇒ D-Erythrose-4-phosphate + D-Fructose
-6-phosphate
tdcA Acetate + ATP ⇐⇒ Acetyl-phosphate + ADP 32.768 0.470
talA Glyceraldehyde-3-phosphate +
Sedoheptulose-7-phosphate ⇐⇒ D-Erythrose-4-
phosphate + D-Fructose-6-phosphate

KO denotes knocked-out. The subscript (e) indicates that the species is extracellular, and the others are in cytoplasm.
182 Process Systems Engineering for Biofuels Development

Figure 7.2 Flux distributions of wild type, drawn using Escher.

Flux distributions for the wildtype that can produce formate (EX_for(e)), acetate
(EX_ac(e)), and ethanol (EX_etoh(e)) simultaneously are shown in Figure 7.2. This
small-scale metabolic map makes clear that the ethanol production rate could be enhanced
if the by-product secretion of formate or acetate was blocked. The optimal strains listed
in Table 7.1 indeed deleted these pathways, e.g. tdcD, pta, and ftsA. For instance, the
wildtype yielded an ethanol production rate of 15.8 mmol/(h gDW). However, the pro-
duction rate could increase twofold higher than the wildtype if the gene pta was knocked
out. We observed that secretion rates for formate and acetate decreased significantly,
i.e. EX_for(e) = 9.91 mmol/(h gDW) and EX_ac(e) = 0 mmol/(h gDW), as shown in
Figure 7.3.
As discussed in Section 7.2, MOMA and ROOM were also applied to predict the flux
distributions for pta knock-out, as shown in Figures 7.4 and 7.5. Both methods were applied
to account for resilient effects so that the improved ethanol production rate ratios were less
than the results for the mutant FBA. As shown Figure 7.4, we observed that MOMA could
yield an ethanol production rate of 18.1 and formate rate of 32.3 mmol/(h gDW), but the
acetate rate was still zero because of pta knock-out. Figures 7.2–7.5 were drawn using
Escher (King et al. 2015).
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 183

Figure 7.3 Flux distributions of mutant FBA, drawn using Escher.

7.4.2 Genome-Scale Metabolic Model of E. coli iAF1260

Second-generation bioethanol production requires the development of economically fea-
sible and sustainable processes that use renewable lignocellulosic biomass as a starting
material. Ethanol production with lignocellulose as a starting material is a complicated
process. Generally, the process consists of at least three steps: pretreatment, hydrolysis,
and fermentation. After the pretreatment and hydrolysis steps, sugars, including glucose,
xylose, galactose, mannose, and arabinose, are yielded from lignocellulose and used for
fermentation. Since xylose is the second major fermentable sugar present in lignocellulosic
hydrolysates, its fermentation is essential for the economic conversion of lignocellulose
to ethanol. However, microorganisms that rapidly ferment xylose to ethanol at high yields
are essential to the development of cost-effective, large-scale xylose to ethanol processes.
Molecular biological methods can be applied to improve production rates. However, ran-
dom mutagenesis and screening are generally difficult to use to achieve this goal because
a microorganism consists of thousands of genes, which must be modulated one by one.
Computer-aided strain design can use CBMs to predict the growth-coupled strains for using
xylose to ferment ethanol.
184 Process Systems Engineering for Biofuels Development

Figure 7.4 Flux distributions of MOMA, drawn using Escher.

As mentioned above, glucose and xylose are the major fermentable sugars present in lig-
nocellulosic hydrolysates. The algorithm 7.1 has been applied to design the mutant strains
for using glucose to produce ethanol (Wang 2016). In this case study, the iAF1260 metabolic
model of E. coli was used to determine the growth-coupled strains for using xylose to
produce ethanol. This mathematical model consists of 1260 genes, 1668 metabolites, and
2383 reactions. The maximum xylose utilization rate was set to 20 mmol/(h gDW), and the
required ATP for non-growth-associated cell maintenance was set to 8.39 mmol/(h gDW).
All optimization problems were solved using the CPLEX solver accessed through GAMS
on a 4.0 GHz Intel Core i7 CPU with 32 GB of RAM. The performance and solution quality
of the NHDE algorithm depended on three settings: the tolerance ratio used in migration,
population size, and maximum number of iterations. The crossover factor CR and toler-
ance ratio 𝜀 were set to 0.5 and 0.05, respectively. A population size of 50 was used, and
the maximum number of iterations was 200.
Optimal growth-coupled strains for ethanol production with one-hit or two hit enzyme
knockout are shown in Table 7.2. For knockout gene rpe, the optimal biomass growth rate
and ethanol production rate achieved were 0.283 h−1 and 23.011 mmol/(h gDW), respec-
tively. The alternative one-hit complex (pntA and pntB) accomplished smaller ethanol pro-
duction rate of 15.941 mmol/(h gDW). However, two-hit knocked-out enzymes achieved
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 185

Figure 7.5 Flux distributions of ROOM, drawn using Escher.

higher ethanol production rates. Wang considered glucose as the carbon source in designing
optimal growth-coupled strains (Wang 2016). We observed that the mutant E. coli using glu-
cose can achieve higher ethanol production rates and biomass growth rates than the optimal
strains using xylose. This observation is consistent with experiments showing that E. coli
prefers to use glucose as a carbon source for fermenting ethanol. The flux envelope of each
optimal strain was computed (Figure 7.6) to explain why each design satisfied the equal-
ity constraint at the maximum biomass growth rate and was a growth-coupled production
strain. Ethanol production rate increases should cause biomass growth rate decreases that
could be observed from these flux envelopes.

7.5 Conclusions
Bioprocess developments can be achieved by process design and integration, metabolic
engineering approaches that mutate microorganisms to enhance production rates. Opti-
mization could be generally applied to bioprocess developments to achieve that goal. From
the molecular biology viewpoint, a microorganism consists of thousands of genes that
constantly undergo random mutagenesis and it is generally difficult to achieve enhanced
production rates via screening. This chapter introduced constraint-based modeling
186 Process Systems Engineering for Biofuels Development

Table 7.2 Optimal growth-coupled strains using xylose to produce ethanol with one-hit or two-hit
knockout enzymes and their encoding genes, and the regulated reactions.
minFVA
vbioeng maxFBA
vcellular
(mmol/ (mmol/
Gene Regulated reaction (h gDW)) (h gDW))
rpe D-Ribulose-5-phosphate ⇐⇒ D-Xylulose-5-phosphate 23.011 0.283
(pntA+pntB) 2 H+ _(p) + Nicotinamide adenine dinucleotide - 15.941 0.383
reduced + Nicotinamide adenine dinucleotide
phosphate → 2 H+ + Nicotinamide adenine
dinucleotide + Nicotinamide adenine dinucleotide
phosphate - reduced
(grcA+pflA+pflB) 2-Oxobutanoate + Coenzyme A → 29.278 0.275
Formate + Propanoyl-CoA
(pflD+pflC) Coenzyme A + Pyruvate → Acetyl-CoA + Formate
(pta) Phosphate + Propanoyl-CoA → Coenzyme A + Propanoyl 27.797 0.301
phosphate
(eutD) Acetyl-CoA + Phosphate ⇐⇒ Acetyl phosphate +
Coenzyme A

The gene association (A+B) denotes gene A and gene B. The subscript (p) indicates the species are in periplasm of cells,
and the others are in cytoplasm.

40

(grcA + pfIA+pfIB),
(pflC+pflD)
30
Ethanol production rate

pta, eutD
rpe

20
(pntA+
pntB)

10 Wildtype

0
0.0 0.1 0.2 0.3 0.4 0.5
Biomass growth rate

Figure 7.6 Flux envelopes for each optimal growth-coupled strain and wild-type E. coli iAF1260.

approaches to screen for feasible mutants. Such mutated strains can enable molecular biol-
ogists to perform experiments to obtain strains for industrial production through adaptive
evolution.
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 187

Appendix 7.A The SBP Program

A program for systems biology studies, analysis, and design of genome-scale metabolic
networks is available for download at www.wiley.com. Please locate the product page for
this book using the search function, then refer to the Downloads section on the product
page. The SBP program (version 1.06) was developed and tested in x86 64 bit MS Win-
dows 7 platform. Download it from the book website, unzip the program package, and run
the SBP.exe execution file directly. The model transformation functions in the “Model”
submenu, GPR parsing functions in the “GPR” submenu, and SBML file parsing and trans-
formation functions in the “Utility” submenu can be executed directly, but the simulation
and analysis functions in the “Simulation” submenu need the support of the GAMS sys-
tem with a valid license. An appropriate license is required for the GAMS Base Module
and for at least LP, mixed integer programming (MIP), and quadratic constrained program-
ming (QCP) GAMS/Solvers. The installation directory of the GAMS system should be set
in the “Path” environment variable. The small-scale metabolic network of E. coli core text
model and genome-scale metabolic model of E. coli iAF1260 are included in the “Example”
subdirectory.

References
Angione, C., Carapezza, G., Nicosia, G. et al. (2012). Robust design of microbial strains. Bioinformatics
28 (23): 3097–3104.
Bautista, E.J., Zinski, J., Szczepanek, S.M. et al. (2013). Semi-automated Curation of metabolic models via
flux balance analysis: a case study with Mycoplasma gallisepticum. PLoS Computational Biology 9 (9):
e1003208.
Becker, S.A., Feist, A.M., Mo, M.L. et al. (2007). Quantitative prediction of cellular metabolism with
constraint-based models: the COBRA toolbox. Nature Protocols 2: 727.
Bordbar, A., Monk, J.M., King, Z.A., and Palsson, B.O. (2014). Constraint-based models predict metabolic
and associated cellular functions. Nature Reviews Genetics 15: 107.
Burgard, A.P., Pharkya, P., and Maranas, C.D. (2003). Optknock: a bilevel programming framework for
identifying gene knockout strategies for microbial strain optimization. Biotechnology and Bioengineer-
ing 84 (6): 647–657.
Chen, M.-L. and Wang, F.-S. (2010). Optimal trade-off design of integrated fermentation processes for
ethanol production using genetically engineered yeast. Chemical Engineering Journal 158 (2): 271–280.
Cheng, H.-C. and Wang, F.-S. (2010). Computer-aided biocompatible solvent design for an integrated
extractive fermentation–separation process. Chemical Engineering Journal 162 (2): 809–820.
Chung, B.K.S., Lee, D.-Y., Koh, G., and Lakshmanan, M. (2012). Software applications for flux balance
analysis. Briefings in Bioinformatics 15 (1): 108–122.
Dräger, A., Planatscher, H., Motsou Wouamba, D. et al. (2009). SBML2L(A)T(E)X: conversion of SBML
files into human-readable reports. Bioinformatics (Oxford, England) 25 (11): 1455–1456.
Dräger, A., Rodriguez, N., Dumousseau, M. et al. (2011). JSBML: a flexible Java library for working with
SBML. Bioinformatics (Oxford, England) 27 (15): 2167–2168.
Feist, A.M., Zielinski, D.C., Orth, J.D. et al. (2010). Model-driven evaluation of the production potential
for growth-coupled products of Escherichia coli. Metabolic Engineering 12 (3): 173–186.
Hsu, K.-C. and Wang, F.-S. (2013). Fuzzy optimization for detecting enzyme targets of human uric acid
metabolism. Bioinformatics (Oxford, England) 29 (24): 3191–3198.
Hyduke, D., Schellenberger, J., Que, R., et al. (2011). COBRA Toolbox 2.0. Protocol Exchange. doi:
10.1038/protex.2011.234.
188 Process Systems Engineering for Biofuels Development

Kim, J. and Reed, J.L. (2010). OptORF: optimal metabolic and regulatory perturbations for metabolic
engineering of microbial strains. BMC Systems Biology 4 (1): 53.
King, Z.A., Dräger, A., Ebrahim, A. et al. (2015). Escher: a web application for building, sharing, and
embedding data-rich visualizations of biological pathways. PLoS Computational Biology 11 (8):
e1004321.
King, Z.A., Lu, J., Dräger, A. et al. (2016). BiGG models: a platform for integrating, standardizing and
sharing genome-scale models. Nucleic Acids Research 44 (D1): D515–D522.
Klamt, S., Saez-Rodriguez, J., and Gilles, E.D. (2007). Structural and functional analysis of cellular net-
works with CellNetAnalyzer. BMC Systems Biology 1: 2–2.
Koop, L., Corazza, M.L., Voll, F.A.P., and Bonilla-Petriciolet, A. (2017). Optimal control of a fermenta-
tion process for xylitol production using differential evolution. In: Differential Evolution in Chemical
Engineering, vol. 6 (eds. G.P. Rangaiah and S. Sharma), 321–351. World Scientific.
Liao, Y.-C., Tsai, M.-H., Chen, F.-C., and Hsiung, C.A. (2012). GEMSiRV: a software platform for
GEnome-scale metabolic model simulation, reconstruction and visualization. Bioinformatics 28 (13):
1752–1758.
Lun, D.S., Rockwell, G., Guido, N.J. et al. (2009). Large-scale identification of genetic design strategies
using local search. Molecular Systems Biology 5: 296–296.
Maranas, C.D. and Zomorrodi, A.R. (2016). Optimization Methods in Metabolic Networks. Wiley.
Ng, C., Jung, M.-y., Lee, J., and Oh, M.-K. (2012). Production of 2,3-butanediol in Saccharomyces cere-
visiae by in silico aided metabolic engineering. Microbial Cell Factories 11 (1): 68.
Orth, J.D., Thiele, I., and Palsson, B.Ø. (2010). What is flux balance analysis? Nature Biotechnology 28
(3): 245–248.
Otero, J.M. and Nielsen, J. (2010). Industrial systems biology. Biotechnology and Bioengineering 105 (3):
439–460.
Patil, K.R., Rocha, I., Förster, J., and Nielsen, J. (2005). Evolutionary programming as a platform for in
silico metabolic engineering. BMC Bioinformatics 6: 308–308.
Pharkya, P. and Maranas, C.D. (2006). An optimization framework for identifying reaction acti-
vation/inhibition or elimination candidates for overproduction in microbial systems. Metabolic
Engineering 8 (1): 1–13.
Pharkya, P., Burgard, A.P., and Maranas, C.D. (2004). OptStrain: a computational framework for redesign
of microbial production systems. Genome Research 14 (11): 2367–2376.
Ranganathan, S., Suthers, P.F., and Maranas, C.D. (2010). OptForce: an optimization procedure for identi-
fying all genetic manipulations leading to targeted overproductions. PLoS Computational Biology 6 (4):
e1000744.
Rashid, A.H., Choon, Y., Mohamad, M. et al. (2013). Producing succinic acid in yeast using a hybrid of
differential evolution and flux balance analysis. International Journal of Bio-Science and BioTechnology
5 (6): 91–100.
Rau, M.H. and Zeidan, A.A. (2018). Constraint-based modeling in microbial food biotechnology. Biochem-
ical Society Transactions 46 (2): 249–260.
Reed, J. (2019). Constraint-Based Workshops, Educational Material. http://reedlab.che.wisc.edu/
educational.php (accessed 27 March 2020).
Rocha, I., Maia, P., Evangelista, P. et al. (2010). OptFlux: an open-source software platform for in silico
metabolic engineering. BMC Systems Biology 4 (1): 45.
Rocha, M., Maia, P., Mendes, R. et al. (2008). Natural computation meta-heuristics for the in silico opti-
mization of microbial strains. BMC Bioinformatics 9 (1): 499.
Segrè, D., Vitkup, D., and Church, G.M. (2002). Analysis of optimality in natural and perturbed metabolic
networks. Proceedings of the National Academy of Sciences 99 (23): 15112.
Shlomi, T., Berkman, O., and Ruppin, E. (2005). Regulatory on/off minimization of metabolic flux changes
after genetic perturbations. Proceedings of the National Academy of Sciences of the United States of
America 102 (21): 7695–7700.
 Computer-Aided Design for Genetic Modulation to Improve Biofuel Production 189

Singh, A. and Rangaiah, G.P. (2017). Process development and optimization of bioethanol recovery and
dehydration by distillation and vapor permeation for multiple objectives. In: Differential Evolution in
Chemical Engineering: Developments and Applications, vol. 6 (eds. G.P. Rangaiah and S. Sharma),
289–320. World Scientific.
Stephanopoulos, G., Aristidou, A., and Nielsen, J. (1998). Metabolic Engineering: Principles and Method-
ologies. Elsevier.
Tepper, N. and Shlomi, T. (2009). Predicting metabolic engineering knockout strategies for chemical pro-
duction: accounting for competing pathways. Bioinformatics 26 (4): 536–543.
Wang, F.-S. (2016). Nested differential evolution for mixed-integer bi-level optimization for genome-scale
metabolic networks. In: Differential Evolution in Chemical Engineering, vol. 6 (eds. G.P. Rangaiah and
S. Sharma), 352–376. World Scientific.
Wang, F.-S. and Wu, W.-H. (2015). Optimal design of growth-coupled production strains using nested
hybrid differential evolution. Journal of the Taiwan Institute of Chemical Engineers 54: 57–63.
Wang, F.-S., Li, C.-C., Lin, Y.-S., and Lee, W.-C. (2013). Enhanced ethanol production by continuous
fermentation in a two-tank system with cell recycling. Process Biochemistry 48 (9): 1425–1428.
Wang, F.-S., Lee, W.-C., and Li, C.-C. (2014). Multi-continuous fermentation with cell recycling for ethanol
production. Japan Patent 5661673.
Xu, Z., Zheng, P., Sun, J., and Ma, Y. (2013). ReacKnock: identifying reaction deletion strategies for micro-
bial strain optimization based on genome-scale metabolic network. PLoS One 8 (12): e72150–e72150.
Yang, L., Cluett, W.R., and Mahadevan, R. (2011). EMILiO: a fast algorithm for genome-scale strain design.
Metabolic Engineering 13 (3): 272–281.
 8
Implementation of Biodiesel
Production Process Using
Enzyme-Catalyzed Routes
Thalles Allan Andrade, Massimiliano Errico, and Knud Villy
Christensen
Department of Chemical Engineering, Biotechnology and Environmental Technology, University of
Southern Denmark, 5230, Odense, Denmark

8.1 Introduction
Due to the continuous human population growth and the increase in the industrial and
transport sectors, the global energy demand has continuously increased. As specified by
the U.S. Energy Information Administration (2017), the total energy consumption in the
world was 575 quadrillion BTU in 2015. This number is expected to have increased by
28% in 2040, reaching a value of 736 quadrillion BTU. This demand is met mainly by
fossil fuels, which includes petrochemicals based on crude oil, coal, and natural gas. At
present, only a minor contribution is from renewable energy sources like biofuels and
hydroelectricity (Srivastava and Prasad 2000).
Environmental concerns due to the escalation in greenhouse gas emissions have increased
over recent decades, as these emissions lead to atmospheric pollution and global warming.
Numerous policies have been implemented aiming to reduce the worldwide emission of
greenhouse gases. As a result, the global energy-related CO2 emissions from fossil fuels
remained approximately flat in 2016, with the increase projected to be only 0.2% after a
growth of 2.2% over the past decade (REN21 2017).

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
192 Process Systems Engineering for Biofuels Development

Natural Gas Biofuels Others

1973 Others 2016
1.6 % 3.0% 1.1%
Coal 1.0% Natural Gas
3.0 % 3.7%

Oil
94.4%
Oil
92.2%

Figure 8.1 World energy consumption percentage by energy source in the transport sector. Source: Data
from International Energy Agency, France.

In addition to climate change and the increasing demand for energy, the increase in the
price of crude oils and the depletion of fossil fuel sources has resulted in a continuous search
for alternative and renewable energy sources as a replacement for fossil fuels. According
to the International Energy Agency (2018), even though crude oil has remained the leading
energy source, a continuous decrease in its market share has been observed, followed by a
reduction in the coal market share.
The transport sector is responsible for around 28% of the total energy consumption and
for 23% of energy-related emissions of greenhouse gases, according to the Renewable
Energy Policy Network (2017). Figure 8.1 shows the world energy source market share
in the transport sector in 1973 and 2016. Oil predominance has been the norm over the past
decades. However, an increase in the biofuel share due to the policies adopted to reduce
fossil fuel dependency can be seen.
Bioenergy is the major contributor for the supply of renewable energy. It is traditionally
divided into three categories: solid biomass that includes wood and harvesting residues;
liquid biofuels, including bio-oils, bioethanol, higher alcohols and biodiesel; and gaseous
biofuels, mainly biogas (World Energy Council 2016; REN21 2017).
Biodiesel is a renewable liquid biofuel based on fatty acid esters produced as a substitute
to the use of mineral diesel. It can be applied as a pure fuel or as a blend with oil-based
diesel. This blend is stable in all ratios. According to the Renewable Energy Policy Net-
work, global production of biodiesel increased 7.5% in 2016 compared with 2015. The
countries leading the production are the United States and Brazil, followed by Indonesia,
Germany, and Argentina. Figure 8.2 shows the global projection of biodiesel consumption
over the next few years. According to the Food and Agriculture Organization of the United
Nations (2017a,b), global biodiesel production is expected to reach 40.5 billion liters by
2026, corresponding to an increase of 12% with respect to the level registered in 2016.
Currently, worldwide biodiesel production is largely from vegetable oils, which mainly
includes rapeseed and sunflower oils in Europe, soybeans oils in the United States, Brazil,
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 193

40

30
Billion liters

20

10

0
2016 2019 2022 2026
Worldwide European Union USA
Figure 8.2 Projected worldwide biodiesel consumption.

and Argentina, lesser shares from palm oil in Indonesia, and other feedstocks such as
coconut and jatropha oils. This biofuel production also uses animal fats and industrial
by-products, such as used cooking oils as raw materials (Cao et al. 2008; Gui et al. 2008).
Among different vegetal oils, castor oil is a low cost non-edible vegetable oil with
unique properties among vegetable oils since it is mainly constituted of ricinoleic acid,
a fatty acid absent in other triglyceride sources (Berman et al. 2011; Bateni et al. 2014).
Biodiesel produced from castor oil has high viscosity and presents excellent lubricating
properties (Maleki et al. 2013). The choice of castor oil not only can decrease the feedstock
expenses, but also reduces concerns regarding food versus fuel usage of vegetable oils and
agricultural lands.
Transesterification is the most common method implemented in the industrial sector to
produce biodiesel. Typically, chemical-catalyzed routes are chosen for the transesterifica-
tion due to the low price of the chemical catalysts. The enzyme-catalyzed route is a less
energy-intensive and more environmentally friendly alternative to the use of chemicals with
many advantages, including higher compatibility with different raw materials and better
product separation (Christopher et al. 2014; Guldhe et al. 2015). On account of the higher
price of enzymes compared with chemical catalysts, the use of this route at present is not
common in industrial biodiesel production.
Enzyme-catalyzed transesterification of castor oil has only been investigated to a limited
extent and still requires further studies. Due to the higher stability and ease of handling of
immobilized enzymes compared with liquid enzymes, studies regarding enzymatic trans-
esterification have focused on the use of immobilized enzymes. The main advantage of
the use of liquid enzymes is their much cheaper price when compared with immobilized
enzymes. As chemical catalysts are still cheaper than liquid enzymes, an in-depth study
of biodiesel production processes using this type of enzymes is necessary to justify using
this process route. To the knowledge of the authors, process simulation and optimization
combined with the reuse of liquid enzymes in biodiesel production from castor oil have not
194 Process Systems Engineering for Biofuels Development

been done before. Therefore, this work is intended to evaluate the industrial applicability
of liquid enzymes for this transesterification.

8.2 Biodiesel Production Routes: Chemical versus Enzymatic Catalysts

Biodiesel is described as the mono alkyl esters of long chain fatty acids that is generally
produced by the transesterification of triacylglycerols (TAGs), the main constituent of veg-
etable oils and animal fats, with a short chain length alcohol, mostly methanol and ethanol
(Al-Zuhair 2007; Wang et al. 2009; Yuan et al. 2011). The transesterification of TAGs with
three moles of alcohol produces three moles of fatty acid alkyl esters (FAAEs), the major
constituent of biodiesel, and 1 mol of glycerol as by-product. The transesterification can be
carried out by various forms of catalysts, including acids, alkalis, and enzymes, or under
supercritical conditions with no addition of catalyst (Cao et al. 2009; Sotoft et al. 2010).
The reaction occurs in three consecutive reversible reactions in which TAGs are ini-
tially converted into diacylglycerols (DAGs); DAGs are converted into monoacylglycerols
(MAGs); and finally, MAGs react to produce glycerol. In each of these steps, 1 mol of
alcohol is consumed while 1 mol of FAAE is released.
Methanol is the most common alcohol used for the transesterification of TAGs, resulting
in the formation of fatty acid methyl esters (FAMEs). The choice of methanol is mainly due
to its low price, volatile nature, and higher reactivity compared with longer chain alcohols
(Christopher et al. 2014).
In the presence of water, hydrolysis of TAGs occurs with formation of 3 mol of free fatty
acids (FFAs) and 1 mol of glycerol per mole of TAG. Like the transesterification, the overall
reaction occurs in a sequence of three reversible reactions in which TAGs are converted into
DAGs, DAGs into MAGs and, finally, MAGs into glycerol. For each step, 1 mol of water
is consumed while 1 mol of FFA is produced. Due to the presence of alcohol, 1 mol of
produced FFA reacts with 1 mol of alcohol, producing 1 mol of FAAE and releasing 1 mol
of water. The reaction is called esterification.
Transesterification reactions are conventionally catalyzed by chemicals i.e. alkali and
acid catalysts. The reaction can be carried out as homogeneous or heterogeneous catalysis,
depending on the chemical catalyst solubility in the reaction mixture. These reactions
can be performed as one-step (alkali or acid catalyzed) or two-step (acid/alkali catalyzed)
processes depending on the FFA content in the raw material (Banković-Ilić et al. 2012).
Transesterification at supercritical conditions as well as enzymatic transesterification
emerge as alternatives to the use of chemical catalysts. Figure 8.3 shows the major
transesterification routes.

TRANSESTERIFICATION

HOMOGENEOUS CATALYST HETEROGENEOUS CATALYST NON CATALYTIC

ALKALI ACID ACID/ALKALI ALKALI ACID ENZYME SUPERCRITICAL

Figure 8.3 Major transesterification routes for biodiesel production.

 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 195

8.2.1 Chemical Catalysts

Alkaline catalysis is the most commonly used reaction route for biodiesel production.
Sodium and potassium hydroxides are the preferred homogeneous alkaline catalysts. They
can catalyze reactions at low temperature and atmospheric pressure (Abbaszaadeh et al.
2012; Nasir et al. 2013). High yield can be achieved in a short reaction time with a small
amount of catalyst that is widely available and economically cheaper than other biodiesel
routes (Lotero et al. 2005). Basic catalysts are also less corrosive to industrial equipment
allowing the use of cheaper reactor materials when compared with acid catalysts.
Kumar et al. (2013) investigated biodiesel production using sodium hydroxide as catalyst
and methanol as acyl acceptor. Linseed oil was chosen as feedstock for the alkali-catalyzed
transesterification. Using different concentrations of catalyst and alcohol-to-oil molar
ratios, the equilibrium conversion was reached in around 40 minutes, with TAG conver-
sion in the range of 88–96%. Al-dobouni et al. (2016) optimized the alkali-catalyzed
transesterification of wild mustard oil. Potassium hydroxide was chosen as catalyst.
Under optimal conditions, a 95.5% FAME yield was obtained after 45 minutes. From the
transesterification of Camellia japonica and Vernicia fordii seed oils, using potassium
hydroxide as catalyst, Chung (2010) produced biodiesel that met the European standard
BD100 with a FAME content higher than 96% in three hours.
The main drawback of the alkali-catalyzed process regards its applicability to feedstock
containing water and FFA. Oils that contain FFA amounts higher than 0.5% cannot be
totally converted into biodiesel. In fact, these FFAs react with the alkaline catalyst to pro-
duce soap that make downstream recovery and purification of biodiesel and glycerol separa-
tion difficult (Christopher et al. 2014; Leung and Guo 2006). Moreover, if the water content
exceeds 0.3 wt%, this may cause hydrolysis of the alkyl esters followed by saponification
under alkaline conditions, reducing the reaction yield.
Acid-catalyzed processes can tolerate up to 5 wt% FFA, and catalyze both the esterifica-
tion and transesterification reactions at the same time. Homogeneous acids such as sulfuric
acid, phosphoric acid, and hydrochloric acid are usually used for the transesterification
(Nasir et al. 2013). Acid catalysts require higher temperatures and pressures than alkali
catalysts as well as higher amounts of alcohol, normally adopting a molar ratio of 30:1
(Zhang et al. 2003a; West et al. 2008). If an excess in alcohol is used, glycerol recover-
ing becomes more difficult. Thus, determination of the optimal alcohol-to-oil molar ratio
appears fundamental (Marchetti et al. 2007). Zheng et al. (2006) studied the biodiesel pro-
duction from waste cooking oil (WCO) using acid-catalyzed methanolysis. Temperature,
methanol-to-oil molar ratio, and acid concentration (sulfuric acid) were investigated. Ini-
tially, the WCO contained 6 wt% FFA that were rapidly converted into FAME in the first
minutes of reaction. The authors obtained above 99% FAME yield at 70 ∘ C, after four hours
of reaction. However, the methanol-to-oil molar ratio was 245:1. When the reaction was car-
ried out at 80 ∘ C, the same yield was obtained using less methanol, 162:4.2:1 and 74:1.9:1
alcohol-to-acid-to-oil molar ratios. For these cases, an increase in the acid concentration
was necessary. Despite its insensitivity to FFA in the composition of the feedstock, transes-
terification catalyzed by acids has been largely disregarded, mostly due to its slower reaction
rate compared with basic-catalyzed reactions (Al-Zuhair 2007).
Economic analysis has demonstrated that acid-catalyzed procedures are more
cost-effective than base-catalyzed ones, since the latter require a step to convert FFA to
196 Process Systems Engineering for Biofuels Development

alkyl esters (Zhang et al. 2003b). To take advantage of both alkali- and acid-catalyzed
processes, a two-step acid/base process that combines both schemes can be used. In
this scenario, oil feedstocks containing high FFA are initially pretreated via alcoholic
esterification in an acidic medium (West et al. 2008). Esterification using an acid catalyst
reduces the FFA content that is converted into biodiesel. After the pretreatment, the TAG
alcoholysis using alkali catalyst is an effective method to obtain a high biodiesel yield
within a shorter reaction time at mild reaction conditions, compared with the one-step
processes. Using acid catalysts in the first step of the process, the risk of soap formation is
disregarded, and the problem of a slow reaction is overcome (Banković-Ilić et al. 2012).
Deng et al. (2010) compared the biodiesel yield obtained from one-step alkali-catalyzed,
one-step acid-catalyzed and two-step acid/alkali-catalyzed processes in an ultrasonic reac-
tor at 60 ∘ C, with methanol and jatropha oil. The transesterification reactions with sodium
hydroxide catalyst carried out with a 6:1 methanol-to-oil molar ratio for one hour reached
only 47.2% biodiesel yield due to soap formation. Acid-catalyzed transesterification using
concentrated sulfuric acid could obtain biodiesel yields up to 92.8% even though a reac-
tion time of four hours was needed. For the two-step process, 96.4% FAME yield was
achieved after 1.5 hours. Jatropha curcas seed oil with a high FFA content of (15%) was
also tested for biodiesel production by a two-step pretreatment process by Berchmans and
Hirata (2008). The FFA content was reduced to less than 1% after esterification with 1 wt%
sulfuric acid for one hour of reaction at 50 ∘ C. Then, after a two-hour transesterification
with 1.4 wt% sodium hydroxide at 65 ∘ C, a 90% FAME yield was achieved.
Since it requires more process units, the two-step process has a higher capital cost when
compared with the chemical catalyzed one-step process.
Numerous disadvantages are related to the chemical biodiesel production. The side reac-
tions of hydrolysis and saponification affect biodiesel yield and purity. The processes are
energy and capital intensive. Recovery and purification of glycerol is expensive and neutral-
ization and wastewater treatment are required (Meher et al. 2006; Christopher et al. 2014;
Parawira 2009).

8.2.2 Enzymatic Catalysts

Enzymatic catalysts are an option to overcome many of the problems related to the chem-
ical catalysts. Enzymes offer a biological route for biodiesel production with a variety of
advantages over the chemical route. Among different alternatives, lipases (defined as the
enzymes that catalyze fats) can catalyze effectively both esterification of FFA and trans-
esterification of TAG at the same time in either aqueous or non-aqueous systems. They
can operate under mild reaction conditions and are more compatible with variations in the
quality of the raw material. Lipases can produce biodiesel in fewer steps using less energy
and allow easier product recovery. They can further be reused, and their stability improved
through enzyme immobilization (Marchetti et al. 2007; Al-Zuhair 2007; Balat and Balat
2010; Fjerbaek et al. 2009).
However, excess alcohol in the enzymatic transesterification is reported to inhibit some
lipases. High concentrations of alcohol can lead to unfolding of the enzymes followed by
deactivation. Some strategies used to minimize this problem are use of an organic solvent
to increase the alcohol solubility or feed the alcohol into the reactor by stepwise additions,
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 197

which reduce the alcohol concentration in the reactor. Glycerol has also negative impact on
the lipase activity. Glycerol molecules form a hydrophilic environment around the lipase
preventing contact between the lipase and the hydrophobic substrate. This inhibition can
be minimized by continuous glycerol removal (Guldhe et al. 2015; Lotti et al. 2015). Other
drawbacks of the enzymatic transesterification include low reaction rate, long reaction time;
activity loss after some days of operation; and the high costs compared with chemical cat-
alysts, all of which make the industrial implementation of this route more difficult (Zhang
et al. 2003a; Fjerbaek et al. 2009).
Enzymes are commercially available in immobilized or liquid forms. Typically, immobi-
lized enzymes have been praised for the enzyme-catalyzed route. They are generally stable
and easily handled, and they can operate at low-water conditions. Immobilized enzymes can
be recovered by simple filtration from a batch reaction. The biggest problem is related to
their high price. This increases significantly the operating costs of the biodiesel production.
These high costs are due to the immobilization process and the carrier used to immobilize
the enzymes. In order to minimize and justify these costs, the enzyme should be reused
many times, maintaining the enzyme activity along the consecutives cycles (Nielsen et al.
2008; Nordblad et al. 2014; Amoah et al. 2016).
Maleki et al. (2013) compared the production of FAME from castor, palm and soybean
oils at the same reaction conditions, using Lipozyme TL IM lipase as catalyst. Castor oil
FAME yield was higher than the yield obtained in the reaction with palm and soybean oils.
This was due to the solubility of castor oil in methanol, which increased the FAME yield
without solvent addition. According to them, this solubility also diminishes the enzyme
deactivation problem and enhances the yield of the enzymatic reaction of methanolysis.
De Oliveira et al. (2004) looked into the transesterification of castor oil using two different
enzymes (Novozym 435 and Lipozyme IM) in a solvent medium. The highest conversion
(98%) was obtained using Lipozyme IM at 65 ∘ C, 20 wt% enzyme, 3:1 ethanol-to-oil molar
ratio and no water addition. Yang et al. (2005) evaluated castor oil methanolysis using enzy-
matic catalysis (Novozym 435 and Lipozym RM, Lipase QL, Lipase PL and Lipase AL)
in a solvent-free medium. Novozym 435 was the lipase with the highest activity among
the enzymes tested obtaining 97% FAME yield at optimal conditions: 6:1 methanol-to-oil
molar ratio, 2 wt% enzyme by weight of oil, at 50 ∘ C, for 24 hours.
The use of liquid lipases is a cheaper alternative to the immobilized enzymes. Liquid
enzymes are distributed in an aqueous solution with stabilizers added to prevent denatura-
tion of the enzyme. The usage of liquid enzymes also avoids the mass transfer limitations
inherent when using immobilized enzymes that often lead to slower reactions. However,
reuse of liquid enzymes is limited by the difficulty of their separation from the reaction
medium. Although these enzymes can be partially recovered in the Glycerol phase, enzyme
inhibition is caused by the presence of glycerol, leading to a decrease in biodiesel yield
(Nielsen et al. 2008). In their investigation of the liquid lipase NS81006 for the biodiesel
production, Li et al. (2014) evaluated the enzyme stability for reuse. They observed that,
at 45 ∘ C, 4.4:1 methanol-to-oil molar ratio, 10 wt% water, and 1.5 ml free lipase per 100 g
of oil, FAME yields above 90% could be achieved for five consecutive batches. At higher
temperatures, the lipases lost their stability.
When using liquid enzymes, water must be added to the system to create an interface
that activates the lipase. Water addition is necessary to preserve the lipase during exposure
to the alcohol (Poppe et al. 2015; Firdaus et al. 2016). As water promotes FFA production
198 Process Systems Engineering for Biofuels Development

Table 8.1 Comparison of different transesterification processes to produce biodiesel.

Process conditions Alkali-catalyzed Acid-catalyzed Lipase-catalyzed

Temperature (∘ C) 60–70 50–80 30–50
Pressure (MPa) 0.14–0.41 0.4 0.1
Reaction time 1–2 h 4 h–3 d 8h
FAAE yield High Medium High
FAAE refining Difficult Difficult Simple
Glycerol recovery Complex, low grade Complex, low grade Easy, high grade
FFA content Saponification Esterification Esterification
Wastewater High High Low
Catalyst reuse Difficult recovery Difficult recovery Reusable
Process costs Cheap Cheap Expensive
Energy requirement Medium High Low
Capital costs High High Low

decreasing the biodiesel yield, the optimal amount of water needs to be determined for
each feedstock.
Even though liquid enzymes are much cheaper than immobilized enzymes, they are still
more expensive than chemical catalysts. Thus, from an economic point of view, enzyme
recovery and reuse are strongly recommended in order to make enzyme-catalyzed routes
competitive with the chemical-catalyzed routes.
A comparison between chemical-catalyzed and lipase-catalyzed processes for biodiesel
production is given in Table 8.1 (Marchetti et al. 2007; Guldhe et al. 2015; Al-Zuhair 2007;
Abbaszaadeh et al. 2012).

8.3 Optimal Reaction Conditions and Kinetic Modeling

In order to be competitive with the biodiesel production from different feedstocks as well as
conventional production routes, the reaction conditions of the castor oil transesterification
catalyzed by liquid enzymes needs to be assessed and optimized.
To increase the TAG conversion into FAAE, numerous variables must be evaluated:
• Reaction temperature: Lipases from different sources have optimum enzyme activity in
the range of 20–70 ∘ C (Guldhe et al. 2015). An increase in the reaction temperature
increases the initial reaction rate thus reducing the reaction time for conversion. However,
above the optimum temperature, lipase activity decreases due to thermal deactivation of
the catalyst (Nordblad et al. 2014; Rodrigues et al. 2008).
• Alcohol-to-oil molar ratio: In order to shift the equilibrium toward product formation,
an excess of alcohol is frequently recommended. However, further increase in the alco-
hol content beyond the optimum concentration does not improve the biodiesel yield.
Conversely, it has a negative impact on the costs of alcohol recovery. Besides, excess of
alcohol inhibits the lipase activity because of insoluble alcohol droplets in the oil, causing
enzyme denaturation (Christopher et al. 2014; Mittelbach and Remschmidt 2006).
• Catalyst loading: Increase in the lipase loading has a positive influence on the reaction
rate. Because of the high costs of enzymes compared with chemical catalysts, large lipase
loadings are not recommended. The optimum loading needs to be determined in order to
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 199

establish a process economically feasible on an industrial scale (Akoh et al. 2007; Tan
et al. 2010).
• Water content: The presence of water during the transesterification may have a positive
influence on the activity and stability of the lipase. Water addition increases the interfacial
area between the aqueous and organic phases where the lipases are active. Excess water
can affect the reaction equilibrium, favoring the hydrolysis of the TAG and reducing the
alkyl esters yield.
• Solvent content: The presence of a solvent can reduce the viscosity and mass transfer
limitations. The solvent increases the solubility between oil and alcohol, providing a
better interaction of the substrates and reducing the lipase inhibition caused by the alco-
hol. Nevertheless, addition of a solvent increases the overall production cost and results
in an additional step for solvent separation (Guldhe et al. 2015; Nasaruddin et al. 2014).
• Reaction time: TAG conversion is generally faster at the beginning of the reaction pro-
cess, but more time is required in order to obtain higher yield (Patel and Sankhavara
2017).
• Rotational speed of mechanical mixing: Increasing the stirrer speed enhances the contact
area between reactants. As the stirrer speed increases, the conversion rate increases due
to the increased contact between enzymes and reactants in the inhomogeneous mixture
(Hsiao et al. 2018).

8.3.1 Evaluation of the Reaction Conditions

Reaction conditions of the castor oil methanolysis have been evaluated using liquid enzyme
Eversa Transform as catalyst. The reactions were carried out at 750 rpm, for eight hours. The
conditions studied were: temperature evaluated between 35 ∘ C and 50 ∘ C; methanol-to-oil
molar ratio, assessed from 4.5:1 to 7.5:1; content of lipase solution, between 1 wt% and
10 wt% by weight of castor oil; and added water content, from 0 to 10 wt% by weight of
castor oil. Methanol was added in four stepwise additions at two-hour intervals in order to
avoid enzyme inhibition.
The composition was characterized using an Agilent 1200 Series HPLC (high-
performance liquid chromatography) system with a UV detector at 205 nm equipped with
a Phenomenex Luna C18, 3 μm, 150 × 4.60 mm column. Samples were taken from the
reactor at predefined reaction times and centrifuged for five minutes at 6000 rpm in a
Spectrafuge™ Mini Laboratory Centrifuge. After the centrifugation, the top biodiesel-rich
phase was collected and diluted in a mixture of isopropanol–hexane (5:4 w/w), systemati-
cally mixed with a vortex mixer, and filtered using a Whatman 0.2 μm filter unit into the
HPLC vials.
Figure 8.4 shows the composition profile of TAG, DAG, MAG, FAME, and FFA for a
reaction at 35 ∘ C, 6:1 alcohol-to-oil molar ratio, 5 wt% enzyme and water contents. TAG
was almost entirely consumed in the first two hours of reaction, being converted into DAG
and MAG while releasing molecules of FAME and FFA. Contents of DAG and MAG
reached a maximum level at the beginning of the reaction, being consumed afterwards.
Stepwise addition of methanol favored the transesterification and esterification, increas-
ing the FAME concentration and decreasing the concentrations of DAG, MAG, and FFA
(Andrade et al. 2017a).
200 Process Systems Engineering for Biofuels Development

100
TAG DAG MAG
FAME FFA
80
Composition (%)

60

40

20

0
0 1 2 3 4 5 6 7 8
Reaction time (h)

Figure 8.4 Composition profile of the oil-biodiesel phase during enzymatic transesterification of castor
oil at 35 ∘ C, 6:1 alcohol-to-oil molar ratio, 5 wt% lipase solution, and 5 wt% water content.

100% 100%

90%
80%
80%
Biodiesel yie

Biodiesel yie

60%
70%

60% 40%
ld

50% 20%
ld

8,0 0 1
5
Al 7, ,0 46 10, 9 0,
8 ,0 0
co 7
ho 6,5 44 Wa 8,0 7,0 ,0
ter ,0 )
l-o 0 42 ) co 6
6 t%
il m 6, ,5 40 e (°C ,0 4 ,0
5 ,0 (w
ola 5 ,0
nte nt
38 ratur nt 4 3, ,0 n te
r r 5 .5 (w ,0 2 0 co
at 4 36 empe t% 2 1 ,0 me
io 4.0 34 T ) 0,0 ,0 zy
E n
(a) (b)

Figure 8.5 Effect of the reaction parameters on the FAME yield for the methanolysis of castor oil catalyzed
by Eversa Transform.

The effect of the reaction conditions on the FAME yield is shown in Figure 8.5. The
temperature had the smallest effect on the FAME yield. However, higher temperatures
decreased the yield due to possible enzyme deactivation (Figure 8.5a). Excess alcohol ini-
tially increased the reaction rate, which resulted in higher FAME yield. Continuous alcohol
increase led to a decrease in the FAME yield, possibly caused by inhibition of the enzyme
(Andrade et al. 2017b).
Increase in the enzyme content had a positive effect on the FAME yield. However, high
concentrations of enzyme solution loading results in higher manufacturing costs, suggesting
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 201

that this increase is not economically interesting. As seen in Figure 8.5b, a lower enzyme
content can be compensated by addition of water. Initially, an increase in water provided
larger interaction between water and oil. This activated the lipase, increasing the FAME
yield. However, larger additions of water decreased the FAME yield, once water started to
favor the hydrolysis.
The use of the lipase Eversa Transform as catalyst for the transesterification of castor
oil with methanol showed high conversion of TAG into FAME. Based on this evaluation
of different sets of reaction parameters, a maximum FAME yield and low FFA content
were obtained at optimal reaction conditions: 35 ∘ C, 6:1 alcohol-to-oil molar ratio, 5 wt%
of enzyme solution and 5 wt% water addition by weight of oil, eight-hour reaction. A yield
of 94% FAME was obtained together with a FFA content of 6.1%.

8.3.2 Kinetic Modeling

Besides the study of the reaction conditions, assessment of a suitable reaction mechanism
is also required for an appropriate reactor design for biodiesel production. The reaction
mechanism makes it possible to obtain the corresponding kinetic parameters to determine
the optimal reactor design as well as describe the reaction performance.
Experimental data were obtained under the optimal reaction conditions previously
obtained. For the sake of extensiveness, the reaction was performed also at 40 and 50 ∘ C.
Matlab v8.5 was used to evaluate the proposed mechanisms and estimate their respec-
tive kinetic parameters. The function ode45 was used to solve the ordinary differential

TAG w FFA
k1 k2
E TAG•E E
k–1 k–2

TAG DAG
FAME MeOH DAG DAG
w

k8 k–8 k–10 k10 k–3 k3

MeOH DAG FAME

MAG
k–11
FFA•E E DAG•E
k11

FAME Gly FAME MeOH w

k–4 k4
k7 k–7
k12 k–12

FFA FFA MeOH MAG FFA

MeOH Gly MAG
k–9 k–6 k–5
MeOH•E E MAG•E E
k9 k6 k5
MeOH FFA w MAG

Figure 8.6 Reaction mechanism of FAME production through enzymatic catalysis of castor oil.
202 Process Systems Engineering for Biofuels Development

equations by numerical integration, and by means of the optimization function lsqcurvefit,

the kinetic parameters were found by minimizing the deviation between the measured and
calculated data.
For the kinetic modeling of the enzymatic reaction, a Ping-Pong Bi Bi mechanism
was used. According to this mechanism, an enzyme-substrate complex is formed from
a reversible reaction between the enzyme and the substrate. This complex generates the
products, while it releases the free enzyme (Li et al. 2009). The formation of the enzymatic
complex with an inhibitor is also taken into account. To derive a rate expression based on
this mechanism, the pseudo-steady-state hypothesis was used by assuming that the net rate
of formation of the enzyme complexes is zero.
The mechanism found to describe best the composition profile of the reaction compo-
nents is shown in Figure 8.6 (Andrade et al. 2017c). According to this mechanism, the free
enzyme E reacts with the substrates TAG, DAG, MAG, and FFA forming the enzymatic
complexes. The inhibition caused by methanol is included by the formation of the inhibitor
complex MeOH⋅E. All the reactions in the mechanism, including the reactions of formation
of complexes, transesterification, hydrolysis, and esterification were found to be reversible.
Besides that, the reactions of transesterification and hydrolysis occur simultaneously with
formation of FAME and FFA, respectively.
In this mechanism, the rate expressions for the reaction components are:

d[TAG]
= −{[(VtTAG [MeOH] + VhTAG [W])[TAG]]
dt
[E]
+ [(V−tTAG [FAME] + V−hTAG [FFA])[DAG]]} (8.1)
[ET ]
d[DAG]
= {[(VtTAG [MeOH] + VhTAG [W])[TAG] − (VtDAG [MeOH] + VhDAG [W])[DAG]
dt
− [(V−tTAG [FAME] + V−hTAG [FFA])[DAG]
[E]
− (V−tDAG [FAME] + V−hDAG [FFA])[MAG]]} (8.2)
[ET ]
d[MAG]
= {[(VtDAG [MeOH] + VhDAG [W])[DAG]
dt
− (VtMAG [MeOH] + VhMAG [W])[MAG]]
− [(V−tDAG [FAME] + V−hDAG [FFA])[MAG]
[E]
− (V−tMAG [FAME] + V−hMAG [FFA])[Gly]]} (8.3)
[ET ]
d[FAME]
= {[(VtTAG [TAG] + VtDAG [DAG]
dt
+ VtMAG [MAG] + VeFFA [FFA])[MeOH]]
− [(V−tTAG [DAG] + V−tDAG [MAG] + V−tMAG [Gly]
[E]
+ V−eFFA [W])[FAME]]} (8.4)
[ET ]
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 203

d[FFA]
= {[(VhTAG [TAG] + VhDAG [DAG] + VhMAG [MAG])[W]
dt
− (VeFFA [FFA][MeOH])] − [(V−hTAG [DAG] + V−hDAG [MAG]
[E]
+ V−hMAG [Gly])[FFA] − (V−eFFA [FAME][W])]} (8.5)
[ET ]
d[Gly]
= {[(VtMAG [MeOH] + VhMAG [W])[MAG]]
dt
[E]
− [(V−tMAG [FAME] + V−hMAG [FFA])[Gly]]} (8.6)
[ET ]
d[MeOH] d[FAME]
=− (8.7)
dt dt
d[W] d[FFA]
=− (8.8)
dt dt
where [ET ] is the total concentration of the enzyme. The ratio [E]/[ET ] is obtained from
[E] 1
=
[ET ] 1 + KTAG [TAG] + KDAG [DAG] + KMAG [MAG] + KFFA [FFA] + [MeOH]∕KMeOH
(8.9)
This mechanism includes 19 kinetic parameters. The equilibrium constants as well as the
inhibition constant for methanol are:
k
KTAG = 1 (8.10a)
k−1
k
KDAG = 3 (8.10b)
k−3
k
KMAG = 5 (8.10c)
k−5
k
KFFA = 7 (8.10d)
k−7
k
KMeOH = −9 (8.10e)
k9
The maximum forward (Eq. 8.11) and reverse (Eq. 8.12) rate constants for transesterifi-
cation are defined as:
k k
VtTAG = 1 10 [ET ] (8.11a)
k−1
k k
VtDAG = 3 11 [ET ] (8.11b)
k−3
k k
VtMAG = 5 12 [ET ] (8.11c)
k−5
k k
V−tTAG = 1 −10 [ET ] (8.12a)
k−1
k k
V−tDAG = 3 −11 [ET ] (8.12b)
k−3
204 Process Systems Engineering for Biofuels Development

k5 k−12
V−tMAG = [ET ] (8.12c)
k−5

The maximum forward (Eq. 8.13) and reverse (Eq. 8.14) rate constants for hydrolysis
are defined as:
k k
VhTAG = 1 2 [ET ] (8.13a)
k−1
k k
VhDAG = 3 4 [ET ] (8.13b)
k−3
k k
VhMAG = 5 6 [ET ] (8.13c)
k−5
k k
V−hTAG = 1 −2 [ET ] (8.14a)
k−1
k k
V−hDAG = 3 −4 [ET ] (8.14b)
k−3
k k
V−hMAG = 5 −6 [ET ] (8.14c)
k−5

While the maximum forward (Eq. 8.15) and reverse (Eq. 8.16) rate constants for esteri-
fication are:
k k
VeFFA = 7 8 [ET ] (8.15)
k−7
k k
V−eFFA = 7 −8 [ET ] (8.16)
k−7

Figure 8.7a compares the experimental and calculated concentration profiles of the main
components in the enzymatic castor oil transesterification. The calculated concentration
was obtained from the mechanism proposed and the sum of squared differences obtained
from the function lsqcurvefit was 0.73 mol2 /l2 .
As a comparison, another mechanism proposed assumed that the reactions of transester-
ification, hydrolysis and esterification were irreversible. The experimental and calculated
concentrations are shown in Figure 8.7b. This led to an insufficient fit with over-prediction
of the concentrations of FAME and FFA, which indicates that this mechanism is unable to
describe the composition profile of the reaction components. The sum of squared differ-
ences was 2.50 mol2 /l2 for the mechanism with irreversible reactions. The higher sum of
squared differences suggests that the incorporation of reverse reaction kinetic parameters is
important to diminish the deviation between the calculated and experimental reaction data.
Table 8.2 displays the kinetic parameters found from the experimental data for differ-
ent reaction temperatures. The maximum forward rate constants for transesterification are
greater than the maximum forward rate constants for hydrolysis, which indicates the pre-
dominance of transesterification over hydrolysis. For the reactions of transesterification
and esterification, forward rate constants were higher than the reverse rate constants. The
reverse was observed for the reactions of hydrolysis, indicating the continuous formation of
glycerides from FFA. Small sums of squared differences for all sets of kinetic parameters
indicate a good fit for the mechanism at different reaction temperatures.
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 205

(a) 2.5
TAG
DAG
2 MAG
FFA
Concentration (mmol/l)

FAME

1.5

0.5

0
0 1 2 3 4 5 6 7 8
Time (h)
(b) 2.5
TAG
DAG
2 MAG
FFA
Concentration (mmol/l)

FAME
1.5

0.5

0
0 1 2 3 4 5 6 7 8
Time (h)

Figure 8.7 Experimental and calculated concentration profiles for the FAME production through enzy-
matic catalysis of castor oil for mechanisms with (a) reversible and (b) irreversible reactions.

8.4 Process Simulation and Economic Evaluation

The optimization of the reaction conditions is not enough to determine the feasibility of
the industrial implementation of this enzymatic route for biodiesel production. In addi-
tion to the transesterification reaction, recovery of the excess alcohol, glycerol separation
and biodiesel refining are fundamental steps in the biodiesel production process. These pro-
cesses also need to be investigated in order to reduce the manufacturing costs while assuring
high quality products.
The process using lipase Eversa Transform as catalyst is presented in Figure 8.8.
The process consists of five main units: feed preparation, transesterification reaction,
methanol recovery, glycerol separation, and biodiesel purification. The process was
206 Process Systems Engineering for Biofuels Development

Table 8.2 Kinetic parameters for the enzymatic transesterification of castor oil
into FAME at different temperatures.

