THESIS - C. JAYASREE - Jayasree Chakkamalayath

STUDY OF CEMENT-SUPERPLASTICIZER INTERACTION AND
ITS IMPLICATIONS FOR CONCRETE PERFORMANCE
A THESIS
submitted by
C. JAYASREE
for the award of the degree
of
DOCTOR OF PHILOSOPHY
BUILDING TECHNOLOGY AND CONSTRUCTION MANAGEMENT DIVISION

DEPARTMENT OF CIVIL ENGINEERING
INDIAN INSTITUTE OF TECHNOLOGY MADRAS
CHENNAI 600 036
DECEMBER 2008
THESIS CERTIFICATE
This is to certify that the thesis entitled “STUDY OF CEMENT-
SUPERPLASTICIZER INTERACTION AND ITS IMPLICATIONS FOR
CONCRETE PERFORMANCE” submitted by C. Jayasree to the Indian Institute
of Technology Madras for the award of the degree of Doctor of Philosophy is a
bonafide record of research work carried out by her under my supervision. The
contents of this thesis, in full or in parts, have not been submitted to any other
Institute or University for the award of any degree or diploma.
Chennai 600 036 Prof. Ravindra Gettu

Date: (Research Guide)
Professor
Department of Civil Engineering
Indian Institute of Technology Madras
ACKNOWLEDGEMENTS
I am deeply indebted to my research guide Prof. Ravindra Gettu, Department of Civil

Engineering for his timely guidance, integral view of research, uncompromising
standards for quality and inestimable help right from the initial stage of the research
having enabled me to complete the thesis within the stipulated time. His calmness,
patience, positive and constructive criticisms, willingness to spend long hours on
discussions and his commitment to the job at hand are qualities to emulate. He always
made me feel welcome in his office with a smiling face, no matter how busy he was,
or how silly my question. The experience has been enriching and enlightening and I
thank God almighty for giving me an opportunity to work as a research scholar with
him at Indian Institute of Technology, Madras.
I also wish to place on record my sincere gratitude to Prof. K. Ramamurthy,

Department of Civil Engineering, who had guided me during the beginning of my
research programme and always kept an eye on the progress of my work by giving
suggestions as a doctoral committee member.
I am extremely thankful to Dr. J. Muralikrishnan, Assistant Professor, Department of

Civil engineering, for the great interest he took to help me with valuable suggestions
at certain stages of the thesis work. My sincere thanks are also due to Dr. Manu
Santhanam, Assistant Professor, Department of Civil engineering for the creative
discussions and suggestions at different stages of the thesis work.
I express my sincere thanks to the present Head of Civil Engineering Department,

Prof. K. Rajagopal and the former head, Prof. S. Mohan for all the administrative
support extended to me for carrying out this work. I would like to thank my doctoral
committee members Dr. Ligy Philip, Dr. G. Sekhar and Dr. Anuradha Banerjee for
the suggestions at different stages of my work.
I take this opportunity to extend my sincere gratitude to all faculty members of the
Building Technology and Construction Management Division, Prof. A.
Ramachandriah, Prof. M. S. Mathews, Prof. K. N. Sathyanarayana, Prof. Koshy
Varghese, Dr. K. Ananthanarayanan and Dr. Ashwin Mahalingam for all the
encouragement and support.
i
I also wish to thank Prof. C. Vijayan, Department of Physics, Dr. Abhijit Deshpande,
Associate Professor, Department of Chemical Engineering and Prof. Kesavan Nair,
Department of Metallurgical Engineering for supporting me with their valuable time
for discussions and permitting me to use the laboratory facilities.
I am pleased to record my gratitude to the Management of Amrita Vishwa Vidya

Peetham University, Coimbatore for sponsoring me for the Ph. D programme under
Q.I.P. scheme. My sincere thanks to Prof. C. R. Ramakrishnan, Prof. T.R.
Padmanabhan, Dr. K.M. Mini, Mrs. Sunitha and other colleagues in the Department
of Civil Engineering, Amrita School of Engineering for extending a helping hand at
crucial times.
The staff of concrete lab will always have a special place in my heart for all the efforts
they have put in towards the successful completion of my experimental work. Ms. A.
Malarvizhi, Messers M. Soundarapandiyan, Murthy, Krishnan, Subramaniam,
Dhanasekaran, Watson all have contributed their might for the smooth conducting of
the experimental work. I am also thankful to all the staff members of the Civil
Engineering Department Office, Stores section and DCF for their excellent
cooperation.
I also wish to thank Palani, Premraj, Premkumar, Vijayakumar, Krishna and Prasad
who admirably and unstintingly helped me physically through all the laboratory work.
I would like to thank the chemical admixture companies, BASF Construction

Chemicals, Sika, Cera-Chem and, Burzin and Leons for giving the product
information that was needed as well as the products that they donated for use in this
thesis. I also wish to thank Grassim Industries for supplying cement for the study.
I have been blessed with a friendly and cheerful group of fellow research scholars-
Cindrela, Glory, Geetha, Indu, Rakesh, Ganesh, Prakash, Anand, Senthil Kumar,
Rajesh, Elson, Paul, Venkatesan, Shajatnan, Rajasekar, Boeing, Liju, Arun, Dr.
Manikandan, Dr. Uma, Dr. Ramesh Babu, Dr. Kanagasabhapathi, Dr. Harikrishnan,
Dr. Nambiar, Dr. Anil kumar and many others. I express my gratitude to all of them
for making my life at IIT pleasant and memorable.
ii
I would like to acknowledge my indebtedness to several of the individuals who have
been instrumental in this research- Simi, Vinoj, Anbudayanidhi and Diwakar had
helped me with suggestions. Many thanks to Jennifer, Sansu, Shanthi and Mohan for
helping me in the laboratory while conducting experiments. I am grateful to
Ms. Nigiya, Dr. Sindhu and their families for their affection and help during my stay
at IIT.
I am deeply indebted to my in-laws who have encouraged me in fulfilling my dream

of doing research at IIT. I am grateful to my husband, Girish who in his own special
way has supported me and my work in the last three years. To my daughter,
Sreelakshmi, for her patience, understanding and interest in my studies. My heartfelt
thanks to Ms. Subhadra, who was with me all the three years for making my life at IIT
pleasant and smooth. Finally, I feel a deep sense of gratitude to my mother, sister and
other family members who have supported me in all possible ways and have made me
what I am today.
C. JAYASREE
iii
ABSTRACT
Chemical admixtures, particularly superplasticizers have become indispensable
because of their beneficial effects in fresh and hardened concrete. However, the
compatibility between the cement and admixture, which is influenced by the
characteristics of cement paste, and the type and dosage of admixtures, could affect
the benefits of incorporating a superplasticizer. It is, therefore, essential to understand
the mechanisms that influence the cement-superplasticizer interaction for selecting an
appropriate type of admixture at an optimum dosage. The current study addresses
these and other specific issues pertaining to the performance of concrete incorporating
chemical and mineral admixtures, especially superplasticizers.
A suitable mixing method for the preparation of paste is first selected so that the paste
prepared will be representative of that of concrete. In order to evaluate the influence
of superplasticizers and cement characteristics, studies are conducted with four
families of superplasticizers and four types of cements on fluidity, loss of fluidity and
setting behaviour. The relative fluidity of different superplasticized pastes are studied
with the Marsh cone, mini-slump and viscometric tests for evaluating the best
combination. The Marsh cone flow time, yield stress and plastic viscosity decreases
whereas the mini-slump spread increases with an increase in the dosage of the
superplasticizer upto the saturation dosage, after which they remain constant. An
experimental and analytical validation of the Marsh cone test is done using
rheological models to understand its effectiveness in capturing the relative fluidity.
The correlation between the flow behaviour of paste with normal concrete and self
compacting concrete shows that the saturation dosages obtained from paste and
mortar are comparable whereas a dosage slightly higher than the saturation dosage of
iv
the paste is required for adequate workability in concrete. The setting behaviour is
studied through Vicat penetration and electrical conductivity methods. Both tests
confirm that superplasticizers retard the setting of cement paste; however, the extent
of retardation varies with the type of superplasticizer.
Another aspect studied in this work is the influence of the superplasticizer on the
hydration processes and on the development of microstructure of hydrated cement
paste. With this aim, cement pastes are analyzed using several characterization
techniques (X-ray diffraction, scanning electron microscopy, thermal analysis and Si
nuclear magnetic resonance). The test results clearly indicate the influence of
superplasticizers in delaying the hydration. The superplasticizer delays the formation
of ettringite as well as the silicate polymerization; however, it produces a dense and
uniform microstructure, enhancing the strength development.
The absorbance study using a UV spectrometer gives insight into the influence of
cement and superplasticizer characteristics on the absorbance of the superplasticizers
and its variation with time. The study of the influence of other admixtures on the
cement-superplasticizer interaction shows that the addition of metakaolin and
viscosity modifying admixtures increases the saturation dosage of the superplasticizer
and the viscosity of the paste.
Finally, based on the test results, a simple methodology is recommended to select a
compatible cement-superplasticizer combination is recommended. The use of Marsh
cone and Vicat penetration tests in the paste phase, and final trials on concrete for
strength and cost evaluation are suggested as part of this methodology.
Keywords: Superplasticizer; Marsh cone; mini-slump; rheology; loss of

fluidity; compatibility.
v
TABLE OF CONTENTS
Title Page No.
ACKNOWLEDGEMENTS ....................................................................................... i
ABSTRACT ............................................................................................................... iv
LIST OF TABLES ................................................................................................... xii
LIST OF FIGURES ................................................................................................ xiv
CHAPTER 1 INTRODUCTION
1.1 General ...........................................................................................................1

1.2 Objectives and Scope .....................................................................................3
1.3 Structure of the Thesis....................................................................................4
CHAPTER 2 STATE-OF THE-ART REPORT ON CEMENT-

SUPERPLASTICIZER INTERACTION
2.1 General ...........................................................................................................7

2.2 Cement Chemistry ..........................................................................................8
2.2.1 Raw Materials Used in Cement Manufacture ................................................8
2.2.2 Cement Composition ......................................................................................9
2.2.3 Hydration of Cement ....................................................................................13
2.2.3.1 Hydration of Cement compounds ................................................................13
2.2.3.2 Hydration of the Portland Cement Grain .....................................................15
2.3 Water Reducers and Superplasticizers .........................................................18
2.3.1 Characteristics and Chemistry of Superplasticizers .....................................19
2.3.2 Mechanisms of Action of Superplasticizers .................................................21
2.3.3 Factors Affecting the Cement-Superplasticizer Interaction ........................28
2.3.3.1 Effect of Chemical Structure of Superplasticizer ........................................28
2.3.3.2 Effect of Calcium Sulphate .........................................................................33
2.3.3.3 Effect of Soluble Alkalis ..............................................................................35
2.3.3.4 Effect of C3A ................................................................................................36
2.3.3.5 Other Factors Affecting the Cement-Superplasticizer Interaction ...............38
2.3.4 Effect of Superplasticizer on the Morphology of Hardened Cement Paste .39
2.3.5 Effect of Superplasticizer on Properties of Concrete .................................41
2.3.6 Admixture–Admixture Interaction ...............................................................43
vi
Table of Contents (Contd.,) Page No.
2.3.6.1 Cement superplasticizer Interaction in the Presence of Mineral

Admixtures ...................................................................................................43
2.3.6.2 Influence of other Chemical Admixtures on Cement-Superplasticizer
Interaction.....................................................................................................44
2.4 Need for the Present Study ...........................................................................44
CHAPTER 3 STUDY OF THE FLOW BEHAVIOUR OF

SUPERPLASTICIZED CEMENT PASTE
3.1 General ................................................................................................ 47

3.2 Materials Used in the Study .........................................................................48
3.2.1 Cement .........................................................................................................48
3.2.1.1 Determination of the Water Demand for Standard Consistency
of Cement ................................................................................................... 50
3.2.1.2 Determination of Initial and Final Setting Times.........................................50
3.2.1.3 Determination of the Fineness of Cement ....................................................51
3.2.2 Superplasticizers ..........................................................................................51
3.3 Test Methods Used for Studying the Flow Behaviour of the
Cement Paste ...............................................................................................52
3.3.1 Marsh Cone Test .........................................................................................53
3.3.2 Mini-slump Test ...........................................................................................56
3.3.3 Viscometer Test............................................................................................57
3.4 Selection of Mixing Method ........................................................................62
3.4.1 Materials Used in the Selection of Mixing Method .....................................63
3.4.2 Experimental Details ...................................................................................63
3.4.3 Ball Milling ..................................................................................................64
3.4.4 Results and Discussions ..............................................................................66
3.4.5 Conclusions ..................................................................................................69
3.5 Influence of Type and Dosage of Superplasticizer on the
Flow Behaviour of Cement Paste .................................................................70
3.5.1 Testing Details..............................................................................................70
3.5.2 Results and discussions ................................................................................71
3.5.3 Correlations between the Results from Empirical Tests and
the Viscometer..............................................................................................78
3.5.4 Influence of Type and Dosage of the Superplasticizer on the
Non-Newtonian Characteristics of Cement Paste ........................................82
vii
3.5.5 Change of Flow Behaviour with Time .........................................................82

3.5.6 Summary ......................................................................................................85
CHAPTER 4 CORRELATIONS BETWEEN THE FRESH AND HARDENED

PROPERTIES OF SUPERPLASTICIZED PASTE, MORTAR
AND CONCRETE
4.1 General .........................................................................................................87

4.2 Experimental Details ....................................................................................88
4.2.1 Materials ......................................................................................................88
4.2.2 Test Procedures ............................................................................................90
4.2.2.1 Tests on Paste ...............................................................................................90
4.2.2.2 Tests on Mortar ............................................................................................90
4.2.2.3 Tests on Concrete .........................................................................................91
4.2.3 Mixing Methods for Paste, Mortar and Concrete .........................................92
4.3 Results and Discussions ...............................................................................93
4.3.1 Comparison of the Flow Behaviour of Paste, Mortar and Concrete ............93
4.3.1.1 Flow Behaviour of Paste ..............................................................................93
4.3.1.2 Flow Behaviour of Mortar............................................................................94
4.3.1.3 Flow Behaviour of Concrete ........................................................................96
4.3.2 Correlation between the Fluidity of Paste, Mortar and Concrete .................98
4.4 Loss of Fluidity ..........................................................................................101
4.4.1 Loss of Fluidity of Paste.............................................................................101
4.4.2 Loss of Fluidity of Mortar ..........................................................................103
4.4.3 Loss of Fluidity of Concrete.......................................................................105
4.5 Effect of Superplasticizer on Setting of Cement Paste and Concrete ........106
4.6 Analysis of Setting Behaviour of Superplasticized Cement Paste Using
Electrical Resistivity Method ....................................................................109
4.6.1 Background ................................................................................................109
4.6.2 Experimental Procedure .............................................................................110
4.6.3 Results and Discussions .............................................................................111
4.7 Effect of Superplasticizer on Compressive Strength of Concrete ..............116
4.8 Summary ....................................................................................................117
viii
CHAPTER 5 EFFECT OF SUPERPLASTICIZER ON PROPERTIESS OF

HARDENED CEMENT PASTE
5.1 General .......................................................................................................120

5.2 X-ray Diffraction (XRD) Study .................................................................121
5.2.1 Background ................................................................................................121
5.2.2 Materials Used and Sample Preparation ....................................................122
5.2.3 Experimental procedure .............................................................................122
5.2.4 Results and discussions ..............................................................................123
5.3 Scanning Electron Microscopy (SEM) ......................................................128
5.3.1 Background .................................................................................................128
5.3.2 Sample Preparation .....................................................................................129
5.3.3 Comparison of the Microstructure of Superplasticized Paste and
Reference paste .........................................................................................130
5.4 Thermal (DTA/TG/ DTG) Analysis ...........................................................135
5.4.1 Background ................................................................................................135
5.5 Nuclear Magnetic Resonance Spectroscopy (Si NMR) .............................142
5.5.1 Background ................................................................................................142
5.5.2 Sample Preparation ....................................................................................143
5.6 Summary ....................................................................................................146
CHAPTER 6 CEMENT-SUPERPLASTICIZER COMPATIBILITY
6.1 General .......................................................................................................148

6.2 Materials Selection and Characterisation ...................................................149
6.2.1 Types of cements ........................................................................................149
6.2.2 Superplasticizers.........................................................................................150
6.3 Experimental Procedure .............................................................................151
6.4 Results and discussions ..............................................................................151
6.4.1 Influence of Cement Composition on the Flow Behaviour of
Superplasticized Pastes ..............................................................................151
6.4.2 Influence of Cement Composition on the Setting of
Superplasticized Pastes ..............................................................................160
ix
6.5 Study of the Adsorption of superplasticizers .............................................162

6.5.1 Background ................................................................................................162
6.5.2 Materials Used............................................................................................163
6.5.3 Experimental Procedure .............................................................................163
6.5.4 Results and Discussions of Absorbance Tests ...........................................165
6.5.4.1 Variation in Superplasticizer Absorbance with type of cement .................168
6.6 Identifying Compatible Cement-Superplasticizer Combinations ..............168
6.7 Summary ....................................................................................................176
CHAPTER 7 FLOW BEHAVIOUR OF SUPERPLASTICIZED PASTES

INCORPORATING OTHER ADMIXTURES
7.1 General .......................................................................................................177

7.2 Materials Used in the Study .......................................................................178
7.3 Flow Behaviour of Pastes with Metakaolin ...............................................179
7.3.1 Effect of Metakaolin in Paste Incorporating PCE based Superplasticizer .180
7.3.2 Effect of Metakaolin in Paste Incorporating SNF based Superplasticizer .183
7.4 Influence of VMA on the Flow Behaviour of Superplasticized Paste .......185
7.4.1 Characteristics of VMA .............................................................................185
7.4.2 Experimental Details ..................................................................................187
7.4.2.1 Study of the Interaction between Superplasticizer and VMA ....................187
7.4.2.2 Effect of Time on the Flow Behaviour of Paste
Incorporating Superplasticizer and VMA .................................................191
7.5 Comparison of the Flow Behaviour of Paste and Self
Compacting Concrete .................................................................................193
7.5.1 Mix Design and Testing of SCC ................................................................193
7.6 Summary ....................................................................................................200
CHAPTER 8 CONCLUSIONS AND RECOMMENDATIONS FOR

FURTHER RESEARCH
8.1 General Conclusions ..................................................................................202

8.2 Specific Conclusions ..................................................................................204
8.2.1 Influence of Mixing Method on the Flow Behaviour of Paste ...................204
8.2.2 Influence of Superplasticizers on the Flow Behaviour of Paste.................205
x
8.2.3 Comparison between Marsh Cone Flow time and

Rheological Parameters ..............................................................................206
8.2.4 Effect of Type and Dosage of superplasticizer on the Retardation of
Setting.........................................................................................................207
8.2.5 Correlation between Paste, Mortar and Concrete.......................................207
8.2.6 Effect of Superplasticizer on the Hydration Process and on the
Microstructural Development ....................................................................209
8.2.7 Influence of Chemical Composition of Cement on Compatibility ............210
8.3 Guidelines for Selecting the Best Combination of Cement and
Superplasticizer ..........................................................................................211
8.4 Recommendations for Further Research ....................................................211
APPENDICES .........................................................................................................214
REFERENCES ........................................................................................................243
PUBLICATIONS BASED ON THE THESIS ......................................................263
xi
LIST OF TABLES

Table No. Title Page No.
2.1 Typical Portland Cement Composition Conforming

to ASTM C 150 .............................................................................................10
2.2 Requirements of Different Grades of Cement as per IS codes ......................11
2.3 Chemical Analyses of Some CEM I Portland Cements of UK......................12
2.4 Chemical Structure of Superplasticizers ........................................................20
2.5 Structural Aspects of the PCE that Affect the Fluidity of Cement Paste .......32
3.1 Chemical properties of the cement and the corresponding clinker ................49
3.2 Bogue composition of the clinker ..................................................................49
3.3 Physical properties of cement 1 ....................................................................50
3.4 Properties of superplasticizers........................................................................52
3.5 Optimum superplasticizer dosage and range of unit weight for different
mixing methods ...............................................................................................67
3.6 Rheological characteristics and Marsh cone flow times ................................77
3.7 Test data for pastes with different superplasticizers at 0 and 60-minutes......83
4.1 Summary of properties of aggregates ............................................................89
4.2 Summary of Marsh cone data and concrete cost............................................94
4.3 Loss of fluidity of mortar .............................................................................104
4.4 Initial and final setting time of paste and concrete ......................................109
4.5 Compressive strength of concrete for different combinations .....................117
5.1 Relative evolution of crystalline phases with age ........................................127
5.2 Summary of mass loss in different pastes ....................................................142
5.3 Ranges of 29 Si chemical shifts of the Qn tetrahedrons.................................143
6.1 Chemical and Bogue composition of different cements ..............................150
6.2 Physical properties of different types of cements ........................................150
6.3 Test data for pastes with different superplasticizers for cement C1 ............155
6.7 Comparison of the saturation dosages at 0-minutes for different cements ..159
6.8 Summary of absorbance test results .............................................................167
6.9 Summary of Superplasticized paste combinations at 0-minutes ..................173
xii
List of Tables (Contd.,) Page No.
6.10 Summary of Superplasticized paste combinations at 60-minutes ................174

7.1 Properties of viscosity modifying agents .....................................................179
7.2 Flow characteristics of pastes with different types and dosages of VMA ...191
7.3 Influence of time on SP-VMA mix ..............................................................192
7.4 Acceptance criteria for SCC ........................................................................195
7.5 Proportion of materials used in the SCC trials .............................................198
7.6 Test results of SCC with various SP-VMA combinations ...........................199
A.1 Test results of three trials in the selection of mixing method ......................214
A.2 Comparison of superplasticizer dosages (data from Marsh cone and viscometer
test results) ....................................................................................................214
C.1 Mix details for steady shear experiments .....................................................224
C.2 Work done in different loading-unloading cycles ........................................227
C.3 Shear thinning indices for different types and dosages of superplasticizers 229
C.4 Mix details for creep and recovery and stress relaxation experiments ........230
C.5 Details of the different data acquisition frequencies ....................................230
C.6 Superposition of separate responses ............................................................237
xiii
LIST OF FIGURES
Figure Title Page No
2.1 Microstructure of hydrated portland cement ..................................................16