Parameter 35 ∘ C 40 ∘ C 50 ∘ C
VtTAG [l/(mol ⋅ h)] 25.5 33.5 33.0
VtDAG [l/(mol ⋅ h)] 36.0 35.8 24.2
VtMAG [l/(mol ⋅ h)] 3.78 3.13 1.66
V−tTAG [l/(mol ⋅ h)] 0.434 0.045 0
V−tDAG [l/(mol ⋅ h)] 4.27 4.04 4.70
V−tMAG [l/(mol ⋅ h)] 0.601 0.503 0.299
VhTAG [l/(mol ⋅ h)] 2.58 3.02 0.836
VhDAG [l/(mol ⋅ h)] 14.4 15.4 17.3
VhMAG [l/(mol ⋅ h)] 3.01 × 10−7 9.36 × 10−7 1.55 × 10−5
V−hTAG [l/(mol ⋅ h)] 20.5 18.9 19.6
V−hDAG [l/(mol ⋅ h)] 45.6 42.9 46.6
V−hMAG [l/(mol ⋅ h)] 2.15 2.04 1.51
VeFFA [l/(mol ⋅ h)] 4.39 3.99 0.356
V−eFFA [l/(mol ⋅ h)] 0.326 0.310 0.063
KTAG (l/mol) 1.62 4.58 6.39
KDAG (l/mol) 35.0 21.3 2.61
KMAG (l/mol) 0 0 0.032
KFFA (l/mol) 3.25 × 10−6 0 3.16 × 10−6
KMeOH (mol/l) 4.83 × 103 4.83 × 103 4.83 × 103
Sum of squared differences (mol2 /l2 ) 0.73 0.93 0.95

simulated in General Algebraic Modeling System (GAMS) as a problem of nonlinear

programming using orthogonal collocation to solve the kinetic equations. Equipment
such as heat exchangers, distillation columns, and phase separators were modeled using
short-cut methods that were validated using the experimental data and the rigorous process
simulator Aspen Plus v8.8. As a basis for the simulation, an annual biodiesel production
of 250 000 tons was assumed, corresponding to the Danish biodiesel production projected
in 2017.
In the process, castor oil, water, and liquid enzyme were mixed in a tank. Water addi-
tion corresponded to 5.0 wt% by weight of castor oil. The enzyme solution content was
varied between 0.3 wt% and 10 wt%. The solution was mixed with methanol. The optimal
alcohol-to-oil molar ratio was defined to be between 3 and 9. Two mixers were used in the
process to merge different streams: in the first, fresh methanol was mixed with methanol
recovered in the distillation column. The second mixer combined the methanol stream with
the mixture from the tank.
The reaction was carried out for eight hours at constant temperature, ensuring that equi-
librium was reached. The reaction temperature was varied between 35 ∘ C and 50 ∘ C. The
reaction was defined to be a semi-batch reaction, with methanol being continuously added
into the reactor. Use of reactors in parallel was chosen to reduce the purchase costs of equip-
ment, and to guarantee a continuous process for the downstream section of the process.
A distillation column was used to recover the unreacted methanol that was going to be
reused in the enzymatic transesterification, as a high methanol recovery is wanted. A max-
imum limit of 150 ∘ C of the bottom temperature of the column was set to avoid thermal
glycerol decomposition (Zhang et al. 2003a).
 Methanol
4

Mixer 1
Castor Oil
6
1
Enzymes
2 TANK
5 HX3
Water Mixer 2
3
7

HX1
11
8

DISTILLATION
COLUMN 1
9 10 HX11
HX2
HX4

REACTOR
HX7 FFA
23 25
12 HX13

HX5 DISTILLATION
19 21 COLUMN 3
13 HX9

HX12
Glycerol DECANTER DECANTER DISTILLATION
15 2 14 1 17 18 COLUMN 2
HX6
FAME
16 HX8
24 26
HX14
Enzymes
TAG, DAG, MAG
20 22
HX10

Figure 8.8 Biodiesel production process through enzymatic catalysis of castor oil.
208 Process Systems Engineering for Biofuels Development

From the distillation column, the product was sent first to a decanter to separate the
organic phase from the remaining components. The top organic phase continues to a second
distillation column, whereas the remaining phases were sent to a second decanter where
glycerol was separated from the residual lipase solution. In order to have market value, the
glycerol phase should have a purity of at least 92 wt%.
Two consecutive distillation columns were implemented to separate the organic phase.
First, unreacted glycerides (including DAG and MAG) were removed from the bottom of
the second distillation column. The temperature in the top should not exceed 250 ∘ C to avoid
decomposition of biodiesel. The temperature in the bottom stream was limited to 350 ∘ C to
avoid oil decomposition. A third distillation column was used for the FFA removal from the
biodiesel phase. Temperature control was also required, assuring the bottom of the column
a maximum of 250 ∘ C to avoid degradation of biodiesel and FFA.
Since castor oil is basically composed of ricinoleic acid, the TAG that characterizes this
oil was taken to be triricinolein. Similarly, DAG and MAG were defined as glyceryl dirici-
noleate and monoricinoleate, respectively. FAME and FFA were, respectively, represented
by methyl ricinoleate and ricinoleic acid. The other components in the model included
methanol, water, glycerol, and propylene glycol, present in the liquid enzyme solution (75%
of water and 25% of propylene glycol, in volume).
Further, streams, decanters, mixers, and tanks in the process were presumed adiabatic.
These units, besides the reactor and heat exchangers were taken to operate at atmospheric
pressure. Pressure drops were not considered in the units and pipes (Andrade et al. 2019).
Due to the high solubility of ricinoleic acid in methanol, mass transfer limitations during
the reaction were neglected (Baron et al. 2014).
Optimal process conditions were obtained by maximizing the profit of the process, con-
sidering the selling price for biodiesel, FFA, and glycerol. The optimization also considered
the costs of the feedstock and catalyst, besides the utility costs.
The optimization of the process showed that the enzyme flow rate should be 0.3 wt% by
weight of castor oil. This low level was because of the high price of enzymes compared
with raw materials and chemicals. According to the manufacturer, Eversa Transform costs
around 15 US$/kg. The optimal alcohol-to-oil molar ratio was 9:1, in which 69% of the alco-
hol mass flow entering is recovered from the distillation column. The reaction temperature
was 36.4 ∘ C. A biodiesel yield of 93.8% was achieved after the reactor.
The decanting section resulted in a glycerol phase consisting of 95.5 wt% glycerol
whereas in the biodiesel purification, FAME and FFA purities of 99.97% and 98.2% were
obtained, respectively. The operating pressure of the distillation columns were 50, 2.3 and
3.5 kPa, respectively.
The economy of the project was evaluated based on the optimal process conditions. Cap-
ital costs were estimated as factors of the purchase costs of equipment. This estimation
followed the procedure presented by Sinnott (2005). As most well agitated jacketed carbon
steel reactors do not exceed 30 m3 , this was chosen as the maximum reactor size. It was
found that 14 reactors in parallel are required for the semi-batch transesterification reac-
tions. Regarding the decanters, horizontal carbon steel decanters with a volume capacity
lower than 100 m3 could be used. Results showed that three decanters at maximum volume
capacity would be necessary for the separation of the organic phase from the remaining
components, while one 35 m3 decanter should be used to separate the glycerol from the
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 209

Table 8.3 Specifications of the distillation columns.

Distillation 1 Distillation 2 Distillation 3

Total number of columns 5 11 2
Number of plates 6 6 40
Feed plate 1 3 28
Column height (m) 9.6 9.6 30
Column diameter (m) 3 3 3

Table 8.4 Estimate costs for an annual biodiesel production of

250 000 tons in millions of US$, for 360 laboring days.

Cost Millions of US$

Heat exchangers 1.1
Distillation columns 10.8
Reactor 1.0
Decanters 0.2
Storage tanks 1.0
Total equipment purchases 14.1
Direct costs 30.5
Total physical plant costs 43.6
Indirect costs 19.6
Total fixed capital 63.2
Working capital 3.2
Total capital cost 66.4
Annualized capital cost 3.3
Fixed costs 15.6
Variable costs 140
Sales, overheads and R&D 38.8
Annual production cost 194

residual lipase solution. Table 8.3 presents the specifications of the distillation columns
required in the process.
Operating costs are represented by fixed costs, which include costs for maintenance,
operating labor, laboratory, supervision, taxes and insurance, and variable costs, including
costs of raw materials, operating materials, and utilities. The approximation of the plant
profit was found by subtracting the annual production cost and the annualized capital cost
from the overall product sales.
The selling price for FAME, glycerol and FFA totalized 8.00 US$/s. Costs of raw
materials and utilities were 4.26 and 0.22 US$/s, respectively. Estimated costs are listed
in Table 8.4. The high capital cost of the process was due to the complexity of the
biodiesel purification process. The annual production cost of 194 million US$ resulted in
a production cost of 0.78 US$/kg. Taking into account the annualized capital investment,
the plant showed a final profit of 51.6 million US$/yr.
Santana et al. (2010) obtained a lower equipment purchase cost of 10.3 million US$ for
their alkali ethanolysis of castor oil, totalizing a production cost of 1.15 US$/kg of biodiesel.
Lower total fixed capital costs were also found for the alkali methanolysis of soybean oil in
210 Process Systems Engineering for Biofuels Development

the investigations by Granjo et al. (2017) and Haas et al. (2006), with around 49.5 million
US$ in both studies. However, the purchase costs of equipment for the process described
by Haas et al. was higher than these costs in the present work: 15.7 million US$. While
Granjo et al. obtained a production cost of 0.70 US$/kg of biodiesel, a cost of 0.76 US$/kg
of biodiesel was obtained by Haas et al. Supercritical methanolysis of WCO as proposed by
Lee et al. (2011) had a total fixed capital cost of 42.5 million US$ and a equipment purchase
cost of 5.7 million US$. The results achieved in the present work are similar to production
costs from the literature. The more complete step of biodiesel purification using two con-
secutive distillation columns to remove unreacted glycerides and FFA which guarantees a
better product quality explains the higher total fixed capital costs necessary for the simu-
lated process proposed using liquid enzymes as catalyst. This indicates that the simulated
enzymatic process could be economically feasible and could compete with alkali-catalyzed
processes. However, if enzyme recycling without a substantial reduction in the enzyme
activity is possible, this could increase the economic profitability of the process.

8.5 Reuse of Enzyme for the Transesterification Reaction

Recovery of Eversa Transform after the reaction can contribute not only to a decrease in
the production costs of the process but also to increase in the sustainability of this route
by decreasing the waste produced. Therefore, the feasibility of recovering the liquid lipase
through centrifugation and filtration with ceramic membranes was evaluated.

8.5.1 Recovery of Eversa Transform by Means of Centrifugation

The initial idea was to recover the enzyme by separation of the liquid lipase from the
oil-biodiesel and glycerol-rich phases using centrifugation. The reaction mixture was cen-
trifuged using a Thermo Scientific Sorvall ST 16R Centrifuge at 4000 rpm, for 30 minutes,
resulting in a three-phase system. The top phase was rich in oil-biodiesel, whereas the inter-
mediate phase mainly consisted of glycerol, and the bottom phase contained the liquid
lipase solution.
The concentration of glycerol and methanol in the top and bottom phases was
determined using an Agilent 1100 Series HPLC system with a Phenomenex Rezex
RHM-Monosaccharide H+, 300 × 7.8 mm column and a refractive index detector. It was
found that the biodiesel-rich phase contained 0.8 wt% of glycerol and 5.2 wt% of methanol.
Likewise, the compositions of glycerol and methanol in the enzyme-rich phase were 56
and 19 wt%, respectively.
The loss of enzyme activity was evaluated for two different scenarios. In the first, only the
recovered lipase was used in the following batch. In the second scenario, a mixture of 50%
fresh lipase and 50% lipase solution recovered from the centrifugation was tested. Eversa
Transform was used for three consecutive reaction cycles. The reaction mixture included
castor oil with a 6:1 methanol-to-oil molar ratio and the liquid lipase (10 wt% by weight
of oil). The reactions were performed at 35 ∘ C, for eight hours, and with a stirrer speed of
750 rpm (Andrade et al. 2017d).
The composition of the organic phase is shown in Table 8.5 for each condition. When
total lipase reuse was used, a decrease of around 27% in the produced FAME was observed
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 211

Table 8.5 Molar composition, in percentage, of the oil-biodiesel phase after

eight hours of reaction at 35 ∘ C with 6:1 methanol-to-oil molar ratio and 10 wt%
lipase by weight of oil using total and partial lipase solution recovered by means
of centrifugation.

Enzyme
reuse Batch TAG DAG MAG FAME FFA
1 0.0 ± 0.0 0.0 ± 0.0 1.3 ± 1.2 94.2 ± 1.3 4.5 ± 0.1
2 2.8 ± 0.3 4.4 ± 0.3 20.6 ± 0.6 68.4 ± 1.0 3.8 ± 0.1
Total
3 20.9 ± 0.6 8.4 ± 0.3 34.3 ± 1.0 33.4 ± 1.0 3.0 ± 0.1
2 0.0 ± 0.0 0.0 ± 0.0 3.5 ± 0.1 92.7 ± 2.8 3.8 ± 0.1
Partial
3 0.0 ± 0.0 1.5 ± 0.1 6.4 ± 0.2 88.4 ± 2.7 3.7 ± 0.1

in the second batch compared with the first batch. A FAME content of only 33.4% was
obtained in the third reaction batch. A smaller reduction in the FAME content was noticed
when a mixture of 50% reused and 50% fresh enzymes was used. A 92.7% FAME content
was achieved in the second batch and 88.4% in the third batch, resulting in only 6% decrease
with regard to the first batch.
Since no addition of water was used in the reactions, a decrease of FFA content was
observed along the batches in both scenarios. This reduction probably happened once part
of the water added originally with the liquid enzymes reacted to produce FFA.
Regarding the enzyme recovery, about 90–95% of the volume was recovered after the
first batch. This recovery fraction was reduced to 80–90% in the centrifugation after the
second batch.
Besides the reduction in the enzyme recovery, the presence of glycerol and methanol in
the enzyme-rich phase caused a decrease in the enzyme activity, indicating that lipase reuse
is not practical after the third batch.

8.5.2 Recovery of Eversa Transform by Means of Ceramic Membranes

Purification and concentration of components by membrane separation achieves high effi-
ciency without addition of chemicals, has low energy requirements and is easy to handle.
Selection of the correct membrane for a specific process separation depends on the charac-
teristics of the substances to be separated. Ultrafiltration (UF) membranes retain solutes like
colloidal solids, viruses, protein, and polysaccharides. They are characterized by a nomi-
nal molecular weight cut-off (MWCO), defined as the smallest globular solute molecular
weight in which at least 90% rejection is obtained by the membrane. UF membranes sep-
arate particles with molar masses between 1 kDa and 300 kDa (de Morais Coutinho et al.
2009; Padaki et al. 2015).
The use of ceramic membranes has several advantages compared with organic polymer
membranes in terms of high-temperature durability, mechanical strength, chemical inert-
ness, organic solvent resistance, unique surface characteristics and less likelihood of bac-
terial contamination. Ceramic membranes can operate at extreme pH values. They can be
cleaned with aggressive chemicals, organic solvents, and hot steam. Further, ceramic mem-
branes are available for industrial UF applications (Cheng et al. 2009; Benfer et al. 2004).
212 Process Systems Engineering for Biofuels Development

As an improvement to centrifugation, recovery of Eversa Transform by centrifugation

and UF was investigated. According to the information from the manufacturer, the molec-
ular weight of the lipase is close to 32 kDa. Based on this, UF membranes were chosen to
recover the liquid enzymes from the biodiesel production. The choice of a ceramic instead of
a polymeric membrane was based on the assumption that polymeric membranes have a ten-
dency to foul faster, which would result in a decline in the permeate flux (Padaki et al. 2015).
The transesterification was carried out at 35 ∘ C with 6:1 methanol-to-oil molar ratio,
5 wt% lipase and 5 wt% water by weight of oil. The reaction was performed for eight
hours at 750 rpm. After the reaction, the reaction mixture was centrifuged at 4000 rpm for
30 minutes and the lipase-rich phase was collected. The centrifugation step was necessary
to avoid castor oil and castor oil esters in the UF step as these would increase the viscosity
of the feed fluid and decrease the permeate flux.
The UF was carried out using a MiniMem Membrane Separation Lab Unit (PS
Prozesstechnik GmbH). The effectiveness of two ceramic membranes with different
MWCO was compared: UF 15 kDa (Batch no.: 267452) and UF 25 kDa (Batch no.:
267450) from Atech Innovations GmbH. These membranes are tubular membranes made
of TiO2 , ZrO2 , and Al2 O3 . The membrane’s length and thickness were 8 cm and 2 mm,
respectively, with an approximate filter surface per element of 15 cm2 . The purpose of the
UF is to eliminate the inhibitors glycerol and methanol from the lipase-rich phase in order
to increase the enzyme activity during the transesterification.
To reduce the glycerol and methanol content in the enzyme-rich membrane retentate,
diafiltration was used. During diafiltration, a diluent, generally water, is added to the reten-
tate in order to washout completely the permeable solutes (Sharma et al. 2017; Roda-Serrat
2017).
The diafiltration was done in batch mode using a control unit for automatic diafiltration
and continuous operation (PS Prozesstechnik GmbH) coupled to the membrane separa-
tion system. Figure 8.9 shows the set-up used for the separation of glycerol and methanol
from the lipase phase. The feed tank contained the lipase-rich phase obtained from the
centrifugation after the transesterification. Feed was continuously mixed with a magnetic
stirrer at 200 rpm. Water was used as the diafiltration liquid and added to the feed tank by
a diaphragm pump. The water flowrate was controlled by a level sensor that ensured a con-
stant feed volume of 40 ml in the feed tank. An HPLC pump was used to pump the feed
solution to the tubular ceramic membrane module. The feed flow rate was kept constant at
10 ml/min. A venting valve was coupled to the system to protect the HPLC valve from over-
pressure. A spring valve was used to regulate the pressure in the membrane module. The
retentate was recycled back to the feed tank through the spring valve, while the permeate
was collected in the permeate tank, with continuous monitoring of the permeate flux. The
filtration was done at room temperature (24 ∘ C). The transmembrane pressure was kept at
10 bar in order to avoid adsorption of solute in the pores, which would lead to membrane
fouling and low permeate flux.
After glycerol and methanol were removed, the process was operated as an ordinary
UF in order to concentrate the feed mixture. The UF reduced the Eversa Transform solu-
tion weight to 30% of the initial feed weight. The mass fraction of glycerol and methanol
removed by each permeate corresponded to the approximate sum of the weight of glycerol
and methanol in the lipase-rich phase after the centrifugation step.
During diafiltration, a nearly constant permeate flux of 5 l/(m2 ⋅h) was obtained.
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 213

DIAPHRAGM PUMP

RETENTATE

LC

LS

WATER VESSEL
FEED

TUBULAR CERAMIC
MEMBRANE
HPLC PUMP

PERMEATE

Figure 8.9 Process set-up for the purification of liquid enzymes by diafiltration and ultrafiltration.

Figure 8.10 shows the amount of glycerol and methanol present in the feed (lipase
phase) for the diafiltration of Eversa Transform. After the centrifugation, the Eversa
Transform-rich phase contained around 52 wt% glycerol and 21 wt% methanol. At least
nine hours of diafiltration was necessary to guarantee high glycerol and methanol removal.
Regarding the reduction of the amount of these substances with time during diafiltra-
tion, comparable behaviors were observed for both membranes. This suggests that ceramic
membranes with MWCO of 15 and 25 kDa have similar efficiencies to separate glycerol
and methanol from the reaction mixture.
Reuse of Eversa Transform was carried out under the same reaction conditions as used
when fresh enzymes catalyzed the reaction. The enzyme reuse was tested under three dif-
ferent separation processes:
1. Centrifugation: lipases recovered from the centrifugation were directly reused in the
transesterification.
2. Diafiltration: the lipase-rich phase obtained from the centrifugation was purified by
diafiltration. The lipase solution free of methanol and glycerol obtained in the feed tank
immediately after the diafiltration was reused as the reaction catalyst.
3. Concentration: lipase solution obtained from the diafiltration was concentrated in the
membrane set-up to remove surplus water prior to its reuse in the reaction.
The molar composition of the oil-biodiesel phase after eight hours of reaction using each
of the three lipase reuse solutions is shown in Table 8.6. After catalysis with the fresh
214 Process Systems Engineering for Biofuels Development

60
Glycerol - 25 kDa Glycerol - 15 kDa
Methanol - 25 kDa Methanol - 15 kDa
Phase composition (%)

40

20

0
0 3 6 9 12
Time (h)

Figure 8.10 Amounts of glycerol and methanol in the lipase-rich phase for the diafiltration.

Table 8.6 Molar composition, in percentage, of the oil-biodiesel phase after eight
hours of reaction at 35 ∘ C with 6:1 methanol-to-oil molar ratio, 5 wt% lipase and 5 wt%
water by weight of oil using lipase solution recovered from different separation
processes.

Separation MWCO Oil-biodiesel composition (%)

(kDa) TAG DAG MAG FAME FFA
Fresh lipase – 0.0 ± 0.0 0.0 ± 0.4 0.0 ± 0.7 94.0 ± 1.1 6.0 ± 0.2
Centrifugation – 0.0 ± 0.0 4.3 ± 0.1 18.5 ± 0.6 67.2 ± 2.0 10.0 ± 0.3
Diafiltration 15 0.0 ± 0.0 3.5 ± 0.1 14.4 ± 0.4 67.6 ± 2.0 14.5 ± 0.4
25 0.0 ± 0.0 3.9 ± 0.1 16.9 ± 0.5 64.1 ± 1.9 15.1 ± 0.5
Concentration 15 0.0 ± 0.0 0.0 ± 0.0 7.1 ± 0.2 82.8 ± 2.5 10.1 ± 0.3
25 0.0 ± 0.0 0.0 ± 0.0 7.1 ± 0.2 82.7 ± 2.5 10.2 ± 0.3

enzyme solution, molar composition of the oil-biodiesel phases was around 94% FAME
and 6% FFA. FAME content was similar to when 10 wt% lipase and no water addition was
used in the reaction with fresh lipases, as seen in Table 8.5.
A lower FAME content was observed when liquid enzyme was reused using centrifu-
gation only. In this case, the FAME content was 67%, while 10% FFA was obtained. In
addition to the higher FFA content, the presence of unreacted glycerides (DAG and MAG)
indicated that transesterification and esterification reactions were not complete. This is
probably because of loss in enzyme activity after the first batch, and enzyme inhibition
due to the presence of glycerol and methanol.
Use of enzyme recovered by centrifugation and diafiltration showed no improvement
in the FAME production. The FAME content was similar to what was obtained when
enzyme recovery by centrifugation only was used. Even though glycerol and methanol
were removed from the solution in the permeate stream, the water content increased,
leading to higher hydrolysis of castor oil. Therefore, the FFA content increased from 10%
to up to 15%.
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 215

100 4-hour reaction 8-hour reaction

80

FAME content (%) 60

40

20

0
15 kDa 25 kDa
Figure 8.11 FAME content for transesterification catalyzed by reused lipase after the concentration
step.

Concentration of the lipase solutions after centrifugation and diafiltration had a positive
effect on the FAME content when the lipase was reused. The FAME content increased
to around 83%. Removal of water in this process increased the lipase concentration in the
solution and reduced the hydrolysis, generating a lower FFA content. As seen in Figure 8.11,
the FAME content increased by around 20% during the last four hours of reaction. Longer
reaction time could probably increase the FAME content to reach values similar to what
was obtained when fresh lipases were used.
Similar FAME contents were obtained when ceramic membranes with MWCO of 15
and 25 kDa were used to recover the enzymes. This indicated that the lipase retention was
similar for both membranes.
The use of membrane technology for recovery of liquid lipases is thus promising. How-
ever, further investigation is required in order to increase the efficiency of the separation.
Thorough investigations are recommended, such as measuring the lipase activity in both
retentate and permeate solutions, and evaluating the UF under different filtration conditions,
including transmembrane pressures, cross flow velocities and membrane pore sizes.

8.6 Environmental Impact and Final Remarks

The optimization of the conditions of the transesterification reaction, the economical evalu-
ation of the process of enzymatic biodiesel production, as well as the recovery and reuse of
the catalyst in the reactions are important parameters to be taken into account for the indus-
trial implementation of enzyme-catalyzed routes for the production of biodiesel. However,
environmental impact assessment is essential to predict the consequences to the environ-
ment of the installation and performance of biodiesel plants.
In chemical processes, separation, treatment and disposal of the residues from the reac-
tion are needed. The more hazardous the substance, the more costly it is to deal with it.
Besides that, the larger the volumes of residues to be treated, the more complex the treat-
ments that are required. Avoiding generation of these residues is cheaper than treating them
(Anastas and Warner 1998). The use of enzyme in the process reduces the wastewater
that needs to be treated compared with the wastewater obtained from the neutralization
216 Process Systems Engineering for Biofuels Development

of catalysts used in chemical-catalyzed routes. The recovery of methanol in the process

also contributes to the reduction of waste.
Further, methods should be designed to use and generate substances that possess little
or no toxicity to human health and the environment. In this way, hazards should be min-
imized or eliminated in the process. The enzymatic route removes the use of hazardous
catalysts, such as sulfuric acid and sodium hydroxide as used in the chemical-catalyzed
routes. Castor oil, water, biodiesel, and glycerol are nonhazardous. In addition, methanol
can be produced from biogas creating a renewable source pathway. Due to their toxicity and
flammability, organic solvents have a negative effect on the environment and hence their
application must be minimal (Guldhe et al. 2015). Here, liquid enzymes are more beneficial
for the environment as solvents are not required.
Energy requirements should be recognized for their environmental and economic
impacts and should be minimized. It is recommended that the process be conducted at
ambient temperature and pressure. Enzymatic catalysis occurs at milder temperatures
than chemical-catalyzed transesterification, minimizing the thermal energy required. The
optimal reactor temperature was found to be 36 ∘ C. One of the most energy intensive
processes is, however, the purification and separation process. The process leading to a
complete separation and purification of FAME is done at temperatures above 200 ∘ C and
pressures lower than 0.04 bar in distillation columns 2 and 3. Therefore, heat integration
should be used to minimize the energy requirements in the process, which would make the
process more economical and environmentally friendly.
To summarize, the use of lipases for transesterification of castor oil combined with
enzyme recycling using centrifugation and membranes for enzyme–product separation has
been investigated.
Optimization of the reaction conditions was done using liquid lipase Eversa Transform
as catalyst. At 35 ∘ C, 6:1 methanol-to-oil molar ratio, 5 wt% water and 5 wt% Eversa Trans-
form solution by weight of castor oil, a FAME content of 94.3% was obtained after eight
hours of reaction. For the transesterification, increase in water addition seemed to compen-
sate for the decrease in the lipase loading. Conversely, high alcohol addition leads to lipase
inhibition, decreasing the FAME yield.
Evaluation of reaction mechanisms and kinetic parameters of the castor oil methanolysis
with 6:1 methanol-to-oil molar ratio, 5 wt% water and 5 wt% Eversa Transform loading,
was performed at 35, 40, and 50 ∘ C. A mechanism was developed that fitted the experimen-
tal data at these temperatures. According to this mechanism, both transesterification and
hydrolysis of TAG into FAME and FFA, respectively, occur simultaneously in the presence
of water. Besides that, all the reactions (transesterification, hydrolysis, esterification, and
formation of enzyme-substrate complex) were found to be reversible.
Process simulation and economical evaluation of biodiesel production were investigated.
From the kinetic parameters obtained in the kinetic evaluation, a process catalyzed by
Eversa Transform was simulated. This process included the steps of transesterification,
alcohol recovery, and product separation and purification. Optimization of the process with
liquid enzyme resulted in a production cost of 0.78 US$/kg, corresponding to a profit of
51.6 million US$/yr.
Operation process costs could be minimized by recovering and reusing the lipases,
increasing the annual profit. Partial reuse of enzymes (50% fresh and 50% reuse enzymes)
showed good FAME yield for three consecutive batches, resulting in a high overall FAME
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 217

content. Ceramic UF membranes combined with centrifugation were used to eliminate

glycerol and methanol from the recycle lipase phase. The content of glycerol and methanol
could be reduced to less than 1 wt% in the lipase solution. Lipases reused after purification
with diafiltration were able to increase the FAME content from 67 to 83% when recycled
Eversa Transform was used for eight-hour batch reactions.
Therefore, castor oil transesterification catalyzed by liquid enzymes seems to be a com-
petitive alternative for biodiesel production, which could replace chemical-catalyzed routes.

Acknowledgments
This project was economically supported by the Brazilian National Council for Scientific
and Technological Development (CNPq) and in collaboration with Novozymes A/S at the
University of Southern Denmark. The GAMS license used was provided by Salamanca
Research, PSEM3 group.

Nomenclature
DAG Diacylglycerol
FAAE Fatty acid alkyl ester
FAME Fatty acid methyl ester
FFA Free fatty acid
HPLC High-performance liquid chromatography
Ki Equilibrium constants for TAG, DAG, MAG and FFA (l/mol)
KMeOH Methanol inhibition constant (mol/l)
MAG Monoacylglycerol
MeOH Methanol
MWCO Molecular weight cut-off
TAG Triacylglycerol
UF Ultrafiltration
Ve, FFA Maximum forward esterification rate constant [l/(mol ⋅ h)]
V−e, FFA Maximum reverse esterification rate constant [l/(mol ⋅ h)]
Vh, i Maximum forward hydrolysis rate constants [l/(mol ⋅ h)]
V−h, i Maximum reverse hydrolysis rate constants [l/(mol ⋅ h)]
Vt, i Maximum forward transesterification rate constants [l/(mol ⋅ h)]
V−t, i Maximum reverse transesterification rate constants [l/(mol ⋅ h)]
W Water
WCO Waste cooking oil

References
Abbaszaadeh, A., Ghobadian, B., Omidkhah, M.R., and Najafi, G. (2012). Current biodiesel production
technologies: a comparative review. Energy Conversion and Management 63: 138–148.
Akoh, C.C., Chang, S.-W., Lee, G.-C., and Shaw, J.-F. (2007). Enzymatic approach to biodiesel production.
Journal of Agricultural and Food Chemistry 55: 8995–9005.
Al-dobouni, I.A., Fadhil, A.B., and Saeed, I.K. (2016). Optimized alkali-catalyzed transesterification of
wild mustard (Brassica juncea L.) seed oil. Energy Sources, Part A: Recovery, Utilization, and Environ-
mental Effects 38: 2319–2325.
218 Process Systems Engineering for Biofuels Development

Al-Zuhair, S. (2007). Review: production of biodiesel: possibilities and challenges. Biofuels, Bioproducts
and Biorefining 1: 57–66.
Amoah, J., Ho, S.H., Hama, S. et al. (2016). Lipase cocktail for efficient conversion of oils containing
phospholipids to biodiesel. Bioresource Technology 211: 224–230.
Anastas, P.T. and Warner, J.C. (1998). Green Chemistry: Theory and Practice. New York, NY: Oxford
University Press.
Andrade, T.A., Errico, M., and Christensen, K.V. (2017a). Influence of the reaction conditions on the
enzyme catalyzed transesterification of castor oil: a possible step in biodiesel production. Bioresource
Technology 243: 366–374.
Andrade, T.A., Errico, M., and Christensen, K.V. (2017b). Transesterification of castor oil catalyzed
by liquid enzymes: optimization of reaction conditions. Computer Aided Chemical Engineering 40:
2863–2868.
Andrade, T.A., Errico, M., and Christensen, K.V. (2017c). Evaluation of reaction mechanisms and kinetic
parameters for the transesterification of castor oil by liquid enzymes. Industrial & Engineering Chemistry
Research 56: 9478–9488.
Andrade, T.A., Errico, M., and Christensen, K.V. (2017d). Castor oil transesterification catalysed by liquid
enzymes: feasibility of reuse under various reaction conditions. Chemical Engineering Transactions 57:
913–918.
Andrade, T.A., Martín, M., Errico, M., and Christensen, K.V. (2019). Biodiesel production catalyzed by
liquid and immobilized enzymes: optimization and economic analysis. Chemical Engineering Research
and Design 141: 1–14.
Balat, M. and Balat, H. (2010). Progress in biodiesel processing. Applied Energy 87: 1815–1835.
Banković-Ilić, I.B., Stamenković, O.S., and Veljković, V.B. (2012). Biodiesel production from non-edible
plant oils. Renewable and Sustainable Energy Reviews 16: 3621–3647.
Baron, A.M., Barouh, N., Barea, B. et al. (2014). Transesterification of castor oil in a solvent-free medium
using the lipase from Burkholderia cepacia LTEB11 immobilized on a hydrophobic support. Fuel 117:
458–462.
Bateni, H., Karimi, K., Zamani, A., and Benakashani, F. (2014). Castor plant for biodiesel, biogas, and
ethanol production with a biorefinery processing perspective. Applied Energy 136: 14–22.
Benfer, S., Árki, P., and Tomandl, G. (2004). Ceramic membranes for filtration applications – preparation
and characterization. Advanced Engineering Materials 6: 495–500.
Berchmans, H.J. and Hirata, S. (2008). Biodiesel production from crude Jatropha curcas L. seed oil with a
high content of free fatty acids. Bioresource Technology 99: 1716–1721.
Berman, P., Nizri, S., and Wiesman, Z. (2011). Castor oil biodiesel and its blends as alternative fuel. Biomass
and Bioenergy 35: 2861–2866.
Cao, P., Dubé, M.A., and Tremblay, A.Y. (2008). High-purity fatty acid methyl ester production from canola,
soybean, palm, and yellow grease lipids by means of a membrane reactor. Biomass and Bioenergy 32:
1028–1036.
Cao, P., Tremblay, A.Y., and Dubé, M.A. (2009). Kinetics of canola oil transesterification in a membrane
reactor. Industrial & Engineering Chemistry Research 48: 2533–2541.
Cheng, L.H., Cheng, Y.F., Yen, S.Y., and Chen, J. (2009). Ultrafiltration of triglyceride from biodiesel using
the phase diagram of oil-FAME-MeOH. Journal of Membrane Science 330: 156–165.
Christopher, L.P., Kumar, H., and Zambare, V.P. (2014). Enzymatic biodiesel: challenges and opportunities.
Applied Energy 119: 497–520.
Chung, K.H. (2010). Transesterification of Camellia japonica and Vernicia fordii seed oils on alkali cata-
lysts for biodiesel production. Journal of Industrial and Engineering Chemistry 16: 506–509.
Deng, X., Fang, Z., and Liu, Y.H. (2010). Ultrasonic transesterification of Jatropha curcas L. oil to biodiesel
by a two-step process. Energy Conversion and Management 51: 2802–2807.
Firdaus, M.Y., Guo, Z., and Fedosov, S.N. (2016). Development of kinetic model for biodiesel production
using liquid lipase as a biocatalyst, esterification step. Biochemical Engineering Journal 105: 52–61.
 Implementation of Biodiesel Production Process Using Enzyme-Catalyzed Routes 219

Fjerbaek, L., Christensen, K.V., and Norddahl, B. (2009). A review of the current state of biodiesel produc-
tion using enzymatic transesterification. Biotechnology and Bioengineering 102: 1298–1315.
Food and Agriculture Organization of the United Nations (2017a). OECD-FAO Agricultural Outlook
2017–2026.
Food and Agriculture Organization of the United Nations (2017b). OECD Agriculture Statistics,
OECD-FAO Agricultural Outlook (Edition 2017).
Granjo, J.F.O., Duarte, B.P.M., and Oliveira, N.M.C. (2017). Integrated production of biodiesel in a soybean
biorefinery: modeling, simulation and economical assessment. Energy 129: 273–291.
Gui, M.M., Lee, K.T., and Bhatia, S. (2008). Feasibility of edible oil vs. non-edible oil vs. waste edible oil
as biodiesel feedstock. Energy 33: 1646–1653.
Guldhe, A., Singh, B., Mutanda, T. et al. (2015). Advances in synthesis of biodiesel via enzyme catalysis:
novel and sustainable approaches. Renewable and Sustainable Energy Reviews 41: 1447–1464.
Haas, M.J., McAloon, A.J., Yee, W.C., and Foglia, T.A. (2006). A processmodel to estimate biodiesel
production costs. Bioresource Technology 97: 671–678.
Hsiao, M.-C., Hou, S.-S., Kuo, J.-Y., and Hsieh, P.H. (2018). Optimized conversion of waste cooking oil to
biodiesel using calcium methoxide as catalyst under homogenizer system conditions. Energies 11 (2622):
1–12.
International Energy Agency (2018). Key World Energy Statistics.
Kumar, R., Tiwari, P., and Garg, S. (2013). Alkali transesterification of linseed oil for biodiesel production.
Fuel 104: 553–560.
Lee, S., Posarac, D., and Ellis, N. (2011). Process simulation and economic analysis of biodiesel produc-
tion processes using fresh and waste vegetable oil and supercritical methanol. Chemical Engineering
Research and Design 89: 2626–2642.
Leung, D.Y.C. and Guo, Y. (2006). Transesterification of neat and used frying oil: optimization for biodiesel
production. Fuel Processing Technology 87: 883–890.
Li, B., Li, B., and Shen, Y. (2009). A novel approach to measure all rate constants in the simplest enzyme
kinetics model. Journal of Mathematical Chemistry 46: 290–301.
Li, Y., Du, W., and Liu, D. (2014). Free lipase-catalyzed biodiesel production from phospholipids-containing
oils. Biomass and Bioenergy 71: 162–169.
Lotero, E., Liu, Y., Lopez, D.E. et al. (2005). Synthesis of biodiesel via acid catalysis. Industrial & Engi-
neering Chemistry Research 44: 5353–5363.
Lotti, M., Pleiss, J., Valero, F., and Ferrer, P. (2015). Effects of methanol on lipases: molecular, kinetic and
process issues in the production of biodiesel. Biotechnology Journal 10: 22–30.
Maleki, E., Aroua, M.K., and Sulaiman, N.M.N. (2013). Castor oil – a more suitable feedstock for enzymatic
production of methyl esters. Fuel Processing Technology 112: 129–132.
Marchetti, J.M., Miguel, V.U., and Errazu, A.F. (2007). Possible methods for biodiesel production. Renew-
able and Sustainable Energy Reviews 11: 1300–1311.
Meher, L.C., Vidya, S.D., and Naik, S.N. (2006). Technical aspects of biodiesel production by transesteri-
fication – a review. Renewable and Sustainable Energy Reviews 10: 248–268.
Mittelbach, M. and Remschmidt, C. (2006). Biodiesel: The Comprehensive Handbook, 3e. Vienna: Boerse-
druck Ges.m.b.H.
de Morais Coutinho, C., Chiu, M.C., Basso, R.C. et al. (2009). State of art of the application of membrane
technology to vegetable oils: a review. Food Research International 42: 536–550.
Nasaruddin, R.R., Alam, M.Z., and Jami, M.S. (2014). Evaluation of solvent system for the enzymatic
synthesis of ethanol-based biodiesel from sludge palm oil (SPO). Bioresource Technology 154:
155–161.
Nasir, N.F., Daud, W.R.W., Kamarudin, S.K., and Yaakob, Z. (2013). Process system engineering in
biodiesel production: a review. Renewable and Sustainable Energy Reviews 22: 631–639.
Nielsen, P.M., Brask, J., and Fjerbaek, L. (2008). Enzymatic biodiesel production: technical and economical
considerations. European Journal of Lipid Science and Technology 110: 692–700.
220 Process Systems Engineering for Biofuels Development

Nordblad, M., Silva, V.T.L., Nielsen, P.M., and Woodley, J.M. (2014). Identification of critical parameters
in liquid enzyme-catalyzed biodiesel production. Biotechnology and Bioengineering 111: 2446–2453.
de Oliveira, D., Di Luccio, M., Faccio, C. et al. (2004). Optimization of enzymatic production of biodiesel
from castor oil in organic solvent medium. Applied Biochemistry and Biotechnology 113–116: 771–780.
Padaki, M., Surya, M.R., Abdullah, M.S. et al. (2015). Membrane technology enhancement in oil-water
separation. A review. Desalination 357: 197–207.
Parawira, W. (2009). Biotechnological production of biodiesel fuel using biocatalysed transesterification:
a review. Critical Reviews in Biotechnology 29: 82–93.
Patel, R.L. and Sankhavara, C.D. (2017). Biodiesel production from Karanja oil and its use in diesel engine:
a review. Renewable and Sustainable Energy Reviews 71: 464–474.
Poppe, J.K., Fernandez-Lafuente, R., Rodrigues, R.C., and Ayub, M.A.Z. (2015). Enzymatic reactors for
biodiesel synthesis: present status and future prospects. Biotechnology Advances 33: 511–525.
Renewable Energy Policy Network (2017). Renewables 2017 Global Status Report.
Roda-Serrat, M.C. (2017). Natural blue food colour: natural food colorant production and membrane sep-
aration. PhD thesis. University of Southern Denmark.
Rodrigues, R.C., Volpato, G., Wada, K., and Ayub, M.A.Z. (2008). Enzymatic synthesis of biodiesel from
transesterification reactions of vegetable oils and short chain alcohols. Journal of the American Oil
Chemists’ Society 85: 925–930.
Santana, G.C.S., Martins, P.F., de Lima da Silva, N. et al. (2010). Simulation and cost estimate for biodiesel
production using castor oil. Chemical Engineering Research and Design 88: 626–632.
Sharma, A., Jelemensky, M., Paulen, R., and Fikar, M. (2017). Modeling and optimal operation of batch
closed-loop diafiltration processes. Chemical Engineering Research and Design 122: 198–210.
Sinnott, R.K. (2005). Coulson & Richardson’s Chemical EngineeringSeries – Chemical Engineering
Design, 4e, vol. 6. Oxford: Elsevier Butterworth-Heinemann.
Sotoft, L.F., Rong, B.G., Christensen, K.V., and Norddahl, B. (2010). Process simulation and economical
evaluation of enzymatic biodiesel production plant. Bioresource Technology 101: 5266–5274.
Srivastava, A. and Prasad, R. (2000). Triglycerides-based diesel fuels. Renewable and Sustainable Energy
Reviews 4 (2): 111–133.
Tan, T., Lu, J., Nie, K. et al. (2010). Biodiesel production with immobilized lipase: a review. Biotechnology
Advances 28: 628–634.
U.S. Energy Information Administration (2017). International Energy Outlook 2017.
Wang, Y., Wang, X., Liu, Y. et al. (2009). Refining of biodiesel by ceramic membrane separation. Fuel
Processing Technology 90: 422–427.
West, A.H., Posarac, D., and Ellis, N. (2008). Assessment of four biodiesel production processes using
HYSYS.Plant. Bioresource Technology 99: 6587–6601.
World Energy Council (2016). World Energy Resources Bioenergy 2016.
Yang, J.-S., Jeon, G.-J., Hur, B.-K., and Yang, J.-W. (2005). Enzymatic methanolysis of castor oil for the
synthesis of methyl ricinoleate in a solvent-free medium. Journal of Microbiology and Biotechnology
15: 1183–1188.
Yuan, H., Yang, B., Zhang, H., and Zhou, X. (2011). Synthesis of biodiesel using castor oil under microwave
radiation. International Journal of Chemical Reactor Engineering 9: 1–11.
Zhang, Y., Dubé, M.A., McLean, D.D., and Kates, M. (2003a). Biodiesel production from waste cooking
oil: 1. Process design and technological assessment. Bioresource Technology 89: 1–16.
Zhang, Y., Dubé, M.A., McLean, D.D., and Kates, M. (2003b). Biodiesel production from waste cooking
oil: 2. Economic assessment and sensitivity analysis. Bioresource Technology 90: 229–240.
Zheng, S., Kates, M., Dubé, M.A., and McLean, D.D. (2006). Acid-catalyzed production of biodiesel from
waste frying oil. Biomass and Bioenergy 30: 267–272.
 9
Process Analysis of Biodiesel
Production – Kinetic Modeling,
Simulation, and Process Design
Bruna Ricetti Margarida, Wanderson Rogerio Giacomin-Junior, Luiz
Fernando de Lima Luz Junior, Fernando Augusto Pedersen Voll, and
Marcos Lucio Corazza
Department of Chemical Engineering, Federal University of Paraná, Polytechnic Center (DEQ/UFPR),
Curitiba 81531-980, Brazil

9.1 Introduction
Biodiesel is a biofuel produced from alkyl esters of long-chain carboxylic acids that can be
obtained from vegetable oils or animal fat. Different routes have been studied and developed
to produce biodiesel from waste oils, reducing raw material costs and contributing to the
generation of biodiesel without the use of edible oils. Since fuels like gasoline and diesel
are derived from petroleum, there are uncertainties about their market in the future and
resource availability. Biodiesel, in turn, can be obtained from renewable resources, which
might contribute to its market stability and it can be considered environmentally friendly
compared with petroleum-based fuels. According to ANP (Agência Nacional do Petróleo,
Gás Natural e Biocombustíveis – the Brazilian agency for fuels and biofuels regulation),
biodiesel should have at least 96.5% (mass) of ester content, a total glycerol mass content
up to 0.25%, and maximum water content of 200 ppm.

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
222 Process Systems Engineering for Biofuels Development

Different possible routes have been proposed for producing biodiesel, as homogeneous-
and heterogeneous-based reactions, enzyme-catalyzed reaction, and supercritical or
non-catalyzed routes, which normally use a short-chain alcohol at supercritical conditions.
The most traditional way to produce it is using an alkaline homogeneous transesterification
with methanol that runs at low temperature (approximately 333 K) and with short reaction
time (approximately 30 minutes). Transesterification (also known as alcoholysis) is usually
the main reaction behind the conventional biodiesel production process. It is defined in
organic chemistry as a reversible reaction in which a carboxylic (or fatty) acid is converted
into an ester of either carboxylic acid or fatty acid by exchanging the organic group of an
ester with the organic group of an alcohol. But what is used to make biodiesel? The most
common materials used are vegetable oils, animal fats, and waste cooking oils. These
materials are widely used in conventional biodiesel production because of their economic
and environmental aspect, as they are composed of triacylglycerol molecules, which are
esters consisting of three fatty acid units linked to a glycerol backbone.
This chapter presents a brief discussion of four main routes for biodiesel production.
The first topic is focused on the thermodynamic analysis and corrections from the Aspen
Plus data bank for pure compounds and mixture parameters, including the input of new
molecules, comparison of parameters between the literature and the standard databank
present in Aspen Plus, and also how one can correct, calculate and estimate properties in
the Aspen Plus framework. The liquid heat capacity is used as a case study for the pure
compounds, and four cases are selected for the mixtures, including phase equilibrium of
binary and ternary mixtures, density calculation, and excess molar enthalpy.
In order to provide a better overview about reactions related to biodiesel processing, both
esterification and transesterification reactions are discussed, as well the use of methanol and
ethanol in both cases. Besides the traditional route, an overview of the non-catalyzed super-
critical route is provided, and a brief explanation of how to regress experimental reaction
data using Aspen Plus. Therefore, this chapter presents a guide to simulate the mostly used
biodiesel production process and some of the Aspen Plus tools and blocks available. Once
all the preparation and project design are studied, the next topics cover some proposed ways
to run simulation and how to improve the results by using both the Aspen Energy Analyzer
and Economic Analyzer.
The goal of this chapter is to guide the reader to better understand the entire process of
biodiesel processing using Aspen Plus as a simulation tool in a more technical and didactic
way. It also compares the information presented in the literature and some of the bottlenecks
and risks involving the incorrect use of some process simulation tools for unconventional
fatty acid-based molecules and systems.

9.1.1 Homogeneous-Based Reactions

In biodiesel production, the use of homogeneous-based reactions indicates that the catalyst
is in the liquid phase, in which usually either a strong acid or an alkali is used to accelerate
the conversion rate. Also, in this type of reaction, excess alcohol must be used to improve the
mixing degree between the reactants and to shift the equilibrium to the product side. A few
steps that are present to produce biodiesel by alcoholysis are: first, it is necessary to combine
the alcohol (typically, methanol or ethanol) with the catalyst, which should be as dry as
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 223

possible since water increases soap formation (Javidialesaadi and Raeissi 2013). It is worth
mentioning that for the alkali-catalyzed reactions, the catalyst (alkoxy) is obtained from
the base and the alcohol used for biodiesel reaction. Secondly, with the catalyst prepared,
the reaction takes place with three moles of alcohol and one mole of triacylglycerol. The
choice of the catalyst is an important reaction step: a reaction with an alkaline catalyst
lasts about 30 minutes, and with an acid catalyst may last up to 48–96 hours, according to
Pisarello et al. (2010) and Schuchardt et al. (1998). In this set of reactions, there are a few
parameters that are also important to control, such as temperature, time, and pressure. This
will be discussed later in this chapter.
It is important to mention that in considering low-quality raw materials such as crude
vegetable oils and waste cooking oils for biodiesel production, attention must be given to
their high content of free fatty acids (FFA), which may act as poison for the catalyst or
lead to saponification during the transesterification. In such cases a pretreatment of the raw
material is required. The most common way to handle raw materials with high levels of
FFA is to separate them using a unit operation such as vacuum distillation. Alternatively,
it is possible to react FFA with a short-chain alcohol in the presence of an acid catalyst.
This reaction is known as esterification, and it allows to enhance the biodiesel (a mixture
of alkyl esters) yield in the process instead of separating the FFA from the raw material.
The acid transesterification (corresponding to the use of acid catalyst) occurs along with
acid esterification of FFA; however, the acid esterification runs faster than the transester-
ification. In addition, during the esterification reaction the formation of water may lead
to the formation of undesirable by-products (Murad et al. 2017), decreasing the reaction
yield.

9.1.2 Heterogeneous-Based Reactions

Heterogeneous catalysis has some advantages when compared with traditional homoge-
neous transesterification because it demands fewer unit operations for biodiesel purification
in the downstream steps with easier ways to separate and catalyst reuse is possible, reduc-
ing wastewater and contributing to a decrease in the process costs (Choudary et al. 2000).
It is also relevant to say that the use of heterogeneous acid catalysts prevents soap forma-
tion, undesirable in the homogeneous base-catalyzed transesterification (Vyas et al. 2010).
Such catalysts might also be modified to be more resistant to the acid present in the reac-
tant oil stream or the water formed during the reaction (Ramos et al. 2017). Even though
the process using heterogeneous catalysts may seem better at first look, this route is con-
siderably slower, requiring severe reaction conditions, and loss of catalyst activity needs
to be considered, which also contributes to increasing the operational costs (Schuchardt
et al. 1996).
In general, these catalysts can be divided into two groups: acid and base. The former
is indifferent to the presence of FFA, avoiding the soap formation that occurs when alkali
catalysts are used (Helwani et al. 2009). On the other hand, reaction using acid catalysts is
slower compared with the alkali-catalyzed ones (Sampaio 2008). Therefore, a careful eco-
nomic analysis is recommended when designing a biodiesel plant and choosing the reaction
approach. Both acid- and alkali-catalyzed approaches present pros and cons in terms of
technical feasibility.
224 Process Systems Engineering for Biofuels Development

9.1.3 Enzyme-Catalyzed Reactions

The interest in enzyme-catalyzed synthesis of biodiesel has increased, especially the use of
lipases that can catalyze with high selectivity reactions of esterification, interesterification,
acidolysis, aminolysis, hydrolysis, and alkanolysis (Sharma and Kanwar 2014). One of the
advantages of using enzymes is that mild temperatures are required (300–315 K); there is
also easier glycerol recovery, as pointed out by Brandão et al. (2006), and fatty acid and
triglycerides conversion to esters is in one stage, as cited by Meher et al. (2006).These cat-
alysts also have high specificity and less reaction stages (Kawashiro et al. 1960). Ramos
et al. (2017) stated that with such technology oils can be used without any pretreatment for
reducing the moisture and FFA content, which implies reducing raw material preparation
costs. Although the use of lipases and other types of enzymes in the biodiesel synthesis
process has significant advantages, there are still many limitations regarding their price and
availability, making it difficult in large-scale processes (Ivana et al. 2016). Another impor-
tant drawback is that the reaction time is much higher when compared with other processes
(from 48 to 72 hours), even though it depends on the enzyme and operating conditions
(Schuchardt et al. 1998).

9.1.4 Supercritical Route Reactions

Supercritical technology is a trendy topic of research, and recent works have shown positive
results. Carvalho dos Santos et al. (2018) have described in their work that the supercritical
route allows simultaneous transesterification of triacylglycerols and esterification of FFA.
Even with high temperature and pressure (above the critical point of the alcohol), it has
received increasing interest as it allows a simple and high yield process. However, there are
some concerns about this method related to decomposition and polymerization of unsatu-
rated fatty acids or even isomerization at very high temperature conditions (Imahara et al.
2008; Carvalho dos Santos et al. 2018).

9.1.5 Methanol or Ethanol for Biodiesel Synthesis

Methanol use for biodiesel production is widespread, as the reaction time is less than half
of that when ethanol is used. Besides, methanol is easier to separate in downstream stages
and it is cheaper than ethanol (Coelho 2011). Even though the use of methanol brings many
advantages to the process, it is derived from petroleum, much more toxic than ethanol, and
there is a higher risk of fires, as this compound is more volatile, and its flame is invisible.
On the other hand, ethanol can produce biofuel with higher cetane number and lubricity,
it is less toxic, and generates 100% renewable fuel when produced from biomass (Murad
et al. 2017).

9.2 Getting Started with Aspen Plus V10

Aspen Plus is a largely used simulation tool, which can calculate many compound prop-
erties, also enabling operational modeling and simulation, besides reaction, safety, and
economic analysis. The software has different data banks separated into groups with dozens
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 225

of compounds for equilibrium data, mixture, and other property analysis and estimation. It
is also possible to evaluate and design a vast number of unit operations in the chemical
industry with details such as sizing, materials, and price involved. This tool can be used in
many cases, and some of them will be presented in this chapter.
This chapter is intended for readers that already have some previous knowledge on the
basic use of Aspen Plus as a simulation tool. Aspen Plus provides online material to help
new users to have first contact with the working environment and its tools. For the best
experience, reading some of the tutorials supplied by AspenTech (such as Getting Started
Building and Running a Process Model, and Getting Started Modeling Processes with Elec-
trolytes) is recommended (Al-Malah 2016).

9.2.1 Pure Compounds

The first step in every simulation is to choose the components and set the property methods.
Even accounting for the massive data bank in Aspen Plus, some molecules do not have
their structures and parameters included. Therefore, it is necessary to add them manually
or assume that the use of incorrect parameters may lead to wrong conclusions.
Moreover, with Aspen Plus, it is possible to draw or import the structure of a new
molecule. As an example, let us consider the methyl oleate compound. Under Components
> Specification > Selection, there is an option called “User Defined.” This option allows
the insertion of a new compound. All that is needed is to write the desired compound’s
name and follow the steps. All property information available should be inserted in Aspen
Plus. After drawing the compound, the simulator will ask if the user wants to search for
the molecule in the NIST data bank or estimate it with Aspen Plus standard correlations.
Therefore, there are three different sources of information provided by Aspen Plus: search
the parameters with the NIST extension, use the default in the Aspen library, or estimate
them with the available standard correlation. It is possible to edit and correct the data input
information. These options allow some flexibility, but can lead to a problem: Is the Aspen
Plus standard library reliable? The answer is “sometimes.”
Table 9.1 presents three columns with data (thermodynamic properties) from different
sources. There are some differences between the experimental values and those in the
Aspen® data bank, and these differences may lead to a wrong conclusion in the simula-
tion or an under design as well. Thus, it is always important to pay attention to which
values are being used for simulations. For an estimation, Aspen uses as default the Joback

Table 9.1 Parameter comparison for methyl oleate.

Property Aspen® estimation Aspen® standard Experimental

ΔHf (kJ/mol) −622.4 −626.0 −720.1∗
ΔGf (kJ/mol) −121.0 −117.0 −117.0∗∗
Tc (K) 768.0 764.0 777.0∗∗∗
Pc (kPa) 1122 1280 1200∗∗∗
w 0.164 1.049 0.919∗∗∗
∗ Brands et al. (2002).
∗∗ Bucalá et al. (2006).
∗∗∗ NIST databank (n.d.).
226 Process Systems Engineering for Biofuels Development

method for the Gibbs free energy of formation, critical temperature, and critical pressure,
and the Benson method for the heat of formation, and the definition of acentric factor to
calculate it. The Aspen standard is defined as the values from the Aspen data bank without
any modification after selecting a molecule.
By comparing some properties values presented in the Aspen Plus data bank with values
available in the literature, some contradictions can be noticed. In the “Review” section, the
user has access to a table of properties. Figure 9.1 shows the values of some properties that
are present in the software and those disagree with values presented in the literature.
It is essential to identify that DGFORM, DHFORM, OMEGA, PC, and TC correspond
to the Gibbs free energy of formation, the enthalpy of formation, the acentric factor, crit-
ical pressure, and critical temperature, respectively. Comparing these values to the ones
presented by Bucalá et al. (2006) and Brands et al. (2002) and others predicted with the
Benson, Constantinou and Gani, Ambrose, and Lee-Kesler methods (Poling et al. 2001), it
is possible to identify differences among the values reported for some of these properties,
which are then corrected in Figure 9.2. For the enthalpy, Gibbs energy, critical pressure and
temperature, and acentric factor, the use of these properties as they are set in the Aspen Plus
data bank may bring significant errors to the simulation results. Therefore, if necessary, it
is possible to modify any component property to the values desired by following the steps
shown above. The following example is shown to exemplify this.

Figure 9.1 Aspen Plus table of properties for methyl oleate (METHY-01), ethyl oleate (ETHYL-01), oleic
acid (OLEIC-01), methyl palmitate (METHY-02), and palmitic acid (N-HEX-01).
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 227

Figure 9.2 Aspen Plus table of corrected properties for methyl oleate (METHY-01), methyl oleate
(ETHYL-01), oleic acid (AC-OLEIC), methyl palmitate (METHY-02), and palmitic acid (AC-PALMIT).

9.2.1.1 Verifying and Calculating the Heat Capacity Values: An Example

As seen before, some properties must be revised, so the simulation can run with corrected
data and generate more reliable results. For this reason, it is crucial to verify another impor-
tant component property, the heat capacity for gas and liquid phases. The software itself
presents the coefficients for the calculation of the ideal gas heat capacity for the compo-
nents shown before, but for the liquid heat capacity the values are missing, indicating that
the user should input them manually.
The coefficients acquired for the ideal gas heat capacity come from the DIPPR equation
107 (NIST n.d.), as shown in Eq. (9.1).
( C3∕T
)2 ( C5∕T
)2
ig
CP = C1 + C2 + C4 (9.1)
sinh ( C3∕T ) cosh ( C5∕T )

The values for the coefficients of each component are presented in Figure 9.3.
Coefficients 1–5 correspond to Eq. (9.1), and the other two correspond to the temperature
range of the equation in Kelvin.
228 Process Systems Engineering for Biofuels Development

Figure 9.3 Coefficients for the calculation of the heat capacity.

For liquid heat capacity, a few more steps are necessary to include the experimental
parameters for regression. The liquid heat capacity values were obtained from Pauly et al.
(2014). So, first, go to Methods > Selected Methods > Routes sheet and, on Subordinate
Property, change the DHL route to DHL09, as it will change how the liquid heat capacity is
calculated. The default mode (DHL00) uses ideal gas enthalpy and the heat of vaporization
values for the liquid heat capacity, but once changed to DHL09, it will use the DIPPR
equation for calculation, based on the experimental data. It is important to say that the
parameter name of the DIPPR equation is CPLDIP, and its base equation is (Eq. (9.2)):

CPl = C1 + C2 T + C4 T 2 + C4 T 3 + C5 T 4 (9.2)

Remember that columns 6 and 7 represent the temperature range of the equation tested.
Therefore, if experimental data are available or an appropriate equation to estimate a spe-
cific parameter, it is possible to insert them on the software and calculate the missing
elements.
The next step is to insert the data itself. For this, go to the Data file and choose the liq-
uid heat capacity property for the desired component, in this case methyl oleate (as the
example considered). Thus, insert the data for each condition and go to the Estimation
mode. Select the “Estimate all missing parameters” option and, on the T-dependent file, the
CPL property must be chosen, determining the component, the Data method, as the exper-
imental data considered for regression, and the temperature range for the analysis. Finally,
run the program and go back to the Analysis mode. Then, a new folder is created with the
CPLDIP-1 label with the regressed parameters in it. Later, it is possible to verify a value
or the parameters’ behavior by clicking on the Pure button in the analysis sheet. Therefore,
choose the CP property, check the liquid option and select the component. After running
the analysis, Aspen will show a graph and a table with the property values. The differ-
ence between the values associated with the regression can be visualized and compared in
Table 9.2.
These new values for the liquid heat capacity are slightly different from the original ones
calculated with the ideal gas enthalpy and the heat of vaporization, giving an error of up to
6.5%, which could impact the estimation of other related properties and simulations. For
this reason, it is essential to compare the Aspen Plus property values with available experi-
mental ones for more accurate estimation and simulation. The example above illustrates the
drawbacks that the user can face when working with and considering compounds related to
biodiesel processing, such as fatty acids, fatty acid alkyl esters, and acylglycerols. For these
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 229

Table 9.2 Comparison between experimental and regressed values for liquid heat
capacity using Aspen Plus.

Temperature (K) Experimental data (kJ/kmol K) Regressed data (kJ/kmol K)

270 564.63 564.63
280 572.02 572.02
290 579.76 579.76
300 587.84 587.84
310 596.27 596.27
320 605.05 605.05
330 614.18 614.18
340 623.65 623.65
350 633.47 633.47
360 643.64 643.64
370 654.15 654.15
380 665.01 665.01
390 676.22 676.22

types of compounds, a careful analysis of pure compounds must be performed to make sure
that the simulator is set and using reliable property values.

9.2.2 Mixture Parameters

Mixtures are an essential topic in any simulation. One needs to know exactly how the com-
pounds interact with each other and therefore their behavior in a mixture or solution. In
order to illustrate the importance of setting the correct parameters, different properties of
example systems will be compared.

9.2.2.1 Phase Equilibrium: Binary System Oleic Acid (1) + Ethanol (2) as a Case Study
For this system, three different thermodynamic model options, and one set of experimen-
tal data of vapor–liquid equilibrium (VLE) are considered: the following thermodynamic
models of UNIFAC, UNIQUAC using UNIFAC model to estimate, and UNIQUAC using
experimental data regression are evaluated comparing with the VLE data experimental pre-
sented by Eduljee and Boyes (1981), considering the temperature of 318.14 K. The results
are shown in Table 9.3.
Since UNIQUAC and UNIFAC are using the same binary interaction parameters, the
results must be quite similar (for the phase equilibrium calculation). The mean deviation
was 6.3% for UNIQUAC and 4.8% for UNIFAC; also, UNIFAC was slightly better and
had a smaller maximum relative error. For this particular system, the maximum error was
small, but this is not standard behavior. Taking, for example, oleic acid (1) + methanol (2)
VLE data, the magnitude of the error comparing the calculated values with the pressure
experimental data, pressure increased from around 12 to 30%, for the UNIFAC predictions.
Therefore, it is essential to know how well the parameters and thermodynamic model can
describe the system.
Aspen Plus allows the use of experimental data to regress binary interaction parameters
for the thermodynamic models and mixing rules available. For this, it is necessary to insert
230 Process Systems Engineering for Biofuels Development

Table 9.3 Comparison between VLE pressure data from oleic acid (1) + ethanol
(2) using different parameter estimators at 318.138 K.

x1 Relative error (%)

P (kPa) UNIFAC UNIQUAC∗ UNIQUAC∗∗
0.000 23.00 −0.3 −0.3 −0.3
0.064 21.67 −1.4 −1.3 0.4
0.186 19.68 −4.1 −3.7 2.0
0.300 17.89 −4.9 −4.2 2.2
0.385 16.35 −5.0 −3.9 1.3
0.507 13.95 −3.4 −1.5 0.3
0.624 11.21 −1.6 1.2 −1.3
0.714 8.96 1.3 4.8 −1.3
0.805 6.49 5.5 9.7 ∼ 0.0
0.869 4.52 8.0 12.6 0.4
0.917 2.96 10.5 15.4 1.5
0.948 1.89 11.8 16.9 1.9
∗ UNIQUAC using UNIFAC as estimator.
∗∗ UNIQUAC using experimental data.

the data set in the data folder and change the run mode to regression. In this case, the
regression will estimate the binary interaction parameters for the UNIQUAC model and
use ordinary least squares (OLS) as the objective function. The Aspen default objective
function is maximum likelihood; however, it is possible to get better results by changing that
for OLS. With the experimental data regression, it is always expected that the model will
present better results. Considering the last example above and regressing the UNIQUAC
parameters, the mean error in pressure changed from 4.8 to 1.1%, and the maximum error
observed decreased from 16.9 to 2.2%.

9.2.2.2 Excess Molar Enthalpy: Ethyl Oleate (1) + Ethanol (2) Mixture
The experimental data of excess molar enthalpy was obtained by Aissa et al. (2017) for a
temperature of 310.14 K, and for the regression, the OLS was used to estimate the binary
interaction parameters for the UNIQUAC model. Table 9.4 presents the results obtained,
considering this example of prediction versus regression. Regarding the maximum error,
it is possible to highlight that the OLS method still presented the smallest relative max-
imum error; however, between UNIQUAC using UNIFAC as estimator and UNIQUAC
regressed to the experimental data, the best parameters cannot be chosen immediately; this
will depend on the molar fraction of ethyl oleate that is used. It is worth mentioning that
UNIFAC presented the worst results.

9.2.2.3 Biodiesel Density: Methyl Oleate, Methyl Palmitate, and Methyl Linoleate
In order to estimate the biodiesel density, it is necessary to determine how many compounds
will be considered, since the biodiesel is a multicomponent mixture of different fatty acid
alkyl esters. First, biodiesel is known as a mixture of methyl esters or ethyl esters and some
of these esters can be found in higher concentration than others, depending on the raw
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 231

Table 9.4 Comparison of excess molar enthalpy of ethyl oleate (1) + ethanol
(2) using different parameter estimators at 310.14 K.

Relative error (%)

x1 ΔHex (kJ/mol) UNIFAC UNIQUAC∗ UNIQUAC∗∗
0.0912 0.588 23.6 9.3 4.8
0.1430 0.896 27.5 9.0 3.4
0.2011 1.139 26.6 4.4 2.9
0.3021 1.435 25.0 −0.5 10.8
0.4194 1.618 24.2 −1.9 15.4
0.5094 1.680 25.6 0.8 14.6
0.5427 1.677 26.1 2.0 14.1
0.6335 1.659 30.3 9.1 7.8
0.7343 1.524 36.3 18.9 1.7
0.7817 1.406 39.6 24.0 7.1
0.8828 0.995 48.2 36.6 21.0
∗ UNIQUAC using UNIFAC as estimator.
∗∗ UNIQUAC using experimental data.

material originating the biodiesel (different vegetable oils and animal fats present different
fatty acid profile). For convenience, in this work, the biodiesel will be considered biodiesel
from palm oil and methyl oleate, methyl palmitate, and methyl linoleate as the representa-
tive compounds. These three compounds together represent 94 wt% of the biodiesel from
palm oil: 0.4245 methyl palmitate, 0.4192 methyl oleate, and 0.098 methyl linoleate, in
mass fraction. These data were obtained by Pratas et al. (2011).
With the experimental data, the binary parameters can be calculated from the Rackett
liquid molar volume model. This parameter can be calculated for all the binary groups,
but since there are two compounds that represent 84% of our sample, a two-compound
approach should be chosen. Table 9.5 shows the different results for both approaches (two
and three compounds). There is not much gain in choosing a three-compound estimation;
in this case, after the estimation, even calculating Rackett without any experimental data,
the error dropped only 0.2%. In general, the binary regression data had almost no effect on
the system.

9.2.2.4 Ternary Mixtures (LLE): Methyl Oleate (1) + Glycerol (2) + Methanol (3)
The first step is to insert the data in Aspen Plus. The ternary data used in this topic were
obtained by Andreatta et al. (2008), and the temperature of 333 K was chosen for the binary
interaction parameter estimation.
The regression was run with the default method of Aspen Plus. Figure 9.4 depicts a
comparison among the experimental data and calculated values using the thermodynamic
model. Figure 9.4d shows the experimental data, and Figure 9.4c the regressed model
(UNIQUAC). Note that UNIFAC (Figure 9.4b) and UNIQUAC using UNIFAC parameters
(Figure 9.4a) presented almost the same results and did not represent the ternary system
well. However, after the regression, the system could be better predicted, mainly with con-
ditions at high concentrations of methanol.
232 Process Systems Engineering for Biofuels Development

Table 9.5 Density estimation for biodiesel at different temperatures (288.15–363.15 K).

Relative error (%)

Two compounds Three compounds
Temperature (K) Density(kg/m3 ) Calculated Regressed∗ Calculated Regressed∗
288.15 877.9 0.7 1.2 0.5 1.2
293.15 874.1 0.7 1.1 0.4 1.1
298.15 870.4 0.6 0.9 0.4 0.9
303.15 866.7 0.6 0.8 0.4 0.8
308.15 863.0 0.6 0.6 0.3 0.6
313.15 859.4 0.6 0.5 0.3 0.5
318.15 855.7 0.5 0.3 0.3 0.3
323.15 852.1 0.5 0.2 0.3 0.2
328.15 848.5 0.5 0.0 0.3 0.0
333.15 844.9 0.5 −0.1 0.2 −0.1
338.15 841.2 0.5 −0.2 0.2 −0.2
343.15 831.6 −0.3 −1.1 −0.5 −1.1
348.15 834.0 0.4 −0.5 0.2 −0.5
353.15 830.4 0.4 −0.7 0.2 −0.7
358.15 826.8 0.4 −0.8 0.2 −0.8
363.15 823.2 0.4 −1.0 0.2 −1.0
∗ Rackett liquid molar volume model using experimental data.

9.3 Kinetic Study

Correct representation of the reaction kinetics is a key factor for the analysis, simulation,
optimization, and mainly designing a chemical plant. For the biodiesel case, as mentioned
before, the reactions involved are the acylglycerol transesterification and fatty acid esteri-
fication with short-chain alcohols. Therefore, in this section, we will discuss some aspects
of using the Aspen Plus simulator to deal with these reactions.

9.3.1 Esterification Reaction

As this chapter focuses on biodiesel production from waste oil, the presence of FFA in
a considerable amount is common. As previously discussed, the direct transesterification
reaction cannot be used for raw material with high acidity (over 0.5% as stated by Clark
et al. [1981] or above 1% as claimed by Canakci and Van Gerpen [1999]). For this reason,
a preliminary step must be performed. The esterification consists of a reaction to transform
the FFA into esters, also reducing the solution acidity (FFA content), enabling further oil
transesterification. Thus, as stated before, the esterification consists of a reaction between
one mole of an organic acid with one mole of alcohol, forming one mole of ester and one
mole of water. The general esterification reaction equation is represented in Eq. (9.3).

rAGL = −k1 [AGL] ⋅ [ROH] + k2 [Est] ⋅ [H2 O] (9.3)

In the experiments carried out by Murad et al. (2017), ethanol was used as the alcohol
for esterification and it was suggested that for the correct kinetic modeling the catalyst
to ethanol mass ratio must be inserted in the model. This equation complementation is
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 233

(a) x2 (b) x2
0.05 0.95 0.05 0.95
0.15 0.85 0.15 0.85
0.25 0.75 0.25 0.75
0.35 0.65 0.35 0.65
0.45 0.55 0.45 0.55
0.55 0.45 0.55 0.45
0.65 0.35 0.65 0.35
0.75 0.25 0.75 0.25
0.85 0.15 0.85 0.15
0.95 0.05 0.95 0.05
x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1 x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1
x2 x2
(c) 0.05 0.95 (d) 0.05 0.95
0.15 0.85 0.15 0.85
0.25 0.75 0.25 0.75
0.35 0.65 0.35 0.65
0.45 0.55 0.45 0.55
0.55 0.45 0.55 0.45
0.65 0.35 0.65 0.35
0.75 0.25 0.75 0.25
0.85 0.15 0.85 0.15
0.95 0.05 0.95 0.05
x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1 x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1

Figure 9.4 Comparison between thermodynamic models/parameters of the ternary system methyl oleate
(1), methanol (2), and glycerol (3) at 333 K using Aspen Plus for (a) UNIQUAC using UNIFAC parameters,
(b) UNIFAC, (c) UNIQUAC with regressed parameters, and (d) experimental data.

performed by adding a parameter in the kinetic constants for the forward and backward
reactions (k1 and k2 ), as shown in Eqs. (9.4) and (9.5).
( E )
− RT1
k1 = a1 ⋅ e (9.4)
( E )
− RT2
k2 = a2 ⋅ e (9.5)
where E1 and E2 are the activation energy, and a1 and a2 the model parameters that depend
on the catalyst to ethanol mass ratio, as shown in Eqs. (9.6) and (9.7).
( )
mH2SO4
a1 = A1 ⋅ (9.6)
m
( EtOH )
mH2SO4
a2 = A2 ⋅ (9.7)
mEtOH
In Eqs. (9.6) and (9.7), A1 and A2 are the adjusted constants by regression. The reaction
constants obtained by Murad et al. (2017) using 0.33 wt% of H2 SO4 as the catalyst, are
presented in Table 9.6.
234 Process Systems Engineering for Biofuels Development

Table 9.6 Kinetic parameters for the

esterification reaction with ethanol.

Parameter Value
A1 (l/mol/s) 9.1 × 107
A2 (l/mol/s) 3.9 × 104
E1 /R (K) 6.56 × 103
E2 /R (K) 3.87 × 103

Source: Murad et al. 2017. Reproduced with

permission of Springer.

Table 9.7 Kinetic parameters for

esterification with methanol at 333.15 K.

Parameter Value
a1 2.869 × 106
a2 37.068
E1 (J/mol) 50 745.2
E2 (J/mol) 31 007.3

Source: Berrios et al. 2007. Reproduced with

permission of Elsevier.

Similarly, a kinetic study using methanol was presented by Berrios et al. (2007). This
time, a catalyst concentration of 5 wt% was used. The parameters obtained by their research
are presented in Table 9.7, where a1 and a2 are the frequency factors.
These parameters will be used in the simulation for the biodiesel production process in
Section 9.3.2.

9.3.2 Experimental Reaction Data Regression

Although some Aspen Plus estimations are reliable, experimental data may improve results.
It is possible to adjust many properties in the Aspen Plus data bank by changing the calcu-
lation parameters. For the reaction properties, for example, it is possible to regress experi-
mental data to adjust an adequate equation describing the reaction behavior.
The first step is to create the reactor block in the simulation and its inlet and outlet
streams. First, if the data were obtained in a batch reactor, use the Rbatch model. Once fin-
ished, go to Model Analysis Tool > Data Fit > Data Set files and create a new Profile-Data.
In this new folder, insert the reactor model and name it. In the Measured block variables
space, choose a variable name and enter what this variable will be, in this case, to obtain
the liquid molar fraction of lauric acid (representing our FFA) through time, choose the
MOLEFRAC-L variable and the lauric acid component. Next, go to the Data file and insert
the obtained data (Table 9.8) and the desired standard deviation value.
Once finished, click on the “Initial conditions” file and add the working temperature
(here consider an isothermal reaction), pressure, and feed composition. In this case study,
the conditions were 343.15 K and 1 bar, and the inlet flow was 118.486 kg/h of ethanol,
11.6102 kg/h of water, 42.882 kg/h of lauric acid, and 0 kg/h of ethyl laurate.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 235

Table 9.8 Experimental data for the lauric acid

esterification with ethanol.

Time (minutes) Lauric acid mol fraction

0 0.0678
15 0.0610
30 0.0533
60 0.0427
120 0.0303
180 0.0221
240 0.0186
300 0.0169
360 0.0159

Source: Murad et al. 2017. Reproduced with permission of Springer.

With this done, go to the Regression folder and create a new file, in the Specification
file, in case more data are available, different weight values can be inserted. Thus, the
main data would be given priority or maintain everything at unitary value for a general
regression with no priority data. In the Vary file, choose four variables for the regres-
sion, activation energy and pre-exponential factor for both direct and inverse reactions.
These properties are reaction variable type and the reaction number is the same as will
be specified in the Reaction folder. Then, go to Reaction folder and create a new Pow-
erlaw reaction. In the newly created reaction, insert a name, and select the reactants and
products involved in the reaction, remembering that reactants have a negative value for
coefficients. It is important to remember that four types of classes are available: Equi-
librium, Powerlaw, Langmuir–Hinshelwood–Hougen–Watson (LHHW), and Generalized
Langmuir–Hinshelwood–Hougen–Watson (GLHHW).
In the Equilibrium class, the user can choose to calculate the equilibrium constant
through Gibbs energy or by a built-in temperature-dependent expression. Powerlaw has
the rate expression based on the kinetic factor and driving force, where the kinetic factor
can be obtained through Eq. (9.8) and the driving force by choosing the concentration
basis. LHHW is identical to the Powerlaw equation, but with an adsorption term. Finally,
the GLHHW is the same as LHHW, but the adsorption term can be manually inserted and
customized. ( )n ( −E )( 1 1 )
T −
r=k e R T T0 (9.8)
T0
The Powerlaw class and mole fraction as concentration basis are used in this work to
illustrate the model fitting. As the regression values are not available yet, a reasonable initial
value of the activation energy and a pre-exponential factor in the file can be considered.
Finally, go back to the batch reactor. There, insert the specifications, the kinetics with the
same name initially chosen, and the batch operation. As the Stop Criteria, set the time or
another parameter of interest. At last, in Operation Time, put the total cycle time as one hour,
which means the software will multiply the flowsheet stream, representing the batch charge
by cycle time. The maximum calculation time to be inserted is at least the same time on the
stop criteria file. The time interval between profile points will depend on how many points
in the profile are wanted to be generated. Finally, run the simulation and see the results.
236 Process Systems Engineering for Biofuels Development

0.090

0.075
Estimated value

0.060

0.045

0.030

0.015
0.015 0.030 0.045 0.060 0.075 0.090
Measured value

Figure 9.5 Comparison between measured and estimated values for an esterification reaction using the
Powerlaw model.

In the Regression file results, a comparison between the experimental data set and
regressed values is available; it is also possible to verify this difference graphically, as
shown in Figure 9.5. On Profiles, available in the batch block, it is possible to access the
reaction behavior by analyzing the composition change of reagents over time. Finally,
regressed parameter values can be obtained by going to the Results file in the created
reaction folder.
It is essential to have in mind that for a more accurate parameter regression, new values
of the parameters should be substituted in the Reaction file, and then the program can be
run once again. By doing this procedure until a constant value of the properties is achieved,
a more precise result can be obtained. Therefore, this presented resource can be used when-
ever experimental data are available instead of using a simple conversion value.

9.3.3 Transesterification Reaction

Transesterification is the most common way to produce biodiesel. This reaction (Figure 9.6)
can occur just by mixing the reactants and waiting until the system reaches equilibrium.
Nevertheless, a process with fast conversion and high purity is always desired. Therefore,
catalyst and excess alcohol are usually used. It is possible to use either acid or alkaline
catalyst, and either ethanol or methanol as the reactant. In this section, the kinetics of
alkali-catalyzed transesterification of soybean oil is discussed.
Much research has already been done in recent years, and now the scientific community
is focusing on the use of waste oil, non-edible oils, and animal fat, mainly because of the
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 237

H2C COOR′ RCOOR′ H2C OH

+
Catalyst
HC COOR″ + 3 ROH RCOOR″ + HC OH
+
H2C COOR‴ RCOOR‴ H2C OH

Triacylglycerols Alcohol Mixture of Glycerol

alkyl esters

Figure 9.6 Transesterification reaction.

k1
Triacylglycerol + ROH Diacylglycerol + RCOOR′
k2
k3
Diacylglycerol + ROH Monoacylglycerol + RCOOR″
k4
k5
Monoacylglycerol + ROH Glycerol + RCOOR‴
k6

Figure 9.7 Reactions involved in the transesterification reaction for biodiesel production, where ROH
represents a short-chain alcohol and RCOOR the fatty acid alkyl esters.

eco-friendly and economic aspects. Still, there are three steps behind the main reaction,
and all three are reversible. Monoacylglycerols and diacylglycerols are the intermediates in
these reactions (Figure 9.7).
Concerning the catalyst, there are several options. Alkaline metal alkoxides have shown
the best performance, high purity (>98%) and short reaction time (30 minutes) at a low
molar concentration (0.5 mol%) (Schuchardt et al. 1998). Moreover, alkaline metal hydrox-
ides have a better cost–benefit ratio. However, the use of hydroxides with alcohol will form
water that can hydrolyze esters and form soap. Even with soap formation being an unde-
sirable effect that makes the recovery of the glycerol difficult, it remains reasonable due to
the cost–benefit aspect.
For the kinetic analysis, experimental data obtained by Noureddini and Zhu (1997)
will be used in this work, where methanol was used as the short-chain alcohol at a
soybean-to-methanol molar ratio of 6:1, as well as sodium hydroxide as catalyst precursor.
In their research, it was stated that the shunt reaction did not improve the fit of the kinetic
parameters. This means that for simulation criteria, it will not be necessary to include the
overall reaction as an extra step to our simulation. To simulate, the activation energy and
the rate constants are needed.
Table 9.9 presents the data of activation energy (in cal/mol) that should be used in the
Arrhenius equation Eq. (9.9) and the average reaction rate constants at 323.15 K. The data
were retrieved from Noureddini and Zhu (1997).

k = A ⋅ exp(E∕RT) (9.9)
238 Process Systems Engineering for Biofuels Development

Table 9.9 Activation energy and rate constants at 323.15 K for

transesterification with methanol.

Reaction Activation energy (cal/mol) Rate constants at 323.15 K

TAG → DAG 13 145 0.050
DAG → TAG 9932 0.110
DAG → MAG 19 860 0.215
MAG → DAG 14 639 1.228
MAG → GLY 6421 0.242
GLY → MAG 9588 0.007

TAG, triacylglycerol; DAG, diacylglycerol; MAG, monoacylglycerol; and GLY, glycerol.

Table 9.10 Activation energy and equilibrium constants at 323.15 K for

sunflower oil transesterification with ethanol.

Activation Equilibrium constants

Reaction energy (kJ/mol) at 323.15 K
TAG ↔ DAG 48.7 3.21
DAG ↔ MAG 49.3 3.18
MAG ↔ GLY 53.9 72.77

Reyero et al. (2015) have obtained data for the transesterification of sunflower oil with
ethanol using sodium hydroxide catalyst. They have also considered soap formation.1 These
data are shown in Table 9.10. It is worth mentioning that the tables present the equilibrium
constants.

9.3.4 Supercritical Route

Reactions at supercritical condition of the alcohol allow simultaneous triacylglycerol trans-
esterification and FFA esterification by working at high temperatures and pressures (above
the critical point of alcohol). The mechanisms in this type of reaction are similar to those
of the traditional transesterification, involving a three-stage reaction, having tri-, di-, and
monoacylglycerols and forming esters from the acylglycerols and alcohol reactions and
production of glycerol as by-product. Thus, one of the main advantages of this method is
that oil pretreatment to remove the FFA is not necessary, as well as the catalyst.
In order to determine the reaction kinetics, Cheng et al. (2008) proposed a transesterifi-
cation reaction of peanut oil under a temperature range of 250–310 ∘ C, pressure of 10 and
16 MPa, and oil-to-methanol ratio of 1:30 and 1:40 through six hours of reaction. Some con-
ditions enabled the reaction to reach the chemical equilibrium condition in much shorter
time. Also, higher molar ratios are used particularly when triacylglycerols contain high FFA
content, where up to an alcohol-to-oil molar ratio of 45:1 can be used (Sprules and Donald
(1950)). As seen in the experiments presented by Cheng et al. (2008), although the pressure
influence was little, the temperature and the reactant molar ratio were important factors to

1
The data are available at: https://doi.org/10.1016/j.fuproc.2014.09.008.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 239

increase both reaction rates and yield. Therefore, for the kinetics calculation Eq. (9.10) was
used.
𝛼 𝛽
rA = k1 ⋅ Coil ⋅ Calcohol (9.10)
where concentrations are in mol/l, and Calcohol can be considered constant as excess alcohol
is used. In addition, Coil may be rewritten as Coil 0 ⋅ (1 − X), where X is the conversion
as a function of time. By manipulating these variables, Eq. (9.11) is obtained, where the
conversion was regressed giving Eq. (9.12).
dx k
ln = 𝛼 ⋅ ln[Coil 0 ⋅ (1 − X)] + ln 1 (9.11)
dt Coil 0
X(t) = 7 ⋅ 10−8 t3 − 5 ⋅ 10−5 t2 + 0.0114t (9.12)
After substituting the experimental data, a table with the calculated rate constant and
reaction order was obtained for different temperatures. With the value of the rate constant
found along the measured temperatures, the activation energy (kJ/mol) and factor frequency
(A) can be obtained. Finally, the kinetic constants are acquired and substituted in Eqs. (9.9)
and (9.10), resulting in Eq. (9.13).
rA = 12.45 ⋅ exp(28.85∕RT) ⋅ Coil
1.5
(9.13)
The reaction has an increasing conversion at higher temperatures, but above 400 ∘ C
decomposition reactions start due to temperature degradation. Therefore, it is recommended
to work around 350 ∘ C, allowing the process to have higher efficiency without degradation
(Kusdiana and Saka 2001).

9.4 Process Design

All biodiesel production process design here presented was developed using Aspen Plus
using the correct experimental property parameters, as shown and discussed before.
Methanol was the alcohol considered for both esterification and transesterification reac-
tions; the oil stream was considered to be 95 wt% triolein and 5 wt% oleic acid. Our goal
in this section is to simulate a process capacity to reach 10% of the Brazilian national
demand estimated to 2025. Therefore, the inlet oil stream will be approximately 48 m3 /h.
The property method used in most of the simulation was UNIQUAC, as very little quantity
of electrolytes is present, and the properties were already corrected. The final simulation
can be observed in Figure 9.8, where the left-hand side shows the esterification section, the
middle shows the transesterification section, and the right-hand side shows the biodiesel
purification section. The explanation of the process is presented in detail next.

9.4.1 Esterification Reaction

Seven reactor models are available in Aspen Plus. One may wonder which one should be
used. The stoichiometric reactor uses a given reaction and its partial conversions. However,
an equilibrium reactor can calculate the product conversion. Different options can be
selected, and each case has its own individual details. Here two of them will be discussed
and compared for esterification: the Gibbs equilibrium reactor and the batch reactor.
 25
1,44
60 74 234

1,20 MKUP 1,44 0,10

59
1,20 30 25
METH MIXX REC WAT
1,74 1,44 25
JOIN
RECYCLE 1,14 234
P06
60 NAOHIN NAOH RECYC2 0,10
62 50
ESTIN 1,20 P02 WATER
60 RECY TRANSIN 1,44
45 NEUTRA 1,04 BIOD
1,20
1,74
TC01 BIO
CSTR D02 EXT1 D03
OIL R01 65 174 50 298
NEUT P07
TOTC 68 LIGHT 1,44 DIN2 0,44 1,14 0,13
60 233 60
2,24
1,20 0,84 1,44 WOIL
NOUT LIQ EXTIN
DIN TOSEP
TC02 51
ESTOUT DOUT ROUT
HEAVY 1,04 46
P01 P03 P04 P05
SEP01 25
65 1,04
1,04 NEUT2
1,44
GLYCEROL
SULF
SALT
P08

Figure 9.8 Flowsheet of the process for biodiesel production simulation with Aspen Plus.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 241

9.4.1.1 Equilibrium Reactor Based on Gibbs Free Energy (RGibbs)

This option minimizes the Gibbs free energy using the atom balances from the specified
components for the reaction. Thus, for a first try, it may give significant base results, but it
is important to notice that it may also drive the reaction to incorrect products due to those
atom balances, which can lead to the occurrence of reactant production instead of products
or different undesirable reactions.
In this reactor, the user can choose the “best” reaction by setting it to calculate only phase
equilibrium and chemical equilibrium on Calculation Option, although the change between
reagents and unexpected formations of products can happen when using this option. For this
reason, the ways to work the correct reactions out will be presented.
For this example, two types of alcohols, methanol and ethanol, were used to test the
esterification and transesterification process. If the option mentioned above was selected,
at 60 ∘ C and 1 bar, considering that the oil stream has 95 wt% triolein and 5 wt% oleic
acid, and a 9:1 alcohol-to-oil molar ratio was used, it would produce methyl/ethyl esters,
glycerol, and water stoichiometrically. Possibly due to the energy of formation, much more
water and less glycerol are produced, indicating that part of the triolein or glycerol formed
may be transformed into oleic acid.
Considering that esterification and transesterification take place at the same time, the
Gibbs reactor gives contradictory results including possible triacylglycerol hydrolysis to
form fatty acids. One possible way to avoid this problem is to separate the simulation in two
reactors and specify which components participate in the reaction and which ones should
be formed. As mentioned before, let us force the desired reaction to occur by using two
reactors for each type of alcohol, one for esterification and one for transesterification. As the
procedure is the same, differing only in the alcohol and ester names, only the step-by-step
process for ethanol will be given in the following.
For the first reactor (esterification) set the Calculation Option as Calculate phase equi-
librium and chemical equilibrium and in the Products file set the option to identify possible
products, as wanted to restrict to the esterification reaction. Then, put all the desired com-
ponents in the product stream: ethyl ester, glycerin, triolein, water, and ethanol (as it enters
in excess in the reaction). In the Inerts file, put all the non-participating components and
the ones that are unreacted as yet (triolein), setting all of them to a unitary fraction.
Here, if the user wants to set how much of the reagents will actually react, it is possible to
settle the fraction or mole flow of feed component that does not participate in the reaction.
When selecting the fraction, it is directly attributed to the feed stream, but the mole flow
can be attributed to the feed stream or from a secondary product formed along with the
reaction, without the need to be present in the feed stream. For the esterification reactor,
this is all that must be done. However, for the transesterification reaction, a few more steps
are needed.
By connecting the first reactor exit stream to a new Gibbs reactor, first, change the Calcu-
lation Option to Restrict chemical equilibrium – specify temperature approach or reactions.
With this option, Aspen will force the desired transesterification reaction to happen. Then,
in the Products file, specify the components that will leave the reactor (the same ones as
for the esterification except for the triolein that will react in this stage). In the Inerts file, set
the same components as for the first reactor, excluding the triolein and adding water and
glycerin.
242 Process Systems Engineering for Biofuels Development

Table 9.11 Stream results using the Gibbs reactor.

Inlet (kmol/h) Outlet (kmol/h)

Components Alcohol Oil Esterification Transesterification
Ethanol 450 – 447.5 305
Triolein – 47.5 47.5 –
Oleic acid – 2.5 – –
Ethyl oleate – – 2.5 145
Water – – 2.5 2.5
Glycerol – – – 47.5

Finally, enter the transesterification reaction on the Restricted equilibrium file in Individ-
ual reaction topic, but remember that three moles of alcohol react with one mole of triolein.
The user can set the temperature approach to the default, as the reaction will occur at a
temperature of 60 ∘ C. Finally, it is just needed to run the program and observe the results,
which are presented in Table 9.11.
Thus, by analyzing the results and the procedure to obtain these results, it is possible to
conclude that the Gibbs reactor may be very useful to predict some reactions involved in
biodiesel production, although it should be very carefully studied as it could show incor-
rect behavior. When this happens, the user may force the desired reaction to occur, as seen
before. For this reason, it is recommended to know how the reaction behaves to be more crit-
ically aware about the results before using the Gibbs reactor. As an alternative, the REquil
model could be used. With this model, only the user specified reactions occur. Both RGibbs
and REquil use the standard Gibbs energy of formation. Therefore, this property should be
carefully verified and compared with those available experimentally in the literature.

9.4.1.2 Batch Reactor (RBatch)

After discussing the Gibbs reactor, it is possible to realize that the Gibbs minimization
reactor can be considered an odd approach, mostly because we do not know if the values
for the Gibbs energy of formation are right or not. Thus, it may end with the full conversion
of one reactant, or it may produce some random compound that has low Gibbs energy of
formation. Therefore, it is better to know the kinetics of reactions and open up a range
of possibilities: for example, a batch reactor (RBatch) or a continuous stirred-tank reactor
(CSTR) can be used. In this section, a batch reactor for the reaction kinetics presented in
Section 9.3.1 is presented and discussed.
For both methanol and ethanol streams, the alcohol-to-oleic acid molar ratio of 9:1 was
maintained, and the sulfuric acid ratio was set as 5 wt% and 0.33 wt% for the reactor using
methanol and ethanol, respectively. Unlike the Gibbs reactor, the only new information that
should be added is the Operation Time, which in this case, will be set as six hours as used
by Murad et al. (2017). After running the simulation, the performance of both alcohols in
the esterification (Table 9.12) can be compared.
A visible difference in the performance concerning the two different alcohols is observed
in Table 9.12, where methanol gave a much better result compared with ethanol. It is now
possible to change the reaction time from the reaction with methanol by setting the desired
conversion to 99% (as Stop Criteria). That should give a reaction time of 1.7 hours.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 243

Table 9.12 Oleic acid conversion in the

esterification (6 hours of reaction).

Alcohol Oleic acid conversion (%)

Ethanol 87.1
Methanol 12.8

9.4.2 Methanol Recycling

In this stage, recovering the excess of methanol that remained after the esterification reactor
is a key factor. For that, use a distillation column, which will separate the methanol for
recycling from the rest of the components that will follow in the process.
To model the distillation column, first, use a shortcut distillation design using the
Winn–Underwood–Gilliland method (DSTWU) to have an estimation for the use of the
rigorous two- or three-phase fraction for single columns (RadFrac). To design a DSTWU
column, insert the light key component, methanol, and the heavy key, where the user can
put the lightest heavy key so that it will not go to the top of the column with the alcohol. In
this case, use water as the heavy key component. As for the recovery, use a high number
in the light key part, once we are interested in recovering most of the methanol and use a
low number in the heavy key part, as this component is not desired in the distillate stream.
Set the desired working pressure and column number of stages or reflux ratio. If there is
no information about these last options, it is common to set the estimation reflux ratio to
−1.5, as the minus symbol is read on the Software as how many times the minimum reflux
ratio will be used. In order to use vapor as the heating utility, let us work with a column
under vacuum (for example 0.8 bar). By analyzing the results obtained, it is possible to use
them as an estimation for the RadFrac design. It is important to mention that there is no
need to always make a DSTWU column before working with the rigorous column, but if
there is no idea of how the separation will be, it may be very helpful for the RadFrac design.
In the RadFrac design, use the minimum number of stages estimated from the DSTWU
column or suppose a reasonable number and modify it later for better adjustments. A
kettle was used as the reboiler type, a total condenser, and the valid phases were set to
vapor–liquid. In the last options, Reflux ratio and Distillate to feed ratio specifications
were used. Here, do the same for the number of stages, use DSTWU results or estimate.
In the stream file, input the feed stage (estimated or not), and in the pressure file insert
the working pressure and, if wanted, the pressure drop along the column. Then, set the
equipment under a pressure of 0.8 bar and pressure drop of 0.007 per stage. This pressure
drop will also be used for the other columns along the process.
It is crucial to remember that it is possible to have more adequate operation conditions
to achieve the desired specification. For that, the Design Specification file is a very useful
distillation column tool to reach the desired specification. Using this option, set the desired
goal and use the Vary folder to modify properties.
Therefore, starting with the Design Specification, create two blocks, one for the methanol
mole purity, and the other for methanol mole recovery in the top stream. In the simulation, a
99.9% methanol mole recovery and 99.7% mole purity for this component as specification
were used, as from other results, the conditions became more operationally expensive. Then,
244 Process Systems Engineering for Biofuels Development

Table 9.13 Methanol recycle distillation column results.

Property Value
Number of stages 7
Feed stage 3 (on stage)
Molar reflux ratio 0.595
Mole distillate to feed ratio 0.131
Methanol mole recovery 99.9%
Methanol mole purity 99.7%

go to the Vary file and set to change the reflux ratio and the distillate to feed ratio to achieve
these specifications. After running the simulation, it will modify our initial number for
these variables and present the found fittable values. The column properties are presented
in Table 9.13. In addition, the Stage Wizard button can be used to modify the total number
of stages, while the optimum feed stage, which minimizes the reflux ratio maintaining both
specifications cited above, will vary according to the number of stages modified.
Sometimes, although the specification is achieved, the column results appear with errors,
especially after resetting the simulation. In this case, deactivate the Design Specification
after substituting the correct values in the column, so that it will return the results obtained
in the design and it will not be necessary to go through all the iterations for the specification
achievement, avoiding possible problems.
After the column design, it is necessary to give the recovered methanol back to the esteri-
fication reaction. As not all the methanol needed will be available only in the recycle stream,
insert a makeup stream to compensate for the methanol reacted and lost in distillation.
To achieve the correct proportion of methanol entering the reactor, use a Design-Spec
block. In the block, create two variables, one representing the methanol that enters the
reactor and the other the oleic acid flow for esterification. For each one, set the stream
category and select the Mole-Flow type for both, as it will be compared directly with the
reactor inlet flow, then input the stream and component name, and the units desired. In
the Spec file, specify the goal; in this case, select that the methanol mole flow entering the
reactor must be nine times that of the oleic acid entering. It is important to remember to use
the same words for the variables created before, including capital letters.
Finally, go to the Vary file and choose to change the methanol flow in the makeup stream
to attain the desired specification. At this point, some attention is required to choose a fea-
sible range for that variable, as it should embrace the actual result. After this, the recycling
distillation column is ready, returning the remaining methanol to the reactor, reducing the
costs of reagents.

9.4.3 Transesterification Reaction

In contrast to the esterification, the transesterification needs an alkaline catalyst. Sodium
hydroxide has the best cost–benefit ratio as the catalyst precursor in this process. For that
reason, it is necessary to neutralize the remaining sulfuric acid from the esterification reac-
tion and insert some amount of sodium hydroxide to act as the catalyst. However, there
is still a problem: how should the sulfuric acid be neutralized without adding a solution
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 245

with water and sodium hydroxide? Following Noureddini and Zhu (1997), a stream of
sodium hydroxide dissolved in methanol can be used. Knowing that the solubility of sodium
hydroxide in methanol is 238 g/l, it is possible to work with a concentration of 230 g/l
(NPCS Board 2012).
In this case, since UNIQUAC is the model used, it will be easier to simulate the neu-
tralization with a reactor block, avoiding the creation of a new thermodynamic section. To
do that, create an RStoic reactor block and add the reaction of neutralization. If the heat of
reaction is known, go to Heat of Reaction and select one of the three options for the cal-
culation. Once the reaction and the entire block are defined, the last challenge is to adjust
the flow of the neutralization stream. There are many options, and the user can opt for the
easiest one: to create a Calculator adding the amount of sulfuric acid as base compound
(see Additional Resources), or using a design spec.
The sodium sulfate formed in the reaction has low solubility in methanol, and even with a
small amount of water the separation of most of it is possible with a filter (Okorafor 1999).
Thus, in Aspen Plus, a good option is to use a Sep block. The Sep block allows the user to
specify the split outputs manually, which means there is no need to simulate this separation,
but instead define the split fraction according to the solubility. In this case, consider the
entire separation of the sulfate by the filter. That cannot be true, we know; however,
considering all the sulfate left (1.7%) in solution is in ionic form, and at low concentration
(approximately 3 wt% in this case), it should not interfere in our simulation results.
The transesterification reactor is the “heart” of the biodiesel production process. For
convenience, methanol was chosen as the reagent. It is important to mind the reaction
conditions. First, the reactor will work isothermally at 333.15 K, avoiding the boiling
temperature of methanol and seeking the best conversion. Secondly, the right amount
of methanol must be present. Noureddini and Zhu (1997) showed that a ratio of 6:1 of
methanol-to-triacylglycerols is a good start, and that the catalyst (sodium hydroxide)
should represent 0.907 wt% of the methanol stream. Those specifications can be easily
reached by using a Design Spec.
By knowing the kinetics, the user must choose between different reactor blocks: RCSTR,
RBatch, or RPlug. In this chapter, RCSTR, a reactor block for simulation of CSTRs, will
be used. Inside the RCSTR block, select the valid phase (liquid only), the residence time
(100 minutes), pressure, and temperature. After choosing the standard values suggested by
Noureddini and Zhu (1997), it should be possible to play with these values and to seek
an optimal point. Following the same steps as for the esterification reactor, add the set of
reactions for the transesterification and run the simulation (Section 9.3.3). Once all those
steps are done, it should be possible to reach around 90% conversion for triolein in products
and sub-products.

9.4.4 Biodiesel Purification

After the transesterification, there is interest in recovering the excess methanol and purify
our future biodiesel. In this first part, a distillation column was used. The strategy used
is the same one as for the first column, setting the specification to have high methanol
recovery for recycling, and high water recovery for the bottom, as water prejudices the
transesterification reaction. The only thing needed to be aware of in this section is that it is
246 Process Systems Engineering for Biofuels Development

Table 9.14 Second methanol recycle distillation column results.

Property Value
Number of stages 6
Feed stage 4 (on stage)
Molar reflux ratio 1.54
Mole distillate to feed ratio 0.423
Methanol mole recovery (top) 98.2%
Water mole recovery (bottom) 84.8%

a three-phase column, as the system has two liquid phases (glycerol-biodiesel) and a vapor
phase (mainly methanol).
The detail in this part is to set the input of the column on the 3-Phase file to test. Thus,
set such assumption from the first stage to the last, and then choose the key component of
the second phase as the biodiesel. If there is any doubt about the three-phase system, it is
possible to go into the stream results and observe if it contains the first and second liquid
phase flows. The column specifications are presented in Table 9.14.
The methanol is recovered from the distillation column and then it returns to the transes-
terification reactor, where there will be a Design-Spec block for the makeup, regulating the
methanol to enter the reactor is six times with the triolein present in the main inlet reactor
stream.
The next step is to separate the water, its ions, and the glycerol from the biodiesel. This
is a critical step, and it must be deeply analyzed. In the previous sections, the pure com-
pound parameters and some other proprieties were discussed and updated. However, the
glycerol–water–biodiesel interaction parameters were purposely not changed. This means
analysis of the ternary diagram of this mixture is necessary to see if it can describe the mix-
ture well. Thus, in Figure 9.9, there are three ternary diagrams: (a) with the experimental
data retrieved from Bell et al. (2013), (b) using UNIQUAC, and (c) using UNIF-DMD. It
is possible to conclude that UNIQUAC cannot describe this system well with the actual
parameters. Therefore, it is necessary to use other thermodynamic models or update these
parameters.
In this case, use UNIF-DMD for the extraction of ions, water, and glycerol. To do that,
change the Property method in the Block Options of the equipment. The separation of these
three compounds can be accomplished by using a decanter or an extractor. The extractor
will often have a better result; however, it will require a new feed stream to work. In this
simulation, an adiabatic extractor with six stages and a water feed of 200 kg/h as the top
stream was used.
As there is the presence of ions and no electrolyte model was used, the user needs to
tell Aspen Plus to treat the sodium hydroxide as water. One option is to change the sodium
hydroxide for water in the Components > Specifications in the Properties environment, but
it would crash the neutralization reactors. Thus, the best option is to create a functional
group for the hydroxide under Components > Molecular Structure > NAOH > Functional
Group tab. Then, select the method (in this case, UNIF-DMD), and the group number that
should be 1300 (group number for water) and one occurrence.
After separating the compounds, it is necessary to neutralize the heavy key component of
the extractor with sulfuric acid or phosphoric acid. It will allow us to reach the specification
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 247

x2 x2
(a) (b) 0.95
0.05
0.05 0.95
0.15 0.85
0.15 0.85
0.25 0.75
0.25 0.75
0.35 0.35 0.65
0.65
0.45 0.45 0.55
0.55
0.55 0.45 0.55 0.45

0.65 0.35 0.65 0.35

0.75 0.25 0.75 0.25

0.85 0.15 0.85 0.15

0.95 0.05 0.95 0.05

x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1 x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1

x2
(c)
0.05 0.95

0.15 0.85

0.25 0.75

0.35 0.65

0.45 0.55

0.55 0.45

0.65 0.35

0.75 0.25

0.85 0.15

0.95 0.05
x3 0.05 0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 x1

Figure 9.9 Comparison between thermodynamic models/parameters of the ternary system methyl oleate
(1), water (2), and glycerol (3) at 333 K, and 0.088 MPa using Aspen Plus for (a) experimental data,
(b) UNIQUAC, and (c) UNIF-DMD.

and to obtain a crude glycerin stream, as well. Finally, the mainstream is directed to a
distillation column to both recover methanol/water and triolein, and to obtain the main
product, biodiesel. In this stage, it is possible to follow the same steps as accomplished in
the first and second columns.
The main difference between this column and the other ones is the methanol and biodiesel
removal in a partial condenser, as methanol and water leave the equipment in the vapor
phase, while biodiesel remains liquid, requiring the modification of the condenser type to
partial vapor–liquid. Another critical detail in this column is that as the components are
very different, the simulation may have some difficulties in finding the correct values. In
this case, set the type of convergence in the Configuration file as Strongly non-ideal liquid
instead of Standard, an option that is recommended when slow convergence is encountered
for the Standard option.
In the case of using Standard, it would be necessary to first change some specifications
for the software to achieve similar estimates, giving the error that it could not converge with
the limit number of iterations, and then put the correct values for such properties (which
248 Process Systems Engineering for Biofuels Development

can be used only after really finding the optimal configuration). This error is common when
trying to optimize a column by changing little by little some properties and then reset and
run the simulation, losing the closer estimative. However, all the effort is simplified by
changing the convergence type.
Thus, for this column, use three specifications, one for water recovery, one for biodiesel
recovery, and one for biodiesel purity, as wanted to achieve the ANP purity specification
of 96.5 wt% and maximum water content of 200 ppm. In these specifications, the first two
should be the highest number possible, while the third one can be set close to the ANP
number (a little more significant to ensure good results), so there will not have to be a much
more expensive column but one that can do the work. As there are three specifications and
three variables, choose to vary reflux ratio, distillate to feed ratio, and distillate to vapor
fraction in order to achieve those specifications. After running the simulation, the results
can be obtained as shown in Table 9.15.
It is important to remember that the esters present in biodiesel start to degrade around
400 ∘ C, so it is necessary to be careful not to surpass this value. In this simulation, to
have a good recovery and a consistent bottom temperature, set the column under a vac-
uum of 0.1 bar. With these conditions, the bottom temperature remains at 298 ∘ C, avoiding
the degradation problem, although the use of steam will not be suitable as the heating utility.
So, with this process, we could produce and purify our biodiesel to ANP’s specification,
maintaining a suitable temperature (without the risk of considerable ester degradation), and
recovering a good amount of the desired product.

9.4.5 Additional Resources

A lot of block and options were discussed in the previous sections, but sometimes the
user can come across something different and need a functional solution. To assist users,
Aspen Plus has manipulators such as Calculators and Design Specs, and analysis tools. This
section summarizes three resources that may be needed at some point during the simulation.

9.4.5.1 Calculator
Calculator is a type of manipulator, just like Design Spec, that can define some variable
in our simulation. To create one, click in Manipulators in the model palette and select

Table 9.15 Biodiesel purification distillation

column results.

Property Value
Number of stages 5
Feed stage 2
Molar reflux ratio 1.25
Mole distillate to feed ratio 0.956
Distillate to vapor fraction 0.055
Water presence in biodiesel 195 ppm
Biodiesel mole recovery 90.0%
Biodiesel mass purity 99.4%
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 249

Figure 9.10 Calculator tool interface from Aspen Plus.

Calculator. After creating the block, the interface must be as in Figure 9.10. Then, create
a new variable and assign its definition. The most important information about this block
is the information flow. The user can choose between import, export, and tear variable.
Importing a variable means that Aspen Plus will be able to use this variable to compute other
variables. However, exporting a variable will overwrite the calculated value to this variable.
After defining the variables, go to the Calculate tab. This tab has a white box to execute
FORTRAN statements. By clicking with the right mouse button in the variable list, a smart
box will appear, allowing the user to drag and drop the names of the variables already
defined. One last hint: start writing the Fortran statement by the seventh column. Columns
1 to 6 are reserved for comments and statement labels.

Example: Calculating the Amount of Sodium Hydroxide

In this example, calculate the amount of sodium hydroxide/methanol stream necessary to
completely neutralize the sulfuric acid from the esterification reactor. The first step is to
define the sampled variables. It will need two or three, depending on how the sodium
hydroxide concentration was defined in the stream. This case requires two import vari-
ables and one export variable. The import variables are the mole flow of sulfuric acid in
the stream that enters the neutralization reactor and the molar fraction of sodium hydrox-
ide in the stream with methanol. The export variable will be the total mole flow for the
stream methanol/sodium hydroxide. After that, starting at column 7, write a code defin-
ing the export variable as two times (reaction stoichiometry) the amount of sulfuric acid
divided by the molar fraction of sodium hydroxide. Now, run the simulation, and the Cal-
culator will do its job. As an exercise, try to calculate the amount of sulfuric acid used as
catalyst necessary for the esterification reactor. By now, it should be easy.
250 Process Systems Engineering for Biofuels Development

Figure 9.11 Sensitivity tool interface.

9.4.5.2 Sensitivity Analysis

“Sensitivity” is a model analysis tool. It is handy when having an already simulated pro-
cess, and the best configuration is being searched. Try, for example, evaluating the change
of reflux ratio or the feed stage. The interface is shown in Figure 9.11. As done for the
Calculator resource, the first step is to define the variables. However, in the Sensitivity tool,
the user needs to define the variables that are to be varied and the results to be seen, which
means it can manipulate the number of stages for a distillation column and have, as results,
the recovery of some compound. The Tabulate tab shows how the results will be displayed.

Example: Evaluating a Distillation Column

In a quick example, let us investigate the feed stage for the defined distiller. The first step is
defining the variable; in other words, create a block variable in the manipulator that will vary
the feed stage (FEED-STAGE). In the next step, define the sampled variables. By choosing
to work with the first distiller in the process, a good variable will be the methanol recovery
in the recycle stream. Thus, define the amount of methanol in the top stream and the feed
stream of the distiller. Now, write in the Tabulate tab, the column number (1) and the recov-
ery expression: the methanol in the top stream divided by the methanol in the feed stream.
Sometimes the manipulated variable is discrete, so, make sure to define the limits having this
in mind. As an exercise, try this with the second distiller, varying the distillate to feed ratio.

9.4.5.3 Optimization Tool

The last additional resource is the Optimization model. Like all the previous resources, first,
define the variables in the Define tab, and then the desired variables range in the Vary tab.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 251

Figure 9.12 Optimization tool interface.

The difference between this tool and the other two is its use. In the Objective & Constraints
tab, it is possible to add the objective (maximize or minimize a variable) and activate some
constraints. The most important detail in this tool is the optimization expression. Since
the model tries to reach the best point, having weak constraints or an opened variable can
easily crash the simulation. That is why the optimization tool can be a double-edged sword.
Figure 9.12 shows the Optimization tool interface.