2.2 Physical Action of Superplasticizers ..............................................................24
2.3 Chemical action of superplasticizers ..............................................................26
3.1 Geometry and dimensions (in millimetres) of the Marsh cone ......................54
3.2 Flow time curve showing the determination of optimum Superplasticizer
dosage (according to the method of Gomes et al., 2001) ...............................55
3.3 Marsh cone test showing continuous and discontinuous flow .......................55
3.4 Mini-slump cone ............................................................................................56
3.5 Mini-slump test ..............................................................................................56
3.6 Brookfield HA DV II +Pro viscometer ..........................................................58
3.7 Details of co-axial cylinder setup ..................................................................58
3.8 Shear history of rheological testing ...............................................................59
3.9 Typical graph showing hysteresis cycles for a superplasticized
cement Paste ...................................................................................................59
3.10 Fits of experimental data with Bingham and Herschel-Bulkley Models ......62
3.11(a) Hobart mixer with B-flat beater ...................................................................64
3.11(b) Hobart mixer with D-wire whip .....................................................................64
3.12 Test set up for ball milling .............................................................................65
3.13 Typical flow time curves obtained with Marsh cone for different mixing
methods .........................................................................................................67
3.14 Typical flow curves obtained with mini-slump for different mixing
methods ..........................................................................................................68
3.15 Marsh cone flow time for different superplasticizers ....................................71
3.16 Mini-slump spread for different superplasticizers .........................................72
3.17 Correlation between Marsh cone flow time and mini-slump spread .............73
3.18 Typical graphs showing hysteresis cycles for a PCE based superplasticized
cement paste ..................................................................................................75
3.19 Comparison of data from different tests ........................................................79
3.20 Prediction of yield stress from mini-slump spread using Roussel-Coussot
model ..............................................................................................................80
xiv
List of Figures (Contd.,) Page No.
3.21 Flow time curve showing the determination of optimum superplasticizer

dosage according to Aïtcin (1998) .................................................................84
4.1 Particle size distribution of fine aggregates (0-4.75mm) ...............................89
4.2 Particle size distribution of coarse aggregates (4.75-20 mm) ........................89
4.3 Flow table test for mortar ...............................................................................90
4.4 Tests for workability of concrete ...................................................................91
4.5 Marsh cone flow time curves for superplasticized cement paste ...................93
4.6 Marsh cone flow times of cement mortar ......................................................95
4.7 Flow table spreads for superplasticized mortars ............................................95
4.8 Variation of concrete slump with dosage of superplasticizer ........................97
4.9 Variation of flow table spread with dosage of superplasticizer .....................97
4.10 Correlation between paste, mortar and concrete in terms of flow time and
slump for SNF-S1 ..........................................................................................99
slump for SNF-D2 ..........................................................................................99
slump for SMF-S1 ........................................................................................100
slump for PCE-D1 ........................................................................................100
4.14 Correlation between slump of concrete, and the yield stress and plastic
viscosity of paste ..........................................................................................101
4.15 Loss of fluidity of paste ...............................................................................102
4.16 Loss in slump of concrete with time ............................................................106
4.17 Vicat penetration measurements for paste with different superplasticizers .107
4.18 Penetration test on concrete .........................................................................108
4.19 Setting of concrete at saturation dosage of superplasticizer ........................108
4.20 Comparative study on setting behaviour of paste with Vicat apparatus and
LCR meter ....................................................................................................111
4.21 Evolution of conductivity for different superplasticizers.............................114
4.22 Electrical conductivity graphs for different pastes along with beginning and
end of setting defined by Vicat penetration measurements .........................115
5.1 X-ray diffractograms of cement paste and superplasticized cement paste ..126
xv
5.2 BSE (250X) Micrograph of pure cement paste and superplasticized cement
paste .............................................................................................................133
5.3 BSE micrograph along with EDAX for a dark grey (C-S-H) particle in the LS
based paste ...................................................................................................134
5.4 DTA analysis of cement paste and superplasticized paste at 3 days............137
5.5 TG and DTG analysis of pastes at 3 days ....................................................139
5.6 TG and DTG analysis of pastes at 28 days ..................................................141
5.7 Details of Si NMR spectrum of different pastes for different ages .............145
5.8 Determination of area of peaks of Si NMR spectrum ..................................145
5.9 Comparison of area of the Q1 peak (A1) with reference to the area of the Q0
peak (A0) at different ages............................................................................146
6.1 Marsh cone flow times of different cement-superplasticizer combinations 154
6.2 Influence of type of cement on setting .........................................................162
6.3 Absorbance spectra for the different superplasticizers ................................165
6.4 Absorbance spectrum of the diluted aqueous solutions of superplasticized
pastes extracted after mixing ........................................................................166
6.5 Absorbance spectrum of cement C2 and LS-C1 ..........................................168
6.6 Methodology to select a compatible cement-superplasticizer combination 172
6.7 Classification of cement based on its chemical composition .......................176
7.1 Marsh cone flow time curves for pastes with metakaolin and PCE based
superplasticizer.............................................................................................180
7.2 Variation of mini-slump spread with dosage of PCE for metakaolin-PCE
combination ..................................................................................................181
7.3 Variation of yield stress with dosage of PCE for metakaolin-PCE
combination ..................................................................................................182
7.4 Variation of plastic viscosity with dosage of PCE for metakaolin-PCE
combination...................................................................................................182
7.5 Marsh cone flow time curves for pastes with metakaolin and SNF-D2 ......184
7.6 Variation of mini-slump spread for pastes with metakaolin and SNF-D2....184
7.7 Variation of yield stress for paste with metakaolin and SNF-D2 ................185
7.8 Variation of plastic viscosity for paste with metakaolin and SNF-D2.........185
7.9 Marsh cone flow time for different viscosity modifying agents ..................188
7.10 Mini slump spread for different viscosity modifying agents .......................189
xvi
7.11 Values of yield stress and plastic viscosity for different viscosity modifying
agents ..........................................................................................................189
7.12 Determination of saturation dosage of superplasticizer at different
dosages of VMA ..........................................................................................190
7.13 Effect of time on Marsh cone flow curves of SP-VMA mix .......................192
7.14 Determination of aggregate proportions for concrete ..................................194
7.15 Slump test .....................................................................................................196
7.16 J-Ring test ....................................................................................................197
7.17 V-Funnel test ................................................................................................197
A.1 Correlation between Marsh cone flow time and rheological parameters for
different compositions ..................................................................................216
A.2 Correlation between Marsh cone flow time and mini-slump spread with
rheological parameters for different compositions ......................................216
C.1 Anton Paar Rheometer with parallel plate attachment .................................224
C.2 Loading-unloading cycles for the superplasticizer, PCE-D1 at the dosage
of 0.1% .........................................................................................................225
C.3 Loading-unloading cycles for the superplasticizer, PCE-D1 at the dosage
of 1% ............................................................................................................226
C.4 Creep and recovery test for pure cement paste ............................................232
C.5 Creep and recovery test result of cement paste and with constant
dosage of different superplasticizers ............................................................233
C.6 Variation in creep and recovery behaviour with different dosages of
PCE-D1 ........................................................................................................234
C.7 Creep test result at saturation dosage of superplasticizers ...........................235
C.8 Test for linear scaling ...................................................................................236
C.9 Stress relaxation of pure cement paste .........................................................238
C.10 Stress relaxation test results for pastes with 0.05% dosage for different
superplasticizers ............................................................................................239
C.11 Stress relaxation at different dosages of PCE-D1 ........................................240
C.12 Stress relaxation of pastes at saturation dosage of superplasticizers ...........241
xvii
CHAPTER 1
INTRODUCTION
1.1 GENERAL
Recent advances in concrete technology have generally resulted in better performance
of concrete in terms of strength, workability and durability over the past few decades.
The enhancement of fresh and hardened state properties has been achieved especially
through the incorporation of chemical and mineral admixtures. The multifunctional
benefits of chemical admixtures, such as the high range water reducers (HRWRs) or
superplasticizers, have made them essential components of modern high quality
concrete. Today’s high performance concretes require superplasticizers in order to
reduce the water content while maintaining high workability, so that better
mechanical integrity and lower permeability are obtained when compared to
conventional concretes.
Though the primary purpose of adding superplasticizers is to provide high
workability, their incorporation in concrete could also affect, sometimes negatively,
the hardening and hardened state properties (Ramachandran, 2002; Mailvaganam,
2001). Moreover, phenomena such as loss in the workability of concrete, rapid/slow
setting, air entrainment, excessive bleeding and segregation could result from the
improper use of superplasticizers. Therefore, a good understanding of the cement-
superplasticizer interaction is required for better utilization of high performance
concrete, and for the further improvement of the properties of the superplasticizers
and concrete.
The performance of a superplasticizer in concrete depends on its properties, as well as
that of the cement, in addition to the concrete mix proportions and environmental
conditions (Jolicoeur and Simard, 1998). The major characteristics of cement that
could affect the performance of chemical admixtures are the C3A/SO3 ratio, alkali
content and nature of sulphates whereas the chemical structure and molecular weight
of polymers are the principal characteristics of superplasticizers (Rixom and
Mailvaganam, 1999; Ramachandran, 2002). The main action of the superplasticizers
is electrostatic repulsion and steric hindrance that result in the dispersion of the
cement particles (Ramachandran et al., 1998). Nevertheless, the four different families
- lignosulphonates, melamines, naphthalenes and polycarboxylates - each has different
distinguishing characteristics and effects (Rixom and Mailvaganam, 1999).
The proposed study focuses on the improvement of the flow behaviour and its
characterization, and the understanding of the microstructure of the superplasticized
cement-based materials. More specifically, the optimization of the superplasticizer
dosage, determination of the rheological behaviour, setting and evolution of the
hydration products are evaluated from tests of cement paste. The study also addresses
the issues of loss of workability and set retardation, which could arise as a result of
cement-superplasticizer incompatibility.
The relevance of the present work in the Indian context can be appreciated by
considering the following points. The types and number of superplasticizers available
in the market have increased significantly over the past few years. The usage of
products without adequate knowledge of their effects, along with the use of high
superplasticizer dosages, has led to significant problems during construction, some of
which can be attributed to cement-admixture incompatibility. The current knowledge
of materials and practices involved in the use of superplasticized concrete is mostly
based on foreign experience, which is not always directly applicable to construction in
India today. For example, the tendency to use synthetic gypsum instead of natural
2
gypsum during cement production results in the increase of superplasticizer dosage
and setting time. Also, ready-mix concrete in most construction sites is expected to be
placeable even 3-5 hours after mixing.
The final aim of the work is to identify the factors that affect the cement-
superplasticizer interaction and to propose remedial measures for the concrete user to
avoid incompatibilities through a materials science approach. Commercially-available
superplasticizers of different types, several cements, mineral admixtures and viscosity
modifying admixtures are used in the study in order to span the wide spectrum of
products used in the manufacture of high performance concrete.
1.2 OBJECTIVES AND SCOPE
The present thesis focuses on the study of the mechanisms that affect the cement-
superplasticizer interaction, which in turn influence the fresh and hardened state
properties of the cement based material. The main objectives of the research are:
• To identify a suitable mixing method for the preparation of pastes such that it
will represent the paste phase of concrete.
• To evaluate the flow behaviour of superplasticized cement paste using simple
methods, and correlating it with rheological characteristics for validating the
use of simple test methods.
• To analyze the correlation between the fluidity of cement paste, mortar and
concrete, in order to validate the optimization of the cement paste composition
as the first step in the mix design of high performance concrete.
• To study the influence of superplasticizer on the hydration processes and on
the microstructural development of cement paste.
3
• To identify the characteristics of cement and superplasticizer that can affect
the behaviour of the cement-superplasticizer system and propose guidelines to
avoid incompatibilities.
The scope of the study is limited to the following with respect to main components
and methods adopted:
• Four types of cements available in India and products from four families of
superplasticizers.
• Empirical correlation between the behaviour of cement paste and concrete is

made through experimental studies.
1.3 STRUCTURE OF THE THESIS
The research strategy that was followed in order to achieve the objectives specified in
the previous section and the results obtained are described in the following chapters of
the thesis.
A detailed review of the influence of superplasticizers on the hydration of cement is
given in Chapter 2, along with appropriate comparisons with the hydration of pure
cement. The characteristics, chemistry and mechanisms of action of superplasticizers
are also presented. The factors affecting the cement-superplasticizer interaction,
mainly, the chemical structure of the superplasticizer and the cement characteristics
are reviewed, and the need and motivation for conducting this study are emphasized
in the same Chapter.
Chapter 3 begins with the selection of a suitable procedure for the preparation of
superplasticized cement paste that simulates the paste phase of concrete. The
determination of the saturation dosage of superplasticizer with the Marsh cone is
explained in detail. The influence of the type and dosage of superplasticizer on the
4
flow behaviour of paste is evaluated through the Marsh cone, mini-slump and
viscometer tests. In order to validate the use of simple engineering-level test methods,
the Marsh cone flow time and mini-slump spread are shown to correlate well with
rheological parameters through appropriate analytical models. This chapter also
analyses the influence of saturation dosage of superplasticizer on the loss of fluidity of
different cement-superplasticizer combinations. Appendix A contains the results for
establishing the correlation between simple tests with rheological parameters. The
matlab programme for predicting the Marsh cone flow time using rheological model
is presented in Appendix B. The viscoelastic characterization of pure cement paste
and superplasticized cement paste, using creep and recovery and stress relaxation
tests, is given in Appendix C. The influence of type and dosage of superplasticizers on
the change in rheological response, from viscoelastic to viscous nature, is also
explained.
Chapter 4 deals with the correlation between the flow behaviour of superplasticized
cement paste, mortar and concrete. The setting behaviour of paste and the set
retardation due to the use of superplasticizers are studied with the Vicat apparatus and
electrical conductivity measurements. The responses of paste, mortar and concrete in
terms of loss of fluidity and setting are compared using appropriate tests. The
influence of superplasticizer dosage on the compressive strength of concrete and its
evolution with age has also been studied.
The study of the effect of the superplasticizer on the hydration processes, as well as
on the microstructural development, through several complementary physico-
chemical characterization techniques like X-ray diffraction, scanning electron
microscopy, thermal methods and nuclear magnetic resonance is presented in
Chapter 5. It is intended to provide an overall picture of the influence of different

5
families of superplasticizers on the types of hydrated products as well as their rate and
amount of formation.
Chapter 6 presents the evaluation of the compatibility between different cements and
superplasticizers through flow and setting tests. The variations of the dosages and loss
of fluidity for different superplasticizers is explained based on an absorbance study
using a UV spectrophotometer. Subsequently, a methodology based on the test results,
for selecting a compatible combination of cement and superplasticizer is
recommended. The studied combinations are evaluated with the recommended
methodology and finally the same is validated with other approaches for selecting the
compatible combination.
Chapter 7 deals with the influence of mineral admixtures and viscosity modifying
admixtures on the cement-superplasticizer interactions. A mix design procedure for
self compacting concrete incorporating VMA is suggested in line with previous
studies. The validation of the use of the saturation superplasticizer dosage obtained
from tests on pastes is illustrated with the characterization of self compacting concrete
through the slump flow, V-Funnel and J-Ring tests.
The conclusions drawn from the research work are summarized and the guidelines for
selecting the superplasticizers are presented in Chapter 8. Also, the scope for further
research is highlighted.

6
CHAPTER 2
STATE-OF THE-ART REPORT ON CEMENT-

SUPERPLASTICIZER INTERACTION
2.1 GENERAL
Admixtures, both mineral and chemical, have become essential components of high
quality concrete and are added to concrete at the mixing stage to modify its properties
in the fresh and hardened state. ACI 116R (2000) defines the admixture as ‘a material
other than water, aggregates, hydraulic cement and fiber reinforcement, used as an
ingredient of concrete or mortar, and added to the batch immediately before or during
its mixing’. Common mineral admixtures are silica fume, fly ash, ground granulated
blast furnace slag and metakaolin, the incorporation of which can improve strength, as
well the durability of the concrete. The incorporation of chemical admixtures can vary
significantly with the application; e.g., air entraining admixtures are very common in
colder regions where the concrete is prone to freeze-thaw damage, set accelerators are
essential in wet-mix shotcreting, and retarders are common when the ready-mix
concrete has to be transported over long distances. However, the superplasticizer or
water reducer is an admixture that has become an important component of all high
quality concretes; the addition of the superplasticizer improves the workability at low
water content, and consequently leading to good fluidity in the fresh state, reduction
in concrete porosity and compositional stability when the concrete is being
transported and placed. The use of superplasticizers/high range water reducers in
concrete has played a central role in the development of high strength and high
performance concretes, without which such concretes would not be technically
feasible.
7
However, the performance of an admixture in concrete is dependent on many factors,
like the nature and amount of admixture, nature of cement and aggregates, water-
cement ratio and environmental conditions. In addition, some problems could
commonly arise as a result of the incompatibility between cement and water reducers,
such as rapid loss of workability, rapid/retardation of setting and low rates of strength
gain. Sometimes, the incompatibility could even occur between a particular admixture
and only some batches of the same brand of cement (probably made at different
plants), suggesting that the problem is complex from the user's point of view.
Therefore, the understanding of the effects of superplasticizers on the properties of
concrete is essential for further improvement of the properties and behaviour of both
the superplasticizer and concrete.
The objective of this chapter is to summarise the various factors responsible for the
physico-chemical reactions in order to have a better understanding of cement-
superplasticizer interaction. Some relevant aspects of cement composition and cement
hydration reactions are also described in order to understand the action of the
superplasticizer in a cementitious system. A detailed review of the different families
of superplasticizers, mechanisms of action and their characteristics that could control
their compatibility with cement is also given.
2.2 CEMENT CHEMISTRY
2.2.1 Raw Materials Used in Cement Manufacture
The raw materials used for the production of cement include calcareous materials and
argillaceous materials. CaO is obtained from limestone, principally calcite; the lime
can be either quick lime, dolomite lime or magnesium lime. In quick lime, the major
part is CaO or MgO. Dolomite lime indicates that there is 35-45% magnesium
8
carbonate while the magnesium lime indicates the presence of 5-35% of magnesium
carbonate in the limestone. Silica (SiO2) exists naturally in the fine state as different
crystalline polymorphs (e.g., α-quartz, cristobalite, tridymite) and as impure poorly
crystallized or amorphous minerals (e.g., opal and flint). It is introduced through the
aluminosilicate minerals of the shale or clay component of kiln feed and combine with
CaO during clinkering process to form the impure calcium silicates, alite and belite.
The other two significant oxides, namely aluminium oxide that occurs in nature as α-
Al2O3 (Corundum) and iron oxide are derived from clay and shale (Hewlett, 2004).
Gypsum and anhydrite (i.e., anhydrous calcium sulphate) are found naturally in many
parts of the World; in Asia, India has the largest reserves of natural gypsum, 95% of
which is found in Rajasthan (Shah, 2006). Natural gypsum used for cement
production in India could be from marine or desert sources. Calcium sulphate is also
available as a byproduct of the acid based manufacture of fertilizers from phosphate
ore. However, when this is used in cement production, it causes an increase in setting
time due to the presence of water soluble fluorine, water soluble phosphates and
phosphates in the crystal lattice of gypsum (Hewlett, 2004). Nevertheless, such
synthetic gypsum of good quality is readily available and, consequently, there is a
growing in tendency to use it instead of natural gypsum in the Indian cement
manufacturing industry.
2.2.2 Cement Composition
Portland cements are multicomponent multiphase inorganic materials and consist of
five major constituents in varying amounts. In the early ages of a cement-based
material, the two calcium silicates – tricalcium silicate (C3S) and dicalcium silicate
(C2S) control the strength and its evolution, and the two aluminates - tricalcium
9
aluminate (C3A) and tetra calcium alumino ferrite (C4AF) and the sulphate phases
(i.e., gypsum) influence the setting behaviour (Hewlett, 2004). In addition to these
main compounds, there are minor compounds, such as MgO, TiO2, Mn2O3, K2O and
Na2O. Two of the minor compounds, i.e., sodium and potassium oxides, known as
alkalis, play an important role in the cementitious system in terms of the rheological
behaviour, and cement-superplasticizer interactions.
Standards such as those of ASTM, BIS and CEN give the acceptable ranges of
components for different classes of cements. For example, the ASTM C 150 (2007)
standard, limits the tricalcium silicates to 35%, tricalcium aluminates to 7%, and tetra
calcium alumina ferrite to 25%, and specifies a minimum dicalcium silicate content of
40%, along with a maximum alkali content of 0.6% (Na2O)eq (Hewlett, 2004); a
typical cement corresponding to this standard could have the composition as shown in
Table 2.1.
Table 2.1 Typical portland cement composition

conforming to ASTM C 150
Type of compound Composition (%)