Example: Optimizing a Distillation Column

In the first distillation column, the most natural idea is to optimize the recovery of methanol
and triolein. Also, it is desired that more triolein goes to the bottom, and more methanol
goes to the top of the distiller. Thus, divide this expression into three stages in the Fortran
tab. Starting in row 1, column 7, the methanol recovery is the amount of methanol that goes
to the top divided by the fed one.
In the same way, in row 2, column 7, define the triolein recovery as the bottom amount
divided by the fed amount of triolein. Finally, in row 3, column 7, write that the objective
function is the sum of the methanol and triolein recoveries. It is also possible to create a
constraint if the amount of water, temperature, or some other variable needs to be specified.
Aspen Plus enables the manipulation of up to 20 variables in the Vary tab. For this
example, only two variables are tested: the distillate to feed ratio and the molar reflux.
To complete all the steps, go into the Objective & Constraints tab, write the name of the
objective function, and select maximize (in this case). If there are any constraints, select
them. Now, run the simulation for the optimization to take place and try to find the optimal
point. Finally, go to the second distiller and vary the same variables, try to optimize the
recovery for the triolein, methyl oleate, methanol, diolein, and monoolein. Note: the second
column must have a temperature constraint, temperatures higher than 380 ∘ C should not be
reached.
252 Process Systems Engineering for Biofuels Development

9.5 Energy and Economic Analysis

For any project in the chemical industry, it is vital to optimize and observe how the process
will behave. Therefore, saving energy and economic evaluating is important for an appropri-
ate estimation. Nowadays, the economic feasibility and the financial return are of increasing
interest to investors. Besides, not only the equipment itself but also the raw materials and
utilities must be included. Thus, in this section, the Aspen Energy Analyzer and Economic
Analyzer will be discussed for the correct estimation and optimization of the operational
costs.
The Aspen Energy Analyzer is an extension of Aspen Plus and is designed to improve
how the energy flows in our plant. Once the simulation is finished, activate the Energy
Analyzer by clicking on the button located in the Activation Dashboard. It may take a
while, but after processing the simulation, Aspen Plus will return the estimated energy
savings.
By changing the view from Simulation to Energy Analysis, the user will get access to
a new project. Thus, select the process type that, in this case, will be the user specified
option and in the Customize box, a temperature approach of 15 ∘ C (Aspen’s® defaults)
should be inserted. Clicking on Analyze on the Home menu, Aspen will generate once
more the data with a new temperature approach. After that, possible designs need to be
evaluated.
To access the Energy Analyzer interface, click on Details in the home menu. Aspen will
redirect the user to a new window with a base case. The standard calculation option of
Energy Analyzer is the cost approach, and the base case calculates the cost with energy
after choosing to work without any recycle and only with utility streams. It is also possible
to generate the best energy design by selecting the scenario that was created and then, in
the bottom menu, the Recommend Design option. After that, click on Solve.
This simulation has four optimized scenarios. To interpret how it works, we have defined
a scenario. As shown in Figure 9.13, the whole concept of this analyzer is the use of the
energy that is already present in the process and stored as temperature. In the scenario
exemplified in Figure 9.13, the stream of the condenser from second distiller column splits
and part goes to an exchanger that uses the stream of the reboiler from the first distiller and
part goes to the reboiler of the first distillation column.
The black lines connecting the black dots represent the integration plan proposed by
Aspen Energy Analyzer. Also, each line represents a heat exchanger and can be interpreted

341°C 244°C 186°C Condenser (2nd Distiller)

HOT
STREAM 239°C 203°C 186°C Exchanger (After 1st Distiller)

239°C 115°C 99°C Reboiler (1st Distiller)

COLD
STREAM 133°C 124°C 60°C Exchanger (After CSTR)

Figure 9.13 Energy scenario obtained by Aspen Energy Analyzer.

 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 253

in the same way as explained before. There are also lines connecting utilities that were
suppressed in our figure but are represented by the white dot with a black outline.
Moreover, each layout has its configuration, and each configuration has its exchangers.
There are plenty of resources available in Energy Analyzer. By generating the scenario,
Aspen also generates an estimative of the total exchange area and its associated costs. It is
also possible to retrofit the Energy Analyzer simulation by clicking on the chosen design
and then on Enter Retrofit Mode; it will export the data to the Energy Analyzer Module in
the Simulation tab.
After running the Energy Analyzer, an overview of what is recommended to be done
can be obtained. By the Energy Analyzer tab in Aspen Plus, it is possible to evaluate some
scenarios with new heat exchangers. Go to Add Scenario on the top menu and then click
on Add Exchanger. Aspen Plus will give some potential changes with additional costs and
payback time to be chosen. Our simulation target is 47.3% of energy cost reduction, and
it can be accomplished by including four exchangers. This process helps to validate the
changes, and what is necessary or what is recommended in the simulation. It is worth noting
that Aspen Plus will not change the simulation, which means the economic analysis will
not consider it, and therefore the changes must be done manually.
After completing the simulation, it is possible to know if the project is economically
viable. As Aspen’s Economic Analysis automatically calculates the equipment price (with
the correct set of equipment types available in the Software to choose), the biggest issue is
to search the correct and current stream product and reagent prices.
In Table 9.16, all the components available in the process, and the price for each one, are
presented. As the waste oil can have different prices depending on the source and purity;
half of the soy oil price was considered for the estimation. The same thing was assumed
for triolein. For the process utilities, Aspen’s default cost values were used. Remember that
part of the values shown will be the cost, and another part will be income, as it is produced,
and it may be sold for other purposes.

Table 9.16 Current prices for the used/produced

components in the simulated industry.

Component Price (US$/ton)

Soy oil∗ 730
Waste oil (considered) 370
Sulfuric acid∗∗ 61
Sodium hydroxide∗∗ 650
Biodiesel∗∗∗ 1040
Methanol† 432
Glycerin‡ 220
Triolein (considered) 370
Sodium sulfate‖ 105
∗ Markets Insider (n.d.).
∗∗ ECHEMI (n.d.).
∗∗∗ NESTE (n.d.).
† Methanex (n.d.).
‡ Landress (2018).
‖ ICIS (2007).
254 Process Systems Engineering for Biofuels Development

Table 9.17 This is shown as from Aspen Plus.

Data Value
Total Capital Cost (USD) 20 699 300
Total Operating Cost (USD/Year) 323 495 000
Total Raw Materials Cost (USD/Year) 291 635 000
Total Product Sales (USD/Year) 596 838 000
Total Utilities Cost (USD/Year) 5 608 870
Desired Rate of Return (Percent/Year) 20
P.O. Period (Year) 1.96
Equipment Cost (USD) 4 252 900
Total Installed Cost (USD) 10 326 100

In order to insert these values in the simulation, go to Setup > Stream Price, and in the
input section, select to add both feed and product streams. Once added, it is necessary to
exclude the stream without any expense or profit, and then put the correct values for each
inlet and outlet stream. The most important detail in this part is that some streams have
different components with distinct price values, so it is necessary to verify precisely which
components are present and the proportion, so that the stream price may be adapted to its
content. Finally, with these values, go to the next step, the process full economic evaluation.
The first step is to activate the Economic Analyzer. As already done for the Energy Sav-
ings, click on the Activation Dashboard (at the View tab), and then on the green box. For
this procedure, it is essential that the Energy Savings is still activated or the savings from
the energy recycle will not be accounted for. After activating the green box, go to the Eco-
nomics tab and press Evaluate. Aspen Plus will ask the user for the mapping options and,
usually, it is a good idea to check both boxes: size equipment and evaluate cost. By map-
ping, choose the distiller configurations, and if desired, change one piece of equipment for
another (like a shell and tube exchanger for a furnace).
In the Summary of the Economic Analyzer, the user has access to a list of information as
shown in Table 9.17. In this case, the plant seems economically viable, since the period of
return is less than two years.
The Total Capital Cost includes the instrumentation, buildings, installation, and con-
struction of the plant, meaning all the one-time payments needed to make the plant ready
for startup, the Total Operation Cost includes the labor and maintenance costs, and the P.O.
Period is the return time of the investment.

9.6 Concluding Remarks

In this chapter, we have proposed and discussed some aspects related to the process sim-
ulation and design of a process for biodiesel production using the Aspen Plus simula-
tor. Some specific aspects about using biodiesel related compounds (acylglycerols and
fatty acid-based molecules) in the simulations were focused on, and how biodiesel can
be obtained with different types of reaction, reagents, and operational conditions. By cor-
recting the pure and mixture components parameters present in Aspen Plus, it was possible
to rely more on our further tests and simulation. Knowing about the most common reac-
tions to obtain the desired product and their advantages and disadvantages, an important
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 255

decision could be made on which would be our biodiesel production route. Using the tra-
ditional transesterification route to produce biodiesel from acid oil, where a pretreatment
step for FFA removal is necessary (esterification), and by having the correct kinetics, it was
possible to design the entire reaction and purification process using Aspen Plus.
Once all the preparation and project design were implemented, the next step was to
demonstrate how to optimize a simulation by using the Aspen Energy Analysis tool, where
we were able to have 47.5% less energy cost with its use. Then, the most important factor to
evaluate the feasibility of an industry was run: the Economic Analysis. In this part, all the
reagent and product prices were searched, and utilities and equipment costs were estimated
to have a reliable source of information.
Finally, besides knowing how to work with Aspen Plus tool, we were able to obtain
biodiesel under ANP’s standards (Brazilian Agency) for both ester and water presence and
with a return period of two years.

Acknowledgment
Reference and screen images from Aspen Hysys®, Aspen Plus®, Aspen Plus Dynamics®,
Aspen Economics Evaluation®, Aspen EDR®, Aspen Energy Analyzer®, and Aspen
Properties® are reprinted with permission from Aspen Technology, Inc. AspenTech®,
Aspen Hysys®, Aspen Plus®, Aspen Plus Dynamics®, Aspen Economics Evaluation®,
Aspen EDR®, Aspen Energy Analyzer®, and Aspen Properties®, Aspen EDR®,
aspenONE®, and the AspenTech leaf logo are trademarks of Aspen Technology,
Inc. All rights reserved.

Exercises

1. Use the paper of Bell et al. (2013) to estimate and regress the UNIQ parameters for the
ternary methyl ester, glycerin, and water mixture as described in Section 9.2.2 and dis-
cuss why UNIQUAC with our given parameters (without this regression) cannot describe
the ELL equilibrium.
2. Compare the overall performance when changing the extractor (Section 9.4.5) for a
decanter with and without the water feed.
3. Use the Aspen regression tool to estimate the reaction parameters: activation energy, and
pre-exponential factor using POWERLAW for the direct and reverse reactions. Condi-
tions: 343.15 K and 1 bar. Inlet flow: ethanol 108.66 kg/h, water 10.53 kg/h, lauric acid
51.94 kg/h, and ethyl laurate 0 kg/h. Use Table 9.18 for the data regression.
4. Use methanol instead of ethanol to build a new simulation and compare the total
biodiesel production for both alcohols and its energy use.
5. From Section 9.3.3 we have found different energy saving designs. Choose one and use
it to adjust and rerun the simulation. Then, rerun the Energy Analyzer and try to improve
energy savings.
6. CHALLENGE. Build a simulation of biodiesel production using a supercritical reactor
(plug flow reactor) and evaluate economic feasibility.
256 Process Systems Engineering for Biofuels Development

Table 9.18 Molar fraction (lauric acid) versus time for

the data regression (Exercise 3).

Molar fraction
Time (minutes) (lauric acid)
0 0.0809
15 0.0712
30 0.0633
60 0.0511
120 0.0348
180 0.0258
240 0.0212
300 0.0188
360 0.0176

References
Aissa, M.A., Ivaniš, G.R., Radović, I.R., and Kijevčanin, M.L. (2017). Experimental investigation and
modeling of thermophysical properties of pure methyl and ethyl esters at high pressures. Energy & Fuels
31: 7110–7122. https://doi.org/10.1021/acs.energyfuels.7b00561.
Al-Malah, K.I. (2016). Introducing Aspen Plus. In: Aspen Plus®, Chapter 1. Wiley https://doi.org/10.1002/
9781119293644.
Andreatta, A.E., Casás, L.M., Hegel, P. et al. (2008). Phase equilibria in ternary mixtures of methyl oleate,
glycerol, and methanol. Industrial and Engineering Chemistry Research 47: 5157–5164. https://doi.org/
10.1021/ie0712885.
Bell, J.C., Messerly, R.A., Gee, R. et al. (2013). Ternary liquid–liquid equilibrium of biodiesel compounds
for systems consisting of a methyl ester + glycerin + water. Journal of Chemical & Engineering Data
58: 1001–1004. https://doi.org/10.1021/je301348z.
Berrios, M., Siles, J., Martín, M.A., and Martín, A. (2007). A kinetic study of the esterification of free fatty
acids (FFA) in sunflower oil. Fuel 86: 2383–2388. https://doi.org/10.1016/j.fuel.2007.02.002.
Brandão, K.S.R., Nascimento, U.M., Sousa, M.C. et al. (2006). Produção de Biodiesel por Transesterifi-
cação do Óleo de Soja com Misturas de Metanol-Etanol. Analysis 1: 141–146.
Brands, D.S., Pontzen, K., Poels, E.K. et al. (2002). Solvent-based fatty alcohol synthesis using supercritical
butane: flowsheet analysis and process design. Journal of the American Oil Chemists Society 79: 85–91.
https://doi.org/10.1007/s11746-002-0439-0.
Bucalá, V., Foresti, M.L., Trubiano, G. et al. (2006). Analysis of solvent-free ethyl oleate enzymatic synthe-
sis at equilibrium conditions. Enzyme and Microbial Technology 38: 914–920. https://doi.org/10.1016/j
.enzmictec.2005.08.017.
Canakci, M. and Van Gerpen, J. (1999). Biodiesel production via acid catalysis. Transactions of the ASAE
42: 1203–1210.
Carvalho dos Santos, K., Pedersen Voll, F.A., and Corazza, M.L. (2018). Thermodynamic analysis of
biodiesel production systems at supercritical conditions. Fluid Phase Equilibria 484: 106–113. https://
doi.org/10.1016/j.fluid.2018.11.029.
Cheng, J., Li, Y., He, S. et al. (2008). Reaction kinetics of transesterification between vegetable oil and
methanol under supercritical conditions. Energy Sources Part A: Recovery, Utilization, and Environmen-
tal Effects 30: 681–688. https://doi.org/10.1080/15567030601082084.
Choudary, B., Lakshmi Kantam, M., Venkat Reddy, C. et al. (2000). Mg–Al–O–t-Bu hydrotalcite: a new
and efficient heterogeneous catalyst for transesterification. Journal of Molecular Catalysis A: Chemical
159: 411–416. https://doi.org/10.1016/S1381-1169(00)00209-0.
 Process Analysis of Biodiesel Production – Kinetic Modeling, Simulation, and Process Design 257

Clark, S., Wagner, L., Piennaar, P. et al. (1981). Hour screening test for alternate fuels in energy notes for,
variables affecting the yields of fatty esters from transesterified vegetable oils 1. American Society of
Agricultural and Biological Engineers 2: 385–390. https://doi.org/10.1007/BF02541649.
Coelho, R.A. (2011). Equilíbrio líquido-vapor de sistemas binários envolvendo ésteres etílicos do biodiesel
(Glicerol ou água) + Etanol. Master thesis. UFPR.
ECHEMI (n.d.). Caustic Soda Price Analysis. https://www.echemi.com/productsInformation/
pd20150901041-caustic-soda-pearls.html (accessed 15 February 2019).
Eduljee, G.H. and Boyes, A.P.J. (1981). Excess Gibbs free energy for eight oleic acid-solvent and
triolein-solvent mixtures at 319.15 K. Journal of Chemical & Engineering Data 26: 55–57.
Helwani, Z., Othman, M.R., Aziz, N. et al. (2009). Solid heterogeneous catalysts for transesterification of
triglycerides with methanol: a review. Applied Catalysis. A, General 363: 1–10. https://doi.org/10.1016/
j.apcata.2009.05.021.
ICIS (2007). Chemical profile: Sodium sulfate. https://www.icis.com/explore/resources/news/2007/09/10/
9060326/chemical-profile-sodium-sulfate (accessed 15 February 2019).
Imahara, H., Minami, E., Hari, S., and Saka, S. (2008). Thermal stability of biodiesel in supercritical
methanol. Fuel 87: 1–6. https://doi.org/10.1016/j.fuel.2007.04.003.
Ivana, L., Kesic, Z., Zdujić, M., and Skala, D. (2016). Vegetable oil as a feedstock for biodiesel synthesis.
In: Vegetable Oils - Properties, Uses and Benefits (ed. B. Holt), 83–128. Nova Science Publishers.
Javidialesaadi, A. and Raeissi, S. (2013). Biodiesel production from high free fatty acid-content oils: exper-
imental investigation of the pretreatment step. APCBEE Procedia 5: 474–478. https://doi.org/10.1016/j
.apcbee.2013.05.080.
Kawashiro, I., Tanabe, H., and Ishii, A. (1960). Applications of gas chromatography to food analysis (I)
studies on fatty acids in butter and cheese. Journal of the Food Hygienic Society of Japan 1: 78–83.
https://doi.org/10.1385/ABAB.
Kusdiana, D. and Saka, S. (2001). Kinetics of transesterification in rapeseed oil to biodiesel fuel as treated
in supercritical methanol. Fuel 80: 693–698. https://doi.org/10.1109/TMAG.2010.2073454.
Landress, L. (2018). US crude glycerine prices could dip as spring nears. https://www.icis.com/explore/
resources/news/2018/02/14/10193613/us-crude-glycerine-prices-could-dip-as-spring-nears (accessed
15 February 2019).
Markets Insider (n.d.). Soybean Oil. https://markets.businessinsider.com/commodities/soybean-oil-price
(accessed 15 February 2019).
Meher, L.C., Vidya Sagar, D., and Naik, S.N. (2006). Technical aspects of biodiesel production by trans-
esterification – a review. Renewable and Sustainable Energy Reviews 10: 248–268. https://doi.org/10
.1016/j.rser.2004.09.002.
Methanex (n.d.). Pricing. https://www.methanex.com/our-business/pricing (accessed 15 February 2019).
Murad, P.C., Hamerski, F., Corazza, M.L. et al. (2017). Acid-catalyzed esterification of free fatty acids
with ethanol: an assessment of acid oil pretreatment, kinetic modeling and simulation. Reaction Kinetics,
Mechanisms and Catalysis 123: 505–515. https://doi.org/10.1007/s11144-017-1335-3.
NESTE (n.d.). Biodiesel prices (SME & FAME). https://www.neste.com/corporate-info/investors/market-
data/biodiesel-prices-sme-fame-0 (accessed 15 February 2019).
NIST (n.d.). Equation Descriptions. https://trc.nist.gov/TDE/Equations/FEquations.html (accessed 29
March 2019).
NIST databank (n.d.). NIST Chemistry WebBook, SRD 69. doi: 10.18434/T4D303.
Noureddini, H. and Zhu, D. (1997). Kinetics of transesterification of soybean oil. Journal of the American
Oil Chemists’ Society 74: 1457–1463. https://doi.org/10.1007/s11746-997-0254-2.
NPCS Board (2012). Detailed Project Profiles on Chemical Industries (Vol II) (2nd Revised Edition). NIIR
Project Consultancy Services.
Okorafor, O.C. (1999). Solubility and density isotherms for the sodium sulfate−water−methanol system.
Journal of Chemical & Engineering Data 44: 488–490. https://doi.org/10.1021/je980243v.
258 Process Systems Engineering for Biofuels Development

Pauly, J., Kouakou, A.C., Habrioux, M., and Le Mapihan, K. (2014). Heat capacity measurements of pure
fatty acid methyl esters and biodiesels from 250 to 390 K. Fuel 137: 21–27. https://doi.org/10.1016/j
.fuel.2014.07.037.
Pisarello, M.L., Dalla Costa, B., Mendow, G., and Querini, C.A. (2010). Esterification with ethanol to
produce biodiesel from high acidity raw materials: kinetic studies and analysis of secondary reactions.
Fuel Processing Technology 91: 1005–1014. https://doi.org/10.1016/j.fuproc.2010.03.001.
Poling, B.E., Prausnitz, J.M., and O’Connell, J.P. (2001). The Properties of Gases and Liquids, vol. 5. New
York: McGraw-Hill.
Pratas, M.J., Freitas, S.V.D., Oliveira, M.B. et al. (2011). Biodiesel density : experimental measurements
and prediction models. Energy & Fuels 25: 2333–2340. https://doi.org/10.1021/ef2002124.
Ramos, L.P., Kothe, V., César-Oliveira, M.A.F. et al. (2017). Biodiesel: matérias-primas, Tecnologias
de Produção e Propriedades Combustíveis. Revista Virtual de Química 9: 317–369. https://doi.org/10
.21577/1984-6835.20170020.
Reyero, I., Arzamendi, G., Zabala, S., and Gandía, L.M. (2015). Kinetics of the NaOH-catalyzed transester-
ification of sunflower oil with ethanol to produce biodiesel. Fuel Processing Technology 129: 147–155.
https://doi.org/10.1016/j.fuproc.2014.09.008.
Sampaio, M.J.F. (2008). Produção de biodiesel por catalise heterogênea. 70. Master thesis. Polytechnic
Institute of Bragança.
Schuchardt, U., Vargas, R.M., and Gelbard, G. (1996). Transesterification of soybean oil catalyzed by
alkylguanidines heterogenized on different substituted polystyrenes. Journal of Molecular Catalysis A:
Chemical 109: 37–44. https://doi.org/10.1016/1381-1169(96)00014-3.
Schuchardt, U., Sercheli, R., Vargas, R.M., and Matheus, R. (1998). Transesterification of vegetable
oils: a review. Journal of the Brazilian Chemical Society 9: 199–210. https://doi.org/10.1590/S0103-
50531998000300002.
Sharma, S. and Kanwar, S.S. (2014). Organic solvent tolerant lipases and applications. Scientific World
Journal 2014 https://doi.org/10.1155/2014/625258.
Sprules, F.J. and Donald, P. (1950).Production of fatty esters. US Patent 2,494,366.
Vyas, A.P., Verma, J.L., and Subrahmanyam, N. (2010). A review on FAME production processes. Fuel
89: 1–9. https://doi.org/10.1016/j.fuel.2009.08.014.
 10
Process Development, Design and
Analysis of Microalgal Biodiesel
Production Aided by Microwave
and Ultrasonication
Dipesh S. Patle1 , Savyasachi Shrikhande2 , and
Gade Pandu Rangaiah2,3
1 Chemical Engineering Department, Motilal Nehru National Institute of Technology, Allahabad 211 004,
India
2 School of Chemical Engineering, Vellore Institute of Technology, Vellore 632 014, India

3 Department of Chemical and Biomolecular Engineering, National University of Singapore,117585,

Singapore

10.1 Introduction
Owing to the increasing population and ever-changing lifestyle, total primary energy con-
sumption (TPEC) is increasing day by day and is projected to increase by 57% from the
year 2010 to 2040, which impacts available fossil fuels since they are non-renewable (Lee
et al. 2010; Kumar and Sharma 2016). The universal energy crisis and environmental con-
cerns have led to extensive research and development programs on biomass utilization (Li
et al. 2015). In the context of the 2012 “International Year for Sustainable Energy for All”
(SE4ALL), the International Renewable Energy Agency launched a global roadmap, named
REMAP 2030, in a bid to double the share of renewable energy by 2030 (Li et al. 2015).
The Chinese National Energy Administration has carried out a “National Twelfth Five-Year

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
260 Process Systems Engineering for Biofuels Development

Plan” on biomass energy (CNEA 2012). In this project, consumption of biofuel (mainly
ethanol and biodiesel) is expected to reach 12 million metric tons by 2020. Also, the U.S.
Department of Energy has set a goal to generate 20% of the transportation fuel from biomass
by 2030 (USDA 2013). Thus, energy security, petroleum price, depletion of fossil fuels, and
environmental concerns have prompted considerable interest in research and development
of biomass-derived fuels such as biodiesel and bioethanol. Of these, biodiesel production
is the topic of this chapter.
Transesterification of oil using alcohol and catalyst yields monoalkyl esters, known as
biodiesel (Khiratkar et al. 2018). First-generation biodiesel is derived from feedstock such
as pure vegetable oil (e.g. soybean oil, corn oil, and palm oil). Second-generation biodiesel
is derived from feedstock such as inedible oil (e.g. jatropha oil and waste/used cooking oil),
and third-generation biodiesel is derived from feedstock such as algae. First-generation
biodiesel synthesis is not feasible and sustainable as it has three significant drawbacks: high
cost of oil, limited availability, and food versus fuel issue. Second-generation biodiesel
production has potential; however, availability of raw material for it is uncertain. Biodiesel
processes based on first- and second-generation feedstock are well researched (e.g. Gogate
2008; Sharma and Rangaiah 2013; Patle et al. 2014) and are also in industrial practice
(Lurgi 2019).
In general, more than 75% of the cost of biodiesel production is for the feedstock or raw
materials (Atabani et al. 2012). At present, plant seed oil (i.e. first- and second-generation
feedstock) is the major source of biodiesel production (Naik et al. 2010). Biodiesel produc-
tion from algae, i.e. third-generation biofuel, needs significant research and development
as it can potentially contribute to the desired biodiesel production. It can be produced from
either wet or dry algal biomass. Production from wet biomass is still under research whereas
production from dry biomass has the drawback of the high energy consumption required to
remove water content.
Sara et al. (2016) reported that microbial oil with accumulated lipids, which in turn acts
as a source of energy, is considered as the best alternative, mainly because it does not alter
the food chain leading to reduced pressure on the land as well as the environment. Algae,
especially microalgae, have better growth rate as compared with terrestrial crops, and oil
yield from algae is 7 to 31 (approximated to be from 20 000 to 80 000 l/acre/yr) times higher
than the most widely used source, i.e. palm oil (Demirbas and Demirbas 2011). Further,
after oil extraction, microalgae produce high cellulose waste biomass that can be hydrolyzed
to form ethanol (John et al. 2011). Approximate water required for microalgal biomass
production varies considerably depending on the cultivation systems; Jorquera et al. (2010)
reported biomass concentration of 0.35, 2.7, and 1.02 g/l of water used for raceway ponds,
tubular photobioreactors, and flat-plate photobioreactors, respectively.
Considering the above advantages, the last decade has seen significant research on
biodiesel production from third-generation feedstock. Various algal species have been
cultured and investigated for their lipid content. The lipid extraction techniques have also
seen great advancements with the inclusion of ultrasonication. Recent research interest has
also been on increasing the concentration of lipids in microalgae and optimization of the
biodiesel process. The cost of biodiesel production from microalgae varies due to the large
range of algae and their lipid content. It is from $10.87 to $13.32/gal of biodiesel (Sun
et al. 2011).
 Process Development, Design and Analysis of Microalgal Biodiesel Production 261

Despite the obvious potential benefits, use of algal biomass for commercial biodiesel
production poses several challenges such as its high water content, wide range of lipid
content depending on the algal species, and cost of production. High water content in wet
microalgae (up to 98% in certain cases) makes lipid extraction more difficult as the water
around algal cells generates a hydrated shell, which acts as an obstacle for both energy and
mass transfer (Martinez-Guerra et al. 2018). In a two-step process for biodiesel, microbial
oils (i.e. single cells oils) are first extracted from the algae and subsequently transesteri-
fied to produce biodiesel. Of late, several researchers (Sara et al. 2016; Zhang et al. 2016;
Martinez-Guerra et al. 2018) have worked on direct, i.e. in situ, transesterification of oils in
the algal biomass or sludge derived biomass, where oil extraction and biodiesel synthesis
occur concurrently. Combining these two steps into one step is expected to reduce the cost
as well as equipment footprint.
As in situ processing means concurrent lipid extraction and its transesterification, it is
clear that the process would have better biodiesel yield if the lipid extraction from the algal
biomass is maximized and the rate of transesterification is enhanced. Such enhancements
are possible using intensification by microwave (MW) radiation and/or ultrasonic irradi-
ation. In microwave intensification, the radiation helps in enhancing the lipid extraction,
whereas the intensification due to ultrasound (US) is due to the traveling of acoustic, i.e.
sound waves through the solvent, resulting in the phenomenon called cavitation. The disrup-
tion of cell walls and enhanced mass transfer are the direct consequences of these intensi-
fications. The continuous development of bubbles generates microturbulence, interparticle
collisions at higher velocity and turbulence in particles of the algal biomass (Paniwnyk
et al. 2009). Therefore, MW power, US power, US frequency, and ultrasonication cycle
play crucial roles in the process enhancement.
Although biodiesel production from algal biomass has significant potential to be used
as renewable fuel, there are many aspects, which need researchers’ attention for it to be
technically and commercially feasible. Ultrasonication–MW intensified in situ synthesis of
biodiesel may be one such way to produce good quality renewable fuel with fewer process-
ing units as conventional alternatives require more processing units. The major challenge
lies in the scale up of MW and US reactors, downstream separation of reactants and prod-
ucts, and the optimization of process parameters. This includes effective transmission of the
acoustic energy, i.e. cavitation into large process volume as well as the engineering aspects
of the design (Gogate 2008). No article in the open literature has focused on these aspects
so far.
In the present study, a continuous process involving an intensified in situ transesterifica-
tion of wet microalgae to produce biodiesel on industrial scale is developed and simulated
in Aspen Plus v8.8. It is based on the process studied experimentally in the laboratory by
Martinez-Guerra et al. (2018). More details on this experimental study are given in the
next section. To the best of our knowledge, the present study is the first that investigated
the complete process development for the in situ biodiesel production from wet microalgal
biomass. Cost analysis is performed for the developed continuous process to understand its
economic feasibility. Comparative analysis is then presented between the developed pro-
cess based on microalgae and other processes using waste cooking oil (WCO) as feedstock.
Later, these processes are discussed to portray the merits and demerits of the developed
process.
262 Process Systems Engineering for Biofuels Development

10.2 Process Development and Modeling

The continuous process developed and studied in this chapter is based on Martinez-Guerra
et al. (2018), who studied experimentally the synergic effect of MW and US radiations
for biodiesel production from microalgae biomass (Nannochloropsis sp.) as raw material.
They also studied reaction kinetics for in situ transesterification of algal biomass to produce
algal biodiesel (i.e. fatty acid methyl ester, FAME). To optimize the process variables and
to understand their parametric interdependence, response surface methodology along with
central composite design was used by Martinez-Guerra et al. (2018).
The microalgae paste used by Martinez-Guerra et al. (2018) in their study, contained
18.4% of biomass (by dry weight) and the rest water, and dry biomass composition was
52% protein, 0.89% chlorophyll, 16% carbohydrates, and 27% lipids. Components in the
remaining fraction (100 − 52 − 0.89 − 16 − 27 = 4.11%) are not stated in the original refer-
ence. Hence, in our simulation, the remaining fraction is equally distributed among proteins
and carbohydrates to make it 100%.
In the present work, all lipids in the above microalgae paste are represented as triolein,
for simplicity. This representation was employed by several researchers such as West et al.
(2008), Morais et al. (2010), and Sharma and Rangaiah (2013), who have used triolein to
represent lipids in WCO. This assumption is justified as the reaction kinetics followed in
this study is for complete microalgal oil. Hence, the product produced will still be realis-
tic. Although the esters produced from these lipids may have different properties such as
density, viscosity, etc. (and hence different separability as well), they all will separate into
the same light phase in phase separators (PS-1 and PS-2). Hence, assuming all lipids as
triolein is reasonable for process simulation. However, it may show some discrepancy in
the heat duties of reactors and distillation columns owing to different properties. In addi-
tion, properties related to combustion, extent of (un)saturation etc. will affect the biodiesel
quality. However, given that the focus of the present study is not on the biodiesel quality but
on the overall process design and techno-economic analysis, this assumption is reasonable.
Further, this study can be modified for more accurate simulation and analysis if contents of
microalgae lipids are available.
The experimental study in Martinez-Guerra et al. (2018) was carried out in the presence
of a base catalyst, namely, NaOH (sodium hydroxide). Twenty grams of microalgae paste
was added to a homogenous mixture of catalyst and methanol (MeOH, CH3 OH), which
was then subjected to the synergic effect of MW and US radiations. The transesterification
reaction with MeOH is described in Eq. (10.1), wherein triglycerides react with alcohol to
form esters and glycerol.

Triglycerides + 3 CH3 OH ↔ 3 R′ COOCH3 + Glycerol (10.1)

where R’ signifies the alkyl group in the triglycerides.

In Martinez-Guerra et al. (2018), reaction time, US and MW power along with catalyst
and methanol flow rates were varied. A reaction time of seven minutes, MW and US power
of 140 W each, wet algal biomass to methanol ratio of 20 g to 30 ml and a catalyst concen-
tration of 1 wt% were found to be the optimum parameters. The obtained product mixture
was washed with 20 ml of hexane and 10 ml of deionized water. Then, the methanol-hexane
phase containing FAME was separated and the procedure was repeated.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 263

The reaction kinetics for the reaction (Eq. (10.1)) was found to be first order, and the rate
constant is given as (Martinez-Guerra et al. 2018):
ln(Xt ) − ln(X0 )
k= (10.2)
t
where Xt is FAME yield at any time t and Xo is FAME yield at t = 0. Activation energy is
found using the Arrhenius equation as:

k = Ae−Ea ∕RT (10.3)

where A is the frequency factor (min−1 ), Ea is the activation energy (J/mol), R is the univer-
sal gas constant (8.314 J/mol/K), and T is the reaction temperature (K). The frequency factor
and activation energy were determined to be 70.52/minutes and 17 298 J/mol, respectively
(Martinez-Guerra et al. 2018).
Based on the experimental procedure in Martinez-Guerra et al. (2018) and our previous
studies on the biodiesel process (Sharma and Rangaiah 2013; Patle et al. 2014), a continuous
process for biodiesel production from microalgae is developed in the present study, and then
it is simulated in Aspen Plus version 8.8. Figure 10.1 depicts the major steps involved in the
present study. A non-random two-liquid (NRTL) model was selected and used for property
estimation of liquid phases and the ideal gas model for vapor phases. Note that this model
was also used in Piemonte et al. (2016). Figure 10.2 shows the flowsheet of the proposed
process, where F is the mass flow rate of the stream in kg/h, T is the temperature in ∘ C,
and square brackets represent the composition of the stream in terms of mass fraction in
the order [xLIPID , xMeOH , xWATER , xHEXANE , xFAME , xGLYCEROL ].
As stated earlier, lipids in the microalgae were approximated as trioleins (triglycerides of
oleic acid); further, proteins are taken to be L-phenylalanine and carbohydrates as sucrose
from the Aspen Plus database. The kinetic parameters (given above) for the transesterifi-
cation reaction are taken from Martinez-Guerra et al. (2018). The extraction of lipids from
microalgal oil is not modeled separately in Aspen Plus simulation as it was not studied sep-
arately in Martinez-Guerra et al. (2018), based on which the current process simulation and
design are carried out. Note that kinetics in this reference is for in situ transesterification
and so they include extraction kinetics.
Suitable unit operation modules, as given in Table 10.1, were chosen for simulating the
process shown in Figure 10.2. Then, process parameters and reaction kinetics were added.
After obtaining simulation results for reactor similar to those in Martinez-Guerra et al.
(2018), optimum design parameters such as number of stages, feed stage, and column
pressure of distillation columns, were determined to minimize total annual cost (TAC) of
the respective distillation column. Thus, the developed process is sufficiently realistic and
optimal.
Since the entire concept of in situ suggests the simultaneous extraction and transester-
ification of biomass, microalgae are fed to the continuous reactor “RTRANS” (simulated
by RCSTR block in Aspen Plus, which is based on the model for a continuous stirred-tank
reactor) at a flowrate of 50 322 kg/h, along with fresh methanol at 663.972 kg/h and cata-
lyst (NaOH) at 26 kg/h (Figure 10.2). Wet microalgae flowrate of 50 322 kg/h was chosen
to process 2500 kg/h of lipids based on the lipid content (= 0.184 × 0.27 × 50 322 kg/h, as
mentioned above). Both extraction and reaction in the continuous reactor are aided by MW
264 Process Systems Engineering for Biofuels Development

Start

Component Module
Addition Selection

Process
Simulation
Process
Property
Parameters
Method
and Reaction
Selection
Kinetics

Results NO
Validation

YES

Variation of Process
Parameters targeting min. TAC
such as: Number of stages,
Feed stage and Pressure of a
Distillation Column

Optimality NO
(min. TAC)

YES

Cost Estimation
of Optimal
Process Model

End

Figure 10.1 Flow chart depicting the steps followed in the present study.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 265

METHANOL
F = 663.997 kg/h
T = 25°C RTRANS
[0,1,0,0,0,0] CATALYST SEP–1 SEP–2
H3PO4

M–1
RNEUT
H–1 PRO-CARB SALT
F = 32.02 kg/h
T = 30°C Qc = –11.72 MW
[0,0,0,1,0,0] Qb = 8.41 MW

5 P–1
PS–1 M-4
M–2
9
Split
10
H–2 FRAC–1
F = 2.66 kg/h
T = 30°C
M–3
[0,0,0,1,0,0] PS–2

P–2
Qc = –47.96 MW
HX–1
Qb = 48.75 MW

Qc = –0.35 MW
Qb = 1.02 MW
4 FRAC–2 P–3 HX–2

7
8 GLYCEROL 7 P–5
F = 283.01 kg/h FRAC–4
METH-REC
P–4 T = 143.25 °C
[0,0,0.041,0,0.05,0.908]
13 HX–3
F = 43983 kg/h
14
T = 63°C
[0,0.89,0.04,0.07,0,0]

Qc = –26.83 MW
Qb = 31.28 MW

7 FRAC–3
HX–4
HEX-REC 12
F = 101277 kg/h P–6
13 WAT-OUT
T = 56°C TG-REC
F = 41462.8 kg/h F = 2297.13 kg/h
[0,0.027,0,0.973,0,0]
T = 30 °C T = 40°C
[0,0.009,0.99,0,0,0] [0.998,0,0,0,0.002,0]

FRESH ALGAE
F = 50322 kg/h
M–5 T = 30°C
[0.05,0,0.816,0,0,0]
WATER OUT
BIODIESEL
F = 2492.81 kg/h WATER IN
T = 25 °C
[0.02,0,0,0,0.98,0] WASH TOWER

Figure 10.2 Biodiesel production from microalgae “Nannochloropsis sp.” Values in square brackets are
the stream composition in mass fraction in the order [xLIPID , xMeOH , xWATER , xHEXANE , xFAME , xGLYCEROL ].
266 Process Systems Engineering for Biofuels Development

Table 10.1 Process parameters of the proposed biodiesel process.

Unit operation and parameters Design value

Distillation columns
FRAC-1 (RADFRAC)
No. of stages 10
Feed stage 5
Pressure (atm) 0.05
Reflux ratio (molar) 0.0143
Qr (MW) 8.414
Qc (MW) −11.728
FRAC-2 (RADFRAC)
No. of stages 8
Feed stage 4
Pressure (atm) 0.4
Reflux ratio (molar) 0.1
Qr (MW) 48.75
Qc (MW) −47.96
FRAC-3 (RADFRAC)
No. of stages 13
Feed stage 7
Pressure (atm) 1.0
Reflux ratio (molar) 1.0
Qr (MW) 31.82
Qc (MW) −26.83
FRAC-4 (RADFRAC)
No. of stages 14
Feed stage 7
Pressure (atm) 0.06
Reflux ratio (molar) 0.77
Qr (MW) 1.02
Qc (MW) −0.35
Reactors
RTRANS (CSTR)
Temperature (∘ C) 55.0
Pressure (atm) 1.0
Residence time (min) 7.0
RNEUT (RStoic)
Temperature (∘ C) 30.0
Pressure (atm) 1.0
Phase separators
PS-1 (DECANTER)
Duty (kW) 0.0
Pressure (atm) 1.0
Volume (m3 ) 149
PS-2 (DECANTER)
Duty (kW) 0.0
Pressure (atm) 1.0
Volume (m3 ) 105
Wash tower (EXTRACT)
No. of stages 5
Pressure (atm) 1.0
Heat exchangers (HEATX)
Temperature approach of HX-1, HX-2, HX-3, and HX-4 10.0
 Process Development, Design and Analysis of Microalgal Biodiesel Production 267

and US. The reactor is maintained at 55 ∘ C and 1 atm, and the residence time of reaction
mixture in the reactor is specified as seven minutes.
The reactor outlet mixture contains FAME (biodiesel), catalyst, proteins, carbohydrates,
chlorophyll, methanol, glycerol, hexane, and unreacted lipids. There are many methods for
separation of proteins/carbohydrates from biomass such as enzymatic hydrolysis (Fleurence
et al. 1995a; Joubert and Fleurence 2008; Harnedy and FitzGerald 2013), physical treatment
(Barbarino and Lourenço 2005; Harnedy and FitzGerald 2013), and chemical extraction
(Fleurence et al. 1995b; Harnedy and FitzGerald 2013; Kadam et al. 2017). For simplicity,
a component separator block (SEP-1) is used in the present simulation (Figure 10.2), to
separate proteins and carbohydrates from the other components.
After SEP-1, the catalyst (NaOH) is neutralized in the neutralization reactor “RNEUT”
(simulated by RStoic block in Aspen Plus, which is based on reaction stoichiometry) by
addition of phosphoric acid (H3 PO4 ). The salt thus formed is removed in SEP-2, which is
also modeled as a component splitter in the present simulation. Component splitter is used
to model both SEP-1 and SEP-2 for simplicity. Complete separation of proteins, carbohy-
drates, etc. in SEP-1, and of salts in SEP-2 is assumed. Modeling the separation of proteins,
carbohydrates, etc. in SEP-1 is difficult due to lack of required properties of components.
The remaining liquid mixture from SEP-2 containing FAME, water, methanol, unreacted
lipids, glycerol, and hexane, is sent to phase separators, PS-1 and PS-2 (each modeled as a
decanter block, i.e. separation of aqueous and organic phases by density difference) oper-
ating at 25 ∘ C and 1 atm.
A certain amount of hexane is fed to each phase separator, for facilitating FAME extrac-
tion into the organic phase. The lighter/organic phase with FAME, unreacted lipids and hex-
ane is collected from both PS-1 and PS-2, and then sent to a distillation column, FRAC-1,
for further separation. The Aspen Plus block used for FRAC-1, FRAC-2, FRAC-3, and
FRAC-4 in Figure 10.2 is RADFRAC, which rigorously models a distillation column via
mass, energy and vapor–liquid equilibrium balances. FRAC-1 has a total of 10 stages with
feed entering on the 5th stage. Note that stages throughout this chapter refer to ideal or equi-
librium stages, and the number of stages includes condenser and reboiler of the column.
Hexane is recovered as the distillate of FRAC-1 for recycling to PS-1 and PS-2.
The bottoms stream of FRAC-1 is fed to the 7th stage of FRAC-4 having 14 stages. The
distillate of FRAC-4, containing primarily FAME with a mass purity of ∼98%, is cooled
using the hexane stream at −12 ∘ C from FRAC-1 and then sent to a wash tower to remove
any other water-soluble impurities. The unreacted lipids are recovered in the bottoms stream
of FRAC-4 and cooled to 40 ∘ C using the hexane stream (used earlier for cooling the dis-
tillate stream), for recycling to the extraction/transesterification reactor, RTRANS.
The heavier/aqueous phase from PS-1 is fed to PS-2 for recovering the remaining FAME
in it by adding some more hexane. The heavier phase from PS-2 containing methanol,
water, and glycerol (plus traces of FAME and hexane), is fed to the 4th stage of FRAC-2
having eight stages. Glycerol is obtained from the bottoms of FRAC-2 along with a small
quantity of FAME (∼17 kg/h out of 2463 kg/h of biodiesel produced from 2500 kg/h of
lipids). A mixture of water, methanol, and traces of hexane is recovered as the distillate, and
it is further separated in another distillation column, FRAC-3, having 13 stages. Methanol
with traces of hexane is obtained as the distillate of FRAC-3, and it is recycled back to the
upstream process units. Water leaves the bottoms of FRAC-3 and also the process.
268 Process Systems Engineering for Biofuels Development

Three distillation columns in the process, namely, FRAC-1, FRAC-2, and FRAC-4
in Figure 10.1, are operated under vacuum (0.1–0.25 atm) in order to avoid deteriora-
tion/decomposition of FAME and glycerol at high temperatures. FRAC-3 does not have
any such restriction, and so it is operated at atmospheric pressure. Note that FAME
decomposes at 250 ∘ C and glycerol decomposes at 150 ∘ C (Morais et al. 2010). Owing
to vacuum operation, the overhead temperature of FRAC-1, FRAC-2, and FRAC-4 is,
respectively, −12, 22, and 212 ∘ C. The high temperature in the condenser of FRAC-4, i.e.
212 ∘ C, allows the use of energy removed in its condenser in some reboiler. However,
the small flow rate of the distillate from FRAC-4 may not provide significant energy.
Nevertheless, this heat integration is considered beyond the scope of the present work.
However, recycled hexane at −12 ∘ C would result in unnecessary heating requirement
in PS-1 and PS-2; to avoid this heating, recycled hexane is used to cool three streams,
namely, TG-REC, Biodiesel, and WAT-OUT in HX-2, HX-3, and HX-4 respectively, to
room temperature, which results in recycled hexane heated to 56 ∘ C. Refrigerant is used
in the condenser of FRAC-1, whereas chilled water is used in the condenser of FRAC-2
and cooling water is used in the condenser of FRAC-3 and FRAC-4. On pumping the
distillate of FRAC-1, FRAC-2, and FRAC-4 to atmospheric pressure from vacuum condi-
tions (0.1–0.25 atm), a temperature increase of 1–2 ∘ C was observed. Taking algal biomass
flowrate as the base, flow rates of all inputs (hexane, water, and methanol) are defined by
the calculator block in Aspen Plus. Data for these and other important streams are given in
Table 10.A1 in the Appendix whereas process parameters used for simulation are listed in
Table 10.1.
The liquid composition and temperature profiles of all the four columns are depicted in
Figure 10.3. In FRAC-1, which is used for the recovery of hexane from the lighter phase
collected from PS-1 and PS-2, the highest temperature is near 250 ∘ C (Figure 10.3a). Note
that this is required to be maintained at/below 250 ∘ C to avoid biodiesel decomposition.
About 99% of hexane is recycled from the distillate, and a mixture of FAME and unreacted
triglycerides is obtained from the bottoms of FRAC-1. In FRAC-2, the maximum temper-
ature is lower than the temperature at which glycerol starts decomposing, i.e. 150 ∘ C. It
can be seen from Figure 10.3b that glycerol is obtained from FRAC-2 bottom with 90%
mass purity along with 6% FAME and 4% water, while a mixture of water and methanol
is obtained from the top of FRAC-2. FRAC-3, which separates a mixture of water and
methanol, is operated at atmospheric pressure (Figure 10.3c). Methanol with 89% mass
purity is obtained from the top of FRAC-3, whereas water is removed from this column
bottom. FRAC-4 separates biodiesel and unreacted lipids, which are fairly easy to separate
due to the large difference in their volatilities; however, it operates at relatively high temper-
atures despite vacuum pressure because of the high boiling point (∼847 ∘ C at atmospheric
pressure) of lipids. Biodiesel i.e. FAME is obtained as the distillate from the top with 98%
purity (as required to meet ASTM and EN standards). The unreacted lipids are obtained
in the FRAC-4 bottoms stream, which is then cooled and recycled back along with fresh
microalgae to the transesterification reactor.
Number of stages in each of the columns (FRAC-1, FRAC-2, FRAC-3, and FRAC-4)
was decided by performing sensitivity analysis. Targeting minimum reboiler duty and TAC
(defined in Section 10.3) of the column, the number of stages was varied in Aspen Plus
simulation. For this, inlet stream flow rate, temperature and composition to each column is
fixed based on preliminary simulations.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 269

Figure 10.4 shows the trend of reboiler duty and TAC with number of stages. For
FRAC-1, reboiler duty decreases and TAC increases as the number of stages increases.
For FRAC-2, reboiler duty and TAC are practically constant within the range of number
of stages examined. With increasing number of stages in FRAC-3, both reboiler duty and
TAC decrease to a certain value and then increase. In the case of FRAC-4, both reboiler
duty and TAC decrease and reach a certain plateau with increasing number of stages. These
trends are due to the combined effect of reboiler duty decrease and capital cost variation
(increase due to a taller column or decrease due to a smaller diameter) with increasing
number of stages. The optimum number of stages is chosen to minimize TAC, ensuring
that height to diameter ratio of the column is below 20 (Seider et al. 2009). Accordingly,
the optimal number of stages for FRAC-1, FRAC-2, FRAC-3, and FRAC-4 is 5, 6, 11, and

1.00
Composition

0.75 TG
FAME
0.50 HEXANE
Temperature
0.25
0.00
Temperature(°C)

213

142

71

0
2 4 6 8 10
No. of stages
(a)

1.0
Composition

0.8 MEOH
WATER
0.6 FAME
GLYCEROL
0.4 HEXANE
Temperature
0.2
0.0
140
Temperature(°C)

120
100
80
60
40
20
2 4 6
No. of stages
(b)

Figure 10.3 Liquid composition (mass fraction) and temperature profiles: (a) FRAC-1, (b) FRAC-2, (c)
FRAC-3, and (d) FRAC-4.
270 Process Systems Engineering for Biofuels Development

1.0

0.8
Composition MEOH
0.6
WATER
HEXANE
0.4 Temperature

0.2

0.0
100
Temperature(°C)

90

80

70

2 4 6 8 10
No. of stages
(c)

1.00

0.75
Composition

TG
0.50 FAME
HEXANE
Temperature
0.25

0.00

300
Temperature(°C)

275

250

225

2 4 6 8 10 12
No. of stages
(d)

Figure 10.3 (Continued)

 Process Development, Design and Analysis of Microalgal Biodiesel Production 271

12, respectively, excluding condenser and reboiler. Although the optimal number of stages
is 5 for FRAC-1 in Figure 10.4a, it is chosen as 10 for subsequent analysis as the height to
diameter ratio is unusually small (i.e. <1) for five stages. Increase in TAC due to this is
quite small.
The biodiesel process presented in this chapter is designed to obtain maximum FAME
with small loss of 0.7% (= 19 kg/h) of total biodiesel produced, in the glycerol stream;
around 1.9% (= 48 kg/h) of lipids goes out in the FAME stream. Further, operating con-
ditions of the extraction/transesterification reactor are kept close to the conditions in the
laboratory experiments, and column operating temperatures are kept below the decompo-
sition temperature of FAME and glycerol.
TAC (Million $) Reboiler duty (MW)

8.500
8.475
8.450
8.425
8.400
5.94
5.67
5.40
5.13
4 6 8 10 12
No. of stages
(a)
48.75940
Reboiler duty (MW)

48.75935

48.75930

48.75925

48.75920
TAC (Million $)

18.3

18.2

18.1
4 6 8
No. of stages
(b)

Figure 10.4 Effect of number of stages on reboiler duty and TAC: (a) FRAC-1, (b) FRAC-2, (c) FRAC-3,
and (d) FRAC-4.
272 Process Systems Engineering for Biofuels Development

TAC (Million $) Reboiler duty (MW)

38

36

34

32

9.0

8.5

8.0
6 8 10 12
No. of stages
(c)
TAC (Million $) Reboiler duty (MW)

1.8
1.6
1.4
1.2
1.0
0.8

2.0

1.5

1.0

8 10 12 14
No. of stages
(d)

Figure 10.4 (Continued)

10.3 Sizing and Cost Analysis

All equipment in the biodiesel process in Figure 10.2 is sized using the procedures in Seider
et al. (2009) and Luyben (2002), except for SEP-1, SEP-2, and all mixers. For sizing the
reactors and reflux drum and sump of distillation columns, the length to diameter (L/D)
aspect ratio is taken to be 2:1. This ratio decides controllability of the process. Typically,
a L/D ratio of 2:1 is used to obtain reasonable control (Luyben 2002). The volume of the
reactor is determined based on the residence time required. The tray sizing option in Aspen
Plus is used for the column diameter calculations, where the column diameter is determined
based on maximum vapor velocity. This was manually verified using the procedure given
in Luyben (2002).
 Process Development, Design and Analysis of Microalgal Biodiesel Production 273

Tray efficiency. Eo is calculated using O’Connell’s correlation (Towler and Sinnott 2013)
given below:
E0 = 51 − 32.4 log (𝜇a 𝛼a ) (10.4)

where 𝜇a is the average liquid viscosity (mNs/m2 ) and 𝛼 a is the average relative volatility
of light key to heavy key in the column. Liquid viscosity was obtained from the column
hydraulics from Aspen plus simulator. The estimation of 𝛼 a requires the partial pressure
of each component in the multicomponent mixture, which was found using the Antoine
equation (Dutta 2009):
′ B′
ln(PVA ) = A − ′ (10.5)
C +𝜃
where PVA is the vapor pressure of component A (mm Hg), 𝜃 is the temperature (∘ C), and A′ ,
B′ , and C′ are Antoine coefficients. Since the tray efficiency is found to be about 90%, the
number of real stages is assumed to be same as the number of ideal stages. Column height
is found by multiplying the number of stages by tray spacing (in the range of 0.609–0.8 m)
and adding 20% of the calculated height to account for space at the top and bottom of the
column (excluding the space for trays).
Table 10.2 shows the dimensions and cost of equipment involved in the biodiesel process.
After sizing each of the units in the process, cost estimation was done using CAPCOST,
an MS-Excel-based program for estimating costs of many types of equipment in process
industries. This program uses cost correlations and procedures from Turton et al. (2009)
for cost estimation, and it is available with this book. The design values such as reactor
volume, distillation column diameter, distillation column height, heat exchanger area, etc.
were obtained as outlined in the previous paragraph. The material of construction was cho-
sen to be carbon steel for all units except for RTRANS and RNEUT, which are assumed to
be made of stainless steel as these two units handle acids and bases. Cost of required MW,
US, SEP-1, and SEP-2 is not included in the cost estimation.
All equipment costs are inflated using Chemical Engineering Plant Cost Index (CEPCI)
of 602 for 2018 (Jenkins 2018). Then, bare module cost (BMC) is calculated. These costs
are shown in Table 10.2. Subsequently, total module cost (TMC) is determined using the
following equation (Turton et al. 2009).

TMC = 1.18 BMC (10.6)

Capital investment is assumed to be equal to the TMC. Based on BMC of $9 423 669 for
all equipment in Table 10.2, TMC for the biodiesel process is $11 308 403. This is as per
procedures and CAPCOST program in Turton et al. (2009).
BMC for the biodiesel process in Figure 10.2 is estimated using two other costing pro-
grams, namely, Aspen Process Economic Analyzer (APEA) available within Aspen Plus
and CCEP (developed by Feng and Rangaiah 2011 based on cost correlations in Seider
et al. 2009) besides CAPCOST (based on cost correlations in Turton et al. 2009). These
results are presented in the last two columns of Table 10.2.
In both CAPCOST and CCEP programs, there are upper limits on tower (column) diam-
eter and heat transfer area for reboiler and condenser. In the case where equipment is larger
than these limits, both programs assume multiple units (of same size), which are equivalent
to the large equipment, and then the cost of multiple units is taken to be the cost of the large
274 Process Systems Engineering for Biofuels Development

Table 10.2 Size and cost of equipment in the proposed biodiesel process.

Bare module Bare module Bare module

Height Diameter cost using cost using sost using
(m) (m) CAPCOST ($) CCEP ($) APEA ($)
Distillation columns FRAC-1 7.68 6.45 2 483 524 3 171 200 1 923 200
(including cost of FRAC-2 5.76 5.13 3 512 219 3 155 800 2 259 100
condenser and FRAC-3 8.04 3.82 1 664 232 1 407 900 1 383 300
reboiler)∗ FRAC-4 8.77 2.20 513 281 797 300 523 100
Subtotal for distillation 8 173 256 8 532 200 6 088 700
columns
RTRANS 17.02 106 688 236 101 305 600
Reactors (volume in m3 )
RNEUT 13.05 87 644 169 750 297 500
Subtotal for reactors 194 332 405 851 603 100
P-1 5.21 19 299 18 300 62 600
P-2 0.47 14 795 13 400 33 700
P-3 2.374 16 292 12 900 58 300
Pumps (power in kW)
P-4 0.014 14 795 16 100 28 100
P-5 0.295 14 795 12 900 33 100
P-6 0.174 14 795 12 900 30 100
Subtotal for pumps 94 771 86 500 245 900
Phase separators∗ PS-1 149 470 000 412 535 205 300
(volume in m3 ) PS-2 105 341 000 365 374 179 600
Subtotal for phase 811 000 777 909 384 900
separators
Wash column WASH TOWER 1.12 20 340 27 521 25 000
(volume in m3 )
HX-1 8.2 11 800 63 700 69 000
Heat exchangers (heat HX-2 1.9 8610 63 600 69 000
transfer area in m2 ) HX-3 104.9 120 000 114 700 106 100
HX-4 3.5 9 900 63 700 69 000
Subtotal for heat 150 310 305 700 313 100
exchangers
Grand total 9 423 669 10 115 560 7 648 640
∗ These
values are found by directly using the cost correlations in Turton et al. (2009) or Seider et al. (2009) instead of
CAPCOST or CCEP programs.

equipment. The cost estimate obtained thus is more than the cost of large equipment found
by directly using the cost correlation (i.e. beyond its stated range). Thus, the latter provides
a lower estimate. Such a lower estimate, i.e. directly using the cost correlations from Turton
et al. (2009) and Seider et al. (2009) instead of CAPCOST and CCEP, respectively, is used
for tower (column), reboiler and condenser in the developed biodiesel process. Also, cost
correlations are used for the cost estimation of phase separators as CAPCOST and CCEP
do not have it. These cost estimates are identified in Table 10.2.
It is seen from Table 10.2 that the BMC estimate generated by CAPCOST and CCEP are
similar for most of the equipment. The major differences arise in the costing of distillation
columns, mainly for FRAC-1 and FRAC-2 probably due to their large diameter of 6.45 and
5.13 m, respectively. BMC estimated by APEA is quite low for FRAC-1 and FRAC-2 in
comparison with that given by CAPCOST and CCEP.
For the developed biodiesel process, the total BMC estimated by CCEP is the highest
(at $10 115 560), followed by that by CAPCOST (at $9 423 669) whereas the estimate
 Process Development, Design and Analysis of Microalgal Biodiesel Production 275

by APEA is the lowest (at $7 648 640). In other words, the BMC estimate can differ
significantly depending on cost correlations; this observation is similar to that in Feng
and Rangaiah (2011) for some case studies. This variation should be kept in mind when
comparing capital cost estimates from different papers unless those studies used exactly
the same correlations and procedures (e.g. from the same research group).
As can be seen from Table 10.2, a large share of BMC (87% as per CAPCOST, 84% as per
CCEP, 80% as per APEA) is for the four distillation columns; in fact, FRAC-1 and FRAC-2
account for 64, 63, and 55% of BMC as per CAPCOST, CCEP, and APEA, respectively. The
main reason for this is column diameter, which is more than 5 m for two of the four columns.
Large column diameter is in turn due to its high feed flowrate. For instance, feed enters
FRAC-1 at 106 100 kg/h which leads to large vapor flowrate inside the column ranging from
41 439 to 102 767 kg/h. To handle this amount of feed and high vapor flowrate in the column,
a column of large diameter is required. The case for FRAC-2 and FRAC-3 is similar. On
the other hand, FRAC-4 processes about 4790 kg/h of feed (FAME and triglycerides), and
so it has a smaller diameter of 2.2 m.
For optimizing each of the distillation columns (presented earlier in Figure 10.4), TAC
is calculated using the following equation.
Capital investment
TAC = Operating cost + (10.7)
Payback period
Here, the operating cost is considered to be the cost of utilities used (i.e. steam, refriger-
ant, and cooling water) as in Babu et al. (2012) and Sharma et al. (2018). This is reasonable
for optimizing a column since other operating costs such as for feed/inlet stream(s), operat-
ing labor, and maintenance are nearly constant. Unit cost of utilities, namely, steam, cooling
water, and refrigerant are taken from Luyben (2011), and they are given in Table 10.3;
payback period of three years is assumed.
For all inclusive cost estimation of biodiesel production, cost of manufacture (COM) is
calculated using the following equation.

COM = 0.28 CTM + 2.73 (operating labor cost)

+ 1.23 (cost of utilities + cost of raw materials) (10.8)

Table 10.3 Utilities and their costs in the proposed biodiesel process.

Unit cost Annual utility

Quantity ($/GJ) cost (for 8000 h)
Utility (GJ/h) (Luyben 2011) ($)
Low pressure steam (6 bar, 160 ∘ C) 290.08 7.78 17 823 034
High pressure steam (42 bar, 250 ∘ C) 30.29 9.88 2 394 360
Very high pressure steam (106 bar, 315 ∘ C) 3.79 24.82∗ 753 129
Refrigerant at −20 ∘ C 42.22 7.89 2 665 268
Cooling water at 25–30 ∘ C 107.91 0.354 305 620
Chilled water at 5–15 ∘ C 172.67 4.43 6 119 576
Electricity 0.03 16.8 4088
Total 30 065 074
∗ This value is taken from Sharma and Rangaiah (2013).
276 Process Systems Engineering for Biofuels Development

Unit cost of raw materials is as follows: wet microalgae (0.034 $/kg), hexane (1.0 $/kg),
methanol (0.4 $/kg), sodium hydroxide (33 wt%) (0.4 $/kg), phosphoric acid (0.034 $/kg),
and glycerol (90 wt%) (1.17 $/kg). These costs were obtained from www.molbase
.com (in May 2019). Quantities and costs of utilities used in the developed process are
summarized in Table 10.3. Nearly 70% of the cost of utilities for the biodiesel process
is for steam (i.e. hot utility), about 29% is for low-temperature cold utilities (namely,
chilled water and refrigerant), and the cost contribution of cooling water and electricity
is very small. The operating labor is calculated following the procedure given in Turton
et al. (2009).
Using Eq. (10.8), COM of $65 054 447 is obtained for the biodiesel process in
Figure 10.2. Two major contributors of COM are microalgal feed and utilities (i.e. the last
term including the multiplier 1.23 in Eq. (10.8)), which account for ∼26% and ∼57% of
COM, respectively. Thus, cost of utilities (i.e. energy consumption) is very significant
in the developed biodiesel process. On the basis of COM obtained and considering the
selling price of glycerol as 1.17 $/kg, breakeven cost of biodiesel comes to approximately
3.13 $/kg (10.2 $/gal), which is very high.1 This cost is, however, comparable with the
range of $10.87–$13.32/gal of biodiesel reported in Sun et al. (2011). COM for the
biodiesel process can be reduced by further optimization, use of energy-efficient separation
processes, heat integration, etc., which should be explored in future work.
Reboiler and condenser duties of distillation columns in the developed biodiesel
process are presented in Figure 10.2; the sum of all reboiler duties is 89.50 MW or
130 MJ/kg or 112 MJ/l assuming biodiesel density of 860 kg/m3 . Higher heating value of
produced biodiesel is 39.81 MJ/kg (Martinez-Guerra et al. 2018) whereas specific energy
for fossil-based diesel is 48.1 MJ/kg (https://en.wikipedia.org/wiki/Energy_content_of_
biofuel [in June 2019]). This means that the energy required for 1 kg of biodiesel is nearly
three times the energy that would be obtained from it. Hence, energy consumption, mainly
for downstream separation, in the developed biodiesel process has to be reduced very
substantially by a factor of 10 or so.
As presented in Table 10.3, the developed process requires a large amount of utilities
(resulting from the large amount of feed with water to be processed and consequently large
amount of hexane and methanol recycle, and separation of large amount of water), which
increases the operating cost. Overall, the increased capital and utilities costs (Tables 10.2
and 10.3) are attributed to the need for processing large amounts of feed and other chemi-
cals, which require separation using distillation columns.
As given in Table 10.A1, methanol recycle is comparable with wet microalgae feed
flowrate whereas hexane recycle is twice the wet microalgae feed flowrate. Such high recy-
cles are due to using the operating conditions (including methanol and hexane quantities)
of laboratory experiments reported by Martinez-Guerra et al. (2018) for the developed
biodiesel process. It may be possible to optimize these conditions and/or develop better
separation processes, and consequently reduce recycle flowrates in large-scale continuous
processes.

1
If the monetary value of glycerol (produced at 283 kg/h) is not considered, then breakeven cost of
biodiesel increases slightly to $3.26/kg.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 277

10.4 Comparison with the WCO-Based Process of the Same Capacity

10.4.1 Biodiesel Process Using WCO as Raw Material
Sharma and Rangaiah (2013) studied biodiesel production from WCO. They simulated the
process in Aspen HYSYS simulator and then optimized it for multiple objectives. Vegetable
oils are mainly a mixture of triglycerides of oleic, linoleic, linolenic, palmitic, stearic and
other acids. Since the physical properties of these triglycerides are more or less similar
(Myint and El-Halwagi 2009), Sharma and Rangaiah (2013) considered triglycerides of
oleic acid (i.e. triolein) as the feedstock. The physical properties were taken from Aspen
Plus database, and the thermodynamic model selected was the UNIQUAC model.
As shown in Figure 10.5, the process studied by Sharma and Rangaiah (2013) has two
main sections, namely, pretreatment (i.e. esterification) of WCO to convert free fatty acids to
biodiesel, and transesterification of treated WCO to produce biodiesel. In the pretreatment
section, free fatty acids present in the WCO react with methanol in the presence of sulfuric
acid in the esterification reactor maintained at 60 ∘ C. The reactor outlet stream is used to
preheat the in-coming WCO stream. Subsequently, the reactor effluent is mixed with fresh
glycerol and sent to a three-phase separator operating at 40 ∘ C. The lighter phase formed
contains biodiesel, unreacted oil, and methanol, whereas the heavier phase contains catalyst,
glycerol, water, and methanol. The lighter phase is sent to a distillation column, BD-C1,
and the heavier phase to another column, G-C1.
The majority of the methanol is recovered as distillate of both these columns, and it
is recycled back to the esterification reactor. Catalyst (sulfuric acid), glycerol, and water
are removed from the bottoms of G-C1. The bottoms stream of BD-C1 mainly contains
biodiesel, oil (i.e. triglycerides), and methanol, and it is sent to the transesterification
section (Figure 10.5), wherein three CSTRs are used in series with phase separators in
between, for the reaction of triglycerides with methanol in the presence of alkali catalyst
to produce biodiesel.
Methanol and alkali catalyst are mixed and distributed to three CSTRs (Figure 10.5). Part
of this stream is mixed with the bottoms stream of BD-C1, and then sent to the first CSTR.
The effluent from the first two CSTRs is sent to a three-phase separator. The lighter phase
containing glycerol and methanol is separated from both the three-phase separators and
then sent to the column G-C2. The heavier phase containing unconverted triglycerides and
basic catalyst is sent to the next CSTR. The product stream from the third CSTR is sent to
column BD-C2. Methanol is recovered in the distillate of the G-C2 and BD-C2 columns and
then recycled. Glycerol and biodiesel are obtained from the bottoms of G-C2 and BD-C2,
respectively. The bottom product of BD-C2 is sent to a neutralization reactor to remove
alkali catalyst, followed by a wash column to produce the desired biodiesel. The process
designed and studied by Sharma and Rangaiah (2013) uses 2500 kg/h of WCO, which is
the same as the lipid content in the wet microalgae used in the present study (Figure 10.2).
Biodiesel is produced at 2512 kg/h from WCO in Sharma and Rangaiah (2013), and at
2493 kg/h from wet microalgae in the present study.

10.4.2 Comparative Analysis

Table 10.4 compares the number of pieces of equipment and capital cost (TMC) for the
developed process (i.e. intensified in situ biodiesel production from microalgae) and the
278 Process Systems Engineering for Biofuels Development

Methanol-1 Recycled Methanol

Acid catalyst
P–1
CSTR0 1

BD-C1 P–3

HE–1 5
Waste oil
P–2 BD-T

P–5

Glycerol 3-phase P–4 HE–2

G-C1
separator–0
4

G-acid (glycerol,
acid and water)

F–1

Water BD (bio-diesel)
CSTR1

3-phase
separator–1 Waste
S1-H F–2 Water
(organic)
wash
column

CSTR2
3-phase Acid
Separator–2
F–3

Neut.
reactor
CSTR3
HE–4
1 1 HE–3

G-C2 BD-C2 P–8

9 14
P–6 P–7 Alkali catalyst

Methanol–2
G-pdt (glycerol)
Recycled Methanol

Figure 10.5 Biodiesel production from WCO. The top part of the process flow diagram is the pretreat-
ment (esterification) section whereas the bottom part is the transesterification section. Source: Sharma and
Rangaiah 2013. Reproduced with permission of Elsevier.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 279

Table 10.4 Comparison between the biodiesel process based on microalgae feedstock and the biodiesel
process using WCO feedstock.

Process using Process using

microalgae feedstock WCO feedstock
(present study) (Sharma and Rangaiah 2013)
Reactors 2 5
Distillation columns 4 4
Number of pieces of
Phase separators 2 3
major equipment
Wash tower 1 1
Total 9 13
Reaction time (min) 7 ∼120
2–10 m3 each (for five
17.02 (for RTRANS
Volume of reactor(s) (m3 ) reactors producing
in Figure 10.2)
biodiesel in Figure 10.5)
∗FCI or TMC (million $) 11.3 2.88
∗ This
cost for the microalgae-based process excludes the cost of the microwave and ultrasound units required.
CAPCOST is used for cost estimation of both processes.

conventional two-step process of biodiesel production from WCO (Sharma and Rangaiah
2013). In addition to the advantage of having faster reaction due to MW–ultrasonication
intensification, the developed process has a certain advantage over the conventional coun-
terpart as the former process uses 9 major processing units whereas the conventional alter-
native uses 13 major processing units. This difference is due to number of reactors and
phase separators. In the study by Sharma and Rangaiah (2013), CEPCI was taken as 600,
whereas it is taken as 602 in the present study as mentioned in Section 10.3. These studies
used the CAPCOST program for capital cost estimation. Hence, for fair comparison, costs
in Sharma and Rangaiah (2013) are inflated to CEPCI of 602. It is clear from Table 10.4
that the developed process is unattractive as its fixed capital investment (FCI, i.e. TMC
excluding the cost of MW, ultrasonication, SEP-1, and SEP-2) is about four times that for
the conventional alternative using WCO. The large investment required is mainly due to the
small lipid content of 5% in microalgae with a lot of water (82%).
To obtain 2500 kg/h of lipids, we need to use 50 322 kg/h of wet microalgae. This results
in an overall increase in size of all equipment. Also, this higher microalgal flow rate requires
a higher amount of chemicals such as methanol, hexane, and catalyst. The FCI of the process
developed in the present study is compared with two other studies, namely the processes
proposed in West et al. (2008) and Patle et al. (2014). West et al. (2008) reported a TMC of
$1.1 million (for CEPCI = 394) for a biodiesel plant to produce 8 kt of biodiesel per annum
using WCO as the feedstock. Further, Patle et al. (2014) reported a TMC of $12.95 million
(for CPECI = 600) for a biodiesel plant processing 120 kt of WCO per annum. Projected
TMC for a plant capacity of 20 kt of lipids per annum, using the six-tenths rule (Turton
et al. 2009) and CEPCI of 602, is $2.9 million (based on West et al. 2008) and $4.43 million
(based on Patle et al. 2014). In the present study, a TMC of $11.3 million is obtained, which
is about 4 and 2.5 times that of the process in West et al. (2008) and Patle et al. (2014),
respectively.
The COM of the present process processing wet microalgae (with ∼20 kt of lipids per
annum capacity) is $65.05 million, which is roughly five times that of the process of the
280 Process Systems Engineering for Biofuels Development

same capacity using WCO feed given in Sharma and Rangaiah (2013), i.e. $13.86 million.
In our earlier work (Patle et al. 2014), a COM of $73.5 million was found for a process
of capacity 120 kt of WCO per annum handling WCO, which translates to $12.25 million
for a plant capacity of 20 kt of lipids per annum. It is important to keep in mind that the
correlations/programs used for costing in West et al. (2008) and Patle et al. (2014) are
different from the one used in this study, which may contribute to some difference in cost
estimates compared in this and the previous paragraph.

10.5 Comparison with the Microalgae-Based Processes

Recently, Heo et al. (2019) analyzed three alternate processes, namely, catalytic, enzymatic
and in situ transesterification for producing microalgal biodiesel. They employed SuperPro
Designer for the simulation of these three processes. The feedstock for these processes
was wet microalgae biomass of Nannochloropsis gaditana at a flow rate of 1000 kg/h; this
feedstock contains 80% water content and lipids equal to 30% of dry cell weight (i.e. with
0.3 × 20 = 6% lipids). Based on their analysis, Heo et al. (2019) concluded that catalytic
transesterification is currently the best (at a production cost of 4.77 $/kg) among the three
processes studied, and that enzymatic and in situ transesterification have great potential to
be improved.
For in situ biodiesel production (without any MW or US intensification), Heo et al. (2019)
reported a FCI of $2.04 million for a plant capacity of 0.48 kt of lipids per annum (based
on 1000 kg/h of feed with 6% lipids for 8000 hours of operation per year). For a plant pro-
cessing 20 kt of lipids per annum (as in the present study) and using the six-tenths rule, this
capital cost translates to $19.12 million. This cost is about two times the estimate obtained
in the present study. Heo et al. (2019) reported a production cost (similar to COM) of
7.62 $/kg of biodiesel by in situ transesterification, and they deduced that there is potential
to reduce this to 4.38 $/kg. These values are higher than the breakeven cost of biodiesel
(3.13 $/kg) found based on the obtained COM in the present study and glycerol value of
1.17 $/kg.
Possible reasons for the above differences in cost estimates of Heo et al. (2019) and the
present study are: (i) Heo et al. (2019), as shown in their Figure 1c, did not consider the
complete process (e.g. for biodiesel/glycerol separation and for methanol recovery/recycle),
(ii) source and cost of feedstock, and (iii) differences in cost correlations, utility costs,
and cost procedures. Further, Figure 1 in Heo et al. (2019) does not present temperature,
pressure, flowrate and composition of streams in the process, which makes it very difficult
to understand the processes analyzed. In the present study, on the other hand, glycerol is
separated and its value at 1.17 $/kg is considered in the biodiesel breakeven cost calculation,
and methanol is separated and recycled for resource conservation. Also, Heo et al. (2019)
assumed reaction yield, whereas realistic reaction kinetics is considered in the present study.

10.6 Conclusions
In the present study, a continuous process is developed for the production of biodiesel from
wet microalgae. In this, in situ transesterification of wet microalgae “Nannochoropsis sp.”
 Process Development, Design and Analysis of Microalgal Biodiesel Production 281

is intensified by MW and US radiations. The developed process is successfully simulated

in Aspen Plus v8.8. Detailed cost analysis is carried out considering the plant to produce
2493 kg/h of biodiesel product from 2500 kg/h of microalgal oil/lipids. Comparative anal-
ysis is then presented between the developed process and a conventional process of the
same capacity, i.e. producing 2512 kg/h of biodiesel from WCO. It is found that the cap-
ital cost of the developed process without considering the cost of US and MW is about
2.5 to 4 times that of the conventional process using WCO despite the attractiveness of the
former process in terms of fewer processing units and faster reaction. The breakeven cost
of biodiesel by the proposed process is found to be 3.13 $/kg. The increased capital cost
and utilities are due to the need for processing a huge amount of microalgae (with sub-
stantial water), hexane and methanol, which require expensive separation using distillation
columns.
Considering the high cost of the developed biodiesel process based on wet microalgae,
future work should focus on (i) increasing the lipid content in the microalgae to obtain
greater amount of biodiesel from lesser wet microalgae, and (ii) optimization of the pro-
duction process to maximize biodiesel yield, minimize utility requirement and minimize
the requirement of chemicals such as hexane, methanol, and catalyst. Heat integration and
heat pump can improve the energy efficiency of the process. Development of suitable het-
erogeneous catalysts for esterification may also be explored as the use of homogeneous
catalyst demands its neutralization and separation.

Acknowledgment
The authors gratefully acknowledge the funding received from DST-SERB through the
research grant ECR/2016/001866 to carry out this work.

Appendix 10.A

Table 10.A1 Data of important streams in the process flow diagram in Figure 10.2.

Stream name Fresh methanol Fresh wet algae Methanol recycle Hexane recycle
Temperature (∘ C) 25 25 63 53
Pressure (atm) 1 1 1 1
Mass flow rate (kg/h) 663.992 50 322.00 43 992.1 101 286
Mass fraction
Trioleate (lipids) 0 0.05 0 0
Methanol 1 0 0.89 0.027
Water 0 0.816 0.04 0
FAME 0 0 0 0
Glycerol 0 0 0 0
Proteins 0 0.1 0 0
Carbohydrate 0 0.034 0 0
Hexane 0 0 0.07 0.973
282 Process Systems Engineering for Biofuels Development

Table 10.A1 (Continued)

Stream name Triglyceride recycle Biodiesel (FAME) Water out Glycerol

Temperature (∘ C) 40 25 30 141
Pressure (atm) 1 1 1 1
Mass flow rate (kg/h) 2 297.13 2 492.81 41 462.9 283.015
Mass fraction
Trioleate 0.998 0.02 0 0
Methanol 0 0 0.009 0
Water 0 0 0.99 0.041
FAME 0.002 0.98 0 0.05
Glycerol 0 0 0 0.908
Proteins 0 0 0 0
Carbohydrate 0 0 0 0
Hexane 0 0 0 0

Exercises

1. List the challenges faced by the biodiesel industry concerning the feedstock.
2. Why is algal biomass considered as a potential feedstock for next generation biodiesel
production?
3. What are the factors that are responsible for algal biodiesel not being commercially
practiced?
4. How does microwave-ultrasonication intensification affect the economics of the in situ
biodiesel production process?
5. Discuss the challenges associated with the in situ biodiesel synthesis from wet microal-
gae.
6. What, in your opinion, can/should be improved in order for microalgal biodiesel to be
commercially feasible?
7. What are the challenges in scaling up the ultrasonic intensification to the industrial-scale
processes?
8. Comment on the merits/demerits of the ultrasonic intensified biodiesel production focus-
ing on the cost and footprint.
9. If the lipid content in wet microalgae is increased by 40% from 5% to 7% at the cost of
proteins in the example process given in this chapter, discuss qualitatively its impact on
the process economics.

References
Atabani, A.E., Silitonga, A.S., Badruddin, I.A. et al. (2012). A comprehensive review on biodiesel as
an alternative energy resource and its characteristics. Renewable and Sustainable Energy Reviews 16:
2070–2093.
Babu, G.U.B., Pal, E.K., and Jana, A.K. (2012). An adaptive vapour recompression scheme for a ternary
batch distillation with a side withdrawal. Industrial and Engineering Chemistry Research 51: 4990–4997.
Barbarino, E. and Lourenço, S.O. (2005). An evaluation of methods for extraction and quantification of
protein from marine macro- and microalgae. Journal of Applied Phycology 17: 447–460.
 Process Development, Design and Analysis of Microalgal Biodiesel Production 283

CNEA (2012). National Twelfth Five-Year Plan on Biomass Energy. Chinese National Energy Adminis-
tration, Beijing.
Demirbas, A. and Demirbas, M.F. (2011). Importance of algae oil as a source of biodiesel. Energy Conver-
sion and Management 52 (1): 163–170.
Dutta, B.K. (2009). Principles of mass transfer and separation processes. The Canadian Journal of Chemical
Engineering 87 (5): 818–819.
Feng, Y. and Rangaiah, G.P. (2011). An evaluation of capital cost estimation methods and programs. Chem-
ical Engineering 118 (8): 22–29.
Fleurence, J., Massiani, L., Guyader, O., and Mabeau, S. (1995a). Use of enzymatic cell wall degradation for
improvement of protein extraction from Chondrus crispus, Gracilaria verrucosa and Palmaria palmata.
Journal of Applied Phycology 7: 393–397.
Fleurence, J., Le Coeur, C., Mabeau, S. et al. (1995b). Comparison of different extractive procedures for
proteins from the edible seaweeds Ulva rigida and Ulva rotundata. Journal of Applied Phycology 7:
577–582.
Gogate, P.R. (2008). Cavitational reactors for process intensification of chemical processing applications:
a critical review. Chemical Engineering and Processing: Process Intensification 47: 515–527.
Harnedy, P.A. and FitzGerald, R.J. (2013). Extraction of protein from the macroalga Palmaria palmata.
LWT - Food Science and Technology 51: 375–382.
Heo, H.Y., Hep, S., and Lee, J. (2019). Comparative techno-economic analysis of transesterification tech-
nologies for microalgal biodiesel production. Industrial & Engineering Chemistry Research https://doi
.org/10.1021/acs.iecr.9b03994.
Jenkins, S. (2018). CEPCI updates. https://www.chemengonline.com/2018-cepci-updates-may-prelim-
and-april-final (accessed October 2018).
John, R.P., Anisha, G.S., Nampoothiri, K.M., and Pandey, A. (2011). Micro and macroalgal biomass: a
renewable source for bioethanol. Bioresource Technology 102: 186–193.
Jorquera, O., Kiperstok, A., Sales, E.A. et al. (2010). Comparative energy life-cycle analyses of microalgal
biomass production in open ponds and photobioreactors. Bioresource Technology 101: 1406–1413.
Joubert, Y. and Fleurence, J. (2008). Simultaneous extraction of proteins and DNA by an enzymatic treat-
ment of the cell wall of Palmaria palmata (Rhodophyta). Journal of Applied Phycology 20: 55–61.
Kadam, S.U., Álvarez, C., Tiwari, B.K., and O’Donnell, C.P. (2017). Extraction and characterization of
protein from Irish brown seaweed Ascophyllum nodosum. Food Research International 99: 1021–1027.
Khiratkar, A.G., Balinge, K.R., Patle, D.S. et al. (2018). Transesterification of castor oil using benzimida-
zolium based Brönsted acidic ionic liquid catalyst. Fuel 231: 458–467.
Kumar, M. and Sharma, M.P. (2016). Selection of potential oils for biodiesel production. Renewable and
Sustainable Energy Reviews 56: 1129–1138.
Lee, J.-Y., Yoo, C., Jun, S.-Y. et al. (2010). Comparison of several methods for effective lipid extraction
from microalgae. Bioresource Technology 101 (1): S75–S77.
Li, C., Zhao, X., Wang, A. et al. (2015). Catalytic transformation of lignin for the production of chemicals
and fuels. Chemical Reviews 115 (21): 11559–11624.
Lurgi (2019). Lurgi biodiesel. https://www.engineering-airliquide.com/lurgi-biodiesel (accessed January
2019).
Luyben, W.L. (2002). Plantwide Dynamic Simulators in Chemical Processing and Control. New York:
Marcel Dekker Chapter 3.
Luyben, W.L. (2011). Principles and Case Studies of Simultaneous Design. Wiley.
Martinez-Guerra, E., Howlader, M.S., Shields-Menard, S. et al. (2018). Optimization of wet micro algal
FAME production from Nannochloropsis sp. under the synergistic microwave and ultrasound effect.
International Journal of Energy Research 42 (5): 1934–1949.
Morais, S., Mata, T.M., Martins, A.A. et al. (2010). Simulation and life cycle assessment of process design
alternatives for biodiesel production from waste vegetable oils. Journal of Cleaner Production 18 (13):
1251–1259.
284 Process Systems Engineering for Biofuels Development

Myint, L.L. and El-Halwagi, M.M. (2009). Process analysis and optimization of biodiesel production from
soybean oil. Clean Technologies and Environmental Policy 11 (3): 263–276.
Naik, S.N., Goud, V.V., Rout, P.K., and Dalai, A.K. (2010). Production of first- and second-generation
biofuels: a comprehensive review. Renewable and Sustainable Energy Reviews 14: 578–597.
Paniwnyk, L., Cai, H., Albu, S. et al. (2009). The enhancement and scale-up of the extraction of
anti-oxidants from Rosmarinus officinalis using ultrasound. Ultrasonics Sonochemistry 16: 287–292.
Patle, D.S., Sharma, S., Ahmad, Z., and Rangaiah, G.P. (2014). Multi-objective optimization of two alkali
catalyzed processes for biodiesel from waste cooking oil. Energy Conversion and Management 85:
361–372.
Piemonte, V., Di Paola, L., Iaquaniello, G., and Prisciandaro, M. (2016). Biodiesel production from microal-
gae: ionic liquid process simulation. Journal of Cleaner Production 111: 62–68.
Sara, M., Brar, S.K., and Blais, J.F. (2016). Comparative study between microwave and ultra-sonication
aided in situ transesterification of microbial lipids. RSC Advances 6: 56009–56017.
Seider, W.D., Seader, J.D., and Lewin, D.R. (2009). Product & Process Design Principles: Synthesis, Anal-
ysis and Evaluation. Wiley.
Sharma, S. and Rangaiah, G. (2013). Multi-objective optimization of a bio-diesel production process. Fuel
103: 269–277.
Sharma, S., Patle, D.S., Gadhamsetti, A.P. et al. (2018). Intensification and performance assessment of the
formic acid production process through a dividing wall reactive distillation column with vapor recom-
pression. Chemical Engineering and Processing: Process Intensification 123: 204–213.
Sun, A., Davis, R., Starbuck, M. et al. (2011). Comparative cost analysis of algal oil production for biofuels.
Energy 36: 5169–5179.
Towler, G. and Sinnott, R. (2013). Chemical Engineering Design: Principles, Practice and Economics of
Plant and Process Design, 2e. Elsevier.
Turton, R., Bailie, R.C., Whiting, W.B., and Shaeiwitz, J.A. (2009). Analysis Synthesis and Design of Chem-
ical Processes, 3e. Prentice Hall.
USDA (2013). Announces investments in bioenergy research and development to spur new markets, inno-
vation, and unlimited opportunity in rural America. Press release. https://www.usda.gov/media/press-
releases (accessed October 2014).
West, A.H., Posarac, D., and Ellis, N. (2008). Assessment of four biodiesel production processes using
HYSYS Plant. Bioresource Technology 99 (14): 6587–6601.
Zhang, X., Yan, S., Tyagi, R.D. et al. (2016). Ultrasonication aided biodiesel production from one-step and
two-step transesterification of sludge derived lipid. Energy 94: 401–408.
 11
Thermochemical Processes for the
Transformation of Biomass into
Biofuels
Carlos J. Durán-Valle
Departamento de Química Orgánica e Inorgánica, IACYS, Universidad de Extremadura, 06006 Badajoz,
Spain

11.1 Introduction
When plants grow, all carbon in the biomass comes from the atmosphere. When plants are
burned, all carbon is liberated, giving a zero net carbon footprint. However, if fossil fuels are
used, it causes a flow of carbon from the soil into the atmosphere, creating a greenhouse
effect that produces global warming. Global warming coupled with declining fossil fuel
reserves has attracted a great deal of attention for sustainable processes using renewable
resources. Among them, obtaining fuels from biomass stands out.
There are two main routes for obtaining biofuels from biomass: biological and thermo-
chemical (Damartzis and Zabaniotou 2011). Other methods such as direct solvent extraction
or chemical hydrolysis are still under investigation or are inefficient in most cases. Depend-
ing on the raw material and the desired target, some methods may be preferable to others.
Figure 11.1 shows a classification of the main methods for obtaining biofuels from biomass
and the main products obtained with each method.
Thermochemical processes for the transformation of biomass are a set of methods char-
acterized for the application of heat to convert biomass into more useful products. This
chapter describes thermochemical methods, which have the following general advantages:

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
286 Process Systems Engineering for Biofuels Development

Methods Products

Extraction Ethanol Biodiesel

Hydrolysis Ethanol

Digestion Methane

Biomass Combustion Heat

Gasification Gas fuel Syngas

Pyrolysis Charcoal Bio-oil Gas fuel

Liquefaction Bio-oil

Chemical methods

Biological methods

Thermochemical methods

Figure 11.1 Processes for the transformation of biomass into biofuels.

– Effective with more biomass, including materials with high water contents (hydrothermal
process).
– Higher productivity than biological methods. Limited generation of waste.
– Pretreatment needed only in a few processes.
– Short process times.
– Non-strict operating conditions.
There are other advantages, but they are particular to some methods and cannot be con-
sidered as general ones.
As for the classification of thermochemical methods, there is some confusion in the lit-
erature. Some authors use pyrolysis for all thermal processes performed on biomass that
result in total or partial decomposition of the biomass. IUPAC, in the Gold Book (IUPAC ),
relates this term to thermolysis, but points out that it is usually associated with high temper-
atures and that it generally refers to processes carried out in an inert environment. In this
chapter we will use the term pyrolysis when the thermochemical treatment is carried out in
an inert atmosphere.
There is less discussion about the term combustion. This can be restricted to the process
in which the biomass reacts with a strong oxidant (usually air) to produce heat and not a
valuable substance.
Liquefaction is a thermal process carried out in the presence of water or another solvent
in order to obtain a liquid product. Gasification is carried out in the presence of a gas
 Thermochemical Processes for the Transformation of Biomass into Biofuels 287

and the objective is to obtain mainly gases, which can be used as fuel or to obtain other
valuable compounds.
It should be borne in mind that all processes except combustion produce gaseous, liq-
uid and solid products. Generally, the methodology is adjusted to obtain a majority of one
of them, but it is difficult to avoid the production of the others, which are considered as
by-products.
The classification proposed in the preceding paragraphs and in Figure 11.1 is simple
and does not include some thermochemical methods. One case of a method that is not
included in this classification is that of carbonization. This can be done by pyrolysis, but
the most traditional and widespread method in the world is that of limited combustion due
to the restriction of air flow. We could consider it as an intermediate term between pyrol-
ysis and complete combustion. For this reason and for its social, historical, and economic
importance, it is discussed in its own section (Section 11.7).
Another difficult process to fit into the above classification is hydrothermal treatment,
which consists of treating the biomass in the presence of a water/steam mixture under pres-
sure. Depending on the working conditions, gases, liquids, or charcoal can be obtained as
a priority. Therefore, it can be said that there is hydrothermal carbonization, hydrothermal
liquefaction, and hydrothermal gasification. Something similar can be said if supercritical
water is used. Figure 11.2 shows a graphical classification of thermochemical processes
used to obtain biofuels from biomass.

Hydrothermal
carbonization

Carbonization
Heating in limited
presence of oxygen

Pyrolysis Combustion
Heating in absence Heating in presence
of oxygen of oxygen

Liquefaction Gasification
Heating in presence Heating in presence
of water/solvent of gases
Hy ique
dr fac
l
ot

al
he tion

m
er ion
rm

h
ot at
al

dr ific
Hy as
g

Figure 11.2 Thermochemical processes for the transformation of biomass into biofuels.
288 Process Systems Engineering for Biofuels Development

This chapter describes the thermochemical methods that are mainly aimed at obtaining
solid, liquid or gaseous fuels, although they are sometimes used to obtain chemicals. In
general, no further treatment of the fuels obtained is described here. An attempt has been
made here to provide a complete overview of the state of the art in the thermochemical
conversion of biomass into fuels.
Numerous reviews have been published on this subject, presenting new results as the
research progresses. They are usually aimed at researchers, and not at teaching or informa-
tive tasks as intended in this book.
Many of the published revisions are more specific and less general than this chapter. Thus,
there are publications dedicated to the rapid pyrolysis and improvement of the products
obtained (Bridgwater 2012), to the use of catalysts in gasification (Sutton et al. 2001), or to
the use of microalgae as raw material (Chen et al. 2014). An interesting review because of
its structure is that of Verma et al. (2012), but it focuses on flash pyrolysis and the types of
reactors used and does not mention hydrothermal methods.
The book Recent Advances in Thermochemical Conversion of Biomass (Pandey et al.
2015) can be cited in which the thermochemical conversion of biomass is studied both to
obtain fuel and to obtain chemical products. It is an extensive text addressed to researchers
on the subject. Some of its chapters are cited in this work.
There are also reviews of very specific aspects related to this chapter. For example, there
is an article (Jiang et al. 2018) that studies the liquefaction processes of wood for the man-
ufacture of adhesives, and there is another on the production of hydrogen from palm oil
(Hossain et al. 2016).
In contrast, there is also a more general study (Canabarro et al. 2013) and therefore more
similar to the objective of this chapter. However, due to their brevity, the most commonly
used reactors and hydrothermal processes are not described.

11.2 Biomass and Biofuels

Most of the biomass used to obtain biofuels consists of lignocellulosic materials. These
materials are mainly composed of three polymers: lignin, cellulose, and hemicellulose.
Cellulose is the main component of lignocellulosic materials and is a linear polymer of
glucose. The monomers are linked by β-1,4-glycosidic bonds. One polymer chain interacts
with nearby chains through van der Waals and hydrogen bond interactions. There are a lot
of interactions in a chain, which makes cellulose relatively stable.
Hemicellulose is a polymer formed by several simple carbohydrates and may be
branched. Its structure is more complex than that of cellulose and is also less stable. It
has an intermediate polarity between the other two polymers and is believed to bind both
components.
Lignin is an aromatic, nonpolar polymer that is not soluble in water. It provides resistance
against physical, biological and chemical attacks. This polymer is three-dimensional and
branched when lignification is advanced. Of these three components, lignin is the one with
the highest resistance to heat.
The first component to be pyrolyzed is hemicellulose, and the most stable one is cellulose.
The production of biofuels has evolved over time, as science and technology have
 Thermochemical Processes for the Transformation of Biomass into Biofuels 289

advanced. Energy efficiency has been improved and the environmental impact of the
process has been reduced.
If we avoid direct biomass combustion, biofuels can be classified according to the
biomass used in three generations.
First-generation biofuels (Balagurumurthy et al. 2015) are obtained from food crops (for
example grains, corn, sugarcanes, or vegetable oils). Carbohydrates are fermented to obtain
ethanol (bioethanol) and the oils are transformed by transesterification into biodiesel fuel.
The success of these biofuels is limited for several reasons:
– Only a limited part of the plant is used. Vegetables are mainly composed of water and
lignocellulosic material, and neither can be used to obtain this type of fuel.
– There is a negative impact on biodiversity.
– There is also a detrimental effect on climate change. In addition, it requires a high use of
arable land, water, and fertilizers per unit of energy obtained.
– There has been a social response due to the fact that its demand has increased the price
of food in some parts of the world.
For these reasons, second-generation biofuels are already being produced. They are based
on the use of lignocellulosic feedstocks, since more tons of usable biomass can be obtained
per hectare than in first-generation biofuels. Agricultural and some industrial residues,
wood, and dedicated crops can be used. Most of the methods described in this chapter
are aimed at the use of this type of biomass.
Third-generation biofuels are being investigated (Chen et al. 2014) and there are a few
commercial installations. They are obtained from algae and can yield biofuels but also
chemicals. These biofuels have several advantages:
– High productivity, including high-value products.
– Neutral carbon dioxide (CO2 ) emissions.
– Use of nonproductive lands and several water sources.
Their disadvantages include the need for tightly controlled environmental conditions and
high expenditure on phosphorus, which is becoming a scarce resource.
The algae can be transformed into biofuel by thermochemical methods. Algae are a
non-lignocellulosic material, like manure, sewage sludge, animal hair, microorganisms, and
others (Li and Jiang 2017). The composition of these materials differs from lignocellulosic
materials in that they contain a higher proportion of heteroatoms, such as nitrogen, sulfur,
or phosphorus, as well as heavy metals. Some may also contain pathogens. This requires
some differences in treatment.

11.3 Combustion
The simplest method of thermochemical conversion is combustion: no transformation is
needed. The biomass is heated in the presence of oxygen (air). Normally, it is necessary to
initiate the reaction; however, as it is exothermic, it runs by itself without energy contribu-
tion. It is a method widely used in the domestic and industrial environment to obtain energy.
The research and development of this process focuses on solving environmental problems.
290 Process Systems Engineering for Biofuels Development

11.4 Gasification
Gasification is a thermal conversion of solid biomass to gases using another gas that acts as
the gasification agent. This one facilitates the production of a mixture of gases that can be
used as fuel and generally contains compounds such as carbon monoxide (CO) and CO2 ,
water, dihydrogen (H2 ), methane, and other organic compounds (Damartzis and Zabaniotou
2011; Canabarro et al. 2013). There are several methods and gasification agents (such as
air, CO2 , water), and it is always carried out at high temperature (800–1300 ∘ C).
This technology was used to power cars and trucks in the twentieth century when there
was a scarce supply of petroleum-derived fuels, e.g. in Sweden since the 1930s and after
World War II (FAO 1993) and in Spain1 after the civil war (1940s) (Montoliú Camps 2005).
Although in the last case it was common to also use charcoal, but with a similar process.
Nowadays, it can be used in several industrial processes, mainly in power generation. One
of the most interesting products is syngas (synthesis gas), which is a mixture of CO and H2 .
To obtain a simple composition, without by-products, it is necessary to optimize the concen-
trations of oxygen and water vapor. The syngas can be converted by the Fischer–Tropsch
process into other chemicals, such as alcohols (mainly methanol and ethanol), dimethyl
ether, and a mixture of alkenes and alkanes, which can be used as fuels (methane, liquefied
petroleum gases, gasoline, and diesel fuel) or in the chemical industry. Hydrocarbons are
the largest fraction, and oxygenated compounds are produced as minor products.
Gasification consists of several steps, which are not always performed entirely within the
reactor:
– Drying. This is the removal of water from the biomass. It can be done before entering
the gasifier.
– Volatilization. It is mainly a pyrolytic process, where matter decomposes into its most
labile parts to produce volatile compounds.
– Gasification. The raw material is completely transformed into gas using the gasifying
agent.
– Combustion. Part of the gases produced and residual char are burned, producing heat
that can be used in other steps. This stage can be carried out in a reactor external to the
gasifier.
The use of air as a gasifying agent is common due to its low cost although the energy
of the final gas is also low (Canabarro et al. 2013). This gas can be used as biofuel, but
not as syngas. Better quality can be achieved by using oxygen or steam, but with a higher
cost and a complex manufacturing plant. This improvement in calorific capacity is due to
two factors: the increased production of H2 when steam is used and the absence of gaseous
dinitrogen (N2 ), since air is not used directly.
An unwanted by-product is tar. Processes are optimized to obtain the lowest possible
amount of tar and the highest amount of biofuels. Tar is viscous and, although volatile at
gasification temperatures, it condenses at low temperature. It produces blockages and other
failures.
Moreover, slag and ashes can be found as by-products. These are mainly composed
of inorganic compounds, and thus gasification can be considered as a method to extract

1
Information obtained orally from the author’s family.
 Thermochemical Processes for the Transformation of Biomass into Biofuels 291

bioenergy free of halogens, metals, and other contaminants. Slag can be corrosive to
the equipment. Ashes are separated by cyclones. Slag and ashes can have different uses,
depending on their composition. Traditionally ashes have been used as fertilizer or as clean-
ing material, and slag as building material. But if the production is in industrial quantities,
these applications are short in volume. The use of ashes as a building material, adsorbents in
aqueous and gaseous media, zeolite synthesis, road base, etc. has been proposed (Ahmaruz-
zaman 2010). The same applications, in general, can be stated (Piatak et al. 2015) for slag
from biomass, although it should be noted that this is only a small amount compared with
slag obtained from metallurgical processes and may have different applications.
There are several gasification techniques, which are described below. In some cases, the
type of reactor used is the differentiating factor. These reactors are also described in the
section on pyrolysis (Section 11.6). The difference between them is the atmosphere of the
process that leads to obtaining the different products.

11.4.1 Fixed Bed Gasification

Fixed bed reactors are the most common gasifiers. They are simple in structure and oper-
ation, and the pretreatment of biomass is not required. The gas passes through the bed of
biomass. They are the preferred gasifiers due to their small-scale plants. The raw material is

Gas, vapor
Biomass feed

Fixed bed
gas

Figure 11.3 Scheme of a fixed bed reactor.

292 Process Systems Engineering for Biofuels Development

usually added through the top of the reactor. The flow of the gasifier can follow an ascending
(updraft type) or descending (downdraft) path, and also a lateral flow (cross-flow gasifiers),
from one side of the reactor to the other. Figure 11.3 schematically shows a fixed bed reactor.

11.4.2 Fluidized Bed Gasification

Figure 11.4 schematically shows a fluidized bed reactor. Small biomass particles are sus-
pended in a stream of the gasification agent. This method improves solid–gas contact,
heat transfer and, thus, achieves a uniform temperature throughout the reactor. Sometimes
biomass is mixed with mineral particles which has several effects: it improves the fluidiza-
tion of the mixture, incorporates a catalyst, and decreases the production of tars.
Frequently, the reactors are equipped with cyclones to remove fine particles, such as
charcoal or ash.
These gasifiers are widely preferred due to the fact that they can be used with different
types of raw materials, including coals and biomass with different properties.

11.4.3 Dual Fluidized Bed Gasification

The plant for this process consists of two reactors (Figure 11.5) between which the biomass
and final products circulate. One of them is a combustion reactor, where the resulting

Gas, vapor, char

Biomass feed

Fluidized
bed
gas

Figure 11.4 Scheme of a fluidized bed reactor.

 Thermochemical Processes for the Transformation of Biomass into Biofuels 293

Gas, vapor

Char, carrier

Combustion
er
Biomass feed

arri
Fluidized dc
ate
bed He

air
gas

Figure 11.5 Scheme of a recirculating or dual fluidized bed reactor.

biomass or gases are partially burned. The heat generated is transmitted to the second
reactor, where gasification takes place. As it is not necessary to generate large amounts
of heat inside this reactor, it is not necessary to use air as a gasifying agent. Steam is more
frequently used; in this way, a gas with a higher calorific capacity is obtained, which can
also be used as syngas or for chemical synthesis.

11.4.4 Hydrothermal Gasification

Most of the thermochemical processes to convert biomass can only be profitable if the raw
material is dry. However, a significant fraction of the biomass has a high water content. This
fraction is often discarded, although hydrothermal treatments are capable of converting this
biomass, obtaining, in some cases, products with different characteristics compared with
the ones obtained with dry biomass.
Hydrothermal processing is the treatment of biomass with water at high pressures and
temperatures. It is carried out at moderate temperature (180–550 ∘ C) with a solvent (gen-
erally water) in a closed reactor. The solvent (e.g. vapor) increases the pressure with the
294 Process Systems Engineering for Biofuels Development

temperature, and a combination of both produces the reactions. As water must be added
in most cases, the moisture of the biomass is not important. This avoids the energy cost
of drying the biomass before conversion. Hydrothermal processing is also an environmen-
tally friendly technology: there are no hazardous products of combustion, such as nitrogen
oxides, aromatic gases, CO, or others. The production of contaminants is limited, and they
are easily handled.
The products obtained depend on temperature (Matsumura 2015). Thus, in treatments
below 200 ∘ C, the main product is solid (charcoal). Between 200 ∘ C and 350 ∘ C, the liq-
uid state dominates, and the process is referred to as hydrothermal liquefaction. Above
350 ∘ C, gasification prevails. It must be remembered that above 374 ∘ C, water behaves like
a supercritical fluid.
Liquid water is mostly responsible for the reactions, since steam is not as reactive. Pres-
sure depends on the temperature and gas composition. If only water is placed in the reactor,
the pressure is very close to saturation pressure at the temperature of the process, with a
little correction due to the air or other gases in the system. However, this is valid only at
subcritical conditions. When supercritical temperature is exceeded, the pressure is deter-
mined by the amount of water and the volume of the reactor. Water at high temperature
has a high ion product and a low dielectric constant, which are favorable for promoting
reactions without the use of catalysts. Thus, large amounts of acid or base catalysts can be
saved, which decreases costs and pollution.
Nevertheless, the hydrothermal conversion can be catalyzed by acids or alkalis. For
instance, alkali-based catalysts such as activated carbon, transition metals, and oxides are
used in hydrothermal liquefaction (Bhaskar and Pandey 2015).
Hydrothermal gasification is carried on at temperatures between 350 ∘ C and 374 ∘ C (sub-
critical conditions) (Matsumura 2015). Due to high temperatures, this process is fast and
complete reactions are achieved. The gas produced contains H2 , CO2 , and light hydro-
carbons. If tar or char are produced, they remain in the reactor, mixed with liquid water.
Moreover, there is no N2 when air is not used. Thus, the gas produced has a high power
value. Furthermore, high temperatures are needed, since tar production decreases.

11.4.5 Supercritical Water Gasification

This method can use wet biomass without the need of a drying pretreatment. In this regard,
it is similar to hydrothermal processes (see Section 11.4.4), with the difference that it uses
supercritical water (temperature >374 ∘ C, pressure >22.1 MPa). This prevents water phase
changes, which require a large consumption of energy. Supercritical water has several roles
in the gasification process: it can participate as medium, reactant, or as catalyst.

11.4.6 Plasma Gasification

Plasma is a mixture of electrons, ions, and neutral molecules, electrically neutral as a whole.
The reactivity of plasma depends on the degree of ionization, which is the proportion of ions
in the gas. Plasma is created by an electrical arc that applies an electric current to a dielectric
gas. The heat generated in this process, together with the reactivity of plasma, produces the
gasification of biomass.
 Thermochemical Processes for the Transformation of Biomass into Biofuels 295

11.4.7 Catalyzed Gasification

One of the problems in gasification is the removal of methane and tars. Methane can affect
the use of syngas, and tars can cause blockages and corrosion. Therefore, the use of cat-
alysts is an area of interest. The catalyst selection criteria can be summarized as follows
(Sutton et al. 2001) and, with small changes, these criteria can be adapted to other catalysts
mentioned in this chapter:
– The catalysts must be effective in the removal of tars.
– If the desired product is syngas, the catalysts must be capable of reforming methane.
– The catalysts should provide a suitable syngas ratio for the intended process.
– The catalysts should be resistant to deactivation as a result of carbon fouling and
sintering.
– The catalysts should be easily regenerated.
– The catalysts should be strong.
– The catalysts should be inexpensive.
Some catalysts for gasification can be mixed with the biomass, whereas others can be
placed in other reactors, where the gas stream generated in the gasification is directed to
undergo a reforming process. Some of the most used catalysts for gasification are dolomite,
alkali metals, and nickel (Sutton et al. 2001).
An example of an external catalysis, which is one of the most commonly used reforming
processes, is employed in the purification of syngas (Damartzis and Zabaniotou 2011). It
consists of reacting the tar formed as an unwanted product with steam to form CO and
H2 at temperatures around 700 ∘ C. Carbon monoxide reacts with steam to produce CO2
and more H2 . Catalysts commonly used in this reaction are combinations of metals with
carbonaceous or silicate minerals.
An alternative to this procedure is the cracking of hydrocarbon tar chains, which is done
at high temperatures, using a catalyst such as olivine or dolomite. Carbon monoxide and
H2 are also obtained, which enrich the syngas mixture.

11.4.8 Fischer–Tropsch Synthesis

A brief description of this well-known industrial process has been included since a large
part of the syngas produced in gasification is used for this purpose. It was developed in
the 1920s and includes several reactions to convert the syngas into medium-mass hydrocar-
bon molecules. The product is similar to a synthetic petroleum. The reactions are carried
out at temperatures of 250–300 ∘ C, and high pressures (2–3 MPa) with catalysts based on
transition metals such as cobalt or iron. These catalysts are very sensitive to sulfur, and the
syngas must be desulfurized. The use of moderate temperatures and high pressures hinders
the formation of hydrocarbons of low molecular mass, such as methane.
A mixture of hydrocarbons is always obtained, the composition of which depends on the
conditions of the process. The final objective is usually to obtain liquid fuels. Therefore,
the formation of volatile hydrocarbons such as methane is not desired, but neither is the
formation of compounds with a high molecular mass. In order to avoid the former, moderate
temperatures and high pressures are used. In the latter case, a hydrocracking reaction is
employed, which causes the large molecules to break down to produce the desired products.
296 Process Systems Engineering for Biofuels Development

11.5 Liquefaction
Liquefaction produces mainly liquid from biomass and the process occurs in an environ-
ment where water or organic solvents are employed. The temperature used is moderate
(200–400 ∘ C) and so is the pressure applied (4–20 MPa) (Jiang et al. 2018). Alkaline cata-
lysts are commonly used. Sometimes, a reactant, such as CO or H2 , improves the process
performance. The catalyst hydrolyzes the macromolecules present in the biomass, such
as lignin and cellulose, into smaller molecules, which undergo decarboxylation, dehydra-
tion, dehydrogenation, and other reactions to yield smaller compounds (Verma et al. 2012).
When water is used as solvent, it can be considered as a hydrothermal process. The main
advantage of this method is that it can use wet biomass when water is used as solvent.
Liquefaction was initially applied in coal liquefaction, in the 1920s in Germany. Later,
biomass was used as raw material. In the 1990s, hydrothermal liquefaction was developed,
and nowadays it is the most common process used in liquefaction.
However, there are some limitations in the use of liquefaction. The yield of bio-oil is
lower than that obtained in the pyrolysis method, the quality is bad (similar to tar), and the
cost is high due to the fact that it requires higher reaction temperature and pressure, cata-
lysts, and reactants. Thus, other methods (moderate acid-catalyzed liquefaction or MACL)
have attracted interest, as they can be carried out at atmospheric pressure and lower tem-
perature, using organic solvents and a catalyst.
The bio-oil yield is low when the biomass has a high lignin content as it is difficult to
degrade and a substantial part remains as solid waste and ash content, since there is less
organic matter and more solid waste.
Catalysts, either acid or alkaline, can be used in hydrothermal liquefaction. The former
are more effective, but are more corrosive to the equipment. The most used solvent is water,
although the product obtained has some drawbacks: bio-oil and water are insoluble, the
yield is low, the bio-oil has a high oxygen content and the heating values are low. Therefore,
organic solvents are used (Jiang et al. 2018).
Hydrothermal liquefaction can be used as an upgrade treatment of biomass for use with
other methods. For example, better gasification performance is achieved.