C3S and C2S 70-80
C3A 0-16
C4AF 1-17
MgO 0.5 -6
Na2SO4, K2 SO4 0.5-3
Free lime 0.2-4
Gypsum 4
Low alkali cement < 0.6 (Na2O)eq
The chemical and physical requirements of different grades of cement as per the IS
codes are given in Table 2.2. It should be noted that the IS codes do not specify the
limits for alkalis (as done by ASTM C 150 (2007))which lead to significantly
10
different rheological behaviour especially if the alkali content varies considerably.
Similarly, the code does not prescribe upper limits of the alumina ratio and specific
surface area, both of which can affect the performance of the cement in the fresh and
setting stages. It can be seen from the table that the only differences between the different
grades of cement are in the rate of compressive strength gain and the 28-day value.
Table 2.2 Requirements of different grades of cement as per IS codes.
Requirements
33 grade (as per 43 grade (as per 53 grade (as per
Characteristics IS 269-1989) IS 8112-1989) IS 12269-1987)
Lime saturation Not greater than Not greater than Not greater than
factor * 1.02 and not less 1.02 and not less 1.02 and not less
than 0.66 than 0.66 than 0.8
Alumina ratio** Not less than 0.66 Not less than 0.66 Not less than 0.66
Insoluble residue Not more than Not more than 2% Not more than 2%
by mass 4%
Magnesia, % by Not more than Not more than 6% Not more than 6%
mass 6%
Total sulphur Not more than 2.5 Not more than 2.5 Not more than 2.5
content calculated and 3.0 when C3A and 3.0 when C3A and 3.0 when C3A
as sulphuric % by mass is 5 or % by mass is 5 or % by mass is 5 or
anhydride (SO3 ), less and greater less and greater less and greater
% by mass than 5, than 5, than 5,
respectively respectively respectively
Total loss on Not more than Not more than 5% Not more than 4%
ignition 5%
Fineness Specific surface Specific surface Specific surface
shall not be less shall not be less shall not be less
than 225 m2/kg than 225 m2/kg than 225 m2/kg
Initial setting time Not less than 30 Not less than 30 Not less than 30
minutes minutes minutes
Final setting time Not more than Not more than 600 Not more than
600 minutes minutes 600 minutes
Compressive For, 72 ± 1h, not For, 72 ± 1h, not For, 72 ± 1h, not
strength less than 16 MPa less than 23 MPa less than 27 MPa
For, 168 ± 2h, not For, 168 ± 2h, not For, 168 ± 2h, not
less than 22 MPa less than 33 MPa less than 37 MPa
For, 672 ± 4h, not For, 672 ± 4h, not For, 672 ± 4h, not
less than 33 MPa less than 43 MPa less than 53 MPa
*Lime saturation factor = Ratio of percentage of lime to percentage of silica,
alumina and iron oxide represented as (CaO - 0.7 SO3)/ (2.8 SiO2 + 1.2 Al2O3 +
0.65 Fe2O3).
**Alumina ratio = Ratio of percentage of alumina to iron oxide.
11
Though cements are classified by national or international standards, as seen earlier,
the chemical composition of a type of cement can vary within certain limits between
brands and even batches, depending on the raw materials and production processes.
For example, in Table 2.3 (from Hewlett, 2004), which gives the chemical analyses of
six different CEM I portland cements available in the UK in 1987, it can be seen that
the compound composition, fineness and particle size distribution vary significantly
from one cement to another, though they all belong to the same “standard” type.
Consequently, different cements of the same type/grade can have significantly
different characteristics that can lead to differences in the interactions between the
cement and admixture.
Table 2.3 Chemical analyses of some CEM I portland cements of UK (Hewlett, 2004)
Type of DO DOX EO FO GO HO
compound (%)
Si O2 22.2 21.8 20.9 20.1 19.5 20.2
Al2 O3 4.2 4.1 4.6 5.0 6.2 5.3
Fe2 O3 2.4 3.3 2.1 2.8 2.0 2.2
Mn2 O3 0.08 0.07 0.06 0.16 0.03 0.04
P 2 O5 0.15 0.1 0.23 0.1 0.17 0.07
Ti O2 0.27 0.23 0.2 0.2 0.31 0.23
Ca O 63.9 64.2 66.3 63.2 64.7 66.4
Mg O 1.0 1.4 0.8 2.2 1.2 0.9
SO3 2.7 2.5 2.8 3.0 3.0 2.5
CO2 0.79 0.3 0.48 ----- ---- ----
K2 O 0.74 0.61 0.8 0.78 0.43 0.56
Na2 O 0.14 0.15 0.16 0.13 0.24 0.23
Free lime 1.5 1.4 1.4 2.1 0.9 1.2
Loss on ignition 1.7 1.5 0.75 1.0 1.1 0.7
Insoluble residue 1.7 0.8 1.1 0.84 0.85 0.5
Na2O equivalent 0.63 0.55 0.69 0.64 0.52 0.6
C3 S 46 51 63 50 58 66
C2 S 29 24 12 20 12 8
C3 A 7.1 5.3 8.6 8.5 13 10.3
C4AF 7.3 10 6.4 8.5 6.1 6.7
Density(kg/m3) 3130 3130 - 3150 3110 3140
Surface area 383 380 - 443 394 344
(m2/kg)
Particles coarser 13.4 12.5 - 9.0 12.3 17.4
than 45µm
12
2.2.3 Hydration of Cement
2.2.3.1 Hydration of Cement Compounds
Hydration of cement is the key to concrete performance in terms of setting, durability
and strength. It includes the hydration of calcium silicates, calcium aluminates and
calcium aluminoferrite. When water is added to cement, each of the compounds
undergoes reactions but with different characteristic contributions to the setting,
strength gain and other properties of the hardened cement paste; e.g., tricalcium
silicate hydration affects the early age response while the hydration of dicalcium
silicate affects the strength at later ages. The relevant aspects of the hydration of the
various components of cement are explained in the following paragraphs. It should be
noted that, by volume, calcium silicate hydrate (C-S-H) constitutes 60-70% of the
hydration products, calcium hydroxide constitutes 20-25% and 5-15% is made up of
other minor phases.
Upon the addition of water, tricalcium silicate (C3S) rapidly reacts to release calcium
ions, hydroxide ions and a large amount of heat. The pH quickly raises to over 12
because of reaction alkaline hydroxide (OH-) ions. This initial hydrolysis slows down
quickly resulting in a decrease in the heat evolved. The reaction continues slowly,
producing calcium and hydroxide ions until the system saturates. Once this occurs,
Ca(OH)2 starts to crystallize. Simultaneously, calcium silicate hydrate begins to form.
Subsequently, calcium and hydroxide ions precipitate out of solution accelerating the
hydration of tricalcium silicate. The C-S-H crystals grow thicker making it more
difficult for water molecules to reach the unhydrated tricalcium silicate.
Consequently, the speed of reaction is controlled by the rate at which water molecules
diffuse through the C-S-H coating on the surface of the unhydrated cement particles.
As the coating thickness increases, the production of C-S-H becomes slower and
13
slower (Jolicoeur and Simard, 1998). The overall chemical reaction and the heat
evolved (∆H) are given in the following equation:
2C3S+6H => C3S2H3+ 3 CH (∆H = - 520 kJ/kg) (2.1)
Dicalcium silicate reacts with water in a similar manner as C3S but much more
slowly, according to the reaction and heat of hydration shown below:
2C2S+4H => C3S2H3+ CH (∆H = - 260 kJ/kg) (2.2)
The hydration of tricalcium aluminate (C3A) in the presence of gypsum occurs as
follows:
2 C3A+ H+ CS => C6AS3 H 32+ 3CH (∆H = - 910 kJ/kg) (2.3)
Calcium aluminate hydrate or ettringite (AFt) needles cover the surface of hydrating
calcium aluminate particle and prevent further hydration. If the supply of sulphate
from gypsum is exhausted before C3A is completely hydrated, a second reaction can
occur resulting in the formation of calcium monosulphoaluminate (AFm). This
reaction may occur before the formation of the ettringite if the availability of gypsum
is lower than that required by the rate of reaction of C3A and sulphate ions. If
monosulphoaluminate is exposed subsequently to sulphate ions, new reactions will
occur leading to the formation of more ettringite.
Calcium aluminoferrite forms the same hydration products as C3A, with a slow rate of
reaction that is decreased further by gypsum and higher iron oxide content. The
reaction is as given by the following equation:
C4AF + CS H2 +12H => C4 (A,F) CSH12 (∆H = - 420 kJ/kg) (2.4)
The relative reactivity of different mineral phases with water can be given as C3A>
C3S> C2S> C4AF; accordingly, the aluminate phases and their hydration products play
an important role in the early hydration processes and cement-admixture interactions.
14
2.2.3.2 Hydration of the Portland Cement Grain
The development of the microstructure of a typical polymineralic cement grain is
shown in Figure 2.1, and the major unhydrated compounds present are shown in
Figure 2.1 (a). After the first contact of cement with water, various reactions occur
that result in a typical heat evolution pattern with time, through which four stages can
be identified in the early hydration process: initial hydration period, dormant period,
acceleration and deceleration period. The basic mechanisms mostly influenced by the
presence of admixtures, particularly those occurring in the early hydration processes
are described below (Mehta and Monteiro, 2005; Jolicoeur and Simard, 1998; Taylor,
1997; Neville, 2000; Hewlett, 2004):
Initial hydration (0-15 minutes): During this period, rapid heat generation, called as
heat of wetting, occurs on mixing cement with water. Easily soluble components like
alkalis, calcium, sulphate phases and free lime are dissolved in the surrounding water
and the solubilization of variety of ionic species like Na+, K+, Ca2+, SO42-, and OH-
ions takes place. Hence, the wetting of cement particles and subsequent dissolution of
ions lead to the formation of an amorphous gel layer over the surface of grains.
Beyond the solubilization process, the formation of solid hydration product is
governed by the nucleation process. The initial nucleation process involves the
formation of calcium sulphoaluminate or ettringite from Ca2+, SO42-and Al(OH)4- ions
on the outer surface of layer of gel, as shown in Figure 2.1 (b). This reaction is the
most important as far as the rheology of the fresh cement paste is concerned. The
initial C3A dissolution may be increased by the presence of more dissolved alkalis
(Spiratos et al., 2003). After some time, the cement grains are coated with a protective
layer of hydration products. At this stage, the rate of reactions sharply reduces
resulting in less heat of hydration.
15
The presence of organic admixtures can interfere with these nucleation and growth
processes resulting in the change of the hydration reactions and hydration products.
Also, after the first few minutes, the chemical admixtures interact with these
hydration products instead of hydrating compounds.
(a) (b) (c) (d) (e) (f)
Fig. 2.1 Microstructure of hydrated portland cement (Taylor, 1997)
Dormant period (2-4 hours): During this period, the mix remains plastic and very
low heat generation takes place, indicating that all the reactions are slow. This is due
to two reasons: (1) the gel formed in the first stage acts as a diffusion barrier around
the C3S; (2) the SiO2 rich electrical double layer formed around C3S and the
supersaturation of Ca2+, OH- and SO42- in solution prevents further dissolution of ions.
As the thickness increases, the time it takes for water to penetrate the coating
increases; thus the reaction becomes diffusion controlled.
In the presence of an adequate concentration of SO42- ions, continued growth of
ettringite crystals takes place. Simultaneously, the reaction of C3S produces ‘outer
product’ C-S-H on the AFt, leaving space between the grain and hydrated shell, as
shown in Figure 2.1 (c). An inadequate concentration of SO42- ions leads to the
16
formation of calcium aluminate hydrate (C-A-H) resulting in flash set. If the SO42-
concentration is too high due to the presence of hemihydrates and alkali sulphates,
massive nucleation and growth of gypsum crystals (false set) occurs.
The interaction of a chemical admixture with any of these reactions or reactive species
significantly affects the behaviour of cement pastes and concretes during the induction
period.
Acceleration period (4-8 hrs): Several theories have been put forward to explain the
end of the dormant period or the start of the acceleration period: Weakening of the
double layer by further hydration, increased diffusion and weakening of ionic strength
around the hydrating particle all result in the end of the dormant period and the
beginning of a new stage of high chemical activity, called the acceleration period.
However, during this period, the suspension loses its plasticity and is converted to a
stiffer matrix, which is no longer fluid. Subsequently, the secondary hydration of C3A
produces long rods of AFt and the intense hydration of C3S during the acceleration
period results in the formation of ‘inner product’ C-S-H and precipitation of C-H, as
shown in Figure 2.1 (d). Later, C2S starts to hydrate and, to a lesser extent, C4AF
continues to hydrate. During the acceleration period, the calcium and sulphate ion
concentrations in the pore water decrease due to the ettringite formation.
The chemical admixtures may influence the formation and properties of the protective
hydrate layer in the acceleration stage. Also, the admixtures remaining in the aqueous
solution further influence the nucleation and growth of the hydration products.
Deceleration period (8-24 hours): The reactions are slow and completely diffusion
dependent in this period, which is reflected by the hardening of the cement paste. Due
17
to the depletion of sulphate ions in the pore water, C3A reacts with the AFt forming
hexagonal plates of monosulphate (AFm), as shown in Figure 2.1 (e). The pore
volume decreases with increasing time and decreasing free water content as hydration
progresses with the formation of C-S-H gel and C-H. This is followed by a steady
state (12-24 hours) in which temperature has little influence on hydration.
Later hydration results in the formation of sufficient ‘inner’ C-S-H to fill the space
between grain and shell, as shown in Figure 2.1 (f). The outer C-S-H becomes more
fibrous during this period and the ettringite is completely converted to
monosulphoaluminate within 1 to 3 days.
2.3. WATER REDUCERS AND SUPERPLASTICIZERS
Water reducers are classified as normal, medium and high range depending on the
level of water reduction possible due to its addition in concrete. Water reducers
belong to the dispersant family; dispersants are long chain organic molecules having a
polar hydrophilic end and non-polar hydrophobic groups that are adsorbed on the
cement particles. Superplasticizers are high range water reducers and are incorporated
in concrete either to maintain the same workability at reduced water cement ratio or to
increase the workability by maintaining constant w/c. These admixtures are synthetic
high molecular weight water-soluble polymers that are effective in dispersing the
flocculated cement grains. Solubility is achieved by the presence of adequate
hydroxyl, sulphonate or carboxylate groups attached to the main organic unit which is
usually anionic (Rixom and Mailvaganam, 1999; Ramachandran, 2002; Bentur,
2002).
After the introduction of the first generation of superplasticizers in 1960s, these
products have become essential for the placement of concrete in areas with low
18
accessibility or with high density of reinforcement, as well as to provide an increase in
the pumpability of concrete. In general, however, the reduction of w/c that can be
obtained with the use of superplasticizers leads to higher strength and durability,
making them common components of high performance and high quality concretes.
Also, newly developed concretes such as self compacting concrete require the
presence of a superplasticizer to achieve the desired properties.
2.3.1 Characteristics and Chemistry of Superplasticizers
Superplasticizers belong to four different families: modified lignosulphonates (LS),
sulphonated naphthalene formaldehyde (SNF), sulphonated melamine formaldehyde
condensates (SMF), and the new generation of superplasticizers, which is made up of
copolymers that include polycarboxylates (PCE), polyphosphonates, polyacrylates
and monovinyl alcohols (Rixom and Mailvaganam, 1999; Ramachandran, 2002). The
different families of superplasticizers are generally based on chemicals of the type
presented in Table 2.4.
The lignosulphonate based admixtures are normal water reducers and at higher
dosages of such admixtures, excessive retardation of set and entrainment of air occurs,
which limits their dosage in practice (Ramachandran, 2002). Nevertheless, modified
lignosulphonates, in which the sugar content is reduced, exhibit less retardation
(Aïtcin, 1998). At higher molecular weights, they can be considered as
superplasticizers and used for obtaining workable concrete, with some retardation of
initial set (Ramachandran, 2002). It has also been found that sodium lignosulphonate
is a better plasticizer than calcium lignosulphonate because the former has higher
solubility.
19
Table 2.4 Chemical structure of superplasticizers (Rixom and Mailvaganam, 1999)
Class Origin Structure (typical repeated unit)

Lignosulphonates Derived from
(LS) neutralization,
precipitation, and CH2OH
fermentation OH
H O CH CH
processes of the
waste liquor SO3M n
O
obtained during Me
production of Me =CH3 , M =Na
paper-making
pulp from wood
Sulphonated Manufactured by
melamine normal H H
formaldehyde resinification of
HO CH N N N CH2 H
(SMF) melamine -
formaldehyde N N
n
HNCH2SO3M
M = Na
Sulphonated Produced from

naphthalene naphthalene by
formaldehyde oleum or SO3 R R
(SNF) sulphonation;
subsequent CH2 H
reaction with
formaldehyde n
leads to
polymerization SO3M SO3M
R =H,CH3,
CH
Polycarboxylic Free radical
ether (PCE) mechanism using
peroxide initiators CH2 CH CH2 CH2
is used for
C=O C=O
polymerization
process in these OCH3 OCH2CH2(EO)12CH2CH2
n
systems
EO: Ethylene oxide
The first admixtures to be classified as superplasticizers were salts of SNF and SMF.
All of these have the same functional groups, namely sulphonates. The use of SNF
and SMF based superplasticizers resulted in significant water reduction at low
20
dosages. The SNF based superplasticizers appear to yield the best fluidizing effect in
the molecular weight range of 4000-40000 gm/mole, while at higher ranges the effect
decreases (Ramachandran, 2002). SMF based superplasticizers generally have poorer
slump retention characteristics, when compared with the SNF based products, and are,
consequently, used more extensively in the pre-cast industry.
In the last few years, several polymeric surfactants with carboxyl, hydroxyl, and
phosphonate functional groups have been introduced in the market as third generation
superplasticizers, and are generically known as polycarboxylates. These are comb
polymers and are characterized by long chains and side chains, in which the
hydrophilic groups are in contact with water and hydrophobic groups are attached to
cement particles (Tanaka et al., 1999; Sakai et al., 2003). The dispersion due to these
superplasticizers depends on the length of the main chain, and the length and number
of the side chains. In general, the polycarboxylates are sufficient in low dosages due
to better dispersing action than first and second generation superplasticizers.
Moreover, it appears that they can be tailored to fulfill the requirements of the
application (Magarotto et al., 2003; Houst et al., 2005; Falikman et al., 2005). The
mechanisms of action of superplasticizers are explained with more detail in the
following section.
2.3.2 Mechanisms of Action of Superplasticizers
The rheology of cement paste is governed by van der Waal's attractive forces between
the cement particles and the electrostatic repulsion due to the surface charges on the
cement particles. However, as the former force is larger than the latter, cement
particles tend to eventually flocculate. The action of the superplasticizer is to prevent
or delay the flocculation and to disperse the cement particles within the paste. The
21
different interactions between the cement and superplasticizer, which are often
classified as physical or chemical interactions, are described in detail in the following
paragraphs.
Flatt and Houst (2001) proposed that the behaviour of the superplasticizer in the
cement paste is made up of three components, which was later supported by Banfill
(2003) and Yamada et al. (2006). The first part of the behaviour is the absorption of
superplasticizer by intercalation (i.e., the incorporation of molecules of
superplasticizer into the precipitates formed as the cement hydrates), co-precipitation
(i.e., precipitation of tiny solid nuclei of the superplasticizer along with the
precipitation of the cement hydrates) or micellization (i.e., groups of polymer
molecules are trapped within the hydrating cement making them unavailable for
active dispersion), leading to the formation of an organo-mineral phase (Banfill,
2003). These phenomena that occur during the formation of ettringite and C-S-H
decrease the amount of admixture available for dispersing the flocs of cement later on.
Consequently, for the same superplasticizer dosage, a cement-admixture combination
with a high degree of absorption will yield less fluidity than another system with a
lower degree of absorption. Nevertheless, the consumption of the superplasticizer by
absorption is generally small, especially when the superplasticizer addition is delayed
(i.e., the superplasticizer is added at least one minute after the addition of water to the
cement) (Banfill, 2003).
The second part of the superplasticizer is adsorbed on the cement particles, resulting
in their dispersal due to the formation of an electrical double layer and consequent
electrostatic repulsion between the particles. The chemical nature of the cement
compounds that form the solid particle makes them undergo ionisation on the surfaces
22
that are in contact with water. The electrical double layer (Uchikawa et al., 1997;
Young, 2008) appears as a consequence of the adsorption of anionic molecules of
superplasticizers on the cationic cement particles. The organic superplasticizer
molecules, having charged groups (SO3-, COO-) interact with the particle surfaces,
imparting negative charges to the mineralogical phases of cement (see Figure 2.2a),
which results in the electrostatic repulsive forces (Aïtcin, 1998; Rixom and
Mailvaganam, 1999; Yoshioka et al., 2002). Mollah et al. (1995) extended the double
layer theory by considering that the cement particle is covered by a layer formed by
Ca2+ ions resulting in a trilayer of a “diffuse ion swarm”, which also delays the
hydration reactions. Due to the dispersive action of the superplasticizer, water that
would be otherwise entrapped within the flocs is released for providing fluidity to the
paste. In other words, the presence of the superplasticizer reduces the amount of water
needed for the effective dispersion of the cement particles. This type of action is
observed in the cases of lignosulphonates, melamines and naphthalenes. The
electrostatic repulsion generated depends on the composition of the solution phase and
the adsorbed amount of the superplasticizer (i.e., greater the adsorption, better the
repulsion and higher the fluidity) (Nakajima and Yamada, 2004). As the electrostatic
potential diminishes, flocculation of cement particles occurs and the cement paste
eventually loses fluidity.
Another type of physical interaction, also based on adsorption, that is observed in the
case of polycarboxylates, in addition to electrostatic repulsive forces, is that the side
chain polymer molecules cause steric hindrance between the cement particles,
preventing flocculation and resulting in their dispersion (Rixom and Mailvaganam,
1999; Corradi et al., 2005; Collepardi, 2005) (see Figure 2.2b). The steric repulsion
depends on the length of main chain, and the length and number of side chains
23
(Yamada et al., 2001; Sugiyama et al., 2003). This action last for a longer time than
the electrostatic repulsion and hence leads to better performance.
Cement particle Cement particle
(a) Electrostatic repulsion
Cement particle Cement particle
(b) Steric hindrance

Fig. 2.2 Physical action of superplasticizers (Aїtcin, 1998)
When electrostatic repulsion is the main mechanism of dispersion, the zeta potential
represents a measure of the electrostatic interaction between the cement particles. The
superplasticizer, being anionic, leads to a higher negative zeta potential that results in
strong superplasticizer adsorption (Krishna, 1993, 1996; Houst et al., 1999a). It
should be noted that the different phases of cement have differences in zeta potential
(Yoshioka et al., 2002) that result in preferential adsorption of the superplasticizers on
some phases such as C3A. Similarly, the zeta potentials of early hydration products
can differ and influence the adsorption of the superplasticizer (Plank and
Hirsch, 2007). For example, the zeta potential of ettringite is more positive than that
of calcium monosulphate (Plank and Hirsch, 2007), which causes more
24
superplasticizer to be adsorbed on the former. Nevertheless, the maximum amount of
superplasticizer adsorbed, in addition to the zeta potential, is also dependent on the
type and morphology of the phase or hydration product. Accordingly, Jolicoeur and
Simard (1998) and Ramachandran (2002) have determined that the amount of
superplasticizer adsorbed on the monosulphate is higher than on the ettringite (though
the zeta potential of the latter is higher), probably due to the higher surface area of the
former.
The scientific literature on superplasticizer actions generally associates the fluidity of
the paste with the amount of superplasticizer adsorbed by the cement grains. It has
been stated that optimal fluidity of the paste occurs when the maximum adsorption is
reached on the cement surfaces (Houst et al., 1999a; Roncero, 2000; Griesser, 2002).
As mentioned earlier, the maximum adsorption depends on the chemical composition
of the cement, as well as the chemical structure of the superplasticizer. It should be
noted that when a higher amount of superplasticizer is adsorbed, there is better initial
fluidity but the fluidity may not be maintained sufficiently over time (Nkinamubanzi
et al., 2000; Li et al., 2003; Corradi et al., 2005). In such cases, the dosage of the
superplasticizer has to be increased to provide an additional amount for maintaining
the fluidity.
In addition to the physical effects of the adsorption, there are chemical reactions
between the adsorbed superplasticizer and cement grains that occur mainly at the most
reactive sites of cement particle surface (Jolicoeur and Simard, 1998). For example,
polynaphthalene sulphonate reacts with C3A in competition with SO42- ions. Jolicoeur
and Simard (1998) further points out that many organic admixtures solubilize the
ionic species, through association or complexation, which delays the precipitation of
25
the hydrated products. Also, the presence of organic molecules in the solid solution
inhibits crystal nucleation, and causes slower growth of the hydration products and a
change in their morphology. This is attributed to the blocking of the pores or gaps in
the initial protective hydrate or the inhibition of the more reactive sites by the low
molecular weight superplasticizer molecules, as shown in Figure 2.3.
High molecular weight Low molecular weight
Surface sites having affinities for RSO4 –-- or SO4--- ions
Fig.2.3 Chemical action of superplasticizers (Aïtcin, 1998)
Furthermore, Gaidis and Gartner (1991) suggest that the strong adsorption of the
superplasticizer on the cement grain affects the surface charge that exerts an
immediate deactivating effect on the C3A, which helps in fluidifying the cement paste.
However, after the admixture is completely removed from the solution, the C3A
resumes a more normal hydration.
The third part of the superplasticizer is non-adsorbed species that may be adsorbed
subsequently maintaining the dispersion of cement particles leading to flow retention.
The unadsorbed molecules also lead to additional mechanisms of action of the
superplasticizer including the dispersion of cement particles by reducing the surface
tension of the mixing water, the decrease in frictional resistance owing to the lining-
up of linear polymers along the flow direction, and the lubrication produced by the
26
low molecular weight polymers between the particles (Uchikawa et al., 1995; Tanaka
et al., 1999).
Furthermore, some additional mechanisms supposedly enhance the performance of
PCEs in terms of flowability and segregation resistance, as discussed by Tanaka et al.
(1999). The ingress of non-adsorbed low molecular weight superplasticizer molecules
between the cement grains causes additional dispersion due to volume repulsion
(depletion effect) whereas the higher molecular weight polymers that cannot ingress
between the cement particles coagulate (depletion coagulating effect) to improve the
segregation resistance. Even the higher molecular weight fractions may not be
completely adsorbed on the cement surface beyond a certain critical concentration
(called as the critical micelle concentration for surfactants; Porter, 1994) giving rise to
the effects given above.
The formation of air bubbles in the cement paste due to the incorporation of the
superplasticizer can also help in its fluidification but also could reduce the strength
and durability if the air content is high. The different types of admixtures have
differing influences on air entrainment in concrete (Mailvaganam, 1999; Rixom and
Mailvaganam, 1999; Ramachandran, 2002). Lignosulphonates incorporate significant
amounts of air, which limits their dosage in concrete, in addition to other reasons
already discussed earlier. The SNF and polycarboxylates can also incorporate air in
the concrete due to which deairing agents are sometimes added to the commercial
product. The SMF based superplasticizer, on the other hand, does not have any
significant air entraining effect. Nevertheless, some of the air entrained due to the
admixture incorporation is released when the concrete is vibrated, especially when it
is highly workable.
27
2.3.3 Factors Affecting the Cement-Superplasticizer Interaction
Though there is significant knowledge about the mechanisms of action of
superplasticizers, there still exist gaps in the comprehension of why occasionally these
chemicals do not work as intended. This is generally termed as incompatibility
between the cement and superplasticizer; the term incompatibility refers to the
adverse effect on the performance when a specific combination of cement and
superplasticizer is used. Common problems of incompatibility include rapid setting,
delayed setting, rapid slump loss, improper strength gain, etc. These problems can in
turn affect the productivity in the construction site and the hardened properties of
concrete, primarily strength and durability.
The cement-superplasticizer interactions have many dimensions. On the one hand,
there is the influence of the type of superplasticizer, and at the other end of the
spectrum are the effects of the composition of cement, particularly the relative
proportions of C3A, alkalis and C3S in the cement. In addition, the type of gypsum
incorporated in the cement (gypsum, hemihydrate or anhydrite) has an important role
to play. The fineness of cement could also affect its compatibility with a particular
admixture. It is, therefore, necessary to understand the factors governing the
interactions and the reasons for incompatibilities. This is all the more important since
the type and number of superplasticizers available in the market has increased, along
with the use of high superplasticizer dosages. The various factors affect cement-
superplasticizer interactions are discussed in further detail in the following sections.
2.3.3.1 Effect of Chemical Structure of Superplasticizer
The type and dosage, degree of polymerization, degree of sulphonation, the position
of functional group in the benzene ring, the molecular weight distribution of the
28
polymer, the addition rate and the time of addition of superplasticizer influences the
interaction between cement and sulphonate based superplasticizers, namely, SMF,
SNF and LS (Aїtcin et al., 2001).
In the case of the lignosulphonates (Mollah et al., 1995; Rixom and Mailvaganam,
1999), the presence of low molecular weight constituents is known to cause excessive
air entrainment leading to loss of strength. In addition, the high sugar content of these
admixtures could cause unnecessary retardation, especially at high dosages. A further
unpredictability might arise depending on whether the lignosulphonate is a sodium or
a calcium salt; Ramachandran (2002) states that the sodium lignosulphonate as a
better plasticizer than calcium lignosulphonate. This difference in effectiveness may
be due to the presence of alkalies or due to the higher solubility of former. In order to
be more effective, lignosulphonates are modified - the sugars are removed by
fermentation, and low molecular weight matter is removed by centrifuging - and used
as superplasticizers. Lignosulphonate admixtures easily produce a complex salt with
Ca2+, thus decreasing the Ca2+ concentration in the liquid phase, resulting in a delay in
the hydration of alite and causing set retardation.
The effect of superplasticizer on the fluidity is related with adsorption and zeta
potential, as mentioned earlier, which in turn depends on the molecular weight of
polymer. There is no consensus, however, on the exact relations among the molecular
weight, adsorption and effectiveness in terms of dispersion. Andersen et al. (1987)
suggested that superplasticizer with smallest molecular weight is the most adsorbed
whereas the polymer that is least adsorbed gives the highest negative zeta potential.
Later, Basile et al. (1989) reported that the effectiveness of SNF in increasing the
fluidity is increased when the monomer content is reduced and molecular weight is
29
increased. They also found that the polymers with higher molecular weight caused
more retardation of the cement hydration at early ages. Similarly, Bonen and Sarkar
(1995) suggested that SNF with higher molecular weight led to more fluidity and
retardation in setting. Rixom and Mailvaganam (1999) have clarified that for SNF, the
influence of its molecular weight on the rheology, as well as the hydration of the
cement paste, depends on the alkali content of the cement. In the case of high alkali
cement, SNF superplasticizers with higher molecular weight (e.g., 16000 g/mol) yield
better fluidity in the cement paste due to the higher surface charge provided by the
larger molecules leading to higher adsorption (Kim et al., 1999; Kim and Aїtcin,
2003; Page et al., 1999). For low alkali cement, an SNF with high molecular weight
does not retard the setting significantly but the incorporation of a lower molecular
weight results in a considerable increase in the induction period, possibly due to lower
adsorption of the latter (Rixom and Mailvaganam, 1999).
Another factor affecting SNF effectiveness is the location of the sulphonate (-HSO3)
group in the naphthalene structure. It is well accepted that the presence of the
sulphonate group in the β-position leads to a high polymer charge and better
electrostatic repulsion (Aïtcin, 1998). So, the degree of sulphonation is an important
parameter for the effectiveness of the SNF based superplasticizer.
The action of SMF based superplasticizers is similar to that of SNF based products.
Nevertheless, there are some differences. The adsorption characteristics of SMF based
superplasticizers are such that they have higher affinity for C3A than C4AF and C3S
(Ramachandran et al., 1998). The stability of SMF appears to vary with the
temperature and could be reason why it is not widely used, especially in countries
with warmer climates. Torresan and Khurana (1998) state that if the SMF is stored at
30
20°C, it could have a shelf life of more than 1.5 years whereas at 40°C, its shelf life is
reduced to 5-7 months. This is attributed to further condensation at higher temperature
and consequent increase in viscosity.
In the case of polycarboxylates, the backbone chain consists of acrylic or methacrylic
copolymers and various functional groups (polar or ionic, carboxyl or hydroxyl
groups) can be grafted on the backbone chain as side chains (e.g., polyethyleneoxide
graft chains). PCE type superplasticizers with high molecular weight (1400-88000
gm/mole) can be obtained with sizes of upto 30-150 nm (Uchikawa, et al.,1997).
However, it appears that the normal length of main chain is 20 nm and that of the side
chain is 7 nm (Ohta et al., 2000; Sakai et al., 2003). The variations in the type and
length of the main and side chains of the PCE type superplasticizers have yielded a
broad variety of new products with very different properties (Yamada et al., 2000;
Maeder and Schober, 2003; Magarotto et al., 2003; Sakai et al., 2003; Sugiyama et al.,
2003; Houst et al., 2005).
The differences in the molecular structure influence the amount of adsorption as well
as the initial rate of adsorption of PCE, which in turn affect the initial fluidity and
flow retention (Uchikawa et al., 1997; Maeder and Schober, 2003; Sugiyama et al.,
2003; Yamada and Hanehara, 2003). In general, higher molecular weight and
reduction of fractions with lower molecular weight leads to better flow in pastes
(Magarotto et al., 2003). Recently, Winnefeld et al. (2007) have proposed the use of
lower side chain densities and shorter side chains to obtain higher fluidity through
better adsorption, and the use of longer side chains for reducing the set retardation,
implying that the optimization of the molecular weight distributions is needed for the
best performance. They have also stated that polymers with higher charge density are
31
adsorbed more strongly and, consequently, result in better initial fluidity. For comb-
type PCE-based superplasticizers, the charge density is higher when more free
carboxylic groups are present (Winnefeld et al., 2007). The Table 2.5 summarises the
relations between the structural factors of the PCE and the dispersibility parameters
(Sakai et al., 2003; Sugiyama et al., 2003).
The adsorption of PCE could also depend on the amount of superplasticizer available
in the solution, and consequently on the dosage of superplasticizer. More precisely, at
low dosages, all fractions of the polymer are adsorbed whereas beyond a certain
dosage, only fractions of superplasticizer with larger molecular weights will be
preferentially adsorbed (Flatt et al., 1998; Winnefeld et al., 2007). This can lead to
differences in the effectiveness of different products since they could have fractions of
varying molecular weights and larger molecular fractions of varying effectiveness,
especially when steric repulsion is the dominant mechanism, as shown in a study of
Na-polycarboxylate-polysulphonate superplasticizers by Flatt et al. (1998).
Table 2.5 Structural aspects of the PCE that affect the fluidity of cement paste
Parameter Structural aspects