11.6 Pyrolysis
Pyrolysis is the process that occurs when heat is applied to a material (biomass) in an inert
atmosphere. The products obtained are solids (e.g. charcoal), liquids (e.g. bio-oils), and
incondensable gases (e.g. methane, H2 , and others). We will tackle those processes carried
out in the absence of gasifying agents or liquids and which are studied in other sections
of this chapter. The products and ratios formed vary depending on the reaction parameters
(Table 11.1).

Table 11.1 Relationship between products and parameters in pyrolysis.

Product Temperature Heating rate Vapor residence time

Solid Low Low Long
Liquid Moderate Moderate Short
Gas High High Moderate
 Thermochemical Processes for the Transformation of Biomass into Biofuels 297

A common problem is the high moisture content of the biomass. It requires a high amount
of energy to produce evaporation. If a liquid fuel is desired, it may contain a high amount
of water, which decreases its energy capacity.
Pyrolysis is an endothermic process in which most of the energy is used to raise the
temperature of the material, especially in the evaporation of water. Some common methods
to provide heat to the reactor are:
– Heating the external surface of the reactor, which is done when they are small in size.
– Heating the carrier gas.
– Heating a solid carrier, such as inert sand or a solid catalyst.

11.6.1 Slow Pyrolysis

This has been used traditionally for the production of charcoal, and today it is utilized in
traditional, semi-industrial and industrial furnaces. The global production of wood charcoal
was estimated at 52 million tons in 2015 (FAO 2017). This method is characterized by slow
heating rates, low-to-medium temperatures, and long processing times (in some cases, sev-
eral days). The main product is charcoal. Due to its importance, and its usual manufacturing
method, which is an intermediate between slow pyrolysis and combustion, it is studied in
this chapter as an independent method.

11.6.2 Fast Pyrolysis

This method is characterized by short times and high heating rates. To achieve a faster pro-
cess, generally the biomass is supplied dry (<10%) and in fine particles (<3 mm) (Bridgwa-
ter 2012; Verma et al. 2012). Liquids are extracted as condensable vapors, which condense
later outside of the reactor. The yield in the liquid can reach 75%, with small quantities of
charcoal and gas.
The bio-oil obtained has several problems: high TAN (total acid number), oxygen and
water content, low heating value, and poor miscibility with crude oils (Bhaskar and Pandey
2015). The high oxygen content is responsible for the high TAN (acidic oxygenated
functional groups) and the oxygen and water contents are responsible for the low heating
value (lower carbon content) and poor miscibility (polar compounds). To overcome these
problems, the process is carried out in the presence of hydrogen (hydropyrolysis). With a
reducing agent, the oxygen content decreases. Thus, the cellulosic biomass can be directly
converted into diesel or gasoline (Marker et al. 2012).
It should also be borne in mind that the biofuel obtained is not thermodynamically stable.
When heated or during storage, it can give rise to reactions that, in general, increase the
molecular mass and produce water.
Another problem related to the quality of biofuel is the presence of small particles of
ash or char. In addition to the difficulties caused by the solid state, fine char catalyzes reac-
tions that increase molecular mass, as mentioned in the previous paragraph, which in turn
increases viscosity.

11.6.3 Flash Pyrolysis

It can be considered an extremely fast pyrolysis. The heating rate is high (near 1000 ∘ C/s),
as well as the process temperature (near 900 ∘ C), with a very short residence time (about
298 Process Systems Engineering for Biofuels Development

0.5 seconds). Necessarily, the biomass must be supplied as very fine, dust-like particles.
Thus, the smaller particle size achieves a much faster heat transfer. As in fast pyrolysis, the
main product is bio-oil.
There are several types of reactor for fast and flash pyrolysis (Verma et al. 2012; Guda
et al. 2015). As general properties, the following can be stated:
– High heat transfer rate.
– Good temperature control.
– Moderate-to-high maximum temperature.
– Fast quenching of gases or vapors (short residence times).
– Fast cooling and condensation of vapors.
Bridgwater (2012) created a table of the fast pyrolysis reaction systems operational in
2012 with capacity and localization, classified by reactor type.

11.6.3.1 Fixed Bed

Fixed bed reactors are not being used commercially for fast or flash pyrolysis. They consist
of a fixed bed of raw materials that is pyrolyzed, heated by external application of heat to
the reactor. The process uses a batch methodology, which renders non-profitable results.
In them, the solid hardly moves, depending on the matter transported by the flow of gases.
Larger particles can be used, although the transport effect (heat and matter) must be care-
fully studied. The yield of bio-oil is comparable with other reactors. However, it can raise
maintenance problems, such as clogging by char and the tar compounds produced, which
increases the resistance to the gas flow. They are used mainly for research, small-scale
applications and in combustion.

11.6.3.2 Fluidized Bed

These are probably the most popular and studied pyrolysis reactors. Particles (of small
size) are maintained as a hydrodynamically stable bed using an inert gas flow. The
biomass heating rate is generally the limiting factor. They are more profitable than fixed
bed reactors, since the biomass input and product output are continuous. Sometimes, a
carrier material, such as sand, is used to produce a good heat transfer, maintain constant
temperature, and increase abrasion of biomass particles. A pressurized gas is employed
to fluidize the mixture. In principle, the lack of mechanical parts allows for a simple and
low-cost design. However, when the scale increases, the cost of energy and gas blowers
also increases and becomes important. Another necessary aspect of the reactor is the
use of cyclones to separate char particles. Char could catalyze the cracking of pyrolysis
vapor. Nowadays, there are reactors that operate with low maintenance needs and are
economically profitable. The char obtained is near 15% of the mass and 25% of the energy
of the biomass, which can be used to provide heating by combustion or separated and sold.
In this case, another fuel is required. A schematic of this reactor is shown in Figure 11.4;
it is similar to a fluidized bed gasification reactor.
 Thermochemical Processes for the Transformation of Biomass into Biofuels 299

11.6.3.3 Recirculating Fluidized Bed

These are a more complex version of fluidized bed reactors. In them, the char obtained is
burned in a second fluidized bed reactor. The heat obtained is used to reheat the carrier par-
ticles. Their main advantage is their higher energy efficiency, becoming a self-sustainable
pyrolysis process. The main problem is the presence of ashes due to the burning of the
charcoal, which can catalyze the cracking of the pyrolysis vapor. Moreover, high energy is
required for gas blowers, since the flow of gas is faster than in fluidized bed reactors. Recir-
culating fluidized bed reactors are useful when the carrier particles are employed also as
catalyst. Combusting the char allows cleaning of the catalyst surface. There is also a similar
reactor for gasification (Figure 11.5).

11.6.3.4 Rotating Cone

This technology does not use an inert/carrier gas. In a rotating cone pyrolyzer (Figure 11.6),
biomass is mixed with hot sand and poured into a high-speed rotating cone in the absence of
oxygen. The rotation causes a better mixing of biomass and sand, which leads to a fast heat
transfer and biomass surface abrasion. The centrifugal force moves the solids to the top of
the cone. When the solids spill over the edge of the cone, vapors are directed to a condenser.

Heated carrier

Gas, vapor

biomass
Combustion

carrier Char, carrier

air

Figure 11.6 Scheme of a rotating cone reactor.

300 Process Systems Engineering for Biofuels Development

The solids drop into a fluid bed and they are sent to a combustor where the organic part is
burned and the carrier is heated in this reaction. The energy cost of this process is smaller
than that of the above. In addition, the requirement of carrier gas is lower in the reactor,
although other gases are required in the combustion reactor and for solid carrier transport.
The high rotating speed of the cone requires continuous monitoring of the operation. It is
also the cause of considerable mechanical wear. An integrated operation is necessary, since
the rotating cone reactor, the combustor and a transport system are required to take the
transporter from the combustor to the pyrolysis reactor.

11.6.3.5 Ablative Reactor

The ablative pyrolyzer requires the biomass to be pressed against a continuously moving hot
surface (600 ∘ C). A combination of mechanical forces and heat produce gasification. One of
the methods proposed is the use of a rotating heated disc against the biomass (Figure 11.7).
Liquids, gases, and aerosols are produced. The liquid, on the hot surface, may undergo
further undesired reactions, and the fast removal of liquids is needed. Another method is
the use of wire-mesh instead of a disc. Sometimes, the biomass must be pretreated, since
small particles or fibrous biomass are not suitable for use in this process. This reactor is
different from those described above since:
– The reaction rates do not depend on heat transfer to biomass particles, but on the rate of
heat transfer to the reactor wall.
– No inert gas is required. The equipment is smaller, the concentration of vapors increases,
and the condensation is more effective.
– No small particles are needed.

Gas, vapor

Hot disc

Pressure
Pelletized biomass

Figure 11.7 Scheme of an ablative reactor.

 Thermochemical Processes for the Transformation of Biomass into Biofuels 301

The process depends on the surface, and thus the scaling is less effective than in other
methods that are dependent on the volume.

11.6.3.6 Screw Reactor

This is also known as an auger reactor. The solids usually require an auger to be trans-
ported. Fortunately, augers can be used to transport biomass through a heated zone. This
design allows to precisely control temperature and residence time. A screw reactor is shown
schematically in Figure 11.8.
Its main advantages are its low energy needs, compact and simple design, and the fact
that it does not need a carrier gas (biomass moves mechanically). Its disadvantage is the
presence of moving parts, which lead to mechanical wear. The use of an inert gas with the
biomass ensures that no oxygen enters the reactor and provides a flow that moves the vapors
through the reactor. When the reactor is narrow, it can be heated from the outside. However,
if the reactor is wide, a solid heat carrier, such as sand or ceramic balls, must be employed.
The residence times are longer than in fluid bed reactors.

11.6.3.7 Vacuum
Pyrolysis can be carried out in vacuum conditions. However, vacuum reactors have some
challenges, such as low mass and heat transfer (due to the absence of gas), expensive
equipment, including a special inlet and outlet for biomass, and pyrolysis by-products.
Nonetheless, they offer control over residence time and reduce the secondary reactions.
The low pressure also allows decreasing the temperature process, involving energy savings.
This method can be used in combination with some of the above described processes.

11.6.3.8 Entrained Flow

The biomass particles are entrained in a carrier gas, with sufficient separation to achieve the
pyrolysis of individual particles. The reactor is heated from the outside, and the particles
Feed

Gas, vapor

Char, carrier

Figure 11.8 Scheme of a screw reactor.

302 Process Systems Engineering for Biofuels Development

Biomass
Gas, vapor

feed

gas

Char

Figure 11.9 Scheme of an entrained-flow reactor.

are introduced through the top. Heat is transferred by the carrier gas or the reactor walls.
Figure 11.9 shows an entrained-flow reactor scheme.
The char particles formed are collected in the bottom of the reactor, and the vapors are
collected in the top and directed to a condensation system. These reactors operate at higher
temperatures (1200–1600 ∘ C) and pressures (2–8 MPa) than other reactors. In addition,
they require a very uniform particle size, which cannot always be achieved with biomass.
Another problem is that ash can melt at operating temperatures. The slag formed can be
corrosive. A second effect of temperature, beneficial in this case, is that it decreases the
formation of tar.

11.6.3.9 PyRos
This is a patented technology (Imran et al. 2018) in which the pyrolysis takes place in a
cyclone. The heat is transferred with a solid inert carrier. The char and vapors are contin-
uously separated in the normal operation of the cyclone. Later, the vapors are condensed
outside of the reactor.
 Thermochemical Processes for the Transformation of Biomass into Biofuels 303

11.6.3.10 Plasma
In this reactor, a plasma zone is created by an electrical field of high voltage. In this zone
there are many electrons and excited molecules, including ions and radicals. In addition,
there is an intense radiation. The raw material is a carbonaceous solid obtained from
biomass, and a previous carbonization process is required. In small particles, the charcoal
is injected in the plasma zone, achieving a fast heating. If medium or large molecules are
released, the plasma can crack them in smaller molecules. Thus, the products obtained are
hydrogen and light hydrocarbons. These reactors need a previous step (carbonization) and
a lot of electrical energy. Besides the fact that the temperatures reached are elevated and it
has a high operating cost, this type of reactor is not widely used.

11.6.4 Catalytic Biomass Pyrolysis

A catalyst is a substance that increases the rate of a reaction without modifying the overall
standard Gibbs energy change in the reaction. The catalyst is both a reactant and a prod-
uct of the reaction (IUPAC). They can be classified as homogeneous (reactants, products,
and catalyst are in the same phase) or heterogeneous (the reaction occurs near an interface
between two phases). If the catalyst is consumed in the reaction, it is called an activator.
Three factors are considered when selecting a catalyst: activity, selectivity, and durabil-
ity. Activity is the ability to produce a reaction, although it cannot be the only factor to be
taken into account, since if many by-products are obtained, it is not profitable. Therefore,
selectivity, which is the ability to obtain the desired product in high percentages, must also
be considered. Finally, in most processes, it is necessary to have a stable, long-lasting cat-
alyst that does not lose its activity. This reduces the cost and facilitates the continuity of
many processes.
Heterogeneous catalysts have some advantages. Since they are (generally) solids they
can be separated from gas or liquids, they are stable, non-corrosive and can be used at
high temperatures, which results in faster reactions. Smaller reactors are needed, and they
can also be reused, which lowers production costs. Frequently, solid catalysts are porous
materials with a large active surface area. Macroporous and mesoporous materials are the
logical choice for reactions that include large molecules, such as biomass (Balagurumurthy
et al. 2015).
Catalysis is often present in biomass heat treatments. This is due to the fact that alkaline
earth metals (frequent in living creatures) catalyze the decomposition of biomass. However,
they generally result in molecules that are too small to be used in some applications (bio-oil,
production of monosaccharides to be used in the biochemical production of other fuels,
etc.). In cases where this is a problem, an acid treatment can be performed to remove the
salts from these metals, although this increases the cost of the process, since, in addition to
the treatment, in many cases it is convenient to carry out a washing process to eliminate the
excess acid.
Catalytic processes can be classified into the following categories:
(a) Vapor catalysis
In this process, the vapors formed in the chemical decomposition of the biomass are
passed through a catalyst bed. This can be in the same reactor that is fed with biomass
304 Process Systems Engineering for Biofuels Development

or in another separate reactor, which allows purifying the vapor stream to remove solids
(carrier, chars) and other unwanted components.
(b) Hydropyrolysis
In this case, pyrolysis is used in the presence of hydrogen gas. It produces bio-oil
with a lower oxygen content.
(c) Direct pyrolysis
In this method, the biomass is previously mixed with the catalyst, or the catalyst is a
component of the heat carrier solids.
A list of catalysts used in the thermochemical conversion of biomass was published by
Dickerson and Soria (2013).
The objectives of using catalysis are very varied. The most immediate one that any chem-
istry student would take into account is the saving of energy in thermochemical processes.
However, this objective, although always interesting, is usually not the most sought after by
researchers, as there are other factors that can decrease economic performance in greater
proportion than energy consumption. It is more interesting to get the reactions going the
right way.
One of the desired objectives is to decrease the oxygen content in the bio-oil obtained.
As already mentioned in Section 11.6.2, as the oxygen content increases, acidity increases,
the calorific power decreases and the polarity increases, reducing the solubility of this prod-
uct with other organic compounds. Another objective may be the selective production of
a compound or a family of compounds. It should be remembered that certain catalysts
with a very narrow pore range, such as zeolites, can result in a size selection of the final
product.

11.6.5 Microwave Heating

Microwave heating can be used to heat biomass. Its main advantage is that the energy is sent
directly to the materials, and that transmission by convection, conduction, and radiation is
not needed. Microwave heating requires a material with a high dielectric constant. Thus,
when microwave heating is used, the biomass is dried quickly. The particles are heated, and
they evolve preferably to form a char.

11.6.6 Product Separation

A final stage of the thermochemical pyrolysis process is the selective condensation of the
formed products. On the one hand, it is desired to eliminate the solid, either because it has
energy value (chars) or because it is the carrier that will be reused. This process is easy to
carry out either with filters, cyclones, or electrostatic separators.
A second issue is to separate water, which is an unwanted component in biofuels. This
can be done by selective condensation of the vapors, controlling the temperature, or with
a liquid spray (water or bio-oil). In the same process, other components can be separated.
This may be due to the existence of organic compounds with a higher value than biofuel,
or because several fractions of different properties are desired.
 Thermochemical Processes for the Transformation of Biomass into Biofuels 305

11.7 Carbonization
The production of charcoal can be carried out by means of slow pyrolysis, as mentioned
in Section 11.6.1, and by means of hydrothermal processes. However, the most common
method is the heating of the biomass with a limited amount of oxygen, in which the par-
tial combustion of the biomass produces the heat needed to obtain charcoal without using
external energy sources. This product is mostly used for heating and cooking in develop-
ing countries, but also for cooking in developed countries (barbecues), in the metallurgical
industry, and to make activated carbon, a widely used adsorbent. Recently, charcoal has
attracted renewed interest due to its potential application as a soil amendment (biochar).
The by-products of carbonization are non-condensable gases (methane, CO and CO2 ,
other hydrocarbons) and condensable vapors. The latter can form two phases: aqueous
and organic. The aqueous phase has water as its main component, but also methanol,
ethanol, acetic and formic acids, acetone and other oxygenated organic compounds of low
or medium molecular masses. The organic phase is constituted by tars, large molecular
fragments that come mainly from lignin. This material can be used as biofuel.
There are changes in the biomass in the process to convert it into charcoal. The main
ones are shown in Table 11.2.
In the carbonization process, the composition of biomass plays an important role. Water,
present in biomass or produced in the process, needs energy to vaporize. Materials with high
water content must be dried before carbonization. Lignin, although it begins to decompose
at a lower temperature than cellulose, has a slower reaction rate. This, together with the
fact that it contains aromatic rings, results in a higher yield if the biomass contains a large
amount of this component.
Another variable of the process is the morphology of the biomass. Large pieces of mate-
rial involve a slow transfer of heat into the interior, thus it is possible that the extent of
carbonization is lower than using smaller particles. The effect is most noticeable if the res-
idence times in the furnace are short. Therefore, if accessible technology does not allow

Table 11.2 Changes in properties in the transformation of biomass into charcoal.

Properties/analysis Concept Variation

%C Increases∗
Elemental (or ultimate) %H Decreases
analysis %O Decreases
%N Increases at medium temperature,
decreases at high temperature
% Fixed carbon Increases∗
% Volatile matter Decreases
Proximate analysis
% Ash Increases
% Water Decreases
Energy (Higher heating value) Increases∗
Mass Decreases
∗ At very high temperatures (>900 ∘ C) and long residence times, the inorganic matter (ash) increases over organic matter,

and these parameters decrease.

306 Process Systems Engineering for Biofuels Development

homogenization or downsizing (which is common in Third World countries), long process

times, sometimes several days, are preferred. It has already been pointed out that in fast
pyrolysis methods, biomass is commonly used in the form of very small particles.
One of the most important uses of charcoal is the manufacture of activated carbon. This
material is an adsorbent widely used in multiple applications. Although the activation pro-
cess is usually carried out at high temperature, given that no fuel is obtained, it will not be
studied here.
Another increasingly interesting use is as “biochar.” This is a charcoal that has been
intentionally produced to be applied as a soil amendment. It has a similar effect to the use
of biomass to improve a crop soil, but with higher resistance (several centuries) to biological
decay. Thus, biochar has high capacity for long-term carbon sequestration. Although it is a
very interesting material due to its ability to extract CO2 from the atmosphere and improve
agricultural productivity, as it is not used as fuel, it will not be studied in this chapter.
There are several types of carbonization furnaces. A first distinction can be made between
batch and continuous furnaces. Some of the most common are described below.
(a) Traditional pits and kilns
These are the most primitive and widely used in the world. They consist of a batch
furnace and the heat required is produced by a limited combustion of biomass. They can
be built on the surface of the ground or by previously creating a hole in the ground. An
example can be seen in Figure 11.10. The wood is stacked to allow air flow, gas exit, and
heat transfer. The wood is covered with a layer of soil, to prevent the free circulation of
air that would result in an uncontrolled combustion. In this layer, several openings are
made to allow the limited entry of air and the exit of gases. Its main advantage is the low
technological level required. Regarding its disadvantages, it is difficult to control the
process, its typical duration is long (several days), and the pollution it produces in the
atmosphere and nearby soil are significant. The yield and the quality are very variable,
depending largely on the ability of the manufacturer.
(b) Brick and steel furnaces
The traditional pits and kilns must be built for every occasion. The logical evolution is
to use a furnace that can be used in multiple cycles. Therefore, the evolution led to using
bricks instead of soil. This results in better thermal insulation and facilitates process
control. Another option for building the furnace is to use steel. Some of these furnaces
can be transported to places rich in raw materials, to lower the cost of transporting
biomass.
(c) Retorts
These are furnaces with external heating. The major advantage is the higher yield,
due to the absence of partial biomass combustion. One possibility is to use as fuel the
vapors produced by the furnace itself. The heat generated by their combustion is used
to heat the retort. It is also possible to use other fuels if they are low in cost.
(d) Lambiotte
This is a continuous process. There are several designs of this furnace, which can be
used as a simple furnace or as a retort. It is a vertical reactor (Figure 11.11), and the bed
moves through the furnace by gravity. Biomass is introduced from the top and charcoal
is obtained in the bottom. The gases move from bottom to top and exit through the upper
zone. In all designs, three vertical zones can be distinguished. In the highest zone, the
 Thermochemical Processes for the Transformation of Biomass into Biofuels 307

exhaust

cover layer

biomass

air inlet

soil

Figure 11.10 Typical structure of a mound kiln.

biomass is dried and preheated due to rising hot gases. Pyrolysis occurs in the middle
zone, where the highest temperatures are reached. In the lower zone, the already formed
charcoal is cooled down. In the simplest design, as a furnace without external heating,
the process is controlled by the air inlet openings, which are located at the bottom of the
middle zone (Figure 11.11). If designed as a retort, the hot gas is introduced in the same
place, but it is also possible to add a cold inert gas flow in the lower zone to help cool
the charcoal. External heating can occur due to the gases extracted during pyrolysis,
although these can also be condensed to obtain valuable products.
(e) Screw reactor
This reactor has already been described in Section 11.6.3.6. If it is used for carboniza-
tion, it is especially interesting in small installations. It can be operated continuously.
(f) Rotary reactor
It is composed of an inclined cylinder, in order to use the force of gravity to move the
biomass. The cylinder rotates slowly to facilitate the movement of solid particles. As in
the case of the Lambiotte furnace, it can be used as a carbonization reactor with limited
air supply or heated externally for use as a retort, and it can be operated continuously.
(g) Hydrothermal reactor
The hydrothermal process, carried out at low temperatures (<200 ∘ C), produces char
as main product. This material can be used for several applications (such as biochar,
adsorbent, and in electrodes) and also as biofuel. It is a solid material with a high oxygen
content, which implies a low energy power. However, it is stable, non-toxic, and easy
to handle, store and transport. In this process, wet biomass such as leaves, manure, or
city waste can be used, solving the problem of their disposal. Thus, it has become an
energy option at a reasonable cost.
Torrefaction is a limited carbonization, performed at low temperatures. There is a drying
and slight decomposition of the biomass (especially the hemicellulose, as it is thermally
308 Process Systems Engineering for Biofuels Development

biomass
gases

drying and
preheating

pyrolysis
air
inlets

cooling

charcoal

Figure 11.11 Scheme of a Lambiotte-type reactor.

unstable compared with cellulose and lignin). Torrefied biomass is drier and has less volatile
matter than raw material. Its calorific value is between that of biomass and charcoal.
Torrefied and carbonized biomass have some favorable properties, such as:
– Hydrophobic nature.
– Higher energy density than biomass.
– Easy to crush or pulverize.
– Stable against moderate heat and most reagents.
– Non-volatile.
Their production makes it possible to reduce transport costs and obtain a more concen-
trated and efficient form of energy than biomass.

11.8 Conclusions
Scientific and technological research have developed a large number of thermochemical
processes for the use of biomass. It is of current interest to obtain fuels from renewable
sources; thus the research and development of these techniques continues. Of the methods
described, some are useful and profitable and are currently used, whereas others have not
 Thermochemical Processes for the Transformation of Biomass into Biofuels 309

been able to reach industrial production due to their technical complexity or lack of perfor-
mance. Therefore, it is necessary to have adequate knowledge of the existing processes in
order to be able to choose the most suitable one. It is also necessary to continue the investi-
gation of these processes, since cheaper fuels and less polluting methods are required, given
the environmental problems that currently exist on our planet.

Acknowledgments
The author thanks the support of Junta de Extremadura and Fondo Europeo para el Desar-
rollo Regional (GRU18035).

References
Ahmaruzzaman, M. (2010). A review on the utilization of fly ash. Progress in Energy and Combustion
Science 36 (3): 327–363. https://doi.org/10.1016/J.PECS.2009.11.003.
Balagurumurthy, B., Singh, R., and Bhaskar, T. (2015). Catalysts for thermochemical conversion of
biomass. In: Recent Advances in Thermochemical Conversion of Biomass (eds. A. Pandey, T. Bhaskar,
M. Stöcker and R.K. Sukumaran) Chapter 4. Elsevier.
Bhaskar, T. and Pandey, A. (2015). Introduction. In: Recent Advances in Thermochemical Conversion of
Biomass (eds. A. Pandey, T. Bhaskar, M. Stöcker and R.K. Sukumaran) Chapter 1. Elsevier.
Bridgwater, A.V. (2012). Review of fast pyrolysis of biomass and product upgrading. Biomass and Bioen-
ergy 38: 68–94. https://doi.org/10.1016/j.biombioe.2011.01.048.
Canabarro, N., Soares, J.F., Anchieta, C.G. et al. (2013). Thermochemical processes for biofuels production
from biomass. Sustainable Chemical Processes 1 (1): 22. https://doi.org/10.1186/2043-7129-1-22.
Chen, W.H., Lin, B.J., Huang, M.Y., and Chang, J.S. (2014). Thermochemical conversion of microalgal
biomass into biofuels: a review. Bioresource Technology 184: 314–327. https://doi.org/10.1016/j.biortech
.2014.11.050.
Damartzis, T. and Zabaniotou, A. (2011). Thermochemical conversion of biomass to second generation
biofuels through integrated process design – a review. Renewable and Sustainable Energy Reviews 15
(1): 366–378. https://doi.org/10.1016/j.rser.2010.08.003.
Dickerson, T. and Soria, J. (2013). Catalytic fast pyrolysis: a review. Energies 6 (1): 514–538. https://doi
.org/10.3390/en6010514.
FAO (1993). El gas de madera como combustible para motores. http://www.fao.org/3/T0512S/t0512s00
.htm#Contents (accessed 12 April 2020).
FAO (2017). The charcoal transition: greening the charcoal value chain to mitigate climate change and
improve local livelihoods. http://www.fao.org/3/a-i6935e.pdf (accessed 12 April 2020).
Guda, V.K., Steele, P.H., Penmetsa, V.K., and Li, Q. (2015). Fast pyrolysis of biomass: recent advances
in fast pyrolysis technology. In: Recent Advances in Thermochemical Conversion of Biomass (eds. A.
Pandey, T. Bhaskar, M. Stöcker and R.K. Sukumaran) Chapter 7. Elsevier.
Hossain, M.A., Jewaratnam, J., and Ganesan, P. (2016). Prospect of hydrogen production from oil palm
biomass by thermochemical process – a review. International Journal of Hydrogen Energy 41 (38):
16637–16655. https://doi.org/10.1016/j.ijhydene.2016.07.104.
Imran, A., Bramer, E.A., Seshan, K., and Brem, G. (2018). An overview of catalysts in biomass pyrolysis
for production of biofuels. Biofuel Research Journal 5 (4): 872–885. https://doi.org/10.18331/BRJ2018
.5.4.2.
IUPAC (n.d.). Compendium of Chemical Terminology (Gold Book). https://goldbook.iupac.org (accessed
12 April 2020).
310 Process Systems Engineering for Biofuels Development

Jiang, W., Kumar, A., and Adamopoulos, S. (2018). Liquefaction of lignocellulosic materials and its appli-
cations in wood adhesives – a review. Industrial Crops and Products 124: 325–342. https://doi.org/10
.1016/j.indcrop.2018.07.053.
Li, D.C. and Jiang, H. (2017). The thermochemical conversion of non-lignocellulosic biomass to form
biochar: a review on characterizations and mechanism elucidation. Bioresource Technology 246: 57–68.
https://doi.org/10.1016/j.biortech.2017.07.029.
Marker, T.L., Felix, L.G., Linck, M.B., and Roberts, M.J. (2012). Integrated hydropyrolysis and hydrocon-
version (IH2) for the direct production of gasoline and diesel fuels or blending components from biomass,
part 1: proof of principle testing. Environmental Progress & Sustainable Energy 31 (2): 191–199. https://
doi.org/10.1002/ep.10629.
Matsumura, Y. (2015). Hydrothermal gasification of biomass. In: Recent Advances in Thermochemical
Conversion of Biomass (eds. A. Pandy, T. Bhaskar, M. Stöcker and R.K. Sukumaran) Chapter 9. Elsevier.
Montoliú Camps, P. (2005). Madrid en la posguerra, 1939–1946: los años de la represión. Silex.
Pandey, A., Bhaskar, T., Stöcker, M., and Sukumaran, R.K. (eds.) (2015). Recent Advances in Thermo-
chemical Conversion of Biomass. Elsevier https://doi.org/10.1016/C2013-0-00403-3.
Piatak, N.M., Parsons, M.B., and Seal, R.R. (2015). Characteristics and environmental aspects of slag: a
review. Applied Geochemistry 57: 236–266. https://doi.org/10.1016/J.APGEOCHEM.2014.04.009.
Sutton, D., Kelleher, B., and Ross, J.R.H. (2001). Review of literature on catalysts for biomass gasification.
Fuel Processing Technology 73 (13): 155–173.
Verma, M., Godbout, S., Brar, S.K. et al. (2012). Biofuels production from biomass by thermochemical
conversion technologies. International Journal of Chemical Engineering https://doi.org/10.1155/2012/
542426.
 12
Intensified Purification Alternative
for Methyl Ethyl Ketone
Production: Economic,
Environmental, Safety and
Control Issues
Eduardo Sánchez-Ramírez1 , Juan José Quiroz-Ramírez2 , and
Juan Gabriel Segovia-Hernández1
1 Departamento
de Ingeniería Química, Universidad de Guanajuato, Noria Alta s/n, Guanajuato,
Guanajuato, 36050, México
2
CONACyT – CIATEC A.C. Centro de Innovación Aplicada en Tecnologías Competitivas, 37545, León,
Guanajuato México

12.1 Introduction
Currently, the environmental disturbances due to global climate change lead to irreversible
damage and transformation of nature/ecosystems (Stocker et al. 2013). Many political
agreements have been proposed to limit this continuous disturbance. For example, The
Paris Agreement (COP 21, 2015) aims to keep the global temperature rise below 2 ∘ C until
the end of the twenty-first century. However, to achieve this commitment, less than 2900
Gt of CO2 should be emitted in this period with 1900 Gt already having been emitted
(Stocker et al. 2013).

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
312 Process Systems Engineering for Biofuels Development

World
1% Other
n/a Electricity
22 % 3% Renewables
4% Natural Gas
8%
2%
2% 29 %

92 % Oil

37 %

transportation industry residential

agriculture other commerce and services

Figure 12.1 Global CO2 emissions by sector.

On a global scale, the amount of energy consumed by the transportation sector is second
highest (29% in 2015) after that consumed by the industry sector (Figure 12.1); besides, this
energy requirement is almost exclusively based on fossil fuels (95.8% in 2015). In order
to achieve the target, two alternatives are forecast: diminished consumption and alternative
fuels. The energy consumption could be reduced by increasing technological efficiency,
for example, reduced car weight and electrification. Even though electricity has been high-
lighted as a promising alternative for CO2 reduction, biofuels are designated as responsible
for reducing CO2 emissions (Yabe et al. 2012). Currently, two alternatives present the
most mature technological response to replace fossil fuels on the market: bioethanol and
biodiesel. However, many prospects are currently under research to be a real alternative. It
is well-known that bioethanol and biobutanol have been proposed as relatively good alter-
natives; however, despite the maturity in the production technology, the main hurdle in their
production is the low concentration and yield for both alternatives. Additionally, the main
limitation for those biofuels is indeed the energy investment applied in the downstream
processes, as aforementioned, the diluted products obtained in fermentation require a large
amount of energy to be invested in their purification. Additionally, many other hurdles are
related to the use of bioethanol or biobutanol (Table 12.1).
On the other hand, when biofuels production is approached, the main issues to overcome
are the energy consumption in the production process in comparison with the energy density
of the pure biofuel, and the energy investment per kilogram of the component of interest.
For example, considering one of the latest biofuel proposals using biobutanol, Table 12.2
shows the energy consumption (represented as MJfuel /kg) necessary to purify biobutanol
depending on the technology used in the downstream process.
Note, as long as the energy requirements decrease, the energy profit in the whole process
increases and consequently the biofuels acquire a sustainable sense. So, the research of
promising biofuels must be oriented in such a way that the energy requirements be as small
as possible.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 313

Table 12.1 Comparison between different frequently discussed alternative fuels (Nakata et al. 2006; Jin
et al. 2011; Hoppe et al. 2016a).

Compound Discussion
Ethanol Advantages
Engine: High RON; increased engine efficiency; increased combustion
efficiency; broader flammability limits; increased flame speed; miscible with
gasoline
Environment: Reduced Nox and HC emissions; reduced CO emissions; reduced
PM emissions
Disadvantages
Engine: Cold start problems; increased specific fuel consumption; low energy
density; corrosive to the fuel system; incompatibility with a classical seal;
increased oil dilution
Environment: Completely miscible with water; increased HC emissions;
increased aldehyde emissions
Health and safety: Low flame luminosity; toxic
Status
Only liquid renewable gasoline alternative in place today
n-Butanol Advantages
Engine: Increased engine efficiency; not hygroscopic; less corrosive than
ethanol; possible drop in fuel; lower tendency to vapor lock (compared with
ethanol); less prone to cold start problems than ethanol; miscible with gasoline
Environment: Reduced Nox emissions; reduced PM emissions; established
industrial process from renewable sources
Health and safety: Low volatility
Disadvantages
Engine: Low heat value; increased oil dilution
Environment: Increased HC emissions
Status
Well-established; numerous experimental studies;
comprehensive review including production pathways
Methyl ethyl ketone Advantages
Engine: Improved cold startability (compared with ethanol); RON close to
ethanol; lower oil dilution than ethanol; increased combustion stability;
miscible with gasoline
Status
Proposed by Hoppe et al. in 2016; already tested in spark ignition engines

Despite many research efforts, a simple comparison between the energy and fuel to purify
the biofuel is sometimes more significant than the hypothetical energy obtained by the com-
bustion of the biofuel. Nevertheless, remarkable efforts have focused on the purification of
bioethanol and biobutanol. For example, for biobutanol purification, Mariano and Filho
(2012) reported an energy consumption of 12.8 MJfuel /kgbutanol using distillation. More-
over, many other separation units have been studied, namely gas stripping (Qureshi et al.
2005), vaccum evaporation (Qureshi et al. 2005), liquid–liquid extraction (Groot et al.
1992), and azeotropic distillation (Atsumi et al. 2008), which have reported 18.9, 21.8,
15.6, and 21.8 MJfuel /kgbutanol , respectively.
Another interesting liquid fuel is bioethanol, however, in the same way as for biobu-
tanol, the energy consumption associated with the downstream process is a big hurdle.
For example, Cardona-Alzate and Sánchez-Toro (2006) reported an energy consumption of
314 Process Systems Engineering for Biofuels Development

Table 12.2 Energy requirements of several recovery systems.

Energy
requirement
Biocatalyst Recovery system (MJfuel /kg product ) References

C. acetobutylicum SolRH Distillation 12.8 Mariano and Filho (2012)

C. beijerinckii P260 Distillation 16.7 Mariano and Filho (2012)
C. beijerinckii BA101 Distillation 15.2 Mariano and Filho (2012)
Steam distillation 21 Qureshi et al. (2005)
Gas stripping 18.9 Qureshi et al. (2005)
Adsorption 7.1 Qureshi et al. (2005)
Pervaporation 11.9 Qureshi et al. (2005)
Liquid–liquid extraction 7.7 Qureshi et al. (2005)
Vacuum evaporation 21.8 Qureshi et al. (2005)
C. saccharoperbutylacetonicum Double effect 8 Hugo et al. (2016)
N1–4 distillation
Pervaporation 9.6 Hugo et al. (2016)
Vacuum evaporation 10.8 Hugo et al. (2016)
C. acetobutylicum Adsorption 36.7 Groot et al. (1992)
Gas stripping 23.3 Groot et al. (1992)
Liquid–liquid extraction 15.6 Groot et al. (1992)
Pervaporation 10 Groot et al. (1992)

23.9 MJ/kgethanol using a process with conventional distillation columns. However, contin-
uous research has promoted the energy investment for this separation below 10 MJ/kgethanol
(Hernández 2008; Errico et al. 2013). Even though many biofuels have been proposed, it
is mandatory to research new alternatives regarding biofuels with a multidisciplinary scope
which further produces either more concentrated biofuels or more options for implementa-
tion as biofuels.
A very useful tool to reduce energy requirements is the process intensification concept.
An intensified process can be classified as enhancements achieved through (i) the integra-
tion of operations, (ii) the integration of functions, and (iii) the integration of phenomena
(Lutze et al. 2010). It may involve a process distinguished by five characteristics: reduced
size of equipment, increased performance of process, reduced equipment inventory, reduc-
tion in using utilities and raw materials, and increased efficiency of process equipment
(Ponce-Ortega et al. 2012).
Among novel alternatives, methyl ethyl ketone (MEK) is identified as a promising bio-
fuel. Note, for example in Table 12.1, the many advantages in comparison with bioethanol
and biobutanol. MEK (2-butanone) is a standard compound used as a solvent, synthesized
from C4-raffinates (Bohnet 2003). However, it has also identified as a hopeful biofuel for
spark ignition engines. With current technologies, and accomplishing the same engine effi-
ciency, MEK provides a higher heat of combustion, less oil dilution, lower hydrocarbon
emissions, and better cold-start properties (Hoppe et al. 2016b).
Despite its application as an industrial chemical being well studied, a possible applica-
tion as biofuel must be comparable with current liquid fuels. Nevertheless, MEK bio-based
production is not yet well based. Currently, several works have proposed biotechnologi-
cal conversion starting from pure sugar as raw material. This best-case assumption would
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 315

be considered as the upper limit for MEK production, however, the reported yields differ
greatly. MEK might be produced by direct fermentation of sugar, however, its production is
quite poor with yields of approximately 0.004 gMEK /gglucose (Yoneda et al. 2014). Another
bio-based alternative is the production of MEK via decarboxylation of levulinic acid, which
might be obtained initially from lignocellulosic material. However, all reported yields are
not relevant even if acetic acid and acetone are produced as by-products (Gong et al. 2010).
A quite promising alternative to produce MEK is utilizing 2,3- butanediol (2,3-BD) as
intermediate. An interesting result of this route is the relatively high yield of 2,3- BD via
fermentation, with yields reaching values near to the theoretical limit of 0.5 g2,3-BD /gglucose
(Ji et al. 2011; Syu 2001). Further, the direct dehydration of 2,3-BD is performed with yields
higher than 95% (Emerson et al. 1982).
Notwithstanding the relatively high yields for 2,3-BD fermentation and further dehy-
dration the downstream process is not well explored so far. Besides, by-products such as
acetoin and isobutyraldehyde (IBA) are valuable products, which may increase valorization
in a bio-based refinery for MEK production since they are important in the food industry
and in the biosynthesis of isobutanol, respectively (Xiao and Lu 2014).
Note, despite those promising thermodynamic properties, MEK purification is still chal-
lenging since two azeotropes are present in the MEK-IBA-acetoin-water mixture.
Distillation is always considered as the first option for this kind of challenging separa-
tion. On account of the maturity of the technology, in the field of biofuels, many authors
have considered distillation as an option for biofuels purification, for either bioethanol
(Cardona-Alzate and Sánchez-Toro 2006) or biobutanol (Mariano and Filho 2012). How-
ever, the main drawbacks of using distillation columns are energy consumption and thermo-
dynamic efficiency. The aforementioned processes may help to overcome those drawbacks;
a clear example of such a task is the hybrid process.
The hybrid process based on liquid–liquid extraction becomes more profitable since the
inclusion of an extracting agent helps in breaking azeotropes, consequently the energy
invested and cost associated with the separation process is reduced. Regarding the hybrid
process, some works have shown the potential of including a liquid–liquid extraction
column in order to mitigate the energy requirements and consequently to improve many
performances indexes (Groot et al. 1992, 1990; Errico et al. 2016). This consideration
would lead MEK to a more sustainable place in comparison with petroleum fuels
(Penner et al. 2017).
On the other hand, when the separation process is approached, many studies are focused
on reducing, initially, the economic impact and sometimes the environmental impact
(Sánchez-Ramírez et al. 2015a). The inclusion of at least two objective functions is
considered nowadays, for example, the total annual cost (TAC) and any environmental
index (Gutiérrez-Antonio 2016). However, other key objectives are commonly set aside
(Govasmark et al. 2011; Law et al. 2011; Pokoo-Aikins et al. 2010). For example, both the
control properties and the inherent risk are not considered in the first design stage. This
two-step methodology may present several drawbacks for both considerations. Regarding
control properties, many separation schemes may present infringement of dynamic
restrictions, over-design, and low performance, so a global performance of any proposed
design cannot be guaranteed (Zhou et al. 2015). Other transitory consequences may be
observed in process schemes with poor flexibility on operative performance. Regarding
the inherent safety, this traditional approach can generate separation alternatives with high
316 Process Systems Engineering for Biofuels Development

risk related to the use of heat duties, explosive liquids, and the size of the column. Note,
for example, the purification of MEK involves handling relatively dangerous components
such as MEK, acetoin, and IBA which according to the Hazardous Materials Identification
System (HMIS) possess a rating scale of health, flammability, and physical hazard of 2-3-0,
2-2-2, and 2-3-0, respectively. In this way, not including risk analysis in the separation
processes can generate a misleading assessment of the risk of the global process.
With this in mind, the aim of this chapter is to design and optimize some separation
schemes to purify MEK. This proposal can generate purification schemes that consider
a balance between several objective functions, in other words, through joint design and
optimization, alternatives will be obtained considering economic, environmental, control-
lability and inherent safety aspects according to the current needs of the industry.

12.2 Problem Statement and Case Study

The quaternary mixture involved in this separation is formed by MEK, water, 2,3-BD, and
IBA. At first sight, it would look not so complicated, however, there are two binary hetero-
geneous azeotropes (MEK-water and IBA-water), both azeotropes at 1 atm (Figure 12.2). In
this work, the composition of components is 65 wt% MEK, 18 wt% water, 10 wt% 2,3-BD,
and 7 wt% IBA (Table 12.3), which is a relatively common outlet stream of a dehydra-
tion reactor (Emerson et al. 1982; Zhao et al. 2016; Tran and Chambers 1987). So, this

0.1 0.9 73.68 C 0.9

0.2 0.8 0.2 0.8

0.3 0.7 0.3 0.7

73.68 C +
0.6
R

0.4 0.6 0.4

ME

ME
TE

TE
WA

WA

+
K

0.5 + 0.5 0.5 0.5

0.6 + 0.4 0.6 + 0.4

+ 0.3 + 0.3
0.7 0.7
+ +
0.8 + 0.2 0.8 0.2
+ +
0.9 0.1 0.9 0.1
+ +
60.97 C v 60.97 C

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
IBA IBA
(a) (b)

Figure 12.2 Heterogeneous azeotropes MEK/water and IBA/water: (a) mole basis and (b) mass basis.

Table 12.3 Feed characterization.

Feed Concentration (wt%) Vapor fraction Flowrate (kg/h) Temperature (K)

Water IBA 2,3-BD MEK
0.18 0.07 0.1 0.65 0 11 764.7 298
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 317

thermodynamic hurdle and the multi-objective evaluation purpose makes this development
relatively complex.
For the separation task, Penner et al.(Penner et al. 2017) showed a conceptual design for
the separation of MEK with four promising sequences based on conventional distillation
(Figure 12.3). Note one drawback of the proposal by Penner et al. (2017) is the recovery of
all components in the feed mixture, which is an interesting issue to be solved. During this
work, initially, the separation alternatives will be revisited; they will be rigorously evalu-
ated by means of multi-objective optimization considering four targets, namely the TAC,
the eco-indicator 99 (EI99), the condition number, and the individual risk (IR) as economic,
environmental, controllability and safety indexes, respectively. Moreover, the role of the
design variables will also be explored. Additionally, an intensified process (hybrid process
based on liquid–liquid extraction columns and distillation columns) was also proposed, con-
sidering p-xylene as entrainer (Murphy 2000). The process schemes were initially modeled
in Aspen Plus 8.8, and the thermodynamic properties were estimated with the NRTL-HOC
model (Penner et al. 2017).
As a general description, in the first four schemes (S1–S4), the heaviest component
(2,3-BD) leaves the downstream process either by the first or second column, always as
a bottoms product. Excluding scheme S2 (Figure 12.3b), the distillate stream of the initial
column is sent to a decanter, with a top phase of MEK jointly with water and IBA. Further,
without a remarkable split between water and IBA and water, the schemes S1 (Figure 12.2a)
and S4 (Figure 12.3d) demand five conventional columns to separate MEK at high purity,
on the other hand, S2 (Figure 12.3b) and S3 (Figure 12.3c) require only four distillation
columns for the same purpose. Note, scheme S3 lacks a distillation column since the IBA
stream is further mixed with the distillation stream of the first column to break up the dis-
tillation boundary. Finally, the last column has the purpose of purifying IBA at the bottom,
while the azeotrope is recycled back to the decanter. On the other hand, in the hybrid process
S5 (Figure 12.3e), the first liquid–liquid extraction column aims to facilitate the azeotrope
separation producing at bottoms a stream with water and 2,3-BD, and at the top a mixture
of IBA, MEK, p-xylene, and traces of water. Furthermore, five columns perform the purifi-
cation of all components. In the case of IBA and MEK, the purification is developed in
two steps since the traces of water result in the use of more columns. In order to use MEK
as biofuel, the product purities in all processes are 99.5 wt% MEK, 99.5 wt% 2,3-BD, and
99 wt% IBA (Hoppe et al. 2016b). Also in the case of the solvent, p-xylene recovery is
obtained above 99.9 wt%.

12.3 Evaluation Indexes and Optimization Problem

For operative purposes, an optimized plant of MEK is mandatory to run a bio-based MEK
industry, in order to contend effectively with current biofuels and, in future terms, with
those oil-derived fuels.
Besides, for current normativity and in a wider scope, any process must accomplish sev-
eral metrics. For example, Jiménez-González et al. (2012) mentioned several metrics that a
sustainable process must consider. In their work, Jiménez-González et al. (2012) highlight
some factors and metrics such as the economics of the process, its environmental impact, the
 (a) (b) (c)

C5 C5
C4
C2 Water(IBA,MEK) C4
C3 Water(IBA)
C3 Water Water
C1 C1
Water Water(IBA,MEK) IBA IBA
C1
IBA MEK MEK
MEK IBA 2,3-BD IBA 2,3-BD
WATER IBA
2,3-BD
WATER C5 C2 MEK
MEK
C2 2,3-BD
Water(IBA,MEK)

S1 S2 S3
MEK
2,3-BD
2,3-BD

IBA(Water
traces)
(d) (e) C3 Water(IBA
C4
and MEK
C1 traces) IBA
C5 Water
IBA C4
C3 Water(IBA) MEK MEK
2,3-BD
Water
C1 Water(IBA,MEK)
IBA
MEK IBA
2,3-BD P-XYLENE MEK
WATER C4 WATER
C2 C2

S4 S5
MEK
2,3-BD 2,3-BD

Figure 12.3 Separation alternatives for the purification of MEK: (a) scheme S1, (b) scheme S2, (c) scheme S3, (d) scheme S4, and (e) scheme S5.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 319

controllability of the process, and also the safety. With this in mind, the indexes considered
in the objective function are mentioned below.

12.3.1 Total Annual Cost Calculation

For the economic evaluation, in this study, the economics are evaluated by means of the
TAC. Initially, this method was reported by Guthrie (1969). The cost estimation is per-
formed considering separated units, in other words, the method calculates the capital cost
as the sum of the annualized cost, plus the operating cost. The TAC calculation is carried
out with the following equation (Turton 2001).
Capital costs
TAC = + Operating costs (12.1)
Payback period
The payback period is considered to be five years with an annual operation time of
8500 hours. In addition, the cost related to heating and cooling water were taken from
Luyben (2011).

12.3.2 Environmental Index Calculation

In this study, the environmental impact is measured by means of EI99. This environmental
indicator was previously proposed by the National Institute of Public Health and Envi-
ronment (RIVM) in Switzerland and the proposed methodology is based on a life cycle
analysis (LCA). The values associated with EI99 are available for different materials, pro-
duction processes, power generation, etc. The absolute value of the eco-indicator has no
relevance since its main purpose is to compare the relative difference between different
products. The scale of the eco-indicator was designed in such a way that the value of 1
point represents one-thousandth of the environmental burden of an average citizen in the
European Union (Goedkoop and Spriensma 2000). Thus, by means of EI99, it is possible to
calculate the environmental load associated with both a process or product, even quantify-
ing a load of a material or energy used. Using this methodology, many works have improved
the environmental performance of several processes (Errico et al. 2016; Penner et al. 2017;
Sánchez-Ramírez et al. 2015a; Gutiérrez-Antonio 2016; Govasmark et al. 2011; Law et al.
2011; Pokoo-Aikins et al. 2010; Zhou et al. 2015; Zhao et al. 2016; Tran and Chambers
1987; Murphy 2000; Jiménez-González et al. 2012; Guthrie 1969; Turton 2001; Luyben
2011; Goedkoop and Spriensma 2000; Alemam et al. 2017; Marcinkowski 2017; Kumar
1996; Kumar and Gayen 2011; Medina-Herrera et al. 2014; Crowl and Louvar 2001; De
Haag and Ale 2005; Bequette and Edgar 1989; Lau and Alvarez 1985; Gómez-Castro et al.
2008; Murrieta-Dueñas et al. 2011; Hovd et al. 1997; Sågfors and Waller 1995; Klema and
Laub 1980; Moore 1986; Segovia-Hernández et al. 2002; Górak and Olujic 2014; Storn and
Price 1997; Madavan 2002; Glover 1989).
The methodology associated with the calculation of EI99 considers 11 impact categories.
These 11 categories are separated into three groups according to the nature of the dam-
age caused: (i) human health, (ii) damage to the ecosystem, and (iii) damage to natural
resources. In this study, for the calculation of EI99, three impact factors were considered:
steam for heating, electricity used for pumps, and steel used in construction of the main
320 Process Systems Engineering for Biofuels Development

Table 12.4 Unit eco-indicator used to measure EI99 in all cases of study.

Impact category Steel (points/kg) Steam (points/kg) Electricity (points/kWh)

Carcinogenics 6.32E-03 1.18E-04 4.36E-04
Climate change 1.31E-02 1.60E-03 3.61E-06
Ionizing radiation 4.51E-04 1.13E-03 8.24E-04
Ozone depletion 4.55E-06 2.10E-06 1.21E-04
Respiratory effects 8.01E-02 7.87E-07 1.35E-06
Acidification 2.71E-03 1.21E-02 2.81E-04
Ecotoxicity 7.45E-02 2.80E-03 1.67E-04
Land occupation 3.73E-03 8.58E-05 4.68E-04
Fossil fuels 5.93E-02 1.25E-02 1.20E-03
Mineral extraction 7.42E-02 8.82E-06 5.7EE-6

Source: Adapted from Goedkoop and Spriensma 2000.

equipment. The values of these three factors are shown in Table 12.4. The data associ-
ated with these factors are commonly taken from standardized databases (Goedkoop and
Spriensma 2000). EI99 is defined as follows:
∑∑∑
EI99 = 𝛿d 𝜔d 𝛽b 𝛼b,k (12.2)
b d k∈K

where 𝛽 b represents the total amount of chemical b released per unit of reference flow
due to direct emissions, 𝛼 b, k is the damage caused by category k per unit of chemical b
released to the environment, 𝜔d is a weighting factor for damage in category d, and 𝛿 d is
the normalization factor for damage of category d.

12.3.3 Individual Risk Index

In order to quantify the inherent safety of a process, it is possible to evaluate the IR of the
same. IR is defined as the risk a person has considering their physical position in the process
plant. Within the calculation of the risk, the damage caused by injury or death, its frequency
and the probability of affectation is considered. The calculation of IR does not depend on
the number of people exposed to risk.
To model the IR, it is necessary to multiply the frequency of an accident (fi ) by the
probability of affectation in a specific position (Pxy ) according to Eq. (12.3).
∑
IR = fi Px,y (12.3)

The frequency and probability of involvement can be determined by quantitative risk

analysis (QRA). This methodology allows to identify potential accidents and evaluate
their consequences and damages. Initially, the QRA identifies possible accidents, which
is any release of matter or energy (Kumar 1996). For distillation columns, incidents can
be grouped as continuous and instantaneous releases. These incidents were previously
determined through a risk and operability analysis (HAZOP). Figure 12.4 shows a tree of
events that includes the possible accidents associated with a distillation column as well
as their respective frequencies (fi ). Similarly, the frequencies were taken from a previous
study of the American Institute of Chemical Engineers (Kumar 1996). The event tree
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 321

INSTANTANEOUS RELEASE fi

BLEVE 5.75x10–6/yr
Immediate
Ignition
P1 = 0.25
UVCE 7.76x10–6/yr
Delayed P3 = 0.5
Ignition
P2 = 0.9
2.3x10–5/yr

P3 = 0.5
Flash Fire 7.76x10–6/yr

No Immediate
Ignition No Ignition
P1 = 0.75 P2 = 0.1
Toxic Release 1.55x10–6/yr

CONTINUOUS RELEASE

jet Fire 3.67x10–5/yr

Immediate
Ignition
P1 = 0.1
Delayed
Ignition
P2 = 0.75
3.67x10–4/yr Flash Fire 2.48x10–4/yr

No Immediate
Ignition
P1 = 0.9 No Ignition
P2 = 0.25
Toxic Release 8.26x10–5/yr

Figure 12.4 Tree of events diagram for distillation columns.

considers the following possible accidents: boiling liquid expanding vapor explosion
(BLEVE), unconfined vapor cloud explosion (UVCE), flash fire, and toxic release. The
continuous releases are: jet fire, flash fire, and toxic release.
Once the possible accidents have been identified, the causative variables must be known
as the second step. For example, for BLEVE, jet fire, and flash fire the causative variable is
thermal radiation (Er ), for UVCE it is the overpressure (Po ), and for toxic release it is the
concentration (C). In this work, a distance of 50 m was considered for all the variables The
calculations of causative variables for each accident have been shown previously by many
authors (Kumar and Gayen 2011; Medina-Herrera et al. 2014).