Relative chain Relative side Relative
length of back chain length number of side
bone polymer chains
Low dispersibility and short Long Short Large
dispersibility retention
High dispersibility Short Long Small
Long dispersibility retention Shorter Long Large
Compared with the polycondensates (SMF and SNF), the PCE based superplasticizer
is adsorbed much less due to its lower charge density (Houst, et al., 2005; Plank and
Hirsch, 2007). This is attributed to the fact that the functional groups of the PCE type
32
superplasticizers are weaker acids compared to sulphonate groups containing SMF
and SNF type superplasticizers, and therefore, PCE type superplasticizers are weaker
electrolytes and show a lower electronic activity. Among the polycarboxylates,
adsorption is more when the charge density is higher, giving better initial fluidity.
Generally, there is a problem in achieving high initial fluidity and maintaining the
same over a long period of time when using products that are currently available in
the market. Consequently, it has been suggested that the combination of polymers
with different rates of adsorption could lead to the retention of high slump in concrete.
Corradi et. al. (2005) have used this concept in the development of new PCE-based
superplasticizers that are supposed to have both water reduction and flow retention
properties. This is attributed to the addition of a new functional monomer that has a
lower adsorption rate than the more conventional component.
Furthermore, the response of the superplasticizers to the variation in chemical
composition of cement also depends on the type of superplasticizer. According to
Nkinamubanzi and Aïtcin (2004), the polycarboxylate based superplasticizer is less
sensitive to the alkali and soluble sulphate content of the cement than sulphonated
superplasticizers such as lignosulphonates and SNF. Further details about the
dependence of the chemical composition of the cement on the cement-superplasticizer
interaction are given in the following sections.
2.3.3.2 Effect of Calcium Sulphate
In order for normal set to occur in cement paste, sufficient amounts of soluble
sulphate and calcium ions should be available in the solution to form calcium
aluminosulphate (ettringite). When gypsum is incorporated in the portland cement
there is, generally, enough soluble sulphate for hydration to proceed normally.
33
However, if the gypsum is replaced partially or completely with natural anhydrite,
compatibility problems like rapid set and accelerated slump loss can occur especially
in pastes with low w/c and with superplasticizers having sulphonate functional groups
and polycarboxylates (Dodson and Hayden, 1989; Jiang et al., 1999; Mailvaganam,
1999; Nkinamubanzi and Aïtcin, 2004). This is because the rate of solution of natural
anhydrite is slower than gypsum and it is further reduced possibly due to the
adsorption of superplasticizer on the surface of natural anhydrite particles (Jolicoeur
and Simard, 1998; Hanna et al., 2000; Ramachandran, 2002). Consequently, the
availability of soluble sulfate ions in the aqueous phase is not sufficient to keep up
with that needed by the C3A, resulting in the stiffening of the cement paste due to
accelerated setting caused by the formation of monosulphate, which is converted later
to ettringite (Jolicoeur and Simard, 1998). Such incompatible behaviour is not
expected to occur when hemihydrate or soluble anhydrite or gypsum is used in the
cement since their solubility rates are not reduced by the addition of superplasticizer
(Dodson and Hayden, 1989). However, Fernon et al. (1997) report that, in lime
saturated solution, Na-SNF enhances the solubility of gypsum and hemihydrates
significantly and that of the anhydrite only slightly, whereas the Ca-SNF does not
affect the solubilities except that of the anhydrite, which is decreased. The increased
solubility of gypsum or hemihydrates in the presence of Na-SNF is attributed to the
formation of calcium ion complexes by the ionized sulphonates in an alkaline
medium.
Since the adsorption of the superplasticizer on the C3A and AFm is much more than
on ettringite, in competition with the SO42- ions, a sufficient amount of
superplasticizer is, consequently, not available in the solution for providing fluidity to
the paste. This results in the loss of fluidity or an increase in the superplasticizer
34
demand to compensate for this effect (Jolicoeur and Simard, 1998; Ramachandran,
2002). Alternatively, the superplasticizer dosage needs to be increased to provide
sufficient superplasticizer in solution so that paste fluidity is adequate. Fernon et al.
(1997) also caution that an excess of dissolved calcium sulphate can lead to rapid
slump loss due to the alteration of the hydration process because the formation of
ettringite is inhibited.
In India, there is an increasing tendency to use synthetic gypsum instead of natural
gypsum during cement production. This is because of the better quality and higher
SO3 content of synthetic gypsum. However, the rate of solubility of synthetic gypsum,
and its effects on the initial adsorption and consequent fluidity, are not clear from
literature and have to be investigated. Moreover, synthetic gypsum could contain
soluble phosphates (in the range of 0.08-0.18 %) that increase the setting time from 3
to 10 hours (Hewlett, 2004). Such retardation can be eliminated if the phosphogypsum
is neutralized with lime, and content of water soluble phosphates and fluorine is
reduced to below 0.02% (Hewlett 2004; Diop et al., 2005).
2.3.3.3 Effect of Soluble Alkalis
The alkali compounds have an active role in the adsorption and consequently in the
interaction. It has been observed that a higher alkali content in the aqueous solution of
cement paste increases the solubility of sulphate ions and hence decreases the loss of
fluidity with SNF due to lower initial adsorption of the superplasticizer (Dodson and
Hayden, 1989; Chandra and Bjornstrom, 2002b). Note that the soluble alkali content
is not necessarily related to the total alkali content of the cement. If cement has low
soluble alkali content, the adsorption of SNF will be more and that remaining in
solution will be less resulting in loss of fluidity (Dodson and Hayden, 1989;
35
Nkinamubanzi et al., 2000). This is because the superplasticizers act as sulphate ion
providers and interact with C3A instead of dispersing it. Consequently, cements with
high alkali contents are expected to have good slump retention and to be more robust
than cements with low alkali contents, when used in combination with SNF
(Nkinamubanzi and Aïtcin, 2004).
The problem of inadequate fluidity for low alkali cement-SNF combinations can be
overcome by adding soluble alkalis, such as Na2SO4 (Jiang et al., 1999; Li et al.,
2003; Spiratos et al., 2003); Jiang et al. (1999) reports that 0.4-0.5% Na2O soluble
equivalent is the optimum soluble alkali content in the cement to maximize initial
fluidity and to minimize the loss of fluidity. However, the addition of sodium sulphate
could increase the incompatibility problem with some polysulphonates, as it alters the
formation of the ettringite (Li et al., 2003).
In the case of the PCE, the excess of alkali sulphates results in the lowering of fluidity
due to the shrinking of the side chains, thereby decreasing the steric repulsion
(Hanehara and Yamada, 1999). Also, efflorescence problems are observed when
naphthalene and melamine based superplasticizers are used with cements having high
alkali oxide contents (Na2O + K2O >0.75%) (Akman and Cavdar, 1999).
2.3.3.4 Effect of C3A
The interaction of the superplasticizer with the C3A phase of the cement has the prime
influence on the rheological properties of the fresh mix due to its higher positive zeta
potential and reactivity compared to other cement compounds. The C3A content or
more specifically SO3/C3A ratio and the reactivity of C3A controls the adsorption of
the superplasticizer, which in turn affects the early hydration behaviour and
workability of cement paste (Aïtcin et al., 2001). The calcium sulphate reactivity
36
determines the availability of soluble sulphates whereas the C3A reactivity determines
the rate of consumption of these ions. Therefore, the amount of soluble sulphates
along with the reactivity of C3A influences the hydration reactions of the aluminate
phases resulting in the setting of cement paste as either normal, quick, flash or false
set as described by several authors (Mehta and Monteiro, 2005; Spiratos, et al., 2003).
The effect of the superplasticizer on C3A hydration depends on the type of
superplasticizer; the adsorption of SNF and SMF based superplasticizers retard C3A
hydration and delay the conversion of ettringite to monosulphonate. This is because
C3A and C4AF surfaces rapidly adsorb large amounts of SNF and SMF. Also, in the
presence of sulphate ions, the SNF forms an organo-mineral compound with C3A and
alters the morphology of the hydration products. However, high molecular weight
SNF shows less retardation of C3A than that with lower molecular weight (Rixom and
Mailvaganam, 1999).
The interaction of PCE with C3A depends on the form of C3A, cubic or orthorhombic,
with the orthorhombic form being more reactive than cubic form. Hence, a high cubic
C3A content results in low flow retention as the PCE will be adsorbed first on the C3A
surface and the ettringite will be formed later. Later, the admixture will be covered by
the hydrated products and will not be available for further flow retention (Magarotto
et al., 2003).
In the case of cements with less C3A, adsorption of the superplasticizer will be more
on the C3S and C2S, preventing their hydration, resulting in the reduction of early
strength (Roberts, 1995). Also, crystallization of gypsum needles and consequent false
set could occur if the C3A content is not enough to consume the dissolved sulphates
(Ramachandran, 2002). The C3A phase in cement has very high surface area
37
compared to its hydration product, ettringite, and consequently adsorbs large amounts
of admixture resulting in loss of fluidity. Hence, the delayed addition of
superplasticizer is advantageous as the hydration of C3A takes place and ettringite is
formed before the superplasticizer is added (Roberts, 1995).
2.3.3.5 Other Factors Affecting the Cement-Superplasticizer Interaction
There are several other factors that can influence the effectiveness of the
superplasticizer and the interaction between the superplasticizer and cement, in
general. Some important factors that have not been discussed earlier are explained in
the following paragraphs.
The specific surface area influences the effectiveness of the superplasticizer since it
increases the water demand, as well as the amount of superplasticizer for a given
workability. Hence, the amount of superplasticizer adsorbed depends on the fineness,
with finer cements leading to higher superplasticizer dosages (Aïtcin, 1998; Jolicoeur
and Simard, 1998; Erdogdu, 2000; Vikan et al., 2007).
As is well known, the w/c ratio affects the performance of superplasticizer in terms of
the flow of the cement paste and workability of concrete. Moreover, most of the
incompatibility problems occur at low w/c, where the lower interparticle separation
between the cement particles makes the system to be more sensitive to the loss of
water by evaporation or hydration reactions resulting in higher loss of fluidity with
time (Collepardi, 1998).
The placing of concrete in different environmental conditions also affects the
superplasticizer performance. Temperature is the major environmental factor affecting
cement-superplasticizer interaction. Generally low temperatures decrease the fluidity,
which cannot be compensated with the addition of superplasticizer beyond the
38
saturation dosage (Gettu et al., 1997; Roncero et al., 1999). On the other hand, high
temperatures increase superplasticizer adsorption, which increases the initial fluidity
and increases the superplasticizer demand for flow retention over a certain range of
temperatures (Roncero et al., 1999; Greisser, 2002). Also, the temperature increase
causes an increase in the reactivity of C3A, which causes a more rapid loss of fluidity
and formation of higher ettringite contents with finer morphology in the presence of
superplasticizer (Greisser, 2002; Spiratos et al., 2003).
The effect of temperature on mixes with PCE based superplasticizers is dependent on
its chemical structure (e.g., the carboxylates have the highest fluidity loss compared to
other functional groups) (Felekoglu and Sarikahya, 2007; Grzeszczyk and Sudol,
2005), with the increase in temperature leading to a lower increase of flow in the paste
and lower flow loss for longer polyethylene oxide graft chains (Nawa et al., 2000).
This is attributed to the higher adsorption of the graft chains with an increase in
temperature, as a result of which the steric repulsion of the longer graft chains is less
affected.
Roncero et al. (1999), Roncero (2000), and Fernandez-Altable and Casanova (2006)
have shown that though the fluidity increases with the temperature, it has no practical
effect on the saturation dosage of superplasticizer, which is attributed to the
dependence of the saturation on the available surface area of the cement particles.
2.3.4 Effect of Superplasticizer on the Morphology of Hardened Cement Paste
The morphology of the hardened cement paste could be affected by the alteration of
the cement hydration reactions by the retarding effect of the superplasticizers as well
as the intercalations with the early hydration products, as described in the previous
sections. The polysulphonates, can combine with the interstitial phase of the cement
39
(C3A+C4AF) to form an organo mineral composite resembling ettringite (Aïtcin et al.,
2001). They could also modify the morphology of ettringite produced during
hydration; i.e., ettringite crystallizes in small and thick crystals rather than
conventional needle shaped crystals (Aïtcin et al., 1994; Hanna et al., 2000; Prince et
al., 2002). The addition of superplasticizer also modifies the morphology of the
portlandite; i.e., the size of portlandite crystals decrease and also their form changes
from block-like to thin plates (Grabiec, 2004).
In addition to modifying the morphology, the delay in the release of SO42- (due to the
presence of anhydrite) results in the formation of monosulphoaluminate, first and then
ettringite is formed (Prince et al., 2003). The addition of lignosulphonates further
reduces the solubility of anhydrite due to the reduction in the rate of diffusion of ions
at solid-liquid interphase, and further promotes this phase transformation (Prince et
al., 2003).
In general, the change in morphology due to the incorporation of superplasticizer
improves the rheological properties by yielding smaller hydrate particles and
preventing hydration products from bridging neighbouring cement particles. This
results in substantial reduction in the yield stress of paste and increase in fluidity.
There is also a difference in porosity and pore size distribution of superplasticized
concrete compared to normal concrete. Cements with high C3A and low SO3 content
have larger pores that decrease in size with higher SO3 contents whereas the pore size
in low C3A cements is independent of the SO3 content. Generally, higher numbers of
smaller pores are produced in superplasticized mixtures (Roncero, 2000; Khatib and
Mangat, 1999). This has been attributed to better silicate polymerization leading to the
refinement of pores and consequent reduction in porosity (Puertas et al., 2005).
40
2.3.5 Effect of Superplasticizer on Properties of Concrete
As intended, the incorporation of superplasticizers increases the workability of
concrete that permits the use of comparatively lower w/c leading to higher strengths.
In addition to strength, the properties of concrete that are affected include slump, air
content, bleeding, segregation and setting time in the fresh state, and properties like
permeability, abrasion resistance and durability in the hardened state. The type of
superplasticizers and cement can cause variations in these properties and in certain
cases the incompatibility between the two leads to adverse effects like loss of fluidity,
retardation in setting, air entrainment and reduction in strength gain, as explained in
the previous sections. Aïtcin et al. (2001) observed that cements with a C3A content of
6% to 10%, a soluble alkali content of 0.4 % to 0.5% give cement-polysulphonate
combinations that are both compatible and robust.
Concrete with high workability containing a superplasticizer generally shows more
rapid slump loss than that with the same workability and without superplasticizer.
This could be attributed to the loss of consistency due to the low w/c during the
dormant stage and the increased adsorption of the superplasticizers by the aluminate
phase (Ramachandran et al., 1998). In addition, the higher slump loss is related with
the C3S phase of cement also; i.e., the hydrate coatings restore charge interactions
between the C3S grains, thereby restoring the tendencies for flocculation
(Ramachandran, 2002). Redosing of the superplasticizer neutralizes these charges and
restores the fluidity, but continued hydration will lead eventually to slump loss.
In a superplasticized concrete, generally no segregation or bleeding is expected to
occur at the saturation dosage of the superplasticizer. When redosing is used to
control the loss of slump, however, bleeding and segregation may occur if the second
41
dosage is too high; overdosage of superplasticizer generally promotes segregation and
formation of a white deposit containing lime, calcium sulphate and calcium carbonate
on the top surface when hardened (Ramachandran, 2002).
When superplasticizers are added to concrete at the appropriate dosage, the 28-day
compressive strength of superplasticized concrete is often equal or greater than the
corresponding strength of the reference concrete (Rixom and Mailvaganam, 1999;
Ramachandran, 2002). This is attributed to the reduction in w/c as well as the higher
workability and better compaction achieved due to the incorporation of
superplasticizers Dhir and Yap (1983) also reports that even though there could be
some initial retardation, superplasticizers produce compressive strengths of the same
order as the normal concrete at 3, 7 and 28-days. This is because, normally, the C3A
and C4AF phases rapidly adsorb large amounts of SNF and SMF, so that only small
amounts of SMF and SNF are available with the aqueous phase to retard the hydration
of C3S.
Nevertheless, there are some reports of strength reduction while using naphthalene
based superplasticizers and melamine based superplasticizers, especially with cements
having low C3A content and high C3S and C2S contents (Ramezanianpour et al.,
1995). This may be either due to higher adsorption on the C3S and C2S phases or
gypsum crystallization and consequent false set in the absence of adequate C3A, as
reported earlier (Roberts, 1995; Ramachandran, 2002)
The creep and shrinkage of superplasticized concrete is comparable to or is less than
that of reference concrete (e.g., Seung-Bum, 1999), although there are several
exceptions. For the same amount of moisture loss, the superplasticized concrete could
exhibit larger shrinkage which may be related to the larger dispersion of cement and
42
cement hydrates (Ramachandran, 2002). Dhir and Yap (1983), and Gettu et al. (2002)
have reported that drying shrinkage is higher for superplasticized concrete with high
workability; with SNF based superplasticizers leading to higher total shrinkage than
PCE based products whereas SMF based shows lower shrinkage than PCE based
superplasticizers. This could be attributed to the refinement of pore structure to more
number of fine pores (Roncero, 2000; Gettu et al., 2002; Roncero et al., 2002).
2.3.6 Admixture – Admixture Interaction
2.3.6.1 Cement-Superplasticizer Interaction in the Presence of Mineral

Admixtures
Supplementary cementitious materials (or mineral admixtures) are generally
incorporated in concrete as a replacement for cement to reduce the heat of hydration
and to enhance strength and durability. It is usually reported that the addition of
mineral admixtures reduces the workability due to the higher surface area and
consequently increases the water and superplasticizer demand (e.g., Agullo et al.,
1999). Nevertheless, there could be some workability enhancement due to the addition
of fly ash and silica fume, in low volumes, which is generally attributed to the
spherical shape of the particles and consequent reduction in the interparticle friction
(Ferraris et al., 2001).
The incorporation of fly ash could also affect the hydration process of cement due to
the additional aluminates (of the fly ash) that suppress the silicate reaction. This
retarding effect is worsened due to the lack of sulphates in the presence of
superplasticizers (Roberts, 1995).
43
2.3.6.2 Influence of Other Chemical Admixtures on Cement-Superplasticizer
Interaction
The other chemical admixtures generally used in concrete include viscosity modifying
admixtures (VMA), air entraining agents, damp proofing compounds and shrinkage
reducing admixtures. A viscosity modifying admixture is usually incorporated with
superplasticizers to produce flowing/self compacting concrete. The interaction of the
superplasticizer with the VMA depends on the mode of action by which the latter
provides higher viscosity in the concrete. Most VMAs tend to increase the viscosity
by forming gel/cement slurries rather than increasing the viscosity of water in the mix.
The VMA molecules could compete with the superplasticizer for active sites on the
cement particles, resulting in lower superplasticizer adsorption (Mailvaganam, 1999;
Bedard and Mailvaganam, 2006).
The addition of a superplasticizer modifies the relationship between the air content
and the spacing factor (air bubble spacing), and sometimes significantly destabilizes
the air-void system (Bedard and Mailvaganam, 2005) caused by the incorporation of
an air-entraining agent. For example, when a lignosulphonate based water reducer is
added to concrete containing an air entraining agent, a substantial increase in air
content occurs and the specific surface of air bubbles are reduced.
The combination of a shrinkage reducing admixture (SRA) with a superplasticizer
reduces the efficiency of the SRA in reducing shrinkage. Also, the combination
increases the retardation in setting more than the superplasticizer alone (Gopinath and
Gettu, 2008).
2.4 Need for the present study
The current level of knowledge worldwide about the factors affecting the cement-
superplasticizer interaction is described in this chapter. Specific issues of
44
incompatibility have also been addressed, namely, loss of workability, alteration of
setting behaviour, reduced rates of strength gain and change in the microstructure.
There are contradictions regarding the influence of molecular weight and the side
chain lengths of PCE on the adsorption behaviour, which have also been brought out.
In the Indian context, cement standards in our country are not very stringent, and
enable manufacturers to adjust their product in many different ways. For example,
while the minimum fineness is specified for different grades of cement, there is no
control on the maximum values. Similarly, wide ranges in the chemical composition
are acceptable. This could result in significant variability in the cement properties.
Also, the alkali content, which has a major influence on cement-superplasticizer
interaction, is not considered as a characteristic of the chemical composition for a
particular grade of cement.
From the viewpoint of the use of superplasticizers, there is insufficient knowledge
among users regarding the limitations of different types of chemicals. Users, who are
unaware of compatibility issues, often suffer when the supply of cement and/or
admixture has to be changed during the course of a project. Problems arising from
compatibility issues are often mistaken for problems with the concrete mixture design
because of the lack of information about the subject amongst practicing engineers.
Admixture manufacturers try to overcome the problem by formulating project-specific
chemicals, which is obviously, only a short term solution.
Cement-superplasticizer interaction in concrete is due to a complex blend of chemical
and physical mechanisms that are interdependent. The complicated nature of the
problem prevents the development of simple solutions to address field related issues
of the application of superplasticizers. Hence, a detailed study on the cement-
45
admixture interaction is required for optimizing different types of admixtures in a
cementitious system.
Studies on cement-superplasticizer interactions in India have been limited to the
workability evaluation of concretes containing these chemicals. There have not been
any investigations to understand the physico-chemical nature of this interaction. Thus,
the results from such studies are not broad-based, i.e. they apply to a small group of
cements and/or chemical admixtures. There is a distinct need for a study of the
interaction of Indian cements and admixtures, as well as to select an appropriate
dosage of the superplasticizer. Moreover, the wide range of cements used, varying
transportation durations and climatic conditions necessitate fundamental studies that
can explain the mechanisms of interaction and help establish methods for identifying
incompatibilities in practical situations.
Simple methodologies are required to identify these incompatibilities due to such
interactions and to further understand the fundamental nature of admixture behaviour
in cement-based systems. The understanding of these interactions should be both at
the applications scale (for example, studying flowability and retention of workability
in pastes and concretes, retardation in setting), as well as at the micromolecular scale,
where some insight can be obtained into the physico-chemical interactions between
cement particles and superplasticizer molecules.
46
CHAPTER 3
STUDY OF THE FLOW BEHAVIOUR OF

SUPERPLASTICIZED CEMENT PASTE
3.1 GENERAL
A brief description about the cement hydration reactions and the influence of
superplasticizers on the hydration reactions, morphology of hydrated products and the
properties of superplasticized concrete has been presented in the previous chapter. In
addition to providing good workability, the superplasticizer also influences the slump
retention and setting process. The interaction between the cement and superplasticizer
governs the flow behaviour of concrete. To study this interaction, it is proposed to
analyse the flow of cement paste with simple methods like Marsh cone and mini-slump
tests, and more fundamental tests like those with a viscometer.
The basic objective of this chapter is to formulate the methodology for studying the
fluidity of superplasticized cement pastes that will be used in the more detailed studies in
the rest of the thesis. The specific objectives are
• To select a suitable mixing method for preparing the superplasticized cement
paste that is representative of the paste phase in concrete.
• To validate the suitability of simple test methods, such as the Marsh cone and
mini slump tests, for studying the flow behaviour
• To understand the significance of results from simple methods with regard to
rheological parameters.
In order to meet the above objectives, the influence of the mixing method on the fluidity
of superplasticized cement paste is studied through the Marsh cone and mini-slump tests.
Subsequently, the influence of type and dosage of superplasticizer on the flow behaviour
of the paste is studied. Finally, the correlations between the results obtained with the
simple test methods and those obtained from viscometric studies are presented which
helps to select an appropriate methodology to study the fluidity of cement paste and to
understand the results from empirical tests better.
3.2 MATERIALS USED IN THE STUDY
3.2.1 Cement
Ordinary 53 grade portland cement (ASTM Type 1 Cement) (denoted as Cement 1)
satisfying the requirements of IS 12269 (2004) was used in the work presented in this
chapter. The chemical properties are given in Table 3.1, as determined by the National
Test House, Chennai, for the sample used.
The compounds present in the cement can be quantified using the results of the chemical
analysis of the clinker corresponding to the cement and the Bogue equation. Accordingly,
the clinker was obtained from the cement manufacturer and tested in Holcim Roodeport
Laboratory in South Africa; the results of the analysis are also given in Table 3.1. These
data were used to determine the proportions of the compounds present in the cement (see
Table 3.2).
From the test results (Tables 3.1 and 3.2), it is confirmed that the cement used can be
considered as a low alkali cement since the (Na2O)eq is less than 0.6% (Hewlett, 2004);
that the SO3 content is within the normal range (i.e., 2–3.5%) for ordinary Portland 53
grade cement and, therefore, is not expected to affect the superplasticizer adsorption
48
significantly (Chandra and Bjornstrom, 2002a); and that the C3A content is in the normal
range (i.e., 4-8%).
Table 3.1 Chemical properties of the cement and the corresponding clinker
Characteristics Quantity
(% by mass)
Cement Clinker
Calcium oxide (CaO) 60.81 63.10
Silica (SiO2) 19.50 20.80
Alumina (Al2O3) 4.12 5.80
Iron oxide (Fe2O3) 6.06 5.30
Magnesia (MgO) 1.52 0.80
Sulphur anhydrite (SO32-) 2.48 0.68
Insoluble residue 1.51 -
Total loss on ignition 3.41 0.40
Total chloride content (Cl-) 0.01 -
Alkali content: Na2O 0.05 -
K2O 0.28 -
(Na2O)eq 0.23 0.20
Lime saturation factor 0.93 0.92
Ratio of alumina to iron oxide 0.68 1.09
Table 3.2 Bogue composition of the clinker
Compound Composition
(%)
C3S 51.7
C2S 20.8
C3A 6.5
C4AF 16.1
In addition to the chemical properties, it is necessary to determine certain physical
properties of the cement used for characterizing its behaviour. In this work, the Blaine
specific surface area, water demand for a paste of standard consistency, and the initial
49
and final setting times were determined experimentally, as detailed in the following
sections.
3.2.1.1 Determination of the Water Demand for Standard Consistency of Cement
The consistency of the cement paste is determined as per ASTM C 187 (2004)and IS
4031: Part 4: (2005), where the quantity of water found that produces a cement paste of
standard consistency, defined as that which permits the Vicat plunger to penetrate upto 5
to 7 mm from the bottom of the Vicat mould. Cement pastes with varying w/c are
prepared (by hand mixing) and the penetration is checked for each of them. The amount
of water in the paste that satisfies the above mentioned condition is expressed as a
percentage by mass of dry cement, and taken as the water demand of the cement. This
can be considered as the minimum w/c needed for the paste to be homogeneously, and
uniformly mixed and moulded. The water demand, obtained as an average result of 3
trials was 31% (see Table 3.3).
Table 3.3. Physical properties of cement 1 (C1)
Property Value (mean and standard deviation)
Water demand for normal consistency of

cement paste 31% (± 1.3%)
Initial setting time 106 minutes (± 7 minutes)
Final setting time 190 minutes(± 6 minutes)
Blaine specific surface area 332 m2/kg (± 10 m2/kg)
3.2.1.2 Determination of Initial and Final Setting Times
The initial and final setting times are determined as per IS 4031: Part 5 (2005). The
cement paste is prepared with 0.85 times the water required to give a paste of standard
consistency (see previous section). It is filled in the Vicat mould and the corresponding
needle is allowed to penetrate through the paste. The period elapsing between the time
50
when water is first added to the cement and the time at which the penetration of the
needle is 35.0 ± 0.5 mm is taken as the initial setting time. Note that this is similar to
procedure of ASTM C-191 (2001a), except that the paste used (i.e., same w/c) is of
standard/normal consistency, and the penetration for the initial setting time is 25 mm.
The test is continued with a standard needle with an annular attachment for determining
the final setting time. The period elapsing between the time when water is added to the
cement and the time at which the needle makes an impression on the surface of the test
block while the attachment fails to do so is taken as the final setting time.
The values obtained for the setting times are given in Table 3.3 as averages from three
trials. As the initial setting time is more than 30 minutes and the final setting time is less
than 10 hours, the IS code requirements for the setting times are satisfied. It should be
noted, however, that the difference between the initial and final setting times is only
about 90 minutes, which is relatively low.
3.2.1.3 Determination of the Fineness of Cement
The fineness of cement has been determined by the Blaine air permeability method as per
IS 4031: Part 2 (2005) and ASTM C 204 (2007). The specific surface area obtained
(332 m2/kg) is within the limits given for a 53 grade OPC.
3.2.2 Superplasticizers
Superplasticizers for this study were selected from the four different families of products
available (i.e., lignosulphonates, naphthalenes, melamines and polycarboxylates) in order
to obtain representative results that would help understand the effect of the different types
on the flow behaviour of cement paste. Table 3.4 gives the details of the commercial
products used; the products are denoted as the abbreviation of the type followed by a
51
number. The values of the solid content, pH, density were obtained in the laboratory as
per IS 9103 (2004), and cross-checked with data from the corresponding supplier. It can
be seen that the active part of superplasticizer varies from 32% to 44 % for different
families as well as for the same family. The recommended dosage of the superplasticizers
is also given in the Table 3.4. The superplasticizer dosages given in this work are all in
terms of the solid content; the water contained in the superplasticizer has been accounted
for in the water content of the mixes.
Table 3.4 Properties of superplasticizers
Density Solid Recommended

Chemical type
Designation content pH dosages
kg/litre (%) (sp/c %)
LS-C1 Lignosulphonate 1.18 38 7.5 0.18-0.27
Polynaphthalene
SNF-S1 1.09 37 6.7 0.2-0.8
sulphonate
Polynaphthalene
SNF-D1 1.16 32 6.8 0.22-0.74
sulphonate
Polynaphthalene
SNF-D2 1.25 44 7.8 0.44-0.66
sulphonate
Polynaphthalene
SNF-C1 1.20 40 7.4 0.2-0.3
sulphonate
SMF-S1 Polymelamine
1.30 40 8.5 0.2-0.68
sulphonate
Polymelamine
SMF-C1 1.28 38 7.0 0.19-0.3
sulphonate
Polycarboxylate
PCE-S1 1.09 38 7.5 0.08-0.33
ether
Polycarboxylate 33
PCE-D1 1.10 7.4 0.07-0.3
ether
3.3 TEST METHODS USED FOR STUDYING THE FLOW BEHAVIOUR OF

THE CEMENT PASTE
Due to the important functions of superplasticizers in concrete, the selection and dosage
criteria are fundamental for determining the optimum composition of concrete. From the
52
practical point of view, it is better to make this choice using the actual concrete at the
construction site under the local conditions. However, tests of concrete imply significant
labour, material and time. Besides, such tests do not reflect the fundamental action of the
superplasticizer with regard to physical and chemical interactions. Therefore, several
methods based on tests of paste and mortar have been developed for determining the
optimum superplasticizer dosage in concrete.
The Marsh cone and mini slump tests are simple empirical methods to characterize the
flow behaviour of cement paste whereas rheological tests with a viscometer provide more
fundamental information about the material but require more complicated equipment and
skilled operators. The fundamentals of these three methodologies, and the procedures
adopted in this work are detailed in the following sections.
3.3.1 Marsh Cone Test
The Marsh cone (Figure 3.1) test has been used previously to evaluate the relative fluidity
and the saturation superplasticizer dosage in cement pastes and mortars (de Larrard et al.,
1998; Aitcin, 1998; Khayat and Yahia,1998; Agullo et al., 1999; Roussel and Roy, 2005;
Royy and Roussel, 2005). In the present study, a metal cone (as per the guidelines of
European standards EN 445, French standards P 18-358, which is similar to ASTM C 939
(1987), with a nozzle of diameter 8 mm was employed. An initial volume of 1000 ml of
paste was poured into the cone and the time required for 500 ml of it to flow out was
measured. The test gives the fluidity of the paste in terms of the flow time; higher the
flow time, lower is the fluidity of the paste. The saturation point is the dosage beyond
which further addition of superplasticizer does not increase fluidity significantly but can
53
produce segregation (Agullo et al., 1999); the saturation dosage can be taken as the
optimum superplasticizer dosage for a given cement paste.
R, H - Radius and length of nozzle

R1, H1 - Radius of the free surface and height of fluid
in the cone at the initial moment
Rt, Ht - Radius of the free surface and height of fluid
in the cone at any time t
R2, H2 - Radius of the free surface and height of fluid
in the cone at the final moment
φ – Angle between the axis and the generator of the
cone
Fig. 3.1 Geometry and dimensions (in millimetres) of the Marsh cone
Gomes et al. (2001) proposed a method for the objective determination of the saturation
dosage based on the Marsh cone flow time curve of pastes, which is illustrated in
Figure 3.2. In this method, the internal angle (α) corresponding to each data point is
calculated and the superplasticizer dosage corresponding to an internal angle of 140°+10°
is taken as the saturation dosage. Interpolation is used to determine the dosage when there
are no data points corresponding to that range of angles. This criterion was proposed
based on about 200 tests on superplasticized cement pastes (Gomes, 2002).
54
1.8 SMF-S1
1.7
log (flow time , sec)

1.6
1.5
1.4 saturation point corresponding to 140

0
0 0
1.3 140 +10
1.2
1.1
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 3.2. Flow time curve showing the determination

of optimum superplasticizer dosage
(according to the method of Gomes et al., 2001)
The Marsh cone test (see Figure 3.3) can also be used to study the loss of fluidity with
time, the fluidity of cement mortar, optimization of cement paste with mineral admixtures
and incompatability between cement and superplasticizer (Agullo et al., 1999).
Fig. 3.3 Marsh cone test showing continuous and discontinuous flow
55
3.3.2 Mini-slump Test
The mini-slump test developed by Kantro (1980) has also been used to study the flow
behavior of superplasticized cement paste (Aitcin, 1998; Khayat and Yahia,1998; Gomes,
2002; Roussel, 2006; Tregger et al., 2008). In this test, a mould in the shape of a
truncated cone, with dimensions proportional to the Abram’s cone, is filled with cement
paste and the spread diameter is measured after lifting the mould. Additionally, the time
taken for the paste to reach a diameter of 115 mm was also determined here. Also, visual
examination helps evaluate the bleeding and segregation of the paste.
3.2 19.0 6.4
15.9
22.2
57.0
38.1
3.0
38.1
All dimensions are in millimetres
Fig. 3.4 Mini-slump cone
(a) Filling of the mini-slump (b) Spread of the Paste
Fig. 3.5 Mini-slump test
56
3.3.3 Viscometer Test
Viscometer tests on cement paste can be used to study the effect of the changes in
cement, type and dosage of admixtures, such as superplasticizers, on paste characteristics
like yield stress and plastic viscosity (Asaga and Roy,1980; Banfill, 1981; Masood and
Agarwal, 1994; Beaupre and Mindess, 1998; Claisse et al., 1999; Cyr et al., 2000; Park et
al., 2005). The basic principle is to apply different shear rates to the paste in a viscometer
and measure the corresponding shear stresses. Generally, a loading-unloading cycle is
applied to the paste, preceded by some pre-shearing, and the response during unloading is
used to determine the rheological parameters (Goisis et al., 2003; Papo and Piani, 2004).
Cement paste normally exhibits shear-thinning behaviour, where the slope of the shear
stress versus shear strain rate curve decreases with an increase in the shear rate. Such
shear-thinning response is attributed to the structural breakdown and rebuilding that takes
place in the paste (Tattersall and Banfill, 1983) with the former mechanism
predominating (Lapasin et al., 1983). There are, however, notable exceptions; e.g., dilute
superplasticized pastes may exhibit Newtonian behaviour while an overdosage of
superplasticizers may result in shear thickening (strain-hardening) that could be attributed
to the agglomeration of excess superplasticizer in the aqueous solution of the paste or
segregation(Cyr et al., 2000; Papo and Piani, 2004; Lootens et al., 2004).
In the present study, a Brookfield HA DV II +Pro viscometer was used with a coaxial
cylinder setup (inner cylinder radius = 16.77 mm, gap width = 1.14 mm) as shown in
Figure 3.6 and Rheocalc software for programming the test. The details of the coaxial
cylinder set up are shown in Figure 3.7. The rheological behaviour depends on the type of
spindle, shear rate, the gap between the cylinders and environmental conditions. The
57
basic principle of the experiment is to apply a known shear rate to the fluid through the
spindle and measure the corresponding shear stress produced.
Fig. 3.6 Brookfield HA DV II +Pro viscometer (Asphalt Laboratory, IITM)
(a) Co-axial cylinder (b) SC4 21 - Spindle (c) HT-2- Sample chamber
Fig. 3.7 Details of co-axial cylinder setup
Three loading-unloading cycles were imposed by increasing, and later decreasing, the
shear rate from 23 to 163 s-1 in seven steps. The shear history of rheological testing for
one hysteresis cycle is shown in Figure 3.8.
58
180
160
140
Shear rate (1/sec)

120
100
80
60
40
20
0
0 20 40 60 80 100 120 140
Time (sec)
Fig. 3.8 Shear history of rheological testing
The typical response is shown in Figure 3.9, where the first two cycles are considered as
pre-shearing and the downward curve of the third cycle is used for determining the
rheological parameters. The pre-shearing is needed for the structural breakdown of
cement paste sample and to create uniform conditions before testing.
35
SNF-S1
Dosage-0.4%
30
Shear stress (Pa)
25
20
First cycle
15 Second cycle
Third cycle
10
20 40 60 80 100 120 140 160 180
Shear rate (1/sec)
Fig. 3.9. Typical graph showing hysteresis cycles for a superplasticized cement paste
59
The shear stress-strain behavior of cement paste has been represented using several
models (Atzeni et al., 1985; Nehdi and Rahman, 2004), the simplest of which is the
Bingham model:
•
τ =τ0 + μp γ (3.1)
•
where τ is the shear stress in (Pa), γ is the shear strain rate (1/sec), μ p is the plastic
viscosity (Pa s) and τ 0 is the yield shear stress (Pa). The Bingham model is generally
fitted to the experimental data (during unloading), as illustrated in Figures 3.10(a-c), to
obtain the yield shear stress and the plastic viscosity of the cement paste. This appears to
give a satisfactory representation of the rheological behaviour in most cases. Even slight
shear thickening response can lead to a negative yield stress when the Bingham model is
fitted to the experimental data. When the τ 0 -value is negative, as in Figure 3.10c, the
data are normally re-fitted by forcing the condition of τ 0 =0 (i.e., Newtonian model). For
these reasons, a more versatile model, namely the Herschel-Bulkley model, has been used
by several researchers (de Larrard et al., 1998):
•
τ =τ0 + kγ n
(3.2)
•
where τ is the shear stress in (Pa), γ is the shear strain rate (1/sec), n is the power index,
k is the consistency (Pa sn), and τ 0 is the yield stress (Pa). Since the consistency and
power index are interdependent, there is no single term that represents the plastic
viscosity of the material.
60
18
Bingham model
Herschel Bulkley model
16
14 SNF-C1 0.1%
Shear stress (Pa)
w/c 0.35
12
10 Bingham model- τ0= 6.724Pa

2
μ=0.057 Pa s, R = 0.99
8 Herschel Bulkley model-τ0=5.122
n 2
k=0.269 Pa s , n= 0.722, R = 0.99
6
0 40 80 120 160 200
Shear rate (1/sec)
(a) Model parameters at less than saturation dosage
22
Bingham model
20 Herschel Bulkley model
18
SMF-S1 0.2%
Shear stress (Pa)
16 w/c 0.35
14
12
Bingham model- τ0=6.895 Pa
10 2
μ=0.0759 Pa s, R =0.99
8 Herschel Bulkley model-τ0=6.427 Pa
n 2
k= 0.117Pa s , n= 0.92, R =0.99
6
0 40 80 120 160 200
Shear rate (1/sec)
(b) Model parameters at saturation dosage
61
7
Bingham model
Herschel Bulkley model
6
5 PCE-S1 1%
w/c 0.35
Shear stress (Pa)

4
2 Bingham model- τ0=-0.264 Pa

2
μ=0.0358 Pa s, R = 0.99
1 Herschel Bulkley model- τ0= 0.388 Pa
n 2
k= 0.0057 Pa s , n= 1.344, R = 0.99
0
0 40 80 120 160 200
Shear rate( 1/sec)
(c) Model parameters at dosage much more than saturation
Figure 3.10 Fits of experimental data with the Bingham and Herschel-Bulkley
models at (a) less than saturation (b) at saturation
(c) dosage much more than saturation
3.4 SELECTION OF MIXING METHOD
The purpose of concrete mixing is to achieve a uniform mixture of all material
components and facilitate hydration. The type of mixer used has an influence on the
mixing efficiency (i.e., degree of dispersion of cement particles) depending on whether it
is intensive or extensive (Yang and Jennings, 1995). Mixing is especially important for
high performance concrete (HPC) of lower binder content. The type and efficiency of
mixer will influence the mixing time needed, the superplasticizer dosage required, and
the quality of HPC produced (Chang and Peng, 2001). Previous studies of Yang and
Jennings (1995), and Diamond (2005) have concentrated on the microstructure and
rheology of pastes mixed in different ways. They have shown that intensive mixing leads
to more homogeneity and better dispersion of cement particles. Also, the viscosity of the
62
paste decreases with intensive mixing, reflecting the effectiveness of such methods
(Williams et al., 1999; Ferraris, 2001).
Studies conducted by Diamond (2005) show that prolonged mixing can induce higher
stiffness in the mix. Therefore, it is better to have an intensive mixing procedure of short
but adequate duration instead of a prolonged mixing time.
Hence, it is important to understand the effect of different mixing methods on the fluidity
of superplasticized cement paste as well as to select the type of mixing method which
leads to a paste that is representative of concrete. In the present work, two simple
methods, the Marsh cone and mini-slump tests, are used for studying the fluidity of
cement paste made with different mixing techniques.
3.4.1 Materials Used in the Selection of Mixing Method
In the present study, tests were conducted on cement paste with water/cement ratio of
0.35. Birla Super 53 grade ordinary portland cement (OPC) and a polycarboxylic ether
based superplasticizer (PCE-D1) was used. The recommended dosage of superplasticizer
is 0.5-1.6 litres per 100 kg of cement. Distilled water was used for mixing the cement
paste. Superplasticizer dosage, denoted as sp/c, is expressed as the ratio of solid content
of superplasticizer to cement content by weight. All the test materials were kept in the
environmental chamber at a temperature of 27 ºC for 24 hours prior to testing.
3.4.2 Experimental Details
The first paste was mixed with a Hobart mixer (1/6 hp) with a B-flat beater (Figure 3.11a)
with low speed (with shaft speed of 139 rpm and planetary speed of 61 rpm), the second
paste with a Hobart mixer with a D-wire whip (Figure 3.11 b) at medium speed (with
shaft speed of 285 rpm and planetary speed of 125 rpm) and third with a high shear rate
63
kitchen blender (550 watts, 16000-18000 rpm). The fourth paste was made by grinder
(1/4 hp, 300 rpm) mixing and fifth by hand mixing. The mixing time of 5 minutes was
selected to minimize any effects due to incomplete mixing with respect to the capability
of the machine. After determining the proportions of constituents, the mixing was
performed in the following sequence for all the cases. The cement and 70 % of the water
required were mixed together in the mixer for one minute, and after that superplasticizer
and the remaining water were added to the cement paste. This sequence of addition is
preferred because the delayed addition of superplasticizer reduces the amount of
superplasticizer required (see Chapter 2). In this case, the water content of the
superplasticizer was deducted from the water added. The paste was mixed for 2 minutes
at the same speed. The sides of mixer bowl were scraped for 15-30 seconds. Again the
paste was mixed for 2 minutes.
Fig. 3.11 (a ) Hobart mixer with B-flat beater (b) Hobart mixer with D-wire whip
3.4.3 Ball Milling
The study for selecting the mixing method is conducted on cement paste instead of
concrete because of the reasons explained in sections 3.3 and 3.4. However to ensure that
the paste is representative of the paste within the concrete, “ball milling” was done for
64
simulating the concrete mixing through the use of steel balls or glass marbles of 27.5 mm,
17.5 mm and 12.5 mm placed in a drum along with the paste (Figure 3.12).
(a) Los Angeles abrasion testing machine (b) Glass marbles
(c) Mixing of paste with glass marbles (d) Sieving of cement paste
Fig. 3.12 Test set up for ball milling
The steel/glass balls were selected instead of stone aggregates because they do not absorb
water or paste and therefore it would be possible to extract the paste for the flow tests.
Glass marbles were selected because the density (2575 kg/m3) is almost same as that of
stone aggregates. Of course, there will be some differences between the effects of the
balls and the aggregates, as the size, shape and surface texture of these are different.
Cement paste and steel balls/glass marbles were used in the ratio of 1:1 by volume.
Different sizes of steel and glass marbles were selected based on the percentage fraction
of aggregates (Mehta and Monteiro, 2005); 30% were of 25mm size, 40% were of
65
17.5mm and 30%, 12.5mm. The paste was mixed by hand for 1 minute before putting it
into the drum and then the steel balls /glass marbles and superplasticizer were added and
the paste mixed for 4 minutes. The mixing was done by rotating the drum at 32 rpm. The
paste was screened through 4.75 mm and 600 micron sieves before the Marsh cone test.
3.4.4. Results and Discussions
Figure 3.13 shows the Marsh cone flow times for the different mixing methods and
dosages of superplasticizers. In all cases, there is an increase in fluidity with an increase
in the dosage of superplasticizer up to a certain dosage, which is taken as the saturation
dosage (Agullo et al., 1999). The saturation superplasticizer dosages in the case of Hobart
mixer with B-flat beater and ball milling with steel balls and glass marbles are
comparable as seen in Table 3.5. The unit weights of paste prepared are almost same, as
shown in Table 3.5, which implies that air entrainment does not vary for different mixing
methods. The blender mixing gives a higher fluidity but the saturation dosage is also
higher. In addition, the paste prepared with the blender had a temperature that was about
10-12 ºC higher than that prepared with the other methods. The saturation dosages
obtained through grinder mixing and Hobart with D wire whip is comparable. Hand
mixing gave the lowest fluidity. The results of three trials for selection of mixing method
are given in Table A.1 of Appendix A.
66
2.2
Hobart with B flat beater
2.0 Hobart with D wire whip
Blender mixing
1.8 Grinder mixing
log (flow time, sec)
Hand mixing
Ball milling with steel balls
1.6
Ball milling with glass marbles
1.4
1.2
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6

sp/c%
Fig. 3.13 Typical flow time curves obtained with

Marsh cone for different mixing methods
Table 3.5 Optimum superplasticizer dosage and range of

unit weight for different mixing methods
Type of mixing Saturation Unit weight of

superplasticizer paste
dosage % (kg/ litre )
Hobart with B flat beater 0.09 1.97-2.03