12.3.3.1 Probability of Affectation (Consequences Analysis)

As a final stage for the calculation of IR, the probability of affectation must be calculated
(probability of injury or death). This probability can be obtained with probit models. A
probit model relates a person’s response to the dose received from a certain incident, such
as heat, pressure, or radiation. In this work, a probit function was used to model the proba-
bility of death due to overpressure and third-degree burns (Kumar 1996; Crowl and Louvar
2001; De Haag and Ale 2005). The parameters for Eq. (12.4) are shown in Table 12.5. The
322 Process Systems Engineering for Biofuels Development

Table 12.5 Probit parameters.

k1 k2 V
( )
4∕
te Er 3
Thermal radiation −14.9 2.56
104

Overpressure −77.1 6.91 po

probability of damage is obtained by replacing probit values in Eq. (12.5) (De Haag and
Ale 2005).
Y = k1 + k2 ln V (12.4)
[ ( )]
Y −5
Px,y = 0.5 1 + erf √ (12.5)
2
Finally, the result from Eq. (12.5) is used, together with data of LC50 , in Eq. (12.3) to
obtain IR.

12.3.4 Controllability Index Calculation

Among several alternatives to calculate the control properties of separation schemes, a rel-
atively well-known technique is the calculation of the condition number using the singular
value decomposition (SVD) of the relative gain matrix of the evaluated separation scheme at
a nominal point, obtained in an open-loop control policy. The application of this techniques
is not novel; some authors have used this kind of methodology for studying conventional
distillation columns (Bequette and Edgar 1989; Lau and Alvarez 1985). Besides, this tech-
nique has also been shown to evaluate more complex distillation systems (Gómez-Castro
et al. 2008; Murrieta-Dueñas et al. 2011). However, those controllability tests were per-
formed as the second task in the design process in contrast to the proposal of this work. In
the studies where separation schemes have been analyzed, similar to those analyzed here,
the calculation of the SVD has shown that, under certain conditions, a multivariable model
can be decomposed into an SVD structure (Hovd et al. 1997). Sågfors and Waller (1995)
have shown that in processes where distillation columns are involved, an SVD system is a
feasible control structure, even with a lack of a quantitative model.
The decomposition into singular values of a given matrix can be represented as the prod-
uct of three components:
K = UΣV T (12.6)
where K is a n × m matrix, U is a n × n orthonormal matrix called the “left singular vector,”
and V is an m × m diagonal of scalars commonly called “singular values.” These values can
be organized, according to their value, for example: 𝜎 1 > 𝜎 2 > 𝜎 3 … 𝜎 m > 0. The product
obtained by dividing the maximum singular value by the minimum singular value is known
as the condition number of the gain matrix. The condition number is a measure of the
relative difficulty that a decoupled multivariable control problem can present (Klema and
Laub 1980). The condition number is calculated as follows:
𝛾 = 𝜎max ∕𝜎min (12.7)
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 323

In terms of the dynamics of a process, a high value of condition number means difficulty
to meet all control objectives (regardless of the appropriate strategy). A large condition
number is evidence of the relative sensitivity of a case study in a multivariable direction
being very weak (Moore 1986). SVD methodology does not predict or solve all the dynamic
problems in real chemical plants, however, it is relatively easy to understand and identify
basic control difficulties.
For the control analysis, each purification alternative provides a relative gain matrix in
its nominal state. To obtain this matrix, the schemes are subjected to a step change in a
manipulable variable (reflux ratio, reboiler duty, etc.). The magnitude of the disturbance
is small enough (0.5%) that a first-order behavior can be assumed according to many pre-
vious works (Murrieta-Dueñas et al. 2011; Segovia-Hernández et al. 2002). To avoid the
SVD dependence of the system unit used (variables limited between 0 and 1, and high
values for reflux ratio and reboiler heat duties), the approach of the proposal used here
is to limit the variables described. Since the maximum opening of the control valves can
be twice the nominal value, the valves are theoretically open by 50%. In this way, to
generate the relative gain matrix, a step change must be applied in the manipulated vari-
able and subsequently, this change must be divided by two. With this consideration, you
get the same range of variation when opening and closing the control valves. The con-
sequence of this consideration is to relate both the amount of change and the magnitude
of change in a range of 0–100%. Moreover, with this form of scaling, and with the term
1/2P in Eq. (12.8), the manipulated variables are simultaneously dimensionless standard-
ized. For example, a relative gain matrix for the purification of three components could be
stated as:
v sp v sp v sp
⎡ xC11 −xC1 xC12 −xC1 xC13 −xC1 ⎤
⎢ 2P 1 1
P 1
P ⎥
⎡K11 K12 K13 ⎤ ⎢ v1 sp v2 sp v3 sp ⎥
2 2

⎢K21 K22 K23 ⎥ = ⎢⎢ xC21−xC2 xC21−xC2 xC21−xC2 ⎥⎥ (12.8)

⎢ ⎥ P P P
⎣K31 K32 K33 ⎦ ⎢ v 2 2 2 ⎥
⎢ x 1 −xsp xv2 −xsp xv2 −xsp ⎥
⎢ C31 C3 C31 C3 C31 C3 ⎥
⎣ 2P 2
P 2
P ⎦

where all elements Kij are the relative gain matrix. The elements of the first row on the
right-hand side correspond to the differences among the mass purity of component A in
sp V
the nominal state xA , and the mass purities after disturbance p. xA1 is the mass purity of a
V
chemical compound after a disturbance in manipulated variable 1, xA2 is the mass purity of
V
a chemical compound after a disturbance in manipulated variable 2, xA3 is the mass purity
of a chemical compound after a disturbance in manipulated variable 3. In this work, the
relative gain matrix was built as N × N, according to the N output streams of the separation
scheme.

12.3.5 Multi-Objective Optimization Problem

The objective function would take into account those four targets already mentioned, how-
ever, since the previous work of Penner et al. (2017) showed some issues about recoveries,
the maximization of the recovery was also introduced as a target in the complete objective
324 Process Systems Engineering for Biofuels Development

Table 12.6 Range and type of variables used in the calculation of

objective functions.

Type of variable Search range

Column stages Discrete 5–100
Feed stages Discrete 4–99
Side stream Discrete 4–99
Reflux ratio Continuous 0.1–75
Distillate rate Continuous 10–248 (kmol/h)
Diameter Continuous 0.9–5 (m)

function. So the objective function is described as:

Min(TAC, EI99, IR, 𝛾, −Rec) = f (Ntn , Nfn , Rrn , Frn , Pcn , FCcn )
→
Subject to xm > y→
m (12.9)

where TAC is the total annual cost, EI99 is the eco-indicator 99, IR is the individual risk,
𝛾 is the condition number, and Rec the recovery for all chemical compounds. Ntn are column
stages, Nfn is the feed stage, Rrn is the reflux ratio, Frn is either distillate or bottoms flux.
Moreover, for IR calculation it is necessary to consider other properties such as explosive
limits (Pcn , Fcn ), LC50 , vapor density and so on; ym and xm are the vectors of obtained and
required purities for the mth component, respectively. In this multi-objective optimization
exercise, about 25 variables, continuous or discrete, were considered. The flows of the com-
pounds of interest and their respective purities were considered as constraints. Table 12.6
shows the type of variables used and the search range in the optimization process. The
variables related to a physical aspect of the distillation columns considered average limits
of industrial distillation columns (Górak and Olujic 2014). For the control study, the vari-
ables to be controlled were the purities of 2,3-BD, IBA, MEK, and water. Additionally, the
distillate flows, and heat duties associated with the output currents of said products, were
considered as manipulable variables.

12.4 Global Optimization Methodology

Once the objective functions are modeled, the optimization procedure was carried out. To
develop the optimization we can use a stochastic hybrid algorithm, Differential Evolution
with Tabu List (DETL). Initially the concept of evolution difference was introduced by
Storn and Price (1997), however, it was applied only for a single objective. Later Madavan
(2002) adapted it to solve problems from a multi-objective perspective. This evolutionary
method uses the classical stages of differential evolution and improves the optimization
process with the Tabu List (TL) concept. The TL concept avoids the revision of previ-
ously evaluated points (Glover 1989). A fairly complete description of this algorithm can
be consulted in the work by Sharma and Rangaiah (2013).
The application of the method was performed by means of a hybrid platform, which
involves the interaction among Microsoft Excel, Aspen Plus, and Visual Basic. As a brief
description, the vector of decision variables (initially proposed by the algorithm) is sent
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 325

to Microsoft Excel, which attributes those values to the process variables in Aspen Plus
to simulate the model. After the simulation is completed, Aspen Plus returns to Microsoft
Excel the results in the form of a vector. Finally, with those data, Microsoft Excel evaluates
the objective function and proposes new values of the decision variables in concordance
with the stochastic optimization method. For the optimization task, the following param-
eters were used in the DETL method: 200 individuals, 1000 generations, a TL of 50% of
total individuals, a tabu radius of 2.5*10−6 , and 0.80 and 0.6 for crossover and mutation
fractions, respectively. These parameters were obtained from the literature and previous
tuning process (Srinivas and Rangaiah 2007).

12.5 Results
In the next paragraphs, all results obtained after robust optimization are discussed and
presented. First, the results obtained from analyzing the pure distillation process are pre-
sented, then a comparison between the best scheme and the intensified process is performed.
Note, despite the fact that four objective functions were jointly evaluated, the results in
Figures 12.5–12.10 are presented in a conventional 2D figure for better understanding.
All Pareto fronts were obtained after 200 000 evaluations. Subsequently, no substantial
improvement was observed, so it was considered that under the evaluation criteria, the
DETL method reached convergence. Thus the results reported here correspond to the best
solution obtained. As purity constraints of the process, we considered 99.5 wt% MEK,
99.5 wt% 2,3-BD, and 99 wt% IBA.
As an initial analysis, the objective function of the pure distillation schemes is evaluated
(S1–S4), having the economic criteria as an initial comparative index. Figures 12.5–12.10
show the difference of this objective function when is evaluated with the other three objec-
tive functions. At first sight, it would be easy to select the best alternative among those four

200000000

180000000

160000000

140000000

120000000
TAC ($/yr)

S4
100000000
S2
80000000 140000000 S3
130000000 S1
60000000 120000000

110000000
40000000
100000000

20000000 90000000
0.0016715 0.0016715 0.0016715 0.0016715 0.0016716

0
0.0013 0.00135 0.0014 0.00145 0.0015 0.00155 0.0016 0.00165 0.0017
IR (probability/yr)

Figure 12.5 Pareto front between TAC and IR for the schemes.
326 Process Systems Engineering for Biofuels Development

180000000

160000000

140000000

120000000
TAC ($/yr)

100000000 S4
130000000
S2
80000000 125000000 S3
120000000 S1
60000000
115000000

40000000 110000000

105000000
20000000
100000000
2800000 2900000 3000000 3100000 3200000 3300000
0
0 2000000 4000000 6000000 8000000 10000000 12000000 14000000 16000000 18000000 20000000
EI99 (points/yr)

Figure 12.6 Pareto front between TAC and EI99 for the schemes.

200000000

180000000

160000000

140000000

120000000
TAC ($/yr)

S1 S3
100000000

80000000
S2 S4
60000000

40000000

20000000

0
0 500 1000 1500 2000 2500
Condition number

Figure 12.7 Pareto front between TAC and condition number for the schemes.

schemes. The cheapest alternative is the S4, and the most expensive is the S2. Moreover,
the TAC difference is high by several magnitude orders. However, taking into consideration
the amount of MEK recovered in the separation process, the view changes completely. In
other words, despite the objective function including a target to maximize the recovery of
all schemes, after optimization, the recoveries were not completely equal. After the opti-
mization process, the MEK recoveries obtained, for such purities already mentioned, were
99.9, 65.4, 56.6, and 44.9 wt% for S2, S1, S3, and S4, respectively. For the by-products
2,3-BD/IBA the recoveries were 99.9/61.3, 99.9/74.4, 99.9/99.7, and 99/98.8 wt% for S2,
S1, S3, and S4, respectively.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 327

20000000
18000000

16000000

14000000
EI99 (points/yr)

S4
12000000
S2
10000000 19000000 S3

8000000 17000000
S1

6000000
15000000
4000000
13000000
0.0013322 0.0013323 0.0013324
2000000

0
0.0013 0.00135 0.0014 0.00145 0.0015 0.00155 0.0016 0.00165 0.0017
IR (probability/yr)

Figure 12.8 Pareto front between EI99 and IR for the schemes.

5000

4500

4000

3500
Condition number

3000
S1
2500 S3
S2
2000 S4

1500

1000

500

0
0 5000000 10000000 15000000 20000000 25000000 30000000
EI99 (points/yr)

Figure 12.9 Pareto front between condition number and EI99 for the schemes.

With this in mind and analyzing Figure 12.5, the difference in both TAC and IR values
are clear for all schemes. Note in Figure 12.5, beyond the clear difference in TAC values
because of the recovery, the difference in IR is reflected immediately because of the dif-
ference in the number of columns. Since schemes S2 and S3 are designed with only four
distillation columns, it is evident that IR decreases due to the number of columns as well. In
other words, the IR of the fifth column is avoided. Moreover, the interesting aspect of this
328 Process Systems Engineering for Biofuels Development

200000

180000

160000

140000
Condition number

120000
S4
100000
S2
60000
80000 50000 S3
40000
60000 S1
30000

20000
40000
10000

20000 0
0.0016715 0.0016715 0.0016715 0.0016715 0.0016716

0
0.0013 0.00135 0.0014 0.00145 0.0015 0.00155 0.0016 0.00165 0.0017
IR (probability/yr)

Figure 12.10 Pareto front between condition number and IR for the schemes.

multi-objective approach is the behavior of both objectives. Figure 12.5 shows the tendency
of both objectives; after optimization, the Pareto front shows the zone where those two tar-
gets find their minimum values. Since the IR calculation considers the instantaneous and
continuous release of chemicals, the higher the number of internal flows, the higher the IR
value as well. However, also consider that the calculation of IR would include some physic-
ochemical properties such as LC50 and vapor density. With this consideration, if most of the
compounds are in solution, the greater the amount of water, the more the IR value decreases.
For example, for the initial four case studies, the first column is mainly responsible for sep-
arating a mixture composed mainly of water, which decreases the concentration of the other
compounds as well as the associated risk. Furthermore, in scheme S2, 2,3-BD is separated
which also promotes the decreasing of IR values since a flammable component is released
from the process. Besides, note that the other three schemes increase the reflux ratio in
order to compensate for the IR value because of the presence of 2,3-BD in the first col-
umn. Consequently, all those modifications in design values such as increase reflux ratio
and diameter have as a consequence an increase in TAC values. In this manner, the best
solution is located at the place where both objective functions reach their minimum values.
In the Pareto front of Figure 12.5, the highlighted point accomplishes those requirements
and recommendations by Wang and Rangaiah (2017) and its design values are shown in
Tables 12.7–12.10.
Regarding the evaluation of TAC and EI99 values, the antagonist behavior of both objec-
tive functions is shown in the Pareto front of Figure 12.6. In the upper zone of the Pareto
front, commonly, there are designs with columns of large size but relatively low heat duties.
This combination produces high TAC values in conjunction with small EI99 values. The
lower zone, on the other hand, is mainly composed of columns of reduced size but large heat
duties. Finally, in the middle of both zones, the objective functions reach their minimum
values (Sánchez-Ramírez et al. 2015b). Some variables affect the trend in both functions.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 329

Table 12.7 Design parameters and performance indexes for scheme S1.

C1 C2 C3 C4 C5
Number of stages 14 13 28 33 21
Reflux ratio 34.175 3.02 6.321 6.232 5.034
Feed stage 2 11 7 25 5
Column diameter (m) 1.501 1.545 1.527 1.429 1.177
Operative pressure (kPa) 101.353 101.353 101.353 101.353 101.353
Distillate flowrate (kmol/h) 117.146 117.811 151.735 14.061 8.298
Condenser duty (kW) 39 983 5149 10 568 933 172 454
Reboiler duty (kW) 40 497 5331 10 981 934 177 767
TAC ($/yr) 104 719 750
Eco-indicator (points/yr) 2 993 581 413
Condition number 3.8
IR (probability/yr) 0.001671555

Table 12.8 Design parameters and performance indexes for scheme S2.

C1 C2 C3 C4
Number of stages 29 25 32 36
Reflux ratio 1.93154 19.476 3.384 1.02469
Feed stage 11 17 14 4
Column diameter (m) 1.34118097 1.1 1.213 0.84682939
Operative pressure (kPa) 101.353 101.353 101.353 101.353
Distillate flowrate (kg/h) 235.02 8.451 5.84518 3325
Condenser duty (kW) 7137 1588 235 60 043
Reboiler duty (kW) 7834 1593 255 65 806
TAC ($/yr) 153 136 510
Eco-indicator (points/yr) 16 200 579
Condition number 4.78
IR (probability/yr) 0.00133414

Table 12.9 Design parameters and performance indexes for scheme S3.

C1 C2 C3 C4
Number of stages 20 15 31 59
Reflux ratio 25.063 4.021 30.112 34.568
Feed stage 4/2 12 7 25
Column diameter (m) 1.273 1.825 1.039
Operative pressure (kPa) 101.353 101.353 101.353 101.353
Distillate flowrate (kg/h) 123.38 133.13 89.665 7.67
Condenser duty (kW) 31 713 8175 26 803 2504
Reboiler duty (kW) 31 191 8270 27 011 2669
TAC ($/yr) 31 011 553
Eco-indicator (points/yr) 14 669 116
Condition number 3.99
IR (probability/yr) 0.0013323
330 Process Systems Engineering for Biofuels Development

Table 12.10 Design parameters and performance indexes for scheme S4.

C1 C2 C3 C4 C5
Number of stages 50 8 39 51 43
Reflux ratio 2.142 0.416 2.523 1.995 2.711
Feed stage 22 7 28 22 17
Column diameter (m) 1.138 1.555 1.684 1.147 1.229
Operative pressure (kPa) 101.353 101.353 101.353 101.353 101.353
Distillate flowrate (kg/h) 117.146 117.811 151.735 69.197 8.298
Condenser duty (kW) 2432 2335 461 1945 332
Reboiler duty (kW) 2956 2508 693 1952 367
TAC ($/yr) 4 435 273
Eco-indicator (points/yr) 891 801 275
Condition number 147.56
IR (probability/yr) 0.001665872

For example, the reflux ratio plays an interesting role for such effects; high reflux ratios
increase directly the reboiler heat duty, cost of services, and the environmental impact.
Regarding the evaluation of TAC and condition number, Figure 12.7 shows the Pareto
front for both objective functions. So far, the TAC values are known, however, in the Pareto
fronts, there is selected a point where both objective functions reach their minimum values.
Note in Figure 12.7 and Tables 12.7–12.10, schemes S1, S2, and S3 present relatively the
same condition number, quite different to scheme S4. Even though the condition number is
not a quantitative measure, the condition number can let us know the expected controllabil-
ity properties of those analyzed schemes. With this in mind, it is easy to claim that schemes
S1, S2, and S3 have relatively good properties in comparison with scheme S4. Even though
the sizing of the columns is one of the design variables that is used to correlate with the
condition number (Vázquez-Castillo et al. 2015), during the optimization, the design vari-
able that shows significant effect on the condition number was the reflux ratio. As long as
reflux increases, the condition number decreases, and so high TAC values are related to low
condition number values (meaning that built-in controllability has its cost). Observing the
parameters from Tables 12.7–12.9, schemes S1, S2, and S3 have larger reflux ratios and
largest heat duty and this allows the design to reject larger disturbances.
When the EI99 is evaluated jointly with IR, a similar tendency is observed. Note in
Figure 12.8 the Pareto front for both objective functions. The behavior observed may be
understood considering that a key parameter for measuring both EI99 and IR is the reboiler
duty. In other words, according to Table 12.4, the heat duty possesses a bigger weighting in
comparison with the other categories, which consequently generates the biggest impact on
EI99. It is the same with IR measurement, to handle a process with high reboiler heat duties
elevate the risk associated with that process. In this manner, to compensate both objectives
it is necessary to eventually obtain a process which accomplishes all constraints with as
low as possible heat duties. Note that this behavior also promotes reducing TAC values as
mentioned before.
The Pareto front in Figure 12.9 shows the tendency observed when EI99 is evaluated
jointly with the condition number. The tendencies observed so far show a clear connec-
tion between the condition number and some design variables such as reflux ratio and heat
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 331

duties. In the same way, the EI99 values can be understood from the perspective of heat
duty. With this in mind, the parabola observed with these two objective functions may be
explained. In other words, the low zone of the Pareto front contains designs with high reflux
ratios and consequently high heat duties, as a result, the condition number decreases other-
wise the EI99.
On the other hand, the upper zone contains schemes designed preferably with low reflux
ratio, diameters, and heat duties. These parameters promote lower EI99 values but higher
condition number. That is why, according to Tables 12.8 and 12.9, schemes S2 and S3 show
the biggest EI99 values but the lowest condition numbers.
So far, it has been denoted that both schemes S2 and S3 showed lower IR values in
comparison with the other two schemes, explaining this behavior because of the design
variables (highlighting a lower number of columns and low reflux ratio). On the other hand,
it has been described that condition number is directly affected by low reflux ratio and low
heat duties. With this in mind, it is easier to understand the curve of the Pareto front in
Figure 12.10. Note, as long as the IR decreases, condition number increases. Transport-
ing this result to the design variables, an immediate correlation is that the lower part of
the Pareto front is formed by designs with high reflux ratios and heat duties, promoting
high IR values but low condition number, otherwise the upper zone is mainly formed by
designs with both low reflux ratios and heat duties, generating a safer process but not well
conditioned for operative process under disturbances.
With the results shown so far, scheme S2 can be considered as the most balanced purifi-
cation process among S1–S4 in Figure 12.3. With this initial point, the separation scheme
S5 was evaluated from the same optimization point of view, and the same objective func-
tions were considered. Unlike schemes S1–S4, the alternative S5 was able to recover 99.29,
99.28, 99.99, and 99.99% wt of IBA, MEK, 2,3-BD, and p-xylene, respectively. Table 12.11
shows the objective functions and design parameters obtained for the hybrid process. Note,
the reduction in both TAC and EI99 is huge. Even though the condition number of scheme
S5 is higher than for the rest of the schemes (S1–S4) it is worth analyzing the role of the
mass entrainer in order to reduce the energy requirements in MEK purification. Note, for

Table 12.11 Design parameters and performance indexes for the intensified scheme.

Liquid–liquid
extraction C2 C3 C4 C5 C6
Number of stages 10 33 45 45 43 38
Reflux ratio 3.483 0.529 16.636 1.995 18.674
Feed stage 1, 10 4 27 5 23 27
Column diameter (m) 1.455 1.285 1.407 1.544 1.098 1.324
Operative pressure (kPa) 101.353 101.353 101.353 101.353 101.353 101.353
Distillate flowrate (kmol/h) 111.997 1693 19.193 11.537 1.386
Condenser duty (kW) 5776 2335 3191 375 236.5
Reboiler duty (kW) 6354 4125 3202 408 236.8
TAC ($/yr) 7 903 251
Eco-indicator (points/yr) 1 338 593
Condition number 88 121
IR (probability/yr) 0.0014087
332 Process Systems Engineering for Biofuels Development

example, the total energy invested in S2 is about 271 764 MJ/h, on the other hand, the best
point obtained for the hybrid process requires 51 580 MJ/h, which represents an energy
reduction of 220 184 MJ/h.
An alternative scenario would be to maximize the purification. It would be interesting to
know what purities would be obtained. However, the impact is direct for those higher con-
centrations on the TAC, specifically to the cost of services. Additionally, since the reboiler
duty increases for higher purities, the eco-indicator will also increase. So, an interesting
view would be to know the behavior of evaluating together in a Pareto front an economic
or environmental objective function and the purities of interest. In this case, it would be
possible to locate a point where because of the purities the process becomes economically
unfeasible. Moreover, many results of the process would be compromised, for example, the
recovery of the components.
Considering the MEK production, scheme S2 consume 35 MJ/kgMEK and scheme S5
consumes 6.7 MJ/kgMEK. In the hypothetical scenario where all MEK produced as fuel was
burned, with 31.45 MJ/kg as energy density, the energy profit of S2 and S5 was −31 266
and 188 918 MJ/h, respectively. Additionally, the reduction in EI99 is also remarkable. For
example, S2 presents an environmental impact of 16 200 579 points/yr, however, scheme S5
presented 1 338 593 points/yr, a reduction above 90%. Regarding inherent safety, scheme S2
is safer than scheme S5 by about 5%. Note as aforementioned, the inherent safety is affected
by many circumstances, in this case the increase in scheme S5 is initially due to the increase
in distillation columns (four columns for scheme S2 and six for scheme S5). Moreover,
note in scheme S2 the presence of water in many columns reduces the concentration of
dangerous components and consequently the inherent safety. However, in scheme S5, most
of the water was initially removed by the mass agent.
Finally, a quite interesting analysis. Some design variables have a key role in the objec-
tive functions. Nowadays, the role of some design variables in the economic and environ-
mental indexes, such as TAC and EI99, has been previously studied (Errico et al. 2016).
However, another interesting perspective is the role of those design variables in the objec-
tive functions. Through the optimization process some tendencies were observed. Note in
Figure 12.11 how it is only possible to obtain low IR values and low condition number
when the reflux ratio increases notably. This behavior is understandable since a high reflux
ratio mitigates the disturbances and increase in some columns to dissolve the flammable
compounds. However, it should be noted that probably this combination affects directly
another performance index such as TAC and EI99. In other words, as long as the reflux
ratio increases, the amount of flow within the column increases, so this amount of matter
allows an increase in disturbance mitigation in comparison with a column with low internal
flows where the disturbance easily affects all the flow.
Figure 12.12 shows the role of the stages for both objective functions. Note, even though
a high number of stages probably promotes further good dynamic behavior and low heat
duties to decrease IR values, it is possible to obtain low IR and condition number values with
a low number of stages as well. This behavior also affects directly the TAC and EI99 values
since the sizing of the column is a key parameter for TAC calculation and EI99 because
of the contribution of the steel. Finally, Figure 12.13 shows a complete mass balance of
schemes S2 and S5 considered as the best alternative for pure distillation schemes and the
hybrid process.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 333

6
34
01
0.0 344
01
0.0 342
IR (probability/yr)

01
0.0 340
01
0.0 338
01
0.0 336
01
0.0 334 1
01 1 20
0.0 332 80 00

r
01

be
6
0.0 40 0

m
0 10

nu
20 3 20
Reflu
x ratio 0 40 50

n
tio
(kmol/ 0

di
60 > 0.0013
kmol)

on
< 0.0013

C
< 0.0013
< 0.0013
< 0.0013
< 0.0013

Figure 12.11 Correlation generated among reflux ratio, IR, and condition number.

45
133
0.00
44
133
0.00
IR (probability/yr)

43
133
0.00
42
133
0.00
41
133
0.00
40
133
0.00
0
00 0
35 000 0
46
44
42

Co 3 0
40

50 00
38

nd
36

itio 2 200 00
34
32

50
30

nn 0 ges
28

um 1 000 Sta
26

00
24
22

be 1
50
20
18
0

r
> 0.0013
< 0.0013
< 0.0013
< 0.0013

Figure 12.12 Correlation generated among number of stages, IR, and condition number.
334 Process Systems Engineering for Biofuels Development

Water 2117 kg/h

IBA 823.5 kg/h
MEK 7647.06 kg/h Water 31.04 kg/hr
2,3-BD 0.3 kg/h IBA 484.69 kg/h
MEK 0.48 kg/h

C3

IBA 482.56 kg/h

Water 2117.64 kg/h C1 C2 MEK 0.48 kg/h
IBA 823.5 kg/h 255 kW
MEK 7647.06 kg/h
2,3-BD 1176.4 kg/h

2,3-BD 1176.1 kg/h

7834 kW Water 2086.6 kg/h
IBA 338.81 kg/h
MEK 7646.57 kg/h C4
2,3-BD 0.3 kg/h IBA 0.5 kg/h
1593 kW MEK 7639.24 kg/h
2,3-BD 0.145 kg/h
65806 kW

IBA = 718.027 kg/h IBA = 99.691 kg/h

Water = 28.4 kg/h Water, MEK (traces)
IBA = 823.529 kg/h
Water = 72.67 kg/h
MEK = 7647.06 kg/h
C4 IBA and MEK traces C5 C6
Water = 2117.65 kg/h
2–3 BD = 1176.47 kg/h
C3 MEK = 155.885 kg/h
MEK = 7489.97 kg/h
Water = 3.8 kg/h IBA (traces)
IBA = 1.1 kg/h

p-Xylene = 36998.1 kg/h p-Xylene = 36994.2 kg/h

2,3-BD, Water (traces)

Water = 2012.67 kg/h

C2 2,3-BD = 19.91 kg/h

2,3-BD = 1146.94 kg/h

Water, p-Xylene
Make Up (traces)
p-Xylene = 3.9 kg/h

Figure 12.13 Complete mass balance for the selected schemes S2 and S5.

Specifically, regarding IR, reflux ratio plays an interesting role, either by helping to
improve or worsen the IR. For example, when the column is associated with a water-diluted
mixture, occasionally if the reflux ratio increases, the water flow may increase too. Con-
sidering the amount of water within the column, the concentration of flammable com-
pounds decreases which minimizes risk. On the other hand, if the column is separating
only flammable components, high reflux ratios increase internal flows and consequently IR
also increases.
The column diameter also plays a main role in all objection functions. Besides the
well-known relation for diameter-control properties (as long as the diameter increases,
the dynamic properties increase), note that if the size associated with each column
increases, the internal flows also increase. The result obtained for IR depends on the kind
of chemicals used. Note, a similar case is found with the stages; as long as the number
of stages increases, the IR is affected because of the internal flows associated with the
column.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 335

Regarding more design variables, the impact of other variables is clear; for example, for
both the reboiler heat duty and column pressure, a bigger duty and pressure will always be
associated with an increase in the IR.

12.6 Conclusions
In this work, an intensified process to separate MEK once the optimization process was
finished was evaluated; the energy requirement of the intensified process was lower than
the process based on conventional distillation columns.
Additionally, this work proposes the evaluation of the inherent risk (IR) and the condition
number (jointly with TAC and EI99) in early design stages in order to generate separation
alternatives which accomplish current global needs. After the optimization process, those
separation schemes showed very interesting results. Scheme S5, a hybrid process based
on liquid–liquid extraction, was the most promissory since it was the only alternative able
to recover and purify the entire feed mixture. Unlike pure distillation schemes, which were
not energetically viable, the hybrid process improves energy consumption and energy profit
in comparison with the scheme based on distillation. Additionally, it showed huge energy
savings which are consequently observed in performance parameters such as TAC, EI99,
and IR. In this manner, process intensification seems the correct alternative for improving
the energy requirements and all indexes evaluated here.

Acknowledgments
The authors acknowledge the financial support provided by CONACYT, Universidad de
Guanajuato and the SEP (UGTO-PTC-668).

Notation
2,3-BD 2,3-Butanediol
BLEVE Boiling liquid expanding vapor explosion
CN Condition Number
DETL Differential Evolution with Tabu List
EI99 Eco-indicator 99
HAZOP Hazard and operability study
IBA Isobutyraldehyde
IEA International Energy Agency
IR Individual risk
LC50 Lethal concentration, 50%
LCA Life cycle analysis
MEK Methyl ethyl ketone
QRA Quantitative risk analysis
SVD Singular value decomposition
TAC Total Annual Cost
TL Tabu List
TS Tabu Search
336 Process Systems Engineering for Biofuels Development

UVECE Unconfined vapor cloud explosion

𝛾 Condition number
𝜎 Singular value
Σ Diagonal matrix

References
Alemam, A., Cheng, X., and Li, S. (2017). Treating design uncertainty in the application of eco-indicator
99 with Monte Carlo simulation and fuzzy intervals. International Journal of Sustainable Engineering
7038: 1–12. https://doi.org/10.1080/19397038.2017.1387824.
Atsumi, S., Cann, A.F., Connor, M.R. et al. (2008). Metabolic engineering of Escherichia coli for 1-butanol
production. Metabolic Engineering 10: 305–311. https://doi.org/10.1016/j.ymben.2007.08.003.
Bequette, B.W. and Edgar, T.F. (1989). Non-interacting control system design methods in distillation. Com-
puters & Chemical Engineering 13: 641–650.
Bohnet, M. (2003). Ullmann’s Encyclopedia of Industrial Chemistry. Wiley-VCH.
Cardona-Alzate, C.A. and Sánchez-Toro, O.J. (2006). Energy consumption analysis of integrated flowsheets
for production of fuel ethanol from lignocellulosic biomass. Energy 31: 2447–2459. https://doi.org/10
.1016/j.energy.2005.10.020.
Crowl, D.A. and Louvar, J.F. (2001). Chemical Process Safety: Fundamentals with Applications. Pearson
Education.
De Haag, P.U. and Ale, B.J.M. (2005). Guidelines for Quantitative Risk Assessment: Purple Book. The
Hague, The Netherlands: Sdu Uitgevers.
Emerson, R.R., Flickinger, M.C., and Tsao, G.T. (1982). Kinetics of dehydration of aqueous 2,3-butanediol
to methyl ethyl ketone. Industrial and Engineering Chemistry Product Research and Development 21:
473–477. https://doi.org/10.1021/i300007a025.
Errico, M., Rong, B., Tola, G., and Spano, M. (2013). Optimal synthesis of distillation Systems for
bioethanol separation. Part 2. Extractive distillation with complex columns. Industrial & Engineering
Chemistry Research 52: 1620–1626.
Errico, M., Sanchez-Ramirez, E., Quiroz-Ramìrez, J.J. et al. (2016). Synthesis and design of new hybrid
configurations for biobutanol purification. Computers & Chemical Engineering 84: 482–492. https://doi
.org/10.1016/j.compchemeng.2015.10.009.
Glover, F. (1989). Tabu search—Part I. ORSA Journal on Computing 1: 190–206. https://doi.org/10.1287/
ijoc.1.3.190.
Goedkoop, M. and Spriensma, R. (2000). Eco-Indicator 99 Manual for Designers. Amersfoort, The Nether-
lands: PRé Consultants.
Gómez-Castro, F.I., Segovia-Hernández, J.G., Hernández, S. et al. (2008). Dividing wall distillation
columns: optimization and control properties. Chemical Engineering & Technology: 1246–1260. https://
doi.org/10.1002/ceat.200800116.
Gong, Y., Lin, L., Shi, J., and Liu, S. (2010). Oxidative decarboxylation of levulinic acid by cupric oxides.
Molecules 15: 7946–7960. https://doi.org/10.3390/molecules15117946.
Górak, A. and Olujic, Z. (2014). Distillation: Equipment and Processes. Academic Press.
Govasmark, E., Stäb, J., Holen, B. et al. (2011). Chemical and microbiological hazards associated with
recycling of anaerobic digested residue intended for agricultural use. Waste Management 31: 2577–2583.
https://doi.org/10.1016/j.wasman.2011.07.025.
Groot, W.J., Soedjak, H.S., Donck, P.B. et al. (1990). Butanol recovery from fermentations by liquid-liquid
extraction and membrane solvent extraction. Bioprocess Engineering 5: 203–216.
Groot, W.J., van der RGJM, L., and Luyben, K.C.A.M. (1992). Review. Technologies for Butanol Recovery
Integrated with fermentations. Process Biochemistry 27: 61–75.
Guthrie, K. (1969). Capital cost estimating. In: Chemical Engineering, 114. New York: McGraw-Hill.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 337

Gutiérrez-Antonio, C. (2016). Multiobjective stochastic optimization of dividing-wall distillation columns

using a surrogate model based on neural networks. Chemical and Biochemical Engineering Quarterly
29: 491–504. https://doi.org/10.15255/CABEQ.2014.2132.
Hernández, S. (2008). Analysis of energy-efficient complex distillation options to purify bioethanol. Chem-
ical Engineering & Technology: 597–603. https://doi.org/10.1002/ceat.200700467.
Hoppe, F., Burke, U., Thewes, M. et al. (2016a). Tailor-made fuels from biomass: potentials of 2-butanone
and 2-methylfuran in direct injection spark ignition engines. Fuel 167: 106–117. https://doi.org/10.1016/
j.fuel.2015.11.039.
Hoppe, F., Heuser, B., Thewes, M. et al. (2016b). Tailor-made fuels for future engine concepts. International
Journal of Engine Research 17: 16–27. https://doi.org/10.1177/1468087415603005.
Hovd, M., Braatz, R.D., and Skogestad, S. (1997). SVD controllers for H2−, H∞− and μ-optimal control.
Automatica 33: 433–439. https://doi.org/10.1016/S0005-1098(96)00167-7.
Hugo, V., Díaz, G., and Tost, G.O. (2016). Butanol production from lignocellulose by simultaneous fer-
mentation, saccharification, and pervaporation or vacuum evaporation. Bioresource Technology https://
doi.org/10.1016/j.biortech.2016.06.091.
Ji, X.J., Huang, H., and Ouyang, P.K. (2011). Microbial 2,3-butanediol production: a state-of-the-art review.
Biotechnology Advances 29: 351–364. https://doi.org/10.1016/j.biotechadv.2011.01.007.
Jiménez-González, C., Constable, D.J.C., and Ponder, C.S. (2012). Evaluating the “greenness” of chemical
processes and products in the pharmaceutical industry - a green metrics primer. Chemical Society Reviews
41: 1485–1498. https://doi.org/10.1039/c1cs15215g.
Jin, C., Yao, M., Liu, H. et al. (2011). Progress in the production and application of n-butanol as a biofuel.
Renewable and Sustainable Energy Reviews 15: 4080–4106. https://doi.org/10.1016/j.rser.2011.06.001.
Klema, V.C. and Laub, A.J. (1980). The singular value decomposition: its computation and some
applications. IEEE Transactions on Automatic Control 25: 164–176. https://doi.org/10.1109/TAC.1980
.1102314.
Kumar, A. (1996). Guidelines for evaluating the characteristics of vapor cloud explosions, flash fires, and
bleves. AIChE Journal 15: S11–S12.
Kumar, M. and Gayen, K. (2011). Developments in biobutanol production: new insights. Applied Energy
88: 1999–2012. https://doi.org/10.1016/j.apenergy.2010.12.055.
Lau, H. and Alvarez, J. (1985). Synthesis of control structures by singular value analysis: dynamic measures
of sensitivity and interaction. AIChE Journal 31: 427–439.
Law, B.F., Pearce, T., and Siegel, P.D. (2011). Safety and chemical exposure evaluation at a small biodiesel
production facility. Journal of Occupational and Environmental Hygiene 8: 37–41. https://doi.org/10
.1080/15459624.2011.584841.
Lutze, P., Gani, R., and Woodley, J.M. (2010). Process intensification: a perspective on process synthesis.
Chemical Engineering and Processing: Process Intensification 49: 547–558. https://doi.org/10.1016/j
.cep.2010.05.002.
Luyben, W.L. (2011). Principles and Case Studies of Simultaneous Design. Hoboken, NJ: Wiley https://
doi.org/10.1002/9781118001653.
Madavan, N.K. (2002). Multiobjective optimization using a Pareto differential evolution approach. In: Pro-
ceedings of the 2002 Congress on Evolutionary Computation, vol. 2, 1145–1150. IEEE.
Marcinkowski, A.Z.K. (2017). The evaluation of efficiency of the use of machine working time in the
industrial company–case study. Management Systems in Production Engineering 25: 251–254. https://
doi.org/10.1515/mspe.
Mariano, A.P. and Filho, R.M. (2012). Improvements in Biobutanol fermentation and their impacts on
distillation energy consumption and wastewater generation. BioEnergy Research: 504–514. https://doi
.org/10.1007/s12155-011-9172-0.
Medina-Herrera, N., Jiménez-Gutiérrez, A., and Mannan, M.S. (2014). Development of inherently safer
distillation systems. Journal of Loss Prevention in the Process Industries 29: 225–239. https://doi.org/
10.1016/j.jlp.2014.03.004.
338 Process Systems Engineering for Biofuels Development

Moore, C. (1986). Application of singular value decomposition to the design, analysis, and control of indus-
trial processes. American Control Conference, 643–50.
Murphy, C.D. (2000). Process of recovering methyl ethyl ketone from an aqueous mixture of methyl ethyl
ketone and ethanol. US Patent 6121497, filed 16 October 1998 and issued 19 September 2000.
Murrieta-Dueñas, R., Gutiérrez-Guerra, R., Segovia-Hernández, J.G., and Hernández, S. (2011). Analy-
sis of control properties of intensified distillation sequences: reactive and extractive cases. Chemical
Engineering Research and Design 89: 2215–2227. https://doi.org/10.1016/j.cherd.2011.02.021.
Nakata, K., Utsumi, S., Ota, A., et al. (2006). The effect of ethanol fuel on a spark ignition engine. SAE
technical paper.
Penner, D., Redepenning, C., Mitsos, A., and Viell, J. (2017). Conceptual design of methyl ethyl ketone
production via 2,3-butanediol for fuels and chemicals. Industrial and Engineering Chemistry Research
56: 3947–3957. https://doi.org/10.1021/acs.iecr.6b03678.
Pokoo-Aikins, G., Heath, A., Mentzer, R.A. et al. (2010). A multi-criteria approach to screening alterna-
tives for converting sewage sludge to biodiesel. Journal of Loss Prevention in the Process Industries 23:
412–420. https://doi.org/10.1016/j.jlp.2010.01.005.
Ponce-Ortega, J.M., Al-Thubaiti, M.M., and El-Halwagi, M.M. (2012). Process intensification: new under-
standing and systematic approach. Chemical Engineering and Processing: Process Intensification 53:
63–75. https://doi.org/10.1016/j.cep.2011.12.010.
Qureshi, N., Hughes, S., Maddox, I.S., and Cotta, M.A. (2005). Energy-efficient recovery of butanol from
model solutions and fermentation broth by adsorption. Bioprocess and Biosystems Engineering 27:
215–222. https://doi.org/10.1007/s00449-005-0402-8.
Sågfors, M.F. and Waller, K.V. (1995). The impact of process directionality on robust control in non-ideal
distillation. IFAC Proceedings Volumes 28 (9): 327–332.
Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., Segovia-Hernández, J.G. et al. (2015a). Process alternatives
for biobutanol purification: design and optimization. Industrial & Engineering Chemistry Research 54:
351–358. https://doi.org/10.1021/ie503975g.
Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., Segovia-Hernández, J.G. et al. (2015b). Economic and envi-
ronmental optimization of the biobutanol purification process. Clean Technologies and Environmental
Policy 18: 395–411. https://doi.org/10.1007/s10098-015-1024-8.
Segovia-Hernández, J.G., Hernández, S., and Jiménez, A. (2002). Control behaviour of thermally coupled
distillation sequences. Chemical Engineering Research and Design 80: 783–789. https://doi.org/10.1205/
026387602320776858.
Sharma, S. and Rangaiah, G.P. (2013). An improved multi-objective differential evolution with a termina-
tion criterion for optimizing chemical processes. Computers and Chemical Engineering 56: 155–173.
https://doi.org/10.1016/j.compchemeng.2013.05.004.
Srinivas, M. and Rangaiah, G.P. (2007). Differential evolution with Tabu list for global optimization and its
application to phase equilibrium and parameter estimation problems. Industrial & Engineering Chem-
istry Research 46: 3410–3421. https://doi.org/10.1021/ie0612459.
Stocker, T.F., Qin, D., Plattner, G.K., et al. (2013). Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to The Fifth Assessment Report of the Intergovernmental Panel on
Climate Change, 1535.
Storn, R. and Price, K. (1997). Differential Evolution – A simple and efficient heuristic for global optimiza-
tion over continuous spaces. Journal of Global Optimization 11: 341–359.
Syu, M.J. (2001). Biological production of 2,3-butanediol. Applied Microbiology and Biotechnology 55:
10–18. https://doi.org/10.1007/s002530000486.
Tran, A.V. and Chambers, R.P. (1987). The dehydration of fermentative 2,3-butanediol into methyl ethyl
ketone preparation of solid acid catalysts properties of catalyst. Biotechnology 29: 343–351.
Turton, R. (2001). Analysis, Synthesis and Design of Chemical Process, 4e, vol. 40. Prentice Hall.
 Intensified Purification Alternative for Methyl Ethyl Ketone Production 339

Vázquez-Castillo, J.A., Segovia-Hernández, J.G., and Ponce-Ortega, J.M. (2015). Multiobjective optimiza-
tion approach for integrating design and control in multicomponent distillation sequences. Industrial &
Engineering Chemistry Research 54: 12320–12330. https://doi.org/10.1021/acs.iecr.5b01611.
Wang, Z. and Rangaiah, G.P. (2017). Application and analysis of methods for selecting an optimal solu-
tion from the Pareto-optimal front obtained by multiobjective optimization. Industrial and Engineering
Chemistry Research 56: 560–574. https://doi.org/10.1021/acs.iecr.6b03453.
Xiao, Z. and Lu, J.R. (2014). Strategies for enhancing fermentative production of acetoin: a review. Biotech-
nology Advances 32: 492–503. https://doi.org/10.1016/j.biotechadv.2014.01.002.
Yabe, K., Shinoda, Y., Seki, T. et al. (2012). Market penetration speed and effects on CO2 reduction of
electric vehicles and plug-in hybrid electric vehicles in Japan. Energy Policy 45: 529–540. https://doi
.org/10.1016/j.enpol.2012.02.068.
Yoneda, H., Tantillo, D.J., and Atsumi, S. (2014). Biological production of 2-butanone in Escherichia coli.
ChemSusChem 7: 92–95. https://doi.org/10.1002/cssc.201300853.
Zhao, J., Yu, D., Zhang, W. et al. (2016). Catalytic dehydration of 2,3-butanediol over P/HZSM-5: effect
of catalyst, reaction temperature and reactant configuration on rearrangement products. RSC Advances
6: 16988–16995. https://doi.org/10.1039/C5RA23251A.
Zhou, M., Li, L., Xie, L. et al. (2015). Preparation of papers for IFAC Conferences & Symposia: integration
of process design and control using hierarchical control structure. IFAC-PapersOnLine 48: 188–192.
https://doi.org/10.1016/j.ifacol.2015.08.179.
 13
Present and Future of Biofuels
Juan Gabriel Segovia-Hernández, César Ramírez-Márquez, and
Eduardo Sánchez-Ramírez
Departamento de Ingeniería Química, Universidad de Guanajuato, Noria Alta s/n, Guanajuato, 36050,
Guanajuato, México

13.1 Introduction
In today’s society and industry, crude oil is the most important source of energy since it
contributes approximately 35% to global energy consumption. Given the growing demand
for oil, and according to several reports, it has been estimated that reserves reached their
maximum production in 2010. Notwithstanding the above, it is expected that petroleum
products will continue to be the main source of energy until 2030, at least. On the other
hand, despite the outlook regarding oil reserves, renewable products have certainly been
revalued and economic income in the countryside has been encouraged (Fortman et al.
2008).
In the same way, research groups have increased their interest in clean and sustainable
energies that come from renewable sources. Thus, technological advances have been stim-
ulated and the profitability of many renewable energies has been improved. In addition,
environmental protection has benefited, as well as the sustainability of conventional pro-
cesses. For example, the transport sector has shown the greatest resistance in the effort
to reduce carbon dioxide emissions due to its great dependence on fossil fuel. In fact,
petroleum derivatives represent approximately 95% of the energy consumed in this sector
(Demirbas 2007). With these considerations, research groups have focused their attention
on other sources of energy, biofuels for example. Additionally, international agreements
have been decreed to favor these trends, for example, the 2 ∘ C proposal is promoted by the

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
342 Process Systems Engineering for Biofuels Development

International Energy Agency (IEA) to mitigate climate change. Briefly, the main objective
of this proposal is to reduce approximately 70% of carbon dioxide emissions compared
with the emissions of 2014. This proposal has as its main objective the transport sector that
currently represents 23% of the total annual global emissions. On the other hand, SO2 and
NOx emissions caused by fuel combustion are expected to be the same (Eggleston et al.
2006).
Although electricity stands out as a very promising alternative to reduce carbon dioxide
emissions, biofuels are considered responsible for reducing carbon dioxide emissions. Since
the predictions of the IEA, the automotive sector could use at least 30.7% of the fuel, which
demonstrates the need to replace fossil fuels in the short term and, at the same time, reduce
oil production. According to several studies, it could be exhausted in a period of 50 years
with the current consumption rate (Demirbas 2007; Reijnders 2006; Ture et al. 1997).
In developed countries, there is a growing trend to use modern and efficient technology
in the field of biofuels, which has led to these compounds becoming more economically
competitive compared with fossil fuels. In general, biofuels have been considered a very
interesting proposal given the advantages they offer, highlighting sustainability, reducing
greenhouse gas emissions, and generating tangible social benefits for regional development
and agriculture. In the same sense, due to the growing interest, biomass is an attractive raw
material for three main reasons: (i) it is up to now a renewable resource; (ii) it seems to
have a positive impact on the environment; and (iii) it seems to have a positive economic
impact if the prices of fossil fuels increase in the future (Demirbas 2000a,b).
In recent times, biofuels have been in vogue. However, this does not mean that they are
a product of new technology, but rather their development and applications have increased.
For example, according to data generated by the IAE, in 2009 the total supply of primary
energy was close to 12 150 toe (tonne of oil equivalent), which represents 10.2% previ-
ously produced from renewable resources (Fortman et al. 2008). It is expected that the
United States and the European Union will replace at least 6% of fossil fuel consumption
by biofuels in the coming years (Demirbas 2000a).
Biofuel can be defined as a fuel of biological origin, i.e. the term covers all fuels derivative
of vegetal biomass. So, biofuel is obtained in a renewable way from organic remains. The
vegetal origin fuels must have similar characteristics to fossil fuels, allowing use in spark
engines without having to make significant modifications (Demirbas 2007). Biofuels can
originate as liquids, gaseous and solid fuels produced mainly from biomass. They can be
produced from a variety of fuels from biomass, such as bioethanol, biobutanol, bioethanol,
biodiesel, and biohydrogen (Cadenas and Cabezudo 1998).
Biofuels can be obtained from several raw materials and are usually used in different
traditional petroleum-based fuel blends (Puppan 2002).
The future of biofuels is in finding solutions and it taking advantage of their benefits. The
success of biofuels resides in the economy and their easy access for everyday use.
Although there are several important factors that must be taken into account for the suc-
cessful implementation of biofuels in our society, there is a certain development level that
is already helping to verify the potential of biofuels in our lives. The biofuels derived from
biomass can be classified as primary, secondary, and so on. Primary biofuels are used in a
way not involving processing, i.e. solid material from forest waste for use as fuel. Secondary
biofuels come from a biomass process and can be used as fuels in vehicles or in an industrial
process. Secondary biofuels are divided into first-, second-, third- and fourth-generation
 Present and Future of Biofuels 343

Biofuels Feedstock and examples

First generation Second generation Third generation Fourth generation

Wheat Straw, Corn,

Sugar, Starch,
Wood, Solid Waste, Vegetable oil,
Vegetable oils, Algae
Energy Crop, Non- biodiesel
Animal Fats
Food Crops

Bioalcohols, bio-oil, bio-

Bioalcohols, vegetable
DMF, biohydrogen, bio-
oil, biodiesel, biosyngas, Vegetable oil, biodiesel Biogasoline
Fischer-Tropsch diesel,
biogas
wood diesel

Figure 13.1 Classification of secondary biofuels. Source: Jefferson 2006. Reproduced with permission of
Elsevier.

biofuels depending on the feedstock used for their production (Figure 13.1). Biofuels are
also classified according to the source and type. They can come from agricultural sources,
municipal waste or waste originated from industry. Biofuels can be solids (firewood, char-
coal, and pellets), liquids (bioethanol, biodiesel, etc.), or gaseous (biogas) (Ture et al. 1997).
Nowadays, there are difficulties in biofuels production development. These difficulties
are technological, political, economic, and/or involving storage and safety. The path sought
in the industrial and academic sector is to reduce these difficulties, and significantly pro-
mote and develop the conventional use of biofuels. Two of the biggest drawbacks are the
production costs and the temporality of the raw materials (Difiglio 1997).
At present, there are two main ways of processing lignocellulosic raw material, namely
biochemical and thermochemical. Certainly, there is no clear candidate for a better
technological pathway than the biochemical and thermochemical pathways (Figure 13.2)

Biomass

Raw Material
Processing

Thermochemical Biochemical
Conversion Conversion

Gasification Liquefaction Pyrolysis Bioethanol Biodiesel Biobutanol Biojet fuel

Biogas Biochemicals Syn-oil

Figure 13.2 Main biomass conversion process.

344 Process Systems Engineering for Biofuels Development

(Tanger et al. 2013). This is why there still needs to be better investment, research, and
development, to guarantee that future biomass raw material production can be carried out
in a sustainable way and that high conversion technologies are identified. Once it has been
tested, there will be a constant transition of biofuels.
Accordingly, several routes for the production of renewable energy have been proposed.
From the alternatives, it is possible to highlight four categories: (i) integration of solar and
wind power to produce fuels and chemicals; (ii) biomass conversion; (iii) a polygenera-
tion process which can produce simultaneously transportation fuels, energy, and high value
added chemicals; and (iv) a combination of syngas and hydrogen to produce chemicals
(Faaij 2006).

13.2 Some Representative Biofuels

A biofuel is produced through contemporary biological processes, such as agriculture and
anaerobic digestion, instead of being produced by geological processes such as the pro-
cesses involved in the formation of fossil fuels. In the particular case of biofuels, if the
biomaterial can grow rapidly, this biofuel is considered as renewable energy. Biofuels can
be derived from plant grains (for example corn) or indirectly from agroindustrial or domes-
tic waste. Additionally, renewable biofuels provide a certain degree of carbon fixation due
to the process of photosynthesis that occurs in all plant matter (Parikka 2004). Biofuels can
be considered as theoretically neutral carbon since the carbon dioxide that is absorbed due
to photosynthesis is also released when the fuel is burned. Other renewable biofuels can
be produced using biomass (considering biomass as living organisms rather than material
derived from plants). In general, biomass can be converted into energy by two different
methods: thermochemical conversion and biochemical conversion. The converted biomass
can result in fuels of solid, liquid or even gaseous nature. In the following, there is a brief
description of the generalities of the most promising biofuels today (Hoekman 2009).