Hobart with D wire whip 0.16 1.97-1.99
Blender mixing 0.16 1.94-2.02
Grinder mixing 0.16 1.97-2.05
Hand mixing 0.20 1.93-1.95
Ball mill effect (steel balls) 0.11 1.96-2.00
Ball mill effect (glass marbles) 0.14 1.94-1.98
The mini-slump spread generally increases with an increase in the dosage of
superplasticizer, as shown in Figure 3.14. The points that have been omitted in the figure
correspond to higher dosages where there is bleeding. Generally, it is seen that bleeding
67
occurs when the dosage is well above the saturation dosage and the tendency for it to
occur increases with the intensity of the mixing.
mini-slump spread (mm) 180
160
140
Hobart with B flat beater
Hobart with D wire whip
120 Blender mixing
Grinder mixing
Hand mixing
Ball milling with steel balls
100 Ball milling with glass marbles
0.0 0.1 0.2 0.3 0.4 0.5

sp/c %
Fig.3.14. Typical flow curves obtained with mini-slump for different mixing methods
Considering the different types of mixing studied here, it can be stated that the blender,
Hobart mixer with the flat beater and ball mill provide intensive mixing, with the blender
being the most intensive of the three while the hand mixing, grinder and whip mixing are
not intensive. In the case of intensive mixing, the mixing is expected to yield better
dispersion of the particles and the breaking up of the initial hydration products (e.g., first
growth of ettringite). The higher mixing intensity also appears to decrease the
superplasticizer demand, in terms of the saturation dosage, with the exception of the
blender, which results in a higher saturation point probably due to the higher surface area
of the particles due to the crushing of the hydration products. The effect of a high dosage
of superplasticizer is also more evident in the case of intensive mixing with bleeding
occurring beyond the saturation point. The values of the unit weight (Table 3.5) of the
paste from different mixing methods show that the variation for different dosages is
68
small. This shows that there is not much segregation of paste for different mixing
methods.
In a concrete mixer, it is expected that the cement paste is mixed in an intensive manner.
This was simulated, to a large extent, by the ball mill. Since the behaviour of the paste
mixed with the Hobart mixer with B-flat beater and the ball mill is similar in terms of
saturation dosage and fluidity, it can be concluded that the Hobart mixer leads to a paste
that is representative of that in concrete.
3.4.5 CONCLUSIONS
• The Marsh cone and mini-slump tests are suitable methods for the objective
comparison of the effects of different mixing methods on the fluidity of
superplasticized cement pastes. Both are simple methods to determine the
comparative fluid behaviour and saturation superplasticizer dosage.
• The mixing with Hobart mixer with a flat beater and a ball mill with glass/metal
can be categorized as intensive as in concrete mixing.
• In the case of blender mixing, the mixing is more intensive, leading to a higher
saturation dosage of superplasticizer and more fluidity. The temperature rise of
the paste prepared with this mixing method is 10-12 ºC higher than that with other
methods and causes rapid hydration leading to change in microstructure.
• The results presented indicate that the paste prepared with the Hobart mixer with
flat beater can be considered as representative of the paste of concrete since the
properties are similar to that of paste mixed in a ball mill, with steel/glass balls, in
terms of saturation dosage as well as flowability. Hence, pastes will be prepared
using the Hobart mixer with flat beater for the rest of work in this thesis.
69
3.5. INFLUENCE OF TYPE AND DOSAGE OF SUPERPLASTICIZER ON THE
FLOW BEHAVIOUR OF CEMENT PASTE
In the previous section, Marsh cone and mini-slump tests were used to study the fluidity
of cement paste and for selecting a suitable mixing method for the preparation of cement
paste. These are simple empirical methods to characterize flow behaviour of cement paste
whereas rheological tests are fundamental but more complicated and require skill for
conducting the tests. Hence, the establishment of correlations between these tests helps to
select an appropriate test to study the fluidity of cement paste and to understand the
results from empirical tests better. This study is more relevant considering
incompatibility problems associated with superplasticized cement paste (see Chapter 2),
which necessitate the use of simple tests for studying different combinations of cement
and admixture. As the admixtures mainly affect the flow behaviour of the cement paste
without significant interaction with the aggregate, it seems reasonable to select
admixtures and evaluate their performance through tests on the cement paste.
3.5.1 Testing Details
In the previous section, it was concluded that cement paste prepared with the Hobart
mixer and B flat beater represents that of concrete satisfactorily. Hence, the same mixing
method, with the exception that the superplasticizer addition was delayed further to two
minutes after the mixing of cement and water to get better efficiency. All the material
components were kept in a humidity chamber at a temperature of 27ºC for one day before
the tests in order to minimize the effect of ambient temperature on the flow behaviour.
Marsh cone, mini-slump and viscometer tests were done as explained in Section 3.3
70
3.5.2 Results and discussions
The Marsh cone flow time curves for different types and dosages of superplasticizers are
shown in Figure 3.15, along with the saturation dosages corresponding to the internal
angle of 140°± 10° for each case (following the method of Gomes et al., 2001); the
individual data are given in Table 3.6. The results show that fluidity increases with an
increase in the dosage of superplasticizer up to the saturation point, after which the curves
remain relatively flat.
1.8
1.7 SP saturation dosage
(sp/c%)
1.6 LS-C1 0.25
SNF-S1 0.20
SNF-D1 0.24
1.5 SNF-D2 0.23
SNF-C1 0.16
1.4 SMF-S1 0.23
SMF-C1 0.20
1.3 PCE-S1 0.20
PCE-D1 0.07
1.2
1.1
1.0
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 3.15 Marsh cone flow time for different superplasticizers
The lignosulphonate-based product LS-C1, though not expected to be as effective as the
other products, yields a response that is similar to those of the naphthalenes and
melamines. Most of the admixtures based on naphthalenes and melamines exhibit similar
flow time curves, except for SNF-C1, which yields lower flow times and a comparatively
low saturation superplasticizer dosage of 0.16%. In comparison, the PCE-based
superplasticizers are more effective, with lower saturation dosages than the other
products; PCE-D1 gives the lowest saturation dosage of 0.07%. These trends are in
71
accordance with the basic actions of the different families of superplasticizers (see
Chapter 2).
The mini-slump test results (see Figure 3.16) show that the spread increases with the
increase in dosage of superplasticizer. However, bleeding was observed in the paste
whenever the saturation superplasticizer dosage was exceeded. The mini-slump spread
(i.e., flow) is least for SNF-D2 and highest for PCE-D1. The time taken for a spread of
115 mm at saturation dosage is less than 1 sec.
200 LS-C1
SNF-S1
180 SNF-D1
Mini-slump spread (mm)
SNF-D2
160 SNF-C1
SMF-S1
SMF-C1
140
PCE-S1
PCE-D1
120
100
80
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 3.16 Mini-slump spread for different superplasticizers
In Figure 3.17 Marsh cone and mini-slump spread are plotted together, showing that as
the flow time increases, the mini-slump spread decreases. A better correlation is observed
for pastes with high fluidity.
72
200
180

160
140
120
100
80
5 10 15 20 25 30 35 40 45 50
Marsh cone flow time (sec)
Fig. 3.17 Correlation between Marsh cone flow time and mini-slump spread
A typical set of shear stress-strain rate curves obtained from the rheological tests are
shown in Figure 3.18 (a-c) for pastes with the superplasticizer PCE-D1. At the low
dosage of sp/c = 0.05%, the area within the loading-unloading loop of first cycle is much
higher than the subsequent cycles, reflecting the energy needed for the structural
breakdown of the paste (see Figure 3.18 a). Moreover, it can be seen that the curves are
nonlinear reflecting the shear-thinning response. At the sp/c dosage of 0.1%, which is just
above the saturation dosage, the shear stress levels are remarkably lower, as seen in
Figure 3.18 b, and the stress-strain rate response is practically linear, especially in the
second and third cycles. At both dosages (see Figures 3.18 a&b), it is observed that the
areas within the loading-unloading loop decrease, along with the stress levels, in
subsequent cycles. At the higher superplasticizer dosage of 0.4% (see Figure 3.18 c), the
curves are similar to that of sp/c = 0.1% but with two significant differences; i.e., the
response of the third cycle is above that of the second cycle and the response in the third
73
cycle shows slight shear-thickening behaviour, indicated by a slight increase in the slope
of the stress-strain rate curve. These two aspects can be attributed to the settlement of the
cement particles due to the over dosage of the superplasticizer leading to the stiffening of
the lower part of the paste, as observed visually in this work and in previous works
(Bhatty and Banfill, 1982; Billberg, 1997; Cyr et al., 2000; Papo and Piani, 2004). At
higher strain rates, more of the settled flocs of cement particles are forced into the
suspension, which would increase the stress resulting in the strain-hardening effect.
Responses similar to those presented in Figures 3.17 a-c were observed for all the
admixtures, except that no shear thickening could be seen for LS-C1 and SNF-C1 even at
the dosage of 1%.
90
SP-PCE-D2
80 Dosage-0.05%
70
Shear stress (Pa)
60
50
40 First cycle
Second cycle
30 Third cycle
20
40 80 120 160
Shear rate (1/sec)
(a) Hysteresis cycles at low dosage
74
8
7 SP-PCE-D2
Dosage-0.1%
6
Shear stress (Pa) 5
3
First cycle
2 Second cycle
Third cycle
1
0
40 80 120 160
Shear rate (1/sec)
(b) Hysteresis cycles at optimum dosage
8 SP-PCE-D2
7 Dosage- 0.4%
6
Shear stress (Pa)
5
4
3
First cycle
2 Second cycle
Third cycle
1
0
40 80 120 160
Shear rate (1/sec)
(b) Hysteresis cycles at excess dosage
Fig. 3.18 Typical graphs showing three hysteresis cycles

for a PCE based superplasticized cement paste at
(a) low (b) optimum and (c) excess dosages
75
The shear stress-strain data were fitted in each case with the Bingham and Herschel-
Bulkley models, and the parameters obtained are given in Table 3.6. It can be observed
that, in general, an increase in the dosage of superplasticizer reduces the yield stress and
plastic viscosity of the Bingham model, which is attributed to the better dispersion of
cement particles (Rixom and Mailvaganam, 1999). The Bingham yield stress tends to
reduce to zero, as the superplasticizer dosage increases; the drop in yield stress is sharpest
in the case of the PCE-based products, confirming the results of previous works (Puertas
et al., 2005). The plastic viscosity decreases initially with an increase in the dosage of
superplasticizer but remains practically constant after a certain point. In the case of the
Herschel-Bulkley model, the yield stress ( τ 0 ) has the same trend as that of the Bingham
model though the values of the former are lower at smaller superplasticizer dosages and
vice versa at higher dosages. These differences are attributed to the change in the
rheological nature of the paste from shear-thinning to slight shear-thickening, as
discussed earlier. The consistency and power index (n) of the Herschel-Bulkley model are
interdependent and hence well defined trends could not be identified for the individual
parameters. Note that for shear-thinning materials, n < 1, and that for shear-thickening
materials, n > 1.
76
Table 3.6 Rheological characteristics and Marsh cone flow times
Type sp/c Bingham Herschel Bulkley Parameters Flow time

of sp % Parameters predicted by the
Nguyen et al.
approach
flow time (sec)

Experimental
τ0 μp τ0 K n
(Pa Sn)
Bing-ham
(Pa) (Pa S) (Pa)
Herschel
Bulk-ley
material
(sec)
(sec)
0.1 20 12.2 0.122 11.0 0.319 0.817 18 16
0.2 14 5.8 0.089 5.0 0.169 0.880 14 13
0.3 13 2.0 0.059 2.0 0.053 1.017 12 12
LS-C1 0.4 13 1.0 0.050 0.7 0.077 0.920 11 10
1 12 0.5 0.068 0 0.135 0.876 12 11
0.2 13 16.2 0.143 12.29 0.668 0.723 20 16
SNF- 0.3 11 1.8 0.055 1.66 0.066 0.965 11 11
S1 0.4 11 0.0 0.033 0.08 0.030 1.020 10 10
1 11 0.1 0.031 0.12 0.029 1.010 10 9
0.1 45 18.31 0.197 14.34 0.657 0.783 27 22
0.2 20 12.13 0.139 11.69 0.265 0.872 20 17
SNF- 0.3 13 2.88 0.074 3.09 0.059 1.041 13 13
D1 0.4 14 0.08 0.049 0.18 0.043 1.027 11 11
0.5 14 0.11 0.043 0.47 0.021 1.132 11 11
1 14 0.09 0.033 0.45 0.013 1.177 10 11
0.1 29 16.99 0.197 20.46 0.034 1.328 27 30
0.2 17 6.94 0.088 7.56 0.100 0.959 14 14
SNF- 0.4 14 0.55 0.042 0.98 0.017 1.168 11 12
D2 0.6 13 0.37 0.041 0.67 0.022 1.113 11 11
1 13 0.13 0.045 0.29 0.034 1.049 11 11
0.1 11 6.72 0.057 5.12 0.269 0.722 12 10
0.2 10 1.25 0.042 1.95 0.008 1.300 10 12
SNF- 0.3 10 0.72 0.029 1.20 0.006 1.303 9 11
C1 1 10 0.24 0.023 0 0.050 0.876 9 8
0.1 62 20.39 0.177 3.24 4.914 0.43 26 16
SMF- 0.2 18 6.89 0.076 6.33 0.103 0.948 14 14
S1 0.25 16 2.36 0.098 1.99 0.128 0.950 15 14
0.3 15 1.54 0.078 1.34 0.109 0.950 13 13
1 15 0.77 0.048 0.98 0.033 1.06 11 11
0.1 >120 33.14 0.199 20.53 2.514 0.559 36 24
0.2 16 13.44 0.128 6.36 1.07 0.638 19 14
SMF- 0.3 13 2.88 0.074 3.09 0.059 1.040 13 13
C1 0.5 12 0 0.039 0.08 0.025 1.090 10 11
1 12 0 0.037 0.13 0.020 1.115 10 11
0.05 26 17.55 0.155 15.93 0.310 0.873 23 20
0.1 16 11.07 0.089 16.01 2.399 0.558 15 20
PCE- 0.2 12 0.51 0.073 0.68 0.010 1.260 13 12
S1 0.3 11 0 0.034 0.44 0.007 1.270 10 11
1 12 0 0.034 0.39 0.006 1.344 10 12
0.05 18 24.45 0.214 1.65 6.910 0.410 32 19
0.1 12 0.01 0.035 0.37 0.014 1.168 10 11
PCE- 0.2 12 0 0.035 0.48 0.006 1.304 10 11
D1 0.4 12 0 0.036 0.53 0.006 1.310 10 11
1 14 0 0.047 0.07 0.020 1.158 10 12
77
3.5.3 Correlations between the Results from Empirical Tests and the Viscometer
In order to study the relationships between the results from empirical tests (i.e., the Marsh
cone and mini-slump tests) and the rheological parameters of cement pastes, the data
presented above are compared in this section. It can be seen in Table 3.6 that the Marsh
cone flow time increases as the yield stress and plastic viscosity increase. More
importantly, the variation of the Marsh cone flow time with the superplasticizer dosage
has the same trend as those exhibited by the yield stress and plastic viscosity, as also
illustrated in the typical plots of Figure 3.19 (shown for SMF-C1). It can be seen that all
three parameters decrease with an increase in superplasticizer dosage until the saturation
dosage and remain practically constant after that. Figure 3.19 also shows the trend of the
mini-slump spread, where it is seen that the spread increases with the superplasticizer
dosage until the saturation dosage, after which it is practically constant. These trends can
be generalized to all the paste systems studied in this work, and are in agreement with the
results of previous works (Gettu et al., 1997; Monte and Figueiredo, 2003;
Schwartzentruber et al., 2006). The correlation between Marsh cone flow time with
rheological parameters for other superplasticized pastes are given in Figures A.1 and A.2
of Appendix A.
78
mini slump spread (mm)
SMF-C1 SMF-C1 200
Yield stress
Plastic viscosity
160
Plastic viscosity (cP)

1.24 200 40

120
1.2 180
30
Yield stress (Pa)

80
1.16 160
20
40
1.12 140
10 0
1.08 120
1.04 100 0
0 0.2 0.4 0.6 0.8 1
sp/c%
Fig. 3.19 Comparison of data from different tests
The Roussel-Coussot model (Roussel and Coussot, 2005) can be used for estimating the
yield stress from the measured conical slump. The model is applicable for different flow
regimes based on the height (H) and radius (R) of the slump cone. According to this
model, the spread distance (R) can be expressed as a function of the yield stress (τc) and
material volume (Ω) and density of material (ρ) as expressed in the equation 3.3.
225 ρ g Ω 2
τc = (3.3)
128 π 2 R 5
The prediction of yield stress from mini-slump spread using the Roussel-Coussot model
is given in Figure 3.20. The result show that even though the graph shows a good
correlation, individual experimental data and predicted values are not comparable.
79
40
Experimental
35 Predicted
30
Yield stress (Pa)

25
20
15
10
5
0
80 100 120 140 160 180 200
Fig. 3.20 Prediction of yield stress from mini-slump spread using Roussel-Coussot model
It is clear that the saturation dosage estimated with the procedure of Gomes et al. (2001)
from the Marsh cone test gives a reliable approximation of the superplasticizer dosage
beyond which the plastic viscosity and yield shear stress remain constant. On the other
hand, the mini-slump test on the tested pastes does not seem to give as reliable an
estimate of the saturation dosage due to bleeding that may occur at higher dosages. From
the data obtained, it does not seem possible to identify unique relations between the
Marsh cone flow time and mini-slump spread with either the plastic viscosity or yield
shear stress, as previously concluded (Ferraris et al., 2001). Nevertheless, it was shown
that the use of flow times from two Marsh cones of slightly different geometries can be
used to determine the rheological characteristics of a paste (Roussel and Roy, 2005).
Conversely, the Marsh cone flow time can be obtained by adequately modeling the flow
and material behaviour, as in the semi-analytical approach, where the Marsh cone flow
80
time is obtained from fluid mechanics principles and finite volume method (Nguyen et
al., 2006).
Nguyen et al. (2006) suggested a semi analytical approach by for relating rheological
parameters to Marsh cone flow time, where it is given that the flow characteristics of
cement paste can be described by Herschel-Bulkley model by three parameters τ 0 , k and
n which relates the shear stress to shear rate. As per this method, the flow time is related
to rheological parameters through the equation 3.4.
H2
( R + H t Tanφ ) 2
t= ∫
H1 Vβ R 2
dH t (3.4)
where R is the radius of the nozzle, Ht is height of fluid in the Marsh cone at any time,
φ is the angle of the Marsh cone, V β is the corrected velocity at the exit of the nozzle.
t is the time taken to lower the level of paste in the Marsh cone from H1 to H2.
The flow times for the different pastes studied here have been predicted using this
approach considering both the Bingham and Herschel-Bulkley models for the paste and
given in Table 3.6. Note that the density of paste was taken as 1.9 kg/litre in the
predictions. A Matlab programme was used for solving the equations and for predicting
the flow time; the programme code is given in Appendix B.
The results in Table 3.6 show that the predictions of the Nguyen et al. approach are, in
general, good for both the rheological models. This confirms that the Marsh cone flow
time can be determined through appropriate modeling of the material response and the
flow behaviour.
81
3.5.4 Influence of Type and Dosage of the Superplasticizer on the Non-Newtonian
Characteristics of Cement Paste
The non-Newtonian characteristics of the superplasticized cement paste depend on the
type of superplasticizer and its dosage and generally rheological model parameters are
used to represent the same, as explained in the previous Sections. However, they do not
represent the transition in the fluid nature based on the dosage of the superplasticizer.
Therefore, steady shear, creep and recovery, and stress relaxation tests were also
conducted in order to develop an understanding of the transition of the behaviour of these
mixtures from viscoelastic to non-Newtonian to linearly viscous, as the superplasticizer
dosage is increased. The test results show that as the superplasticizer dosage increases, in
general, there is a transition in the behaviour from viscoelastic to viscous fluid behaviour.
It was found that the dosages determined based on the Marsh cone tests were close to the
dosages at the transition to the linearly viscous behaviour. The experimental details and
the test results are given in detail in Appendix C.
3.5.5 Change of Flow behaviour with Time
The flow properties of cement paste, especially that of superplasticized cement paste, can
change significantly after mixing. This is attributed to hydration, intercalation of
superplasticizers, chemical degradation of the polymer due to the high pH of the aqueous
solution and physical relaxation or change in molecular structure due to adsorption
(Tattersall and Banfill, 1983; Justnes et al., 2003). In general, both yield stress and plastic
viscosity tend to increase with time (Roy and Asaga, 1980). Consequently, the Marsh
cone flow time tends to increase with time (Agullo et al., 1999) whereas the mini-slump
spread decreases (Kantro, 1980; Lim et al., 1999).
82
The change in paste flow behaviour has been studied here by comparing the test results
obtained for the same materials just after mixing and 60 minutes later. During the interim
period the paste was maintained in an air-tight container at 27ºC. The paste was taken out
of the container and mixed for 15 seconds before performing the 60-minute tests. The
variation of the flow parameters with time using Bingham model can be seen from the
results given in Table 3.7.
Table 3.7 Test data for pastes with different superplasticizers at 0 and 60-minutes
Flow time Mini-slump spread Yield stress Plastic viscosity