13.2.1 Bioethanol
Bioethanol or ethyl alcohol is a colorless, biodegradable and low toxicity liquid, so that the
environmental pollution caused in the case of spillage is really small. In the case of combus-
tion, bioethanol produces carbon dioxide and water. The octane rating of bioethanol is high;
it can even be mixed with gasoline to oxygenate the mixture and promote more complete
combustion to reduce gas emissions. Currently, ethanol is widely sold in the United States
for its mixture with gasoline; the most common mixture is 10% bioethanol and 90% gaso-
line (E10). Gasoline vehicles do not require modifications to run under the E10 mixture and
their warranty is also not affected. Only flexfuel vehicles can operate with mixtures above
85% ethanol and 15% gasoline (E85) (McMillan 1997).
Bioethanol can be produced by biological pathways; however, it can also be obtained by
the reaction of ethylene and vapor. Mainly, sugar is considered as raw material to produce
bioethanol from grains, for example wheat crops, sawdust, corn, etc. (Gray et al. 2006).
The material considered as feedstock is in principle, any plant; in practice, the selection of
the feedstock depends on the speed of growth of the plant, as well as the sugar content and its
ease of availability in the plant. As a result, a wide variety of raw materials and consequently
 Present and Future of Biofuels 345

the production processes come from sugar- and starch-containing raw materials. However,
various available types of lignocellulosic biomass such as agricultural and forestry residues,
and herbaceous energy crops could serve as feedstocks for the production of bioethanol,
energy, heat, and value-added chemicals.
Globally, the largest amount of bioethanol is obtained in Brazil from sugar cane. In the
United States it is produced from molasses and corn; however, other sources of sugar can
also be used, for example wheat, barley, and even rye can be used. Grains containing starch
must first be converted to sugar. Approximately 3 tons of grains are required to produce
1 ton of bioethanol. In Europe, the most common matter for producing bioethanol is wheat
and sugar beet. Sugar beet grows in most of the European Union countries and has a higher
growth per hectare than wheat (Chen and Qiu 2010).
Currently, research and development activities in the field of bioethanol are focused on
lignocellulosic material and woody materials. Within said material may be included willow,
miscanthus and eucalyptus, agricultural residues and municipal solid waste. About 2–4 tons
of woody material are required to produce 1 ton of bioethanol (Cardona et al. 2010).
There are many reasons to switch to bioethanol production of lignocellulosic material.
Lignocellulosic material is more abundant and less expensive than the aforementioned
grains due to the direct competition with the food market. In addition, the energy balance
is larger, making the environmental impact lower. In fact, lignocellulosic material has the
potential to accumulate up to 90% of greenhouse gas emissions; on the other hand, this type
of biomass has greater difficulty in converting sugar due to its relative low accessibility in
the biomass structure.
The most used technology to purify and produce bioethanol is fermentation and distilla-
tion. Fermentation is a process of biochemical conversion in which matter is decomposed
by microorganisms. The microorganisms can act in different types of raw material. Among
the microorganisms used, bread yeast (Saccharomyces cerevisiae) is the most used since
it requires only monomeric sugar as raw material. Conventional bioethanol fermentation
can produce 0.51 kg of bioethanol from 1 kg of any six-carbon sugar. However, not all
raw materials contain sugars of that nature. Starch and lignocellulose are polymers, so
hydrolysis is required to break the bonds between the monomers to generate six-carbon
sugars.
For example, with grain for the production of bioethanol, the first step for the conversion
is a mechanical process that includes grinding the grain to release its starch. Subsequently
the mass generated must be diluted to adjust the amount of sugar. Adjusting the amount
of sugar allows to manipulate the mass generated in the grinding. Then you must cook
to dissolve all the soluble starches in the aqueous solution. The starch is converted into
sugars by enzymatic action or acid hydrolysis. In the case of acid hydrolysis, dilute mineral
acid is added to the suspension before cooking. The short chain hydrocarbons that result
from this process can be fermented by microorganisms. For the microorganisms to carry
out the biological process, the solution must be slightly acidic, that is, a pH between 4.8
and 5.0. During the fermentation, bioethanol dissolved in water as well as carbon dioxide
is produced. Subsequently, such effluent must be treated in a purification process to obtain
concentrated solutions (Segovia-Hernández et al. 2014).
The lignocellulosic material can be converted to bioethanol only differing from the pro-
cess described above in the sugars to be fermented. Hydrolysis of this kind of material is
more difficult than for energy drops (for example starch) since lignocellulosic material is
346 Process Systems Engineering for Biofuels Development

composed of biopolymers (cellulose 40–60%, hemicellulose 20–40% dry weight). On one

hand, cellulose is composed of long chains of glucose linked together. On the other hand,
hemicellulose is composed of arabinose, galactose, mannose or xylose. However, the main
problem is that both biopolymers are not soluble in water. The remaining fraction (lignin
10–25% dry weight) cannot be fermented for it is resistant to biodegradation. However,
lignin can be used for the production of electricity and/or heat (Rass-Hansen et al. 2007).
To be used as fuel, ethanol must have a purity of almost 100%. This means that the water
content must be much lower compared with the bioethanol produced by current industrial
technology. For the purification of bioethanol, there are several technologies, for example
the use of molecular sieves and membranes.
Despite its great advantages, bioethanol does not possess a higher energy density in com-
parison with gasoline. In practical terms for a car, this means that a tank full of bioethanol
will provide less movement in proportion to the energy density. Bioethanol has a higher
octane rating than gasoline; hence, bioethanol has better antiknock characteristics. This
characteristic of bioethanol can be exploited in motors with slight modification. This would
increase the fuel efficiency of the engine. The oxygen content of bioethanol also gener-
ates greater efficiency, resulting in an eco-friendlier combustion process at relatively low
temperatures.
Reid’s vapor pressure, a measure of fuel volatility, is very low for bioethanol, which
indicates a slow evaporation. This reduces the risk of explosions. However, the low vapor
pressure of bioethanol is disadvantageous in comparison with starting the engine at low
temperatures. Without any help, engines that use bioethanol cannot start at temperatures
below 20 0∘ C. The difficulty of cold start is a main problem for future application of alcohols
such as biofuels (Torres-Ortega et al. 2018).
In summary, the general process for bioethanol production is shown in Figure 13.3.
Bioethanol can be used as:
• Fuel for possible fossil fuels substitution.
• Fuel for power generation.
• Fuel for fuel cells used in thermochemical reactions.
• Fuel in systems of energy cogeneration.
• Raw material in the chemicals industry.

Biomass Enzyme
Bioethanol
Handling Production

Biomass Cellulose Glucose Bioethanol

Pretreatment Hydrolysis Fermentation Recovery

Glucose
Fermentation

Figure 13.3 Bioethanol production process diagram.

 Present and Future of Biofuels 347

Bioethanol has better results in spark ignition engines due to its high octane rating. How-
ever, it is not recommended for use in diesel engines. In general, it is not practical to use
pure bioethanol in spark ignition engines because of its low vapor pressure and high latent
heat of vaporization which makes cold starting problematic.

13.2.2 Biodiesel
This fuel is from triglyceride transesterification; its nature is very similar to fossil diesel.
The elementary raw material for its production is vegetable or animal oil and tallow, being
a conventional source from the residual oils from culinary work (restaurants and industrial
food plants). Vegetable oils commonly used are from oil crops especially sunflower, canola,
soy and palm. The UK represents one of the largest biodiesel producers in Europe, with
the canola seed being the most used. The percentage of biodiesel production would be
much higher, except for the high cost of raw material of some oils. For the above, used
oil recycling is an attractive option, which despite requiring special treatment to eliminate
impurities, competes adequately with fossil diesel. Barnwal and Sharma (2005) present
work on the cost of biodiesel production and information on its production.
Biodiesel, like most biofuels, being a renewable fuel that does not originate from fossil
sources results in high benefits for the environment. Biodiesel aims to reduce the amount
of carbon dioxide emitted. Nevertheless, there are a number of elements that increase and
unbalance the carbon dioxide production and absorption in biodiesel production: the first
aspect to consider is the carbon dioxide originated from fertilizer production for culti-
vation. The other aspects that result in a high percentage of carbon dioxide production
are: the processes of esterification, the oil extraction with solvent, and the refining, dry-
ing and transportation. A methodology that results in adequate estimation of contaminants
of this fuel source is life cycle analysis, which basically evaluates from the biofuel ori-
gin until its final use. The biodiesel results are environmentally friendly since the toxicity
is very low and in the case of spillage it does not represent a danger to the environment
(Demirbas 2009).
Biofuel production is based on three conventional routes starting from oils and fats. The
first two routes involve oil transesterification, with the first route being catalyzed by a base,
and the second route catalyzed by an acid. The third route involves oil conversion into
fatty acids for further transformation into biodiesel. It should be mentioned that higher
production of biodiesel is carried out by the base-catalyzed transesterification, since is a
much less expensive process, with affordable operating conditions and with very high con-
version yields (above 98%) (Leung et al. 2010). Therefore, it is useful to describe this
process.
For the base-catalyzed transesterification, through the esterification process, triglyc-
erides are reacted with alcohols (bioethanol is commonly used), using a catalyst (alkaline),
such as sodium hydroxide or potassium hydroxide. Once the reaction is done, monoesters
are formed, commonly called biodiesel, and in turn the production of glycerol occurs
(Figure 13.4). The most recommended catalyst for ethyl ester biodiesel production is
potassium hydroxide since it has the better properties. Likewise, for methyl ester biodiesel
production, either sodium hydroxide or potassium hydroxide are convenient (Leung et al.
2010). The transesterification reaction can be said to be successful, if it gives the separation
348 Process Systems Engineering for Biofuels Development

Methanol and Biodiesel

Oil and Fats
Catalyst

Washing and
Pretreatment Transesterification Purification
Drying

Glycerin

Figure 13.4 Biodiesel production process diagram.

of glycerol from the esters, after the reaction time. Glycerol is much heavier, and begins
to separate naturally, settling in the background. Once separated, glycerol can be sold to
other industries, for example the cosmetics industry (Leung et al. 2010).
Biodiesel is compatible with fossil diesel, which can be mixed in any proportion (0 to
100%), making it suitable for use in any diesel engine or in any oil-fired furnace. There are
many studies on the effect of biodiesel on these engines, and the conclusion is that it works
the same and many times better than the fossil fuel ones. The transesterification process
of the oil is beneficial for the engine, giving the following characteristics to biodiesel: low
viscosity, complete elimination of glycerides, and the boiling, swelling and fluidity points
are reduced (Gerpen 2005). Certain analyzes are required for the use of biodiesel as a com-
mercial fuel, to ensure that it meets the required specifications. The most important points
to ensure trouble-free operation in diesel engines are: that there is complete elimination of
glycerin, catalyst, and alcohol; that the reaction is carried out completely; and that there are
no free fatty acids.

13.2.3 Biobutanol
This compound is an alcohol, colorless, and flammable. Not only is it used as a biofuel, but
some industrial sectors use it as a solvent. Like all biofuels, it is expected that biobutanol
could be an adequate substitute for fossil fuels thus causing a reduction in greenhouse gases.
It is evident that many biofuels cannot be used directly in internal combustion engines, but
that does not prevent them from being used in mixtures with fossil fuels, such is the case of
biobutanol. To convert butanol as an additive to fossil fuel and improve its conditions, there
are the following useful properties to consider: the energy density (29.2 MJ/dm3 ), the low
melting point (−89.5 ∘ C), the adequate boiling point (117.2 ∘ C), the low flash point (36 ∘ C),
and a beneficial autoignition temperature (340 ∘ C). Currently there is no suitable motor for
using only bioalcohols, this is why all research is carried out with the use of biobutanol as
a powerful additive to fossil fuels (No 2016).
Biobutanol production can be carried out in different ways. An elemental process in
biobutanol production is fermentation. This process is usually carried out with bacteria
of the genus Clostridium acetobutylicum under anaerobic conditions. It is called the ABE
 Present and Future of Biofuels 349

process, in which acetone, butanol, and ethanol are formed, in typical proportions of 3:6:1,
with a final approximate biobutanol concentration of 3% (Quiroz-Ramírez et al. 2017). If
it is decided to produce biobutanol by a fermentative route, the following factors should be
considered: pretreatment cost, raw material cost, process profitability, biobutanol purifica-
tion cost (very low amount with respect to the other compounds), and the toxicity of the
process. The raw material that is better for biobutanol production is waste of agricultural
origin (straw, grass, grains and fruits in poor condition, etc.), since it is cheaper making
the process more profitable. The above is in comparison with grain in good condition for
the fermentation. Another alternative that has been proposed in recent years is the use of
vegetable origin biomass, as in the case of algae, since it is not labor intensive and does not
have high production costs. Some microalgae contain a high percentage of sugars (Chlorella
contains around 30–40% sugars), which increase the biobutanol production. This has led to
the genetic alteration of some bacteria, such as Clostridium acetobutylicum and Clostrid-
ium beijerinckii, where the resistance to the concentration of biobutanol in the fermentation
broth was increased (Sánchez-Ramírez et al. 2017a).
Once the fermentation broth with a low biobutanol concentration is obtained, the dis-
tillation method required for the purification is expensive. This has led to biobutanol not
being economically competitive in comparison with other biofuels. Therefore, other sep-
aration techniques have been explored, such as membrane use, adsorption, liquid–liquid
extraction, pervaporation, and reverse osmosis. Especially, pervaporation is a promising
technology in the recovery of biobutanol, since it allows the separation and concentration
of the product during a single process (Figure 13.5) (Sánchez-Ramírez et al. 2017b).

13.2.4 Biojet Fuel

This biofuel, like its name suggests, is used by the aviation industry (to boost gas turbine
engines). Therefore, it has to conform to certain requirements, more strictly than for other
fuels. On par with the other biofuels, biojet fuel is thought to reduce the dependence on fossil
fuels and the greenhouse gas emissions that these generate. It is estimated that the commer-
cial air industry contributes around 6% of total carbon emissions globally (Krammer et al.
2013). Biojet fuel offers the possibility to reduce greenhouse gas emissions in the aviation
industry.

Biomass
Biobutanol
Handling

Biomass
Hydrolysis Fermentation Recovery
Pretreatment

Lignin Solids

Acetone and Ethanol

Figure 13.5 Biobutanol production process diagram.

350 Process Systems Engineering for Biofuels Development

Biojet fuel

Isomerization/
Biomass Deoxygenation Recovery
Hydrocracking

Figure 13.6 Biojet fuel production process diagram.

There are several technologies for aviation fuel production from biomass. Some of these
technologies are in the research stage and others are already available for use on a commer-
cial scale. The production processes are dependent on the raw material. Hydroprocessing
technologies, such as hydrotreating, deoxygenation, and isomerization and hydrocracking,
are based on oils being converted into biojet fuel. Likewise, there are processes such as
catalytic hydrothermolysis to treat the oils based on triglycerides (Nygren et al. 2009). Tech-
nologies such as biomass gasification are often used on solid raw material, transforming the
matter in alcohols (with the use of biochemical and thermochemical processes), in bio-oils
(through pyrolysis processes), in sugars (through biochemical processes), and in synthesis
gases (Figure 13.6). The previous raw materials can be transformed into biojet fuel, through
synthesis processes, catalytic or fermentative.
Nowadays, the biomass synthesis processes by Fischer–Tropsch to biojet fuel are
approved by the D7566 ASTM International Method, achieving successful mixes up to
50%. At a commercial level (large scale), only hydroprocessing using vegetal oils and
waste is documented, being the only cost-effective conversion path (Nygren et al. 2009).
Fuel use in the aviation sector for the United States constitutes consumption of 20 billion
gallons of fuel per year. This is why a minimum cost reduction in such fuel, results in a large
financial saving. At the beginning of 2012, it was estimated that the annual cost of fuel was
around 50 billion dollars. Nevertheless, for the year 2030, it is estimated that aviation fuel
cost can be reduced by introducing biojet fuel at $2.5/gal. This is due to the improvement in
fuel production technologies for aviation, making it possible to reduce the total annual cost
up to 30%. The parameters to improve the cost of biojet fuel are: raw material costs, the
improvement of the production equipment, the increase in the conversion and product per-
formance, the generation of high value-added products, and generation of adequate energy
integration systems (Hari et al. 2015). Agricultural and forestry raw materials, as well as
algae biomass, are important raw materials for production of alcohol fuels and are able to
provide large quantities of alcohols to convert into biojet fuel (Hari et al. 2015).
Other specifications to comply with biojet fuel are: an acceptable minimum energy den-
sity; a maximum freezing temperature allowed; a maximum allowable viscosity; a max-
imum permissible sulfur content; a minimum aromatic compound content; a minimum
electrical conductivity of the fuel; and a minimum allowable flash point. At the same time,
biojet fuel must have a lower freezing point for flight purposes (Chuck 2016). Among the
biofuels used for the aviation sector are the following:
• Alcohol-to-Jet (ATJ) Fuel: This is the fuel coming from alcohols (bioethanol, biobu-
tanol, and the long chain fatty alcohols). The percentage permitted in the mix with other
fuel is approximately 15% (Chuck 2016).
 Present and Future of Biofuels 351

• Oil-to-Jet (OTJ) Fuel: There are three processes for this route: hydroprocessed renew-
able jet (HRJ), catalytic hydrothermolysis, and pyrolysis (also known as hydrotreated
depolymerized cellulosic jet [HDCJ]). Nowadays, just the HRJ route products are con-
sidered for mixtures and have according to ASTM specification (Chuck 2016).
• Gas-to-Jet (GTJ) Fuel: This route details the conversion processes of biogas, natural
gas, or synthesis gas into biojet fuel (Chuck 2016).
• Sugar-to-Jet (STJ) Fuel: There are two ways to generate biojet fuel starting from inter-
mediate sugar raw materials. The first consists in the catalytic improvement of sugars and
hydrocarbons. The second is linked to the biological conversion of sugars into hydrocar-
bons (Chuck 2016).
Agricultural and forestry raw material, as well as algae biomass, are important raw mate-
rials for the production of alcohol fuels and may provide a considerable quantity of alcohols
to convert into fuel for aircraft. Vegetable oils, animal fats, waste cooking oils, algae oil and
pyrolysis oils are the predominant raw materials for the conversion processes related to oil
(Nair and Paulose 2014).

13.2.5 Biogas
Global warming and its effects are of major concern today. The main purpose of the climate
objectives is to limit the average warming to a temperature of 2 ∘ C compared with the aver-
age temperature in the nineteenth century. Biogas originates from biogenic material and
it is a type of biofuel, typically referred to as a gas produced by bacterial fermentation of
organic material under anaerobic condition. In general, biogas is a contributor toward the
growing interest in the wider use of biofuels.
The composition of biogas is mostly methane and carbon dioxide, typically 40–95% and
5–55% of the blend, respectively. Biogas is approximately 20% lighter than air. In terms of
physical properties, it has an ignition temperature between 50 ∘ C and 750 ∘ C; the calorific
value of 1 m3 is approximately 22 MJ; and it is an odorless and colorless gas that burns with
a clear blue flame.
Biogas can be produced from a wide variety of available organic materials; for example,
wastes, including animal manure, sewage sludge, and municipal organic waste. Anaero-
bic digestion is a simple technology widely used for processing the biodegradable, organic
waste for biogas production (Figure 13.7). Animal manure is used as inoculum, pretreat-
ment of a substrate. The estimation of anaerobically digested substrate growth is around
25%. In this sense, the biogas industry has the potential to generate a substantial amount of
energy. Upon completion of the anaerobic digestion process, the biomass is converted into
biogas (Wieland 2010).
Biogas

Biomass Hydrolysis Acidogenesis Acetogenesis Methanogenesis

Figure 13.7 Biogas production process diagram.

352 Process Systems Engineering for Biofuels Development

13.2.5.1 Biochemical Process

Anaerobic digestion can be divided into four stages: (i) hydrolysis, (ii) acidogenesis,
(iii) acetogenesis, and (iv) methanogenesis. These four stages are developed by different
microorganisms in different conditions.
The microorganisms that hydrolyze and ferment are responsible in the first stage for
attacking polymers and monomers and produce hydrogen and acetate and several volatile
fatty acids (i.e. propionate and butyrate, for example).
Cellulase, cellobiase, xylanase, amylase, lipase, and protease (hydrolytic enzymes) are
excreted by hydrolytic microorganisms. Most bacteria are anaerobic, but some facultative
anaerobes, for example, Enterobacteria and Streptococci, participate in the process. The
fatty acids with higher volatility are converted into acetate and hydrogen by the acetogenic
bacteria. In this process some important details can be highlighted; the accumulation of
hydrogen plays an important role, since it can inhibit the microorganism’s own metabolism.
For this reason, the concentration of hydrogen remains relatively low. When the degrada-
tion stage ends, two groups of methanogenic bacteria generate methane from acetate or
hydrogen and carbon dioxide (Woon and Lo 2016).
In addition, with the decomposition of organic or biological materials, several gases are
discharged. Organic decomposition can occur in two ways: aerobic decomposition (in the
presence of oxygen) and anaerobic decomposition (oxygen is not present). The decomposi-
tion products are quite different: (i) carbon dioxide, ammonia and some other gases in small
amounts are produced in aerobic decomposition or fermentation. Also, very little heat and a
final product (with a higher nitrogen content) is produced in this step; (ii) methane, carbon
dioxide and traces of other gases are produced in anaerobic decomposition.
Anaerobic decomposition is a two-stage process. In the first step, acid bacteria break
down complex organic molecules into peptides, glycerol, alcohol and the simplest
sugars. When these compounds are produced in sufficient quantities (second step), these
simpler compounds are converted into methane with a second type of bacteria. The
methane-producing bacteria are influenced by environmental conditions, which can slow
or finish the process (Achinas et al. 2017). The resulting biogas contains 55–80% methane,
depending on the type of waste. The main gases produced are methane and carbon dioxide.
The composition of the gas varies with the raw material used. The typical composition is:
methane, 50–75%; carbon dioxide, 25–50%; nitrogen, 0–10%; hydrogen, 0–1%; hydrogen
sulfide, traces; steam from water, traces; oxygen, 0–2% (Kaparaju et al. 2009).
Anaerobic digestion is a simple and efficient technology that has been used for a long
time for the production of biogas and it is ready for use in domestic and agricultural applica-
tions. This technology can contribute substantially to the generation of energy from organic
waste, particularly that from agriculture and municipal waste. This type of waste is gen-
erated in very large quantities worldwide. The use of anaerobic digestion technology, in
addition to being an important source of energy generation, also allows the integral use of
biomass residues and impacts on agriculture, livestock, forestry and fishing, thus controlling
pollution and protecting the environment.

13.3 Perspectives and Future of Biofuels

According to the discussion in this chapter, only 1% of energy is generated from biomass.
With this in mind, it is clear that there is a great opportunity for the use of renewable fuels
(Wieland 2010). In recent years there has been an increase in the use of fuels in cars and
 Present and Future of Biofuels 353

urban transport. Currently, the transport sector represents about 20% of the total emissions
in the atmosphere. Additionally, it is estimated that more than 650 million tons of carbon
dioxide have been released per year, with an emission equivalent to that released by 136
million cars. This behavior predicts a constant increase in carbon dioxide emissions and
greenhouse gases; adding to what is already issued by air transport that is often neglected
in investigations (Hill et al. 2015). With this in mind, biofuels from renewable sources
are considered a good alternative to reduce carbon emissions. Several biofuels have been
considered for conventional use in the United States and the European Union with several
intentions: to reduce the emissions, to reduce the use of fossil fuels, and to increase the
inherent safety of fuels.
Several countries are giving this transition impetus. The leaders of the Group of Seven
(G7) countries declared in 2015 to totally decarbonize their economies by dispensing with
fossil fuels by 2040 and using other alternatives (biofuels) as a universal source of energy
(Figure 13.8). Likewise, in the United Nations Climate Change Conference in Paris, coun-
tries approved a new climate protection policy to complement the Kyoto Protocol. It is a
vital step in the right direction. Nevertheless, countless efforts will still be required to get a
change in emissions by 2040 so that climate change does not surpass the 2 ∘ C aim. There-
fore, biofuels are an opportunity for the human race, nature, and the economy. It is time to
take action (United Nations 2016).
The main obstacle to biofuels is the current price of fossil fuels, and the main incen-
tive for biofuels is the growing global population and the need to increase food supplies
(Becken 2002). The biofuels industry helps greatly in the reduction of greenhouse gases,
as well as reducing the dependence on petroleum derivatives, and increases diversity in
renewable sources of energy, and in the generation of jobs. For these reasons, the value
of biofuels is beyond their use as substitutes for fossil fuels; their economic and envi-
ronmental impact must also be considered. At present, the long-term success of biofuels
requires essential information. By the year 2030, there is major potential for biofuel pro-
duction from inedible vegetable and waste. In addition to bioethanol, other chemicals such
as biobutanol, levulinic acid and various carboxylic acids have an enormous potential to
increase the value and usefulness of biofuels (Eggleston and Lima 2015; Ding et al. 2016;
Ramos et al. 2016).
Electricity generation mix worldwide (TWh)

39 444

33 214
Coal
27 222
23 318 Oil

Gas

Nuclear

Hydro

Renewables (Biofuels)

2013 2020 2030 2040

Figure 13.8 Demand of power. Electricity generation mix worldwide (TWh) Source: Adapted from Birol
2015
.
354 Process Systems Engineering for Biofuels Development

References
Achinas, S., Achinas, V., and Euverink, G.J.W. (2017). A technological overview of biogas production from
biowaste. Engineering 3: 299–307.
Barnwal, B.K. and Sharma, M.P. (2005). Prospects of biodiesel production from vegetable oils in India.
Renewable and Sustainable Energy Reviews 9: 363–378.
Becken, S. (2002). Analyzing international transit flow to estimate energy use associated to air travel. Jour-
nal of Sustainable Tourism 10: 114–131.
Birol, F. (2015). World Energy Outlook 2015. International Energy Agency, 1(3).
Cadenas, A. and Cabezudo, S. (1998). Biofuels as sustainable technologies: perspectives for less developed
countries. Technological Forecasting and Social Change 58: 83–103.
Cardona, C.A., Quintero, J.A., and Paz, I.C. (2010). Production of bioethanol from sugarcane bagasse:
status and perspectives. Bioresource Technology 101: 4754–4766.
Chen, H. and Qiu, W. (2010). Key technologies for bioethanol production from lignocellulose. Biotechnol-
ogy Advances 28: 556–562.
Chuck, C.J. (2016). Biofuels for Aviation: Feedstocks, Technology and Implementation. Elsevier.
Demirbas, A. (2000a). Biomass resources for energy and chemical industry. Energy Education Science and
Technology 5: 21–45.
Demirbas, A. (2000b). Recent advances in biomass conversion technologies. Energy Education Science
and Technology 6: 19–40.
Demirbas, A. (2007). Progress and recent trends in biofuels. Progress in Energy and Combustion Science
33: 1–18.
Demirbas, A. (2009). Progress and recent trends in biodiesel fuels. Energy Conversion and Management
50: 14–34.
Difiglio, C. (1997). Using advanced technologies to reduce motor vehicle greenhouse gas emissions. Energy
Policy 25: 1173–1178.
Ding, J.C., Xu, G.C., Han, R.Z., and Ni, Y. (2016). Biobioethanol production from corn stover pretreated
with recycled ionic liquid by Clostridium saccharobutylicum. Bioresource Technology 199: 228–234.
Eggleston, G. and Lima, I. (2015). Sustainability issues and opportunities in the sugars and sugar-bioproduct
industries. Sustainability 7: 12209–12235.
Eggleston, S., Buendia, L., Miwa, K. et al. (eds.) (2006). 2006 IPCC Guidelines for National Greenhouse
Gas Inventories, vol. 5. Hayama, Japan: Institute for Global Environmental Strategies.
Faaij, A. (2006). Modern biomass conversion technologies. Mitigation and Adaptation Strategies for Global
Change 11 (2): 343–375.
Fortman, J.L., Chhabra, S., Mukhopadhyay, A. et al. (2008). Biofuel alternatives to ethanol: pumping the
microbial well. Trends in Biotechnology 26 (7): 375–381.
Gerpen, J.V. (2005). Biodiesel processing and production. Fuel Processing Technology 86: 1097–1107.
Gray, K.A., Zhao, L., and Empatge, M. (2006). Bioethanol. Current Opinion in Chemical Biology 10:
141–146.
Hari, T.K., Yaakob, Z., and Binitha, N.N. (2015). Aviation biofuel from renewable resources: routes, oppor-
tunities and challenges. Renewable and Sustainable Energy Reviews 42: 1234–1244.
Hill, J., Nelson, E., Tilman, D. et al. (2015). Environmental, economic, and energetic costs and benefits of
biodiesel and bioethanol fuels. Proceedings of the National Academy of Sciences of the United States of
America 30: 11206–11210.
Hoekman, S.K. (2009). Biofuels in the U.S. – challenges and opportunities. Renewable Energy 34: 14–22.
Jefferson, M. (2006). Sustainable energy development: performance and prospects. Renewable Energy 31:
571–582.
Kaparaju, P., Serrano, M., Thomsen, A.B. et al. (2009). Bioethanol, biohydrogen and biogas production
from wheat straw in a biorefinery concept. Bioresource Technology 100: 2562–2568.
 Present and Future of Biofuels 355

Krammer, P., Dray, L., and Kohler, M.O. (2013). Climate-neutrality versus carbon-neutrality for aviation
biofuel policy. Transportation Research Part D 23: 64–72.
Leung, D.Y.C., Wu, X., and Leung, M.K.H. (2010). A review on biodiesel production using catalyzed
transesterification. Applied Energy 87: 1083–1095.
McMillan, J.D. (1997). Bioethanol production: status and prospects. Renewable Energy 10: 295–302.
Nair, S. and Paulose, H. (2014). Emergence of green business models: the case of algae biofuel for aviation.
Energy Policy 65: 175–184.
No, S.Y. (2016). Application of biobutanol in advanced CI engines – a review. Fuel 183: 641–658.
Nygren, E., Aleklett, K., and Hook, M. (2009). Aviation fuel and future oil production scenarios. Energy
Policy 37: 4003–4010.
Parikka, M. (2004). Global biomass fuel resources. Biomass and Bioenergy 27: 613–620.
Puppan, D. (2002). Environmental evaluation of biofuels. Periodica Polytechnica Social and Management
Sciences 10: 95–116.
Quiroz-Ramírez, J.J., Sánchez-Ramírez, E., Hernández-Castro, S. et al. (2017). Multi-objective stochastic
optimization approach applied to a hybrid process production-separation in the production of Biobutanol.
Industrial and Engineering Chemistry Research 56: 1823–1833.
Ramos, J.L., Udaondo, Z., Fernández, B. et al. (2016). First-and second-generation biochemicals from
sugars: biosyntheis of itaconic acid. Microbial Biotechnology 9: 8–10.
Rass-Hansen, J., Falsig, H., Jorgensen, B., and Christensen, C.H. (2007). Bioethanol: fuel or feedstock?
Journal of Chemical Technology and Biotechnology 82: 329–333.
Reijnders, L. (2006). Conditions for the sustainability of biomass based fuel use. Energy Policy 34:
863–876.
Sánchez-Ramírez, E., Alcocer-García, H., Quiroz-Ramírez, J.J. et al. (2017a). Control properties of hybrid
distillation processes for the separation of biobutanol. Journal of Chemical Technology & Biotechnology
92: 959–970.
Sánchez-Ramírez, E., Quiroz-Ramírez, J.J., Hernández, S. et al. (2017b). Optimal hybrid separations for
intensified downstream processing of biobutanol. Separation and Purification Technology 185: 149–159.
Segovia-Hernández, J.G., Vázquez-Ojeda, M., Gómez-Castro, F.I. et al. (2014). Process control analysis for
intensified bioethanol separation systems. Chemical Engineering and Processing: Process Intensification
72: 119–125.
Tanger, P., Field, J.L., Jahn, C.E. et al. (2013). Biomass for thermochemical conversion: targets and chal-
lenges. Frontiers in Plant Science 4: 218.
Torres-Ortega, C.E., Ramírez-Márquez, C., Sánchez-Ramírez, E. et al. (2018). Effects of intensification on
process features and control properties of lignocellulosic bioethanol separation and dehydration systems.
Chemical Engineering and Processing: Process Intensification 128: 188–198.
Ture, S., Uzun, D., and Ture, I.E. (1997). The potential use of sweet sorghum as a non-polluting source of
energy. Energy 22: 17–19.
United Nations (2016). Framework Convention on Climate Change. FCCC/CP/2015/10. https://unfccc.int/
resource/docs/2015/cop21/eng/10.pdf (accessed 7 April 2020).
Wieland, P. (2010). Biogas production: current state and perspectives. Applied Microbiology and Biotech-
nology 85: 849–860.
Woon, K.S. and Lo, I.M.C. (2016). A proposed framework of food waste collection and recycling for renew-
able biogas fuel production in Hong Kong. Waste Management 47: 3–10.
 Index

Ablative pyrolyzer, 300 Aspen Plus Economic Analyzer

Ablative reactor, 300 (APEA), 273
Acentric factor, 91, 99, 101, 226 ASTM D7566, 150
Acid catalysts, 125, 127, 128, 137, 194–196, Aviation sector, 9, 149, 167, 350
223, 278
Activity coefficient, 89, 90, 92, 94–97 Bagasse, 4, 22, 25, 37, 159
Adaptive evolution, 175, 186 Bare module cost, 273, 274
Added value products, 21, 22, 26, 33–36, 69, 86, Batch reactor, 138, 141, 153, 154, 157, 234, 235,
110 239, 242
Adsorbents, 32, 38–40, 291 Bifunctional oxides, 120
Agricultural activities, 18 Bi-level optimization, 178
AlCl3 ⋅ 6H2 O, 134–140 Bioalcohols, 2, 6, 7, 111, 343, 348
Alcohol production plus oligomerization, 159 Biobutanol, 2, 10, 54, 312, 314, 315, 342, 343,
Alcohol-to-jet, 150, 350 348–350, 353
Algal biomass, 260–262, 268, 282 Biochemical conversion, 343–345
Alkaline catalysts, 125, 128, 140, 195, 223, 236, Biodiesel, 2, 10, 15, 16, 25, 28, 31, 32, 37, 38,
244, 296 54, 55, 63, 67, 71, 87, 88, 100–103,
Anaerobic digestion, 1, 38, 54, 55, 154, 344, 351, 109–111, 113, 121–133, 137–140, 166,
352 191–201, 205–217, 221–224, 228,
ANP (Agência Nacional do Petróleo, Gás Natural 230–232, 234, 236, 237, 239–241,
e Biocombustíveis), 221, 248, 255 245–248, 253–255, 259–263, 265–268,
Antoine equation, 55, 273 271–281, 286, 289, 312, 342, 343, 347,
Arabinose, 183, 346 348
Artificial neural network, 5, 58, 106 decomposition, 268
Aspen Energy Analyzer, 222, 252 first-generation, 121, 260
Aspen Plus, 10, 50, 206, 222, 224–229, 231–234, process using waste cooking oil, 10
239, 240, 245–249, 251–255, 261, 263, production cost, 124
267, 268, 272, 273, 277, 281, 317, 324, second-generation, 260
325 third-generation, 260

Process Systems Engineering for Biofuels Development, First Edition.

Edited by Adrián Bonilla-Petriciolet and Gade Pandu Rangaiah.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
358 Index

Biodiesel from microalgae 3rd generation, 2

alternate processes, 280 4th generation, 2
process flowsheet, 265 Biorefinery, 8, 9, 22, 50, 56, 59, 65, 67, 68, 72,
process parameters, 266 74, 86, 101, 109, 112, 161, 165, 166
Biodiesel plants, 8, 126, 138, 215, 223, 279 smart, 86, 109, 112
Biodiesel process, 10 Biotechnology, 5, 173, 174
Biodiesel production process design, 239 Blending problem, 68
Biodiesel purification, 205, 208–210, 223, 239, Byproducts
245, 248 green diesel, 151, 152, 159, 160, 164, 165
Biodiesel yield, 31, 128, 196–198, 200, 208, 261, light gases, 151–153, 156, 159, 161, 164, 165
281 naphtha, 151–153, 157, 159, 161, 164, 165
Bioenergy, 3, 6, 16, 18, 109, 165, 192, 291
Bioethanol, 2, 7, 10, 15, 16, 18, 22–24, 35, 36, CAPCOST program, 273, 279
54, 62, 67, 71, 159, 160, 165–167, 174, Carbon dioxide emissions, 56, 149, 191, 289,
183, 192, 260, 289, 312–315, 342–347, 312, 341, 342, 353
350, 353 Carbonization, 10, 287, 303, 305–307
common mixture, 344 Carbonization furnaces, 306
fermentation, 345 Castor oil, 130, 193, 197–201, 204–210, 212,
Biofuel 214, 216, 217
perspectives and future of, 352 methanolysis, 199
climate change, 342, 353 transesterification, 204
transport sector, 341, 342, 353 Catalytic biomass Pyrolysis, 303
producing microorganisms, 6 Catalyzed gasification, 295
supply chains, 3–5, 9 Cavitation, 261
used for aviation sector, 350 CCEP program, 273
Biofuel feedstocks, 19, 38 Cell-recycle reactors, 35
Biofuel process design, 99, 101, 102, 104, 109, Centrifugation, 35, 199, 210–217
110, 112 Ceramic membrane, 210–213, 215
Biofuel production, 2–11, 16–20, 24, 26, 28–32, tubular membrane, 212
34, 35, 40, 62, 68, 86, 88, 94, 114, 193, Chemical biodiesel production, 196
313, 343, 347, 353 Chemical Engineering, 2, 8, 11, 98, 99, 103, 112,
Biogas, 10, 15, 35–38, 40, 52, 54, 73, 109, 154, 273
192, 216, 343, 351, 352 Chemical Engineering Plant Cost Index
anaerobic decomposition, 352 (CEPCI), 273
biochemical process, 352 Chemical equilibrium, 87, 238, 241
composition, 351 Chemical potential, 89
Biogases, 111 Chicken fat, 38, 153
Biogasoline, 2, 343 Citrus waste, 34, 35
Biojet fuel, 2, 149–156, 159–162, 164–167, 343, Co-culture fermentation, 35
349–351 Cold flow properties, 38
cost, 350 Combustion, 10, 19, 28, 30, 32, 35, 37, 51, 68,
production, 9, 152, 159–161, 350 262, 286–290, 292–294, 297–300, 305,
specifications, 350 306, 313, 314, 342, 344, 346, 348
Biokerosene, 150 Computational fluid dynamics (CFD), 8, 55
Biomass, 1, 2, 5, 9, 20, 25, 29, 101, 139 Computer-aided design, 10, 173, 180
feedstocks, 2, 19, 86 Computer-aided methodologies, 5, 7, 9
transformation routes, 2, 5, 8 Constraint-based modeling, 175, 176, 179,
Biomass feedstock 185
1st generation, 2 controllability and inherent safety, 316
2nd generation, 2 Corn waste biomass, 24
 Index 359

Cost Enzymes, 10, 35, 54, 59, 60, 124, 125, 129, 130,
biodiesel production, 10, 124, 260, 275, 347 133, 140, 178, 181, 184, 186, 191, 193,
raw materials, 34, 275, 276 194, 196–203, 206–208, 210–217, 222,
Cost analysis, 10, 261, 272, 281 224, 346, 352
Cost of manufacture, 173, 275 Equation of state, 5, 90
Critical properties, 99–101, 104, 112 Equilibrium reactor, 239, 241
Escherichia coli, 181, 183–187
Esterases, 128
Design Institute for Physical Properties (DIPPR),
Esterification, 6, 10, 28, 31, 125–129, 131–140,
99, 227, 228
194–196, 198, 199, 202, 204, 215, 216,
Diacylglycerols, 194, 237
222–224, 232, 234–236, 238, 239,
Diafiltration, 212–215, 217
241–245, 249, 255, 277, 278, 281, 347
Differential evolution, 173, 174, 179, 324
Ethanol, 6, 22, 23, 31, 32, 35–37, 59, 62, 67–69,
Direct esterification, 125, 127, 128, 131, 132, 101, 115, 129, 131, 132, 159, 160, 174,
134–141 181–186, 194, 197, 222, 224, 229–236,
Discrete element method, 55 238, 241–243, 260, 286, 289, 290, 305,
Distillation, 6, 7, 11, 34, 35, 37, 54–56, 62, 63, 313, 344, 346, 349
103, 113, 139, 151, 152, 159–161, 164, Evaporation, 123, 124, 139, 297, 313, 314, 346
165, 206–210, 216, 223, 243–248, Eversa transform, 129, 130, 199, 200, 205, 208,
250–252, 262–264, 266–268, 272–277, 210–213, 216, 217
279, 281, 313–315, 317, 320–322, 324, Evolutionary algorithms, 178
325, 327, 332, 335, 345, 349 Extraction, 6, 7, 246, 260, 261, 267, 271, 285,
D-limonene, 35 286, 313–315, 317, 320, 331, 335, 347,
Double-counting rule, 120 349
DSTWU, 243
Dual fluidized bed gasification, 292 Fatty acid alkyl esters (FAAEs), 194, 228, 230,
Dual fluidized bed reactor, 293 237
Duality theory, 178, 179 Fatty acid methyl esters (FAMEs), 121–126, 128,
129, 132–135, 137–139, 194–197,
Economic analysis, 124, 164, 195, 223, 224, 252, 199–202, 204–206, 208–211, 214–217,
253, 255, 262 262, 263, 267, 268, 271, 275
Economic evaluation, 10, 86, 88, 205, 254, 319 Feedstock
Economic impact, 216, 315, 342 biomass-based, 1
first-generation, 122, 260
Energy crops, 4, 7, 16, 24, 86, 343, 345
second-generation, 122, 123, 139, 260
Energy integration, 2, 61, 65, 150, 166, 167, 350
Fermentation, 1, 2, 6, 7, 11, 19, 29, 35, 36, 40,
Enterprise wide optimization, 68
55, 59, 62, 69, 159, 165, 173, 174, 183,
Enthalpy, 32, 50, 51, 102, 222, 226, 228, 230,
312, 315, 345, 346, 348, 349, 351, 352
231
Filtration, 123, 124, 129, 132, 197, 210, 212, 215
Entrained flow, 301, 302
First-generation biofuels, 29, 289
Environmental impact, 5, 20, 28, 29, 36, 61, 65, Fischer–Tropsch synthesis, 150, 161, 164, 295
86, 122, 133, 139, 215, 289, 315, 317, Fixed bed
319, 330, 332, 345, 353 gasification, 291
Enzymatic biodiesel production, 128, 130, 215 reactors, 30, 152, 291, 292, 298
Enzymatic catalysts, 6, 10, 139, 194, 196 Fluidized bed
Enzymatic transesterification, 193, 194, 196, gasification, 292, 298
197, 200, 206 reactor, 37, 292, 293, 299
Enzyme-catalyzed routes, 10, 191, 198, 215 Flux balance analysis, 175
Enzyme inhibition, 197, 199, 214 Flux envelop analysis, 180
Enzyme recovery, 198, 211, 214 Flux variability analysis, 176
360 Index

Food and non-food crop, 17 Hydrolysis, 19, 24, 34–36, 55, 62, 69, 128, 135,
Food industry waste, 19 159, 183, 194–196, 199, 201, 202, 204,
Fossil fuel 214–216, 224, 241, 267, 285, 286, 345,
consumption, 342 346, 349, 351, 352
replacement, 16 Hydroprocessing, 150, 151, 154, 155, 159–161,
Fossil fuels, 1, 16, 18, 20, 29–31, 49, 68, 85, 86, 350
121, 191, 192, 259, 260, 287, 312, 320, cracking, 151–153, 158, 160, 161, 164, 295,
341, 342, 344, 346, 348, 349, 353 298, 299, 350
Four-pillar strategy, 149 deoxygenation, 152–154, 161, 350
Free fatty acids (FFAs), 21, 121–129, 131, 132, Hydropyrolysis, 298, 304
134, 137, 138, 140, 153, 194–201, 208, Hydrothermal, 154, 286–288, 293, 294, 296,
210, 211, 214–216, 223, 224, 232, 234, 305, 307
238, 255, 277, 348 Hydrothermal carbonization, 287
Fugacity, 89–91, 93–95 Hydrothermal gasification, 287, 293, 294
Fuzzy equal metabolic adjustment (FEMA), 181 Hydrothermal liquefaction, 154, 287, 294, 296

Galactose, 183, 346 Indexes/objectives

Gasification, 1, 11, 19, 28, 52, 56, 62, 67, 101, condition number, 317, 322–324, 326–332,
150, 151, 159, 164, 286–288, 290–296, 335
298–300, 344, 350 controllability, 316, 317, 319, 322, 330
General algebraic modeling system (GAMS), economic performance, 304
180, 206
environmental impact, 315, 317, 319, 330, 332
Genetic modulation, 173, 180
individual risk, 317, 320, 324
Genome-scale metabolic network, 179, 180, 187
minimum singular value, 322
Gibbs energy, 87, 93, 102, 226, 235, 242, 303
relative gain matrix, 322, 323
Glucose, 67, 159, 164, 181, 183–185, 288, 346
tree of events, 320, 321
Glycerol decomposition, 206
Instant coffee, 22
Glycerolysis, 126, 127
Integrated process and product design, 109
Grape pomace, 36, 37
Intensification
Grape skins, 36
microwave, 6, 7, 259, 261, 304
Greenhouse gases, 52, 65, 173, 191, 192, 342,
ultrasound, 6, 7, 261
345, 348, 349, 353
Intensified process, 2, 11, 314, 317, 325, 335
Group contribution methods, 5, 103, 115
International Energy Agency (IEA), 192, 342
Growth-coupled Strains, 178, 181, 183–186
Ionic liquids, 7, 9, 124
Isomerization, 152–155, 161, 164, 224, 350
Hazard and Operability Study (HAZOP), 66, 320
Heat exchanger networks (HEN), 62
Heterogeneous azeotropes, 316 Kinetic modeling, 10, 198, 201, 202, 221, 232
Heterogeneous catalysis, 9, 127, 128, 194, 223 Kinetic models, 53, 54, 175
cation exchange resins, 127 Kinetic parameters, 201–204, 206, 216, 234, 237,
FeSO4 , 128 263
H-form zeolites, 127 Kriging, 57–59
metal oxides, 127, 160
Higher level simulation algorithms, 94 Lewis acids, 125, 126, 132–134
Homogeneous catalysis, 125, 140 Life cycle analysis, 8, 29, 65, 319, 347
metal acetates, 126 Lignocellulosic material, 1, 9, 10, 159, 161, 166,
stearates, 126 288, 289, 315, 345
sulfamic acid, 125, 126 Linear programming, 62, 176, 177
Husks, 18, 22–24, 38, 39, 162 Lipases, 128–130, 196–199, 213–217, 224
 Index 361

Liquefaction, 1, 10, 153, 154, 161, 286–288, 294, Optimization, 2–8, 10, 50, 54, 56, 59–70, 86, 90,
296, 343 92, 93, 97, 98, 104, 125, 149, 166, 173,
Liquid–liquid equilibria (LLE), 88, 92, 95, 97, 175–181, 184, 185, 193, 202, 205, 208,
231 215, 216, 232, 250–252, 260, 261, 276,
281, 316, 317, 323–326, 328, 330–332,
Macauba, 38 335
Mango seeds, 39 Optimization methodology
Mannose, 183, 346 differential evolution with tabu list, 324
Mass and energy balances, 50, 51, 55, 71 Pareto front, 325–328, 330–332
Mathematical modeling, 135 purities, 326, 332
Mechanistic models, 51 purity constraints, 325
Metabolic engineering, 6, 174, 185 recoveries, 326
Metaheuristics, 8 Orange peel, 33–35
Metal hydrated salts, 133 Organic wastes, 16, 18, 19, 351, 353
Methane from citrus, 34
Methanol, 31, 38, 52, 67, 68, 71, 101, 115, 121, Parameter regression, 236
126, 133, 140, 194–197, 199, 201–203, Perennial biomass, 25
205, 206, 208, 210–214, 216, 217, 222, Petroleum, 6, 11, 121, 152, 156, 221, 224, 260,
224, 229, 231, 234, 236, 237, 239, 290, 295, 315, 341, 342, 352
241–247, 249–251, 262, 263, 267, 268, Phase equilibrium, 4, 5, 9, 54, 55, 85, 93, 94,
276, 277, 279–281, 290, 305, 348 100, 102, 114, 222, 229, 241
Methyl ethyl ketone, 11, 100, 115, 311, 313, 314 Phase separators, 206, 262, 267, 274, 277–279
Microalgae, 2, 7, 10, 55, 152, 154, 155, 260–263, Physicochemical pretreatments, 123, 124
268, 276, 277, 279–281, 288, 349 Physicochemical properties, 4, 5, 33, 328
Microbial fermentation, 6 Pinch technology, 61
Microbial genome engineering, 6 Plantwide control, 8
Microbial oils, 260, 261 Plasma gasification, 294
Microorganisms, 2, 6, 174, 178, 180, 183, 185, Polyphenolic compounds, 37
289, 345, 348, 349, 352 Primary biofuels, 342
Process controllability, 8
Minimization of metabolic adjustment, 176
Process design, 2, 4, 5, 7–10, 17, 33, 50, 63,
Mixed integer linear programming (MILP), 62,
66–68, 73, 85, 86, 98, 99, 101–105,
177
109–113, 185, 221, 239, 262, 277
Mixed integer programming (MIP), 187
Process development, 173, 174, 185, 259, 261,
Monoacylglycerols (MAGs), 126, 194, 237, 238
262, 281
Multi-effect columns, 63
Processing pathways, 150
Multi-objective optimization (MOO), 8, 17, 69,
Process intensification, 2, 5, 6, 50, 150, 166, 167,
173, 317, 323, 324
314, 335
Multiscale, 9, 49, 71–74
Process simulation, 10, 114, 193, 205, 216, 222,
254, 262, 263
Nano-sized solid acid, 128 Process synthesis, 2, 9, 50, 60, 66, 99
Nested hybrid differential evolution (NHDE), Production rate, 174, 177, 178, 181–183, 185
179 Product separation, 193, 216, 304
Nonlinear programming, 62, 65, 206 Pyrolysis, 1, 10, 19, 24, 25, 28, 29, 32, 37, 56,
Non-random two-liquid (NRTL) model, 92, 95, 69, 101, 150, 151, 159–161, 286–288,
263, 317 291, 296–307, 343, 350, 351
fast, 160, 297, 298, 306
Oil residues, 31 flash, 288, 297, 298
Oligomerization, 159, 160, 165 hydroprocessing, 160
Olive tree, 26, 27, 29 slow, 25, 29, 32, 297, 305
362 Index

Quadratic constrained programming (QCP), 187 Syngas, 31, 55, 62, 67–69, 71, 151, 164, 286,
Quadratic programming (QP), 59, 177 290, 293, 295, 344
Quantitative structure-property relationships Synthetic paraffinic kerosene, 150
(QSPRs), 103, 105 Systems biology, 174, 175, 180, 187
makeup language, 179
Reaction data regression, 234
Reaction kinetics, 232, 239, 242, 262, 263, 280 Thermochemical conversion, 288, 289, 304, 343,
Reaction mechanism, 201, 216 344
Reaction rates, 121, 128, 129, 137, 195, 197, Thermochemical processes, 29, 285–289, 293,
198, 200, 237, 239, 300, 305 304, 308, 350
Reaction temperature, 127, 198, 204, 206, 208, Thermodynamic modeling framework-TMF, 9,
263, 296 86
Reactor models, 10, 234, 239 Thermodynamic models, 4, 5, 9, 87, 88, 90,
Recirculating fluidized bed, 299 92–94, 114, 229, 231, 246, 277
Regulatory on–off minimization (ROOM), 177 binary interaction parameters, 92, 93, 97, 98,
Renewable aviation fuel, 149, 150, 152, 155, 113, 229, 230
161, 165–167 equation of state (EoSs), 5, 90
Renewable energy, 1, 9, 10, 16, 191, 192, 259, gamma-phi method, 89
344 perturbed-chain statistical associating fluid
Renewable sources, 16, 308, 341, 353 theory (PC-SAFT), 92
Rice residues, 24 phi-phi method, 89, 92
Rotating cone, 299, 300 Soave–Redlich–Kwong (SEK), 90, 99
Route reactions, 224 statistical associating fluid theory (SAFT), 92
Thermodynamic properties, 4, 9, 10, 85, 94, 111,
Rules of thumb, 56, 60, 61
225, 315, 317
Thermophysical properties, 9, 86–88, 94, 98, 99,
Screw reactor, 301, 307
103, 112–114
Secondary biofuels, 342, 343
Third-generation biofuels, 260, 289
Second-generation biofuels, 289 Top-to-bottom, bottom-to-top coupled method,
Separation, proteins/carbohydrates, 267 17
Sewage scum, 122, 133, 139 Torrefaction, 28–30, 307
Sewage sludge, 139, 140, 289, 351 Total annual cost, 263, 315, 319, 324, 350
Sewer grease, 122 Total capital cost, 254
Short cut methods, 50, 55, 206 Total module cost, 273
Six-tenths rule, 279, 280 Total operation Cost, 254
Sizing, 10, 225, 272, 273, 306, 330, 332 Transesterification, 1, 5, 7, 10, 19, 31, 55, 63, 87,
Soap formation, 196, 223, 237, 238 101, 103, 121, 124–129, 131–133,
Stage wizard, 244 137–140, 193–199, 201–206, 208, 210,
Stochastic optimizers, 8 212–217, 222–224, 232, 236, 238, 239,
Stoichiometry Matrix, 175 241, 242, 244–246, 255, 260–263, 267,
Straw of cereals, 22 268, 271, 277, 280, 289, 347, 348
Sulfuric acid, 125, 127, 136, 195, 196, 216, 242, Transport and distribution of biofuels, 29
244–246, 249, 253, 277 Transport fuels, 16, 20, 122
Supercritical fluid extraction, 36, 37 Trap grease, 122, 139
Supercritical water gasification, 294 Triacylglycerols, 126, 194, 222–224, 238, 241,
Superstructure, 59, 61, 62, 64, 65, 70, 71 245
Supply chain, 2–5, 7–9, 11, 28, 49, 50, 68–74, Triolein, 113, 115, 239, 241, 242, 245–247, 251,
150, 166 253, 262, 277
Surrogate models, 56–59, 70, 71 Triple-level mixed-integer linear optimization
Sustainable development, 2, 29, 49, 65, 149, 167 problem, 179
 Index 363

Ultrafiltration, 211 solid urban waste, 159

permeate, 212, 214, 215 textile residues, 159
Ultrasound, 6, 7, 261 Wood waste, 28, 159
UNIQUAC, 95–97, 229–231, 239, 245, 246, 255, Waste mass valorization, 30
277 Waste sugar and starchy feedstock
Urban primary sludge, 139 Fruit waste, 164
Waste bakeries, 165
Vacuum reactors, 301 Waste candy, 165
Valorization of biomass, 16, 18 Waste sweets, 165
Value-added functional products, 86 Waste triglyceride feedstock
Value of bioenergy, 18 animal fats, 1, 4, 7, 129, 131, 133, 150, 152,
Van der Waals mixing rule, 91 154, 167, 193, 194, 221, 222, 231, 236,
Vapor catalysis, 303 343, 351
Vapor–liquid equilibria (VLE), 88, 92, 95, 100, bio-oil, 31, 32, 37, 69, 150, 151, 192, 286,
229, 267 296–298, 303, 304, 343, 350
Vegetable oils, 1, 4, 15, 28, 109, 121, 122, 126, cooking oil, 9, 10, 25, 37, 38, 69, 109, 111,
129, 133, 152, 153, 192–194, 221–223, 122, 150–154, 156, 157, 166, 167, 193,
231, 260, 277, 289, 343, 347, 351 195, 222, 223, 260, 261, 351
Water networks, 64
Waste animal grease, 122 Wheat straw, 24, 343
Waste collection and management, 30
Waste cooking oils, 9, 10, 38, 69, 122, 150–154, Xylose, 67, 183–185, 346
156, 166, 167, 195, 222, 223, 261, 351
Waste lignocellulosic feedstock
agricultural residues, 2, 7, 159, 345