Type of (sec) (mm) (Pa) (Pa s)
sp sp/c %
0 60 0 60 0 60 0 60
min min min min min min min min
0.1 20 21 93 90 12.20 13.50 0.122 0.132
0.2 14 16 153 123 5.81 7.59 0.089 0.097
LS-C1 0.3 13 13 162 143 2.0 3.15 0.059 0.075
0.4 13 14 163 163 1.0 0.35 0.050 0.064
1 12 13 164 162 0.5 0.23 0.068 0.049
0.2 13 26 143 90 16.21 24.61 0.143 0.185
SNF-S1 0.3 11 11 151 150 1.80 5.91 0.055 0.066
0.4 11 11 165 163 0.0 0.21 0.033 0.054
1 11 11 170 178 0.09 0.10 0.031 0.025
0.1 45 * 82 * 18.31 20.23 0.197 0.241
0.2 20 70 96 * 12.13 16.61 0.139 0.178
SNF-D1 0.3 13 13 127 115 2.88 5.98 0.074 0.089
0.4 14 17 bleeding 141 0.08 0.09 0.049 0.083
0.5 14 18 bleeding 164 0.11 0 0.043 0.045
1 14 13 bleeding 163 0.09 0.12 0.033 0.035
SNF- D2 0.1 29 * 87 * 16.99 25.52 0.197 0.194

0.2 17 39 106 87 6.94 12.10 0.088 0.149
0.4 14 14 bleeding 148 0.55 0 0.042 0.043
0.6 13 13 bleeding 176 0.37 0.25 0.041 0.039
1 13 12 bleeding 160 0.13 0.07 0.045 0.038
SNF-C1 0.1 11 14 126 95 6.72 7.24 0.057 0.046

0.2 10 11 152 128 1.25 6.69 0.042 0.064
0.3 10 10 168 168 0.72 0.77 0.029 0.046
1 10 9 173 178 0.24 0.17 0.023 0.035
0.1 62 * * 20.39 * 0.177 *
SMF-S1 0.2 18 44 102 90 6.89 2.39 0.076 0.092
0.25 16 22 115 109 2.36 3.77 0.098 0.101
0.3 15 18 122 113 1.54 8.97 0.078 0.065
1 15 18 167 161 0.77 0 0.048 0.043
83
Table 3.7 (Contd..)
0.1 >120 * * * 33.14 * 0.199 *
SMF-C1 0.2 16 26 108 103 13.44 30.92 0.128 0.139
0.3 13 16 132 112 2.88 7.71 0.074 0.100
0.5 12 12 181 169 0 0 0.039 0.042
1 12 11 177 172 0 0 0.037 0.041
0.05 26 * 86.3 * 17.55 * 0.155 *

PCE-S1 0.1 16 * 110 * 11.07 23.67 0.089 0.162
0.2 12 14 bleeding 157 0.51 0.36 0.073 0.068
0.3 11 13 bleeding 179 0 0.08 0.034
0.5 11 14 bleeding 192 0 0 * 0.035
1 12 12 bleeding bleeding 0 0.034 0.034
0.05 18 * 97 * 24.45 * 0.214 *

PCE-D1 0.1 12 16 166 147 0.01 2.37 0.035 0.079
0.2 12 14 170 bleeding 0 0.09 0.035 0.043
0.4 12 14 199 bleeding 0 0 0.036 0.036
1 14 16 193 bleeding 0 0 0.047 0.054
* could not be determined
Generally, in all the compositions tested, the rheological parameters increase with time
after mixing, with the increase in yield stress being more pronounced than that of the
plastic viscosity. Also, the rate of increase is more significant at superplasticizer dosages
less than the saturation point. Similarly, when the dosages are low, the Marsh cone flow
time increases with age as seen when the 60-minute data are compared with those
obtained immediately after mixing (See Figure 3.21).
1.8 SMF-S1
flow time (0-minutes)
1.7 flow time 60-minutes
1.6
1.5
1.4
1.3 Saturation dosage

1.2
1.1
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 3.21 Flow time curve showing the determination of optimum
superplasticizer dosage according to Aïtcin, 1998
84
Note that when the flow behaviour at 60 minutes is more significant (for a particular
application) than the immediate behaviour, the saturation point could be defined based on
the flow time curves at 60 minutes, or any other critical time (See Table A.2 of Appendix
A). Nevertheless, Aïtcin (1998) suggests that the flow time and saturation dosage at 60
minutes should not differ significantly from those obtained immediately for a cement-
superplasticizer combination to be considered compatible.
3.5.6 Summary
The following conclusions can be made from the study performed on the flow-related
properties of superplasticized cement pastes.
• The flow time determined from Marsh cone test is a good indication of the
relative fluidity of superplasticized paste and the saturation dosage obtained from
the same test is appropriate for the selection of the superplasticizer.
• Both the Herschel-Bulkley and Bingham models fit the experimental data from
the viscometer study satisfactorily. It is observed that the nature of flow in
superplasticized paste varies slightly with the dosage of superplasticizer, as
follows. At lower dosages, nonlinear shear thinning is generally observed; around
the saturation dosage the response follows the Bingham model; and at higher
dosages, there may be some shear thickening. The yield stress values obtained
with both the models have the same trend with respect to superplasticizer dosage,
though the values from the Bingham model tend to be higher at smaller
superplasticizer dosages. Even though shear-thickening nature of paste is better
represented through the Herschel Bulkley model, the Bingham model represents
the behaviour of normal pastes well. The Marsh cone flow times could be
85
predicted satisfactorily through the Nguyen et al. approach with the parameters of
both models.
• An increase in superplasticizer dosage leads to a decrease in the yield stress,
plastic viscosity and Marsh cone flow time, and an increase in mini-slump spread,
as long as the dosages are below the saturation point. Beyond the saturation
dosage, these parameters are practically constant.
• All the nine superplasticizers tested resulted in pastes with good fluidity over 60
minutes when the corresponding saturation dosages were used. It is concluded
that the cement-superplasticizer combinations studied here are all compatible as
far as the flow behaviour is concerned.
86
CHAPTER 4
CORRELATIONS BETWEEN THE FRESH AND

HARDENED PROPERTIES OF SUPERPLASTICIZED
PASTE, MORTAR AND CONCRETE
4.1 GENERAL
Optimization and characterisation of the superplasticized paste is considered as a first
step in the design of high performance concrete, as the paste phase is mainly
responsible for workability, mechanical properties and durability of concrete. The
incorporation of superplasticizers results in the production of high performance
concrete due to several beneficial effects associated with it in the fresh and hardened
state of concrete. However, the characteristics of cement paste and type and dosage of
superplasticizers could sometimes compromise the benefits of incorporating the
superplasticizer because of compatibility problems, leading to slump loss, retardation
in setting time or excessive heat of hydration (Rixom and Mailvaganam, 1999).
Therefore, for the optimization of the composition of high performance concretes
(HPC), the optimization of the paste and mortar phases are essential and should
include the choice of superplasticizer type and dosage in order to obtain the desired
fresh and hardened properties of concrete. The methodology for the determination of
the saturation dosage of superplasticizer through the Marsh cone test and its influence
on the loss of fluidity has been explained in the third chapter.
Tests involving the paste are simpler and require much less effort than tests of
concrete. Nevertheless, results obtained from paste tests need to be correlated to the
behaviour of concrete to take the possible influence of aggregates and the aggregate-
paste interfaces into account (Ferraris and Gaidis, 1992; Toralles-Carbonari et al.,
1996; Roncero et al., 2000; Tang et al., 2001; Giaccio and Zerbino, 2002; Banfill,
2003; Laskar and Talukdar, 2007). For a particular particle size distribution and
volumetric fraction of aggregate, the flow properties of concrete are related with
rheology of paste, which is generally studied through simple flow or viscometer tests
as explained in Chapter 3 (Ferraris and Gaidis, 1992; Toralles-Carbonari et al., 1996;
Roncero et al., 2000; Lachemi et al., 2007).
The work presented in this chapter deals with the comparison of the results of tests on
superplasticized cement pastes with the behaviour of mortars and concretes. The
fluidity of mortar is studied through Marsh cone and flow table spread. Finally, an
attempt is made to correlate the cement paste flow behaviour with that of concrete as
characterized through the slump, setting time and compressive strength.
4.2 EXPERIMENTAL DETAILS
4.2.1 Materials
The physical and chemical characteristics of cement and four types of
superplasticizers used for the correlation study are same as given in Section 3.2. River
sand (of 0-4.75 mm grain size range) and crushed granite coarse aggregate (of 4.75-10
mm and 10-20 mm grain size ranges), satisfying the requirements of IS 2386 (2007),
are employed in the preparation of concrete.
The sieve analysis was done according to the Indian Standard IS 2386 (2007) and the
test results on fine and coarse aggregates are given in Figures 4.1 and 4.2. The sand
has a fineness modulus of 2.29 and conforms to grading zone II as per IS 383 (2002).
The properties of aggregates are given in Table 4.1. The specific gravity, bulk density
and coefficient of water absorption were determined for each aggregate according to
IS 2386 (2007). The absorption coefficients for the 10-20 mm, 4.75-10 mm and 0-
4.75 mm aggregates are 0.4%, 0.5% and 1.5%, respectively.
88
100
Percentage passing
80
60
40
20
0
0.1 1 10
Sieve size (mm)
Fig. 4.1 Particle size distribution of fine aggregates (0-4.75 mm)
10 mm down
20 mm down
100
Percentage passing
80
60
40
20
0
1 10 100
Sieve size (mm)
Fig. 4.2 Particle size distribution of coarse aggregates (4.75-20 mm)
Table 4.1 Summary of properties of aggregates
Properties Sand Coarse aggregates
4.75-10 mm 10-20 mm
Specific gravity 2.62 2.70 2.72
Water absorption (%) 1.5 0.5 0.4
Fineness modulus 2.29 5.71 7.51
Bulk density (kg/ m3) 1670 1568 1618
Flakiness index (%) - - 16.1
Elongation index (%) - - 29
Angularity number - - 8
Crushing Value (%) - - 28
Impact Value (%) - - 30
89
4.2.2 Test Procedures
4.2.2.1 Tests on Paste
The Marsh cone, mini-slump and viscometer tests, as explained in Chapter 3, were
conducted for understanding the flow behaviour of the paste. The Marsh cone test
results for four superplasticized pastes, already presented in Chapter 3, are
summarized in Section 4.3.1.1 for making the comparison of flow behaviours of
paste, mortar and concrete possible.
4.2.2.2 Tests on Mortar
Marsh cone and flow table tests (as per IS 1199 (2004) and ASTM C109 (1998)) were
used for understanding the fluidity of mortar. A Marsh cone with an aperture diameter
of 12.5 mm was used for the mortars. In addition, the consistency of mortars has been
determined using a flow table. Here, a hollow truncated metallic cone with a base
diameter of 100 mm, top diameter of 70 mm and a height of 50 mm was used. It was
placed on the table and filled with mortar, the cone was removed and the mortar was
jolted 25 times in 25 seconds causing it to spread. The mean final base diameter was
taken as the flow of consistency. In order to account for the higher fluidity of
superplasticized cement mortar, the flow table was modified by increasing its
diameter using a polycarbonate plate, as shown Figure 4.3. This is in accordance with
the test procedure of Domone (2006) for studying the flow behaviour of
superplasticized mortars.
Fig. 4.3 Flow table test for mortar

90
4.2.2.3 Tests on Concrete
In order to correlate the results obtained from paste tests to the behaviour of concrete,
tests were conducted on concrete with four superplasticizers studied for paste and
mortar. The fresh behaviour of concrete has been characterized using the slump test
and the flow table test, as per IS 1199 (2004), as shown in Figure 4.4. For concrete,
the following mix proportions were used per m3: 450 kg cement, 675 kg sand, 454 kg
of 20 mm coarse aggregate and 680 kg of 10 mm coarse aggregate with a w/c of 0.35
(mix design as per IS 10262 (2004)). Note that the sand-cement ratio was maintained
as 1.5 in both the mortar and concrete. In all the cases, the aggregates were assumed
to be in a saturated surface condition and, therefore, the water needed for saturating
them was added according to the absorption coefficients given in Table 4.1. The
humidity of the aggregates was compensated for, after its determination using the
ASTM microwave method (ASTM D 4643 (1987)). The fresh density of concrete
with different superplasticizers was also measured to account for the probable air
entrainment due to superplasticizers. In addition to the slump and flow table tests, for
all superplasticizer dosages, the loss of slump and setting time of concrete at the
saturation dosage of superplasticizer were determined. Also, the compressive
strengths at 3, 7 and 28 days were evaluated for different dosages of superplasticizers.
(a) Slump test
(b) Flow table test on concrete

Fig. 4.4 Tests for workability of concrete
91
4.2.3 Mixing Methods for Paste, Mortar and Concrete
The pastes and mortars were prepared in the Hobart mixer (1/6 hp) with the B flat
beater at low speed (i.e., shaft speed of 139 rpm and planetary speed of 61 rpm),
based on the study explained in Chapter 3. For mixing of the paste, the cement and
70% of the water required were mixed for 2 minutes; after that, the superplasticizer,
along with the remaining water (to ensure better dispersion of superplasticizer), was
added to the cement paste; the mixing was stopped and the sides of mixer bowl were
scraped (for 15-30 seconds); and the paste was again mixed for 3 minutes.
In the case of mortar mixing, oven-dried sand was mixed with the water necessary for
its saturation for 30 seconds. Then the cement was added and mixed with the wetted
sand for 15 seconds to ensure the homogeneity of solid particles before the
incorporation of water. The same sequence as in cement paste was followed for the
addition of water and superplasticizer.
For the concrete, the materials were dry mixed and the subsequent mixing sequence
for the addition of water and superplasticizer was the same as that for paste; the total
mixing time was 5 minutes.
A water-cement ratio (w/c) of 0.35 was employed throughout this study. Also, note
that all materials, including distilled water, that were used in the paste and mortar
mixes, were kept at a temperature of 27° C in an environmental chamber for at least
24 hours before mixing. This was done to limit the influence of the ambient
temperature on the results.
92
4.3 RESULTS AND DISCUSSIONS
4.3.1 Comparison of the Flow Behaviour of Paste, Mortar and Concrete
4.3.1.1 Flow Behaviour of Paste
Figure 4.5 shows the flow curves for different superplasticized pastes obtained with
the Marsh cone immediately after mixing (0 minutes), along with the corresponding
saturation dosages; when the flow time is higher, fluidity is less. The saturation
dosages and the corresponding flow times are given in Table 4.2. The PCE based
superplasticizer performs better than the other superplasticizers in terms of fluidity as
well as the saturation dosage (0.07%), as expected. The pastes with low dosages of
SMF-S1 exhibit high flow times indicating poorer dispersion of the cement particles
at such dosages. Also, at saturation, superplasticizers have different flow times
indicating their different effectiveness.
1.8
Type of SP saturation dosage
1.7 (sp/c%)
SNF-S1 0.20
1.6
SNF-D2 0.23
1.5 SMF-S1 0.23
PCE-D1 0.07
1.4
1.3
1.2
1.1
1.0
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 4.5 Marsh cone flow time curves for superplasticized cement paste
93
Table 4.2 Summary of Marsh cone data and concrete cost
Type Marsh cone test results *Dosage Unit *Cost of

of sp of sp cost/kg materials
Paste Mortar needed (Rs./kg) for
Saturation Flow Saturation Flow (kg/m3) concrete
dosage of time at dosage of time at (Rs./m3)
sp saturation sp saturation
(sp/c%) (sec) (sp/c%) (sec)
SNF- 0.20 13 0.22 13 0.23 35 2538
S1
SNF- 0.23 17 0.23 22 0.25 50 2570
D2
SMF- 0.23 16 0.30 18 0.37 60 2692
S1
PCE- 0.07 12 0.07 11 0.08 160 2616
D1
*for a mix with a slump of 100 mm
4.3.1.2 Flow Behaviour of Mortar
The results obtained from Marsh cone test of mortar are shown in Figure 4.6 and in
Table 4.2. The fluidity of the PCE based superplasticized cement mortar at low
dosages is higher than that with the SNF and SMF based superplasticizers. The PCE
yields the highest fluidity with a relatively low saturation dosage whereas the SMF
gives the least fluidity and the highest saturation dosage compared to the other
products. This is because, even at low dosages, better dispersion is produced due to
steric repulsion in the case of the former. Also, it has been observed that high dosages
(i.e., above the saturation point) of the SNF with retarder (i.e., SNF-D2) lead to
bleeding due to the loss of cohesion in the paste. The flow table test results are given
in Figure 4.7, where it can be seen that the spread generally increases with an increase
in the dosage of superplasticizer up to the saturation dosage and remains practically
constant beyond that; all the mortars exhibited spreads of 250-300 mm at saturation.
94
1.5 SNF-S1
SNF-D2
SMF-S1
1.4

PCE-D1
1.3
1.2
1.1
0.0 0.2 0.4 0.6 0.8

sp/c %
Fig. 4.6 Marsh cone flow times of cement mortar
300
flow table spread (mm)
SNF-S1
SNF-D2
250 SMF-S1
PCE-D1
200
150
100
0.0 0.2 0.4 0.6 0.8
sp/c %
Fig. 4.7 Flow table spreads for superplasticized mortars
Considering that both the flow table spread and the inverse of the Marsh cone flow
time are measures of fluidity, it can be seen that the trends obtained in the results with
increasing dosage are similar for all the mortars studied. Moreover, the saturation
dosages obtained in the mortars are comparable with those obtained from the test
results of paste.
95
4.3.1.3 Flow Behaviour of Concrete
In order to correlate the behaviour of cement paste and mortar to that of concrete and
to validate the test of cement paste as a tool for the mix design of high performance
concrete, tests were also conducted on concrete with four superplasticizers. The flow
behaviour of concrete, with the same paste and mortar compositions as those
discussed in the previous sections, has been characterized using the slump and flow
table tests as per IS 1199 (2004). The flow table test, which is similar in principle to
the flow table test for mortar, was used to measure the flowability of concrete and to
observe segregation. Concrete was moulded within the 120 mm high hollow frustum
of a cone with bottom and top diameters of 250 mm and 170 mm, respectively. After
removing the mould, the spread diameter was measured in four perpendicular
directions after jolting for 15 times in 15 seconds. The results of the two tests on fresh
concrete are shown in Figures 4.8 and 4.9. The data for all the concretes reflect an
increase in slump (see Fig. 4.8) as the superplasticizer dosage increases, as expected.
In the case of the naphthalenes, it can be seen that SNF-S1 gives a relatively low
slump at its saturation dosage (0.2%) but the slump increases rapidly just beyond this
dosage, whereas SNF-D2 gives a steady increase in slump with an increase in dosage
though the workability achieved at higher dosages is not as much as with SNF-S1. In
comparison, the slump values obtained with SMF-S1 are lower for similar dosages of
the other superplasticizers. On the other hand, a fluid concrete with a slump of more
than 100 mm is obtained for PCE-D1 even at the dosage of 0.1%.
It can be observed in Figure 4.8 that in order to obtain a slump of 100 mm, the
required dosage of superplasticizer is 0.08% for the PCE, about 0.25% for the
naphthalenes and 0.37% for the SMF based superplasticizer. The material costs for
concretes possessing this slump value have been estimated and are presented in
96
Table 4.2. The comparison of the costs shows that a low-cost superplasticizer need not
always result in cost-effective concrete as the dosage of the superplasticizer for
optimum fluidity is also an important factor (Gettu et al., 2006). The results of the
flow table spread tests are generally in accordance with those of the slump tests,
exhibiting an increase in spread with an increase in the dosage of superplasticizer (see
Figure 4.9). Again, the PCE based product performs best and the SMF based product
is the least effective.
200
Concrete slump (mm)
150
100
SNF-S1
SNF-D2
50
SMF-S1
PCE-D1
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
sp/c %
Fig. 4.8 Variation of concrete slump with dosage of superplasticizer
420
400
Flow table spread (mm)
380
360
340
SNF-S1
SNF-D2
320 SMF-S1
PCE-D1
300
0.0 0.1 0.2 0.3 0.4 0.5 0.6
sp/c %
Fig. 4.9 Variation of flow table spread with dosage of superplasticizer
97
Note that the data from the slump and flow table tests have not been plotted for cases
where segregation or bleeding is observed. It is seen that there is a high tendency for
segregation and/or bleeding when the superplasticizer dosage is much higher than the
saturation dosage, and almost certain to occur when the dosage is close to twice that
of saturation. For example, the concrete with a 0.3% dosage of the PCE showed
severe bleeding, segregation and stiffening within 5 minutes after mixing. The
stiffening seen when the superplasticiser is overdosed could be attributed to shear
thickening caused by the excess of superplasticizer in the aqueous phase that could
interlock leading to the stiffening (Papo and Piani, 2004). This has also been observed
in cement pastes in the present study, especially at high dosages.
4.3.2 Correlation between the Fluidity of Paste, Mortar and Concrete
In Figures 4.10-4.13, the results from the Marsh cone tests of pastes and mortars are
plotted along with the slump values of concrete as a function of the superplasticizer
dosage. It can be seen that the Marsh cone flow time curves of paste and mortar are
generally parallel and the saturation dosages are comparable, as previously observed
by other researchers (Roncero et al., 1999; Banfill, 2003; Lachemi et al., 2007). The
test results of concrete with SNF-S1 and SMF-S1 indicate a low workability at the
saturation dosage of paste obtained from the Marsh cone tests whereas in the cases of
SNF-D2 and PCE-D1 the saturation dosages yield medium workability in the
concrete. In all the cases, a dosage slightly higher than saturation provides good
workability. This can be attributed to the absorption of part of the superplasticizer by
the fines in the crushed coarse aggregate This is also confirmed through the
adsorption studies; the absorbance of the superplasticized paste with aggregate was
higher than that of superplasticized paste (the details of the test is given in Chapter 6).
Consequently, the amount of superplasticizer available in the cement paste decreases
and maximum dispersion of cement particles is not achieved in the concrete at the
98
same dosage as in the paste tests. Therefore, further addition of superplasticizer is
required for high slump values in the concrete. Nevertheless, the saturation dosage
obtained from paste studies can be used as a guideline for the selection of
superplasticizers and the dosage needed in concrete, especially when the gravel or
sand does not have large amounts of fines, where it can be expected that the saturation
dosages obtained from the paste tests will yield good workability in the concrete, as
reported by Roncero et al. (1999), and Hidalgo et al. (2008).
1.40
SNF-S1
200
1.35
1.30
Paste flow time 150
Slump (mm)
Mortar flow time
1.25 Concrete slump
100
1.20
1.15 50
1.10
0
0.1 0.2 0.3 0.4
sp/c %
Fig.4.10 Correlation between paste, mortar and concrete
in terms of flow time and slump for SNF-S1
1.50 180
SNF-D2
1.45 160
1.40 140
1.35 Paste flow time 120

Mortar flow time
Slump (mm)
1.30 concrete slump 100

80
1.25
60
1.20
40
1.15
20
1.10
0
0.0 0.2 0.4 0.6 0.8
sp/c%
Fig. 4.11 Correlation between paste, mortar and concrete
in terms of flow time and slump for SNF-D2
99
1.8 SMF-S1
200
1.7
1.6

150
Paste flow time
Slump (mm)
1.5 Mortar flow time
Concrete slump 100
1.4
1.3
50
1.2
1.1
0
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
in terms of flow time and slump for SMF-S1
200
1.25 PCE-D1 180
Paste flow time
Mortar flow time 160
Concrete slump
1.20 140
120
Slump (mm)
1.15 100
80
1.10 60
40
1.05 20
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6
sp/c %

in terms of flow time and slump for PCE-D1
Figure 4.14 shows that no clear correlation exists between the yield stress and plastic
viscosity of paste with the slump of concrete. Nevertheless, some observations can be
made regarding the data obtained. It appears that a yield stress of less than 5 Pa and
plastic viscosity of less than 0.05 Pa s is required to achieve a slump of more than
100 mm. This is in accordance with the reported results of Lachemi et al. (2007), and
100
Hidalgo et al. (2008). It also appears that the concrete can be categorized as low
(<25mm) medium (25-50mm) and high slump (>100 mm) where the values of the
yield stress and plastic viscosity of paste have different ranges as indicated in
Figure 4.14.
24 Slump vs Yield stress

Slump vs Plastic viscosity 0.20
Plastic viscosity (Pa s)

20
0.16
Yield stress (Pa)
16
0.12
12
0.08
8
4 0.04
0 0.00
0 50 100 150 200
Concrete slump (mm)
Fig. 4.14 Correlation between slump of concrete, and the

yield stress and plastic viscosity of paste
4.4. LOSS OF FLUIDITY
4.4.1. Loss of Fluidity of Paste
The change in the flow behaviour of paste with time has been studied by comparing
the results obtained just after mixing and 60 minutes later. During the interim period,
the paste was maintained in an air-tight container at 27ºC, after which it is taken out
of the container and re-mixed for 15 seconds before performing the 60-minute tests.
The Marsh cone flow-time curves obtained after 60 minutes are given in Figure 4.15
(along with the data obtained immediately after mixing plotted with dashed lines for
reference). It can be seen that the curves for the 60-minute tests have the same trends
as the curves obtained immediately after mixing (denoted hereafter as 0-minute tests),
with a decrease in flow time with increase in the superplasticizer dosage upto the
101
saturation dosage, beyond which the flow time does not change much or may even
increase slightly (as in the cases of the SMF and PCE based superplasticizers) due to
settling of the cement particles in the paste. As before (see Section 3.5.2), it can be
observed that the saturation dosages differ with the type of superplasticizer (see
values in the inset of the figure).
More importantly, when the 60-minute plots are compared with those obtained at 0
minutes, it can be observed that there is a shift to the right for all the combinations
due to higher flow times at 60 minutes, indicating loss of fluidity over this period. The
saturation dosage of superplasticizers also increases with time with saturation at
60 minutes occurring at dosages significantly higher than at 0 minutes. This can be
attributed to the increase in surface area of the hydrated cement particles and the
covering up of the superplasticizer by the newly formed hydrates (see Chapter 2 for
further explanations).
1.8
Type of Sp (time) Saturation dosage
SNF-S1 (0) 0.20
1.6 SNF-S1 (60) 0.30
SNF-D2 (0) 0.23

SNF-D2 (60) 0.40
SMF-S1 (0) 0.23
SMF-S1 (60) 0.30
1.4 PCE-D1 (0) 0.07
PCE-D1 (60) 0.17
1.2
1.0
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig.4.15 Loss of fluidity of paste
102
The plots in Figure 4.15 also show that the difference in flow time between the 0- and
60-minute curves is higher at lower superplasticizer dosages, with comparable levels
of fluidity being achieved at 60 minutes only with initial overdosage of the
superplasticizer (i.e., dosage higher than the 0-minute saturation point), except in the
case of the PCE, where the 60-minute curve is always above that at 0 minutes. The
high fluidity loss at low dosages can be attributed to more rapid flocculation due to
inadequate dispersion in pastes with less than saturation dosage. In the context of
readymix concrete, the results suggest that when the fluidity of the paste after 60
minutes is more important than the immediate (0-minute) fluidity, the saturation
dosage corresponding to 60 minutes should be used in the mix design rather than
arbitrarily overdosing the superplasticizer, as is often done in practice.
4.4.2 Loss of Fluidity for Mortar
To evaluate the change in the flow behaviour of mortar with time, tests were
conducted with the SMF and PCE based superplasticizers that led to higher loss of
fluidity in the paste tests than the others; Marsh cone flow time and flow table spread
were measured immediately after mixing and after 60 minutes, and the test results are
presented in Table 4.3. It is observed that the flow time increases and the flow table
spread decreases with time for both the superplasticizers, especially at lower dosages.
The loss of fluidity is higher for the SMF than the PCE, as expected (Chandra and
Bjornstrom, 2002b). Note that in some cases the mortar did not pass through the
Marsh cone or there were clear signs of bleeding indicating the segregation of the
mortar.
103
Table 4.3 Loss of fluidity of mortar
Type of sp sp/c % Marsh cone flow time (sec) Flow table spread (mm)
0 60 0 60
0.2 22 no flow 242 193
0.3 21 no flow 265 198
0.4 13 29 bleeding 260
SMF-S1 0.5 11 15 bleeding 305
0.8 11 14 bleeding bleeding
0.05 13 no flow 310 203
0.1 11 19 bleeding 281
PCE-D1 0.15 11 15 bleeding 290
0.2 11 12 bleeding 313
0.3 11 12 bleeding 315
In the case of SMF based mortar, low dosages of superplasticizers are not sufficient
even to initiate the flow after 60 minutes. Moreover, loss of fluidity occurs in the case
of the SMF even at the saturation dosage corresponding to the 60-minute tests of
pastes whereas for the PCE, there is no appreciable loss at the 60-minute saturation
dosage. These results imply that the PCE is more effective in maintaining the fluidity
than the SMF, especially at high dosages. However, in both cases the loss of fluidity
in the mortar is more than that of the paste.
The results clearly show that when overdosing of the superplasticizer is done to obtain
flowable mixes after one hour or so after mixing, the initial mix may segregate. This
has to be taken into account during mixing, discharging of the material from the mixer
into the truck and transportation. Another important factor is that any material
delivered ahead of the estimated delivery time could be segregated and unusable.
The comparison of the paste and mortar test results show that the incorporation of
sand does not affect the fluidity or saturation superplasticizer dosage in the 0-minute
tests but it seems to increase the loss of fluidity due to the absorption of water and
superplasticizer.
104
4.4.3 Loss of Fluidity of Concrete
To understand the change in flow behaviour of concrete with time, four mixes with
the corresponding saturation dosages of superplasticizer (corresponding to 0-minute
tests on pastes; see Section 3.5.2) were prepared. The slump was observed at intervals
of 30 minutes and the concrete was kept at rest in the pan mixer between
measurements; just before each slump test, the concrete was mixed for 15 seconds.
The results are shown in Figure 4.16, where it can be observed that the slump
decreases with time in all the combinations.
The loss of slump for SNF-S1 is 17% over 30 minutes and 54% over 60 minutes. The
loss of slump for SNF-D2 is the least compared to other superplasticizers, as seen in
Figure 4.16; it yields the lowest loss in slump of 18% and 26% over 30 and 60
minutes, respectively, reflecting its retarding nature. Slump losses of 54% over 30
minutes and 87% over 60 minutes are obtained in the case of the SMF based
superplasticizer, and slump losses of 57 % over 30 minutes and 82% over 60 minutes
for the PCE based superplasticizer. The SMF gives a higher loss of slump than the
PCE, as already seen in the case of mortar. This loss of slump with time in all cases
can be either due to cement hydration or drying due to water evaporation or
absorption of superplasticizer and water by the aggregates. The fines in the aggregates
also seem to increase the superplasticizer demand over the period of 60 minutes. This
indicates that in such cases, the saturation dosage determined at 0 minutes in the paste
will not be able to maintain the slump over a long period of time, say 60 minutes, in
concrete. When this occurs, it is recommended that the dosage of superplasticizer in
the concrete be taken as the saturation dosage obtained in the paste at 60 minutes.
105
180
SNF-S1
160 SNF-D2
SMF-S1
140 PCE-D1
120
Slump (mm)
100
80
60
40
20
0
0 20 40 60 80 100 120
Time (minutes)
Fig.4.16 Loss in slump of concrete with time
4.5 EFFECT OF SUPERPLASTICIZER ON SETTING OF CEMENT

PASTE AND CONCRETE
As already discussed, the superplasticizer slows down the hydration and dissolution of
ions, setting of superplasticized cement paste can be retarded (Torrents et al., 1998;
Roncero et al., 1999; Brooks et al., 2000). In this study, the Vicat apparatus was used
to determine the setting time of superplasticized cement paste. The cement paste was
prepared using a Hobart mixer, with a w/c of 0.35 and the saturation dosage of
superplasticizer (sp/c%) of each case. In Figure 4.17, the beginning and end of setting
(defined as the first incomplete Vicat penetration reading and the first zero penetration
reading are shown for each superplasticizer at saturation dosage). The graph shows
that retardation produced due to SNF-S1 and PCE-D1 is more than that of the SMF
based superplasticizer. This is because the higher molecular weight SMF will be
adsorbed more than the lower molecular weight SNF whereas, more of the SNF
remains in the solution and retards more than that of SMF. In the case of PCE-D1, the
retardation is because of the higher concentration of ionic functional groups
(sulphonic and carbonic) in the aqueous phase or due to shorter mean backbone chain
106
lengths in the PCE, in accordance with the reported results of Yamada et al. (2000).
The paste with the SNF with the retarder sets more slowly, as expected.
It can be observed that the time difference between the beginning and end of setting
varies approximately from 1 to 3 hrs. This shows that for the tested cement paste
systems, the superplasticizers lead to a delay in the beginning of setting due to delay
in the hydration of alite. However, it does not prolong the setting process because the
hydration is not affected further by the superplasticizer, after the beginning of alite
hydration (Yamada et al., 2000).
Control paste
40 SNF-S1
SNF-D2
Initial setting
Depth of penetration (mm)
SMF-S1
PCE-D1
30
20
10
0
200 400 600 800 1000 1200
Time (minutes)
Fig. 4.17 Vicat penetration measurements for pastes with different superplasticizers
The setting time of concrete was determined using a standard penetrometer as per IS
8142 (2002) as shown in Figure 4.18. The concrete was sieved through a 4.75 mm
mesh. The passing through mortar was tested, and the results obtained are shown in
Figure 4.19. The times at the penetration resistances of 3.43 N/mm2 and 26.97 N/mm2
are taken as the initial and final setting times of the concrete, respectively; the setting
times of paste and concrete at the saturation dosage of superplasticizer are shown in
Table 4.4, along with the paste flow time and slump values. The initial setting time
(IST) for the PCE based concrete was 8 hrs and final setting time was 10 hrs. The
107
SNF-S1 shows a similar trend whereas the SMF sets much earlier and the SNF with
retarder (SNF-D2) sets much later. However, except in the case of PCE-D1, the
difference between initial and final setting times (FST) is approximately 2 hours in
both the paste and concrete. It should be noted that both the Vicat penetration test on
paste and the penetration resistance test on mortar measure setting through the
determination of the degree of stiffening. Since the initial fluidity of the paste or
mortar also affects the stiffness and its evolution, the comparison of the different
results is not straightforward, especially since each of the pastes has a different initial
consistency. However, it appears that the superplasticizer that causes more retardation
of setting in the paste also tends to retard the setting of the concrete having the same
paste proportions.
Fig.4.18 Penetration test on concrete; (a) Pocket penetrometer,

(b) Penetration test on specimen (c) Impressions
made on the surface during measurements
30 SNF-S1
Final setting SNF-D2
25 SMF-S1
PCE-D1
Pressure (N/mm )
2
20
15
10
5 Initial setting
0
300 350 400 450 500 550 600 650
Time (minutes)
Fig. 4.19 Setting of concrete at saturation dosage of superplasticizer
108
Table 4.4 Initial and final setting time of paste and concrete
Type of Paste Concrete

superplasticizer Flow time IST FST Slump IST FST
(sec) (hrs) (hrs) (mm) (hrs) (hrs)
Control mix * 3.6 5.2 0 * *
SNF-S1 13 9.5 11.9 40 7.9 9.6
SNF-D2 17 16.3 18.7 90 8.2 10.5
SMF-S1 16 6.6 8.3 45 6.7 8.8
PCE-D1 12 10.5 14 73 8.0 9.9
4.6 ANALYSIS OF SETTING BEHAVIOUR OF SUPERPLASTICIZED

CEMENT PASTE USING ELECTRICAL RESISTIVITY METHOD
4.6.1 Background
The setting of cement paste and concrete is usually determined through penetration
resistance, as seen in the previous section. Electrical property measurements are being
proposed as an alternative to the penetration measurements for studying the setting
and hardening processes within the portland cement pastes over the initial 24 hours of
hydration through continuous real time monitoring (McCarter and Curran, 1984;
McCarter and Tran, 1996; McCarter et al., 1999, 2003). Calleja (1952a, 1952b) has
used electrical resistance measurements for estimating the initial and final setting time
of hydraulic materials. Torrents et al. (1998) and Roncero (2000) have used AC
impedance spectroscopy over a frequency range of 1 kHz to1 MHz for understanding
the influence of superplasticizers on the electrical resistivity of cement paste. More
recently, Xiao et al. (2007) have used electrical resistance methods to compare the
effect of superplasticizers at the saturation dosage and have correlated resistivity to
strength gain and setting time. The previous studies of Srinath (2007) on electrical
conductivity of cement mortar at the frequencies of 1, 10 and 100 kHz show that a
lower conductivity is obtained at 1 kHz whereas the values are similar at 10 and 100
kHz, indicating that frequencies of 10 kHz and above are adequate for the
109
conductivity measurements on fresh pastes since electrode polarization effects are
reduced to negligible proportions (McCarter and Tran, 1996).
The objective of this study is to understand the influence of superplasticizers on the
different stages of setting of cement paste based on the electrical conductivity of
paste. Tests were conducted with pure cement paste and with 4 types of
superplasticizers. The results are compared with standard Vicat test measurements.
4.6.2 Experimental Procedure
Aplab MT 4080A high accuracy LCR meter (hand held AC/DC impedance
measurement instrument), which can measure the inductance, capacitance and
resistance with imposed currents with frequencies of up to 100 kHz, was used in the
present study. A mould with internal dimensions of 50 × 50 × 50 mm was made with
polycarbonate plates of 2 mm thickness. This sample size was considered large
enough to obtain a representative bulk electrical measurement on paste (following
McCarter and Tran, 1996). Stainless steel plates of 50×50 mm were placed on
opposite walls of the mould. The cement paste with the saturation dosage of
superplasticizer was prepared, as explained in Section 3.4. The sample (cement paste)
was placed inside the mould and the steel plates were connected to the H and L ports
of the LCR meter to serve as the electrodes (Figure 4.20). Impedance values were
determined at the frequency of 10 kHz and the voltage of 1 Vrms over a period of 24
hours. The resistance values were converted to electrical resistivity using equation
(4.1).
RA
ρ= ohm metres (4.1)
d
where d is the length of the specimen, A is the cross sectional area of specimen, ρ is
the electrical resistivity. Due to sample preparation and setting-up procedures, test
110
measurements were taken 15 minutes after mixing. So the zero minute reading
presented in the graph in this section corresponds actually to 15 minutes after casting.
The resistance values were taken subsequently every 15 minutes over a test period of
24 hours. The top surface of the sample was covered with a polycarbonate plate to
avoid evaporation. All tests were carried out in ambient conditions.
LCR meter
Fig. 4.20 Comparative study on setting behaviour of

paste with Vicat apparatus and LCR meter
4.6.3 Results and Discussions
The conductivity and its derivative for pure cement paste and the superplasticized
pastes are plotted as a function of time in Figure 4.21. The variation in conductivity
can be correlated with the different stages of hydration of cement paste. Stage 1
corresponds to the initial stage in which the conductivity increases due to the
dissolution of ions, in all the cases. However, the increase in conductivity is more for
the pure cement paste than that of superplasticized pastes. This confirms that the
presence of superplasticizer prevents further dissolution of ions, which slows down
the increase in conductivity in the initial stage. Stage 2 corresponds to the dormant
stage in which the conductivity remains almost constant. In the case of pure cement
111
paste, the dormant stage is caused due to the supersaturation of solution and
consequent inhibition of dissolution of ions. The superplasticizer appears to mainly
extend this dormant stage and it could be attributed to; (1) the lowering of the rate of
chemical reactions further by inhibiting the dissolution of ions due to the presence of
superplasticizers in solution; and (2) the formation of an adsorption layer on the
cement particle and the formation of complexes with Ca2+. In stage 3, the conductivity
decreases as the paste sets and the resistance increases due to the decrease in the
mobility of Ca2+, SO42- and OH- ions. The study confirms that the addition of
superplasticizer generally extends the dormant stage of hydration process, retarding
the onset of acceleration stage, in accordance with the results of earlier studies.
(Torrents et al., 1998; Roncero, 2000; Xiao et al., 2007). However, the evolution of
the derivative of conductivity clearly reflects the rate of chemical reactions only in the
case of pure cement paste.
1.8 0.015
Conductivity
1.6 0.010
derivative of conductivity
Derivative of conductivity
Conductivity (siemens)
0.005
1.4
0.000
1.2 2
1 3
-0.005
1.0
-0.010
0.8 -0.015
0 50 100 150 200 250 300 350

Time (minutes)
(a) Pure cement paste with w/c = 0.35
112
2.0
Conductivity 0.003
1 derivative of conductivity
1.6
0.002
2
1.2 0.001
3
0.8 0.000
-0.001
0.4
-0.002
0.0
0 500 1000 1500 2000
Time (minutes)
(b) Cement paste with LS-C1
2.0
1 2
0.002
Conductivity
1.5 Derivative of conductivity 0.001
3
0.000
1.0
-0.001
0.5 -0.002
-0.003
0.0
0 500 1000 1500 2000
Time (minutes)
(c) Cement paste with SNF-S1
113
2.0
1 2 0.0010
Conductivity
0.0005
Derivative of Conductivity
1.5
Conductivity (Siemens)
3 0.0000
1.0 -0.0005
-0.0010
0.5
-0.0015
-0.0020
0.0
0 500 1000 1500 2000
Time (minutes)
(d) Cement paste with SMF-S1
0.8 Conductivity 0.0015

1 0.0010
2
0.6 3
0.0005
0.4 0.0000
-0.0005
0.2
-0.0010
-0.0015
0 400 800 1200
Time (minutes)
(e) Cement paste with PCE-D1
Fig. 4.21 Evolution of conductivity for different superplasticizers
The conductivity curves for the different pastes, along with the beginning and end of
setting (IST and FST) obtained from Vicat penetration tests, are shown in Figure 4.22.
Compared to pure cement paste, the initial conductivity is much lower for the PCE
and similar for the other superplasticizers. As discussed earlier, the results confirm
that the dormant stage lasts longer for the superplasticized pastes.
114
The comparison of the evolution of the conductivity with the Vicat test results shows
that for pure cement paste, as well as paste with LS, SNF and PCE based
superplasticizers, both the IST and FST fall in the acceleration stage. For the SMF-
based superplasticizer, setting begins in the dormant stage whereas the final setting
time falls in the acceleration stage. This is in accordance with the penetration test
results of both paste and concrete where the retardation of setting was seen to be the
least for the SMF-based superplasticizer. It is observed that the superplasticizer that
extends the dormant stage of hydration most, yields the highest IST and FST in
concrete. Hence, the test results obtained from electrical resistivity method, which is
closely related to the chemical activity in the paste, indicate that the setting time is in
the following order: Cement paste < Paste with SMF< Paste with SNF< Paste with
PCE< Paste with LS.
2.0
Control paste
LS-C1
1.6 SNF-S1
SMF-S1
IST PCE-D1
1.2
FST
0.8
0.4
0.0
0 500 1000 1500 2000
Time (minutes)
Figure 4.22 Electrical conductivity graphs for different pastes
along with beginning and end of setting defined by
Vicat penetration measurements
Another point to be noted is that even though the tests are conducted at the saturation
dosage for different superplasticizers, the maximum conductivity values varies with
115
the type of superplasticizer. This indicates that the concentration of ions in the
aqueous solution is also modified by the type of superplasticizer, especially in the
case of PCE-based superplasticizer. This further confirms that the influence of the
superplasticizer on setting, as well as on fluidity, varies with its mechanism of action,
as explained in Chapter 2. It can be observed from Figure 4.22 that the time difference
between the beginning and end of setting is approximately the same in all cases.
Hence, it can be concluded that for the present cement paste system, the
superplasticizers considered delay the beginning of setting but do not prolong the
setting process.
4.7 EFFECT OF SUPERPLASTICIZER ON COMPRESSIVE STRENGTH

OF CONCRETE
The compressive strengths of concretes incorporating different types of
superplasticizers are shown in Table 4.5. It can be observed that the strength of
superplasticized concrete at 28 days is generally more than that of the reference
concrete, as also observed by other researchers such as Dhir and Yap (1983), Seung-
Bum (1999), Ramachandran (2002) and Kapalko (2006). It can be observed that the
concretes with the saturation dosages of the superplasticizer exhibit slightly higher
strengths than the reference concrete, probably due to better compaction.
Nevertheless, superplasticizer dosages much above saturation lead to bleeding and
immediate stiffening, which cause strength reduction.
116
Table 4.5 Compressive strength of concrete for different combinations
Type of Dosage of Saturation Slump Compressive strength

superplasticizer superplasticizer dosage (mm) 3 days 7 days 28 days
(sp/c %) of sp N/mm2 N/mm2 N/mm2
(sp/c %)
0 (Reference) 0 39 45 56
0.1 0 38 47 56
SNF-S1 0.2 0.20 25 39 49 58
0.3 195 34 54 69
0.4 bleeding 31 40 50
0.1 10 47 57 62
0.2 67 46 57 64
SNF-D2 0.25 0.23 100 40 55 63
0.3 130 43 59 65
0.4 165 37 56 67
0.1 20 39 - 57
0.2 32 45 55 69
SMF-S1 0.3 0.23 70 40 53 67
0.5 158 41 52 73
0.6 bleeding 40 52 60
1 bleeding 45 - 53
0.05 35 36 46 58
0.1 120 42 47 54
PCE-D1 0.2 0.07 180 46 55 67
0.3 bleeding 34 41 45
4.8 SUMMARY
An extensive study on paste, mortar and concrete aimed at establishing correlation
between their behaviour has led to the following conclusions:
• The saturation dosages obtained for paste and mortars are comparable whereas
a dosage slightly higher than the saturation dosage of paste was required for
adequate workability in concrete due to the adsorption of some
superplasticizer by the fines present in the coarse aggregate used here. The
comparison of the cost of superplasticized concretes shows that a low-cost
superplasticizer need not always result in cost-effective concrete, as the dosage
of the superplasticizer for the optimum fluidity is also an important factor.
117
• The study of the flow behaviour of pastes at 60 minutes after mixing shows
that loss of fluidity occurs for all the superplasticizers and dosages considered.
If the fluidity of superplasticized paste at 60 minutes or at any other duration is
more important than the immediate fluidity, the saturation dosage
corresponding to that duration should, therefore be used in the mix design of
concrete rather than arbitrarily overdosing the superplasticizer. In the case of
mortar, the addition of sand results in higher loss of fluidity than in the paste,
probably because of the absorption of superplasticizers by the fine aggregates.
Generally, it was found that the SMF-S1 and PCE-D1 led to more slump loss
than the SNF-S1 based concrete as observed in the case of corresponding
paste. Also, as expected, the SNF with retarder was effective in limiting the
slump loss of concrete.
• The Vicat penetration test results shows that the setting behaviour depends on
the type of superplasticizer, and that the setting time is higher for the
lignosulphonates and polycarboxylate than the sulphonated naphthalene-and
melamine-based superplasticizers. A comparison of the penetration tests on
paste and concrete indicates that the trend in the setting with different types of
superplasticizers is almost same in both cases, and hence the test of paste can
be used as a guideline in the selection for the superplasticizers for concrete in
terms of setting time. Also, it is confirmed through the study that
superplasticizers tend to retard the beginning of setting and do not necessarily
prolong the setting process, i.e., the difference between the initial and final
setting times are not affected.
• Electrical conductivity measurements on the superplasticized pastes have been
used to characterize the setting behaviour, along with Vicat penetration
measurements. It has been observed that the superplasticizer delays the initial
118
setting without prolonging the setting process, which is mainly manifested as
an increase in the duration of the dormant stage.
• The study shows that the compressive strength can increase slightly due to the
addition of superplasticizers, probably due to better compaction, especially at
the saturation dosage of superplasticizer.
Generally, it can be concluded that the selection of the type and dosage of
superplasticizer in concrete for high performance can be based on the test results from
paste. This could lead to substantial savings in material and effort within the mix
optimization process.
119
CHAPTER 5
EFFECT OF SUPERPLASTICIZER ON
PROPERTIES OF HARDENED CEMENT PASTE
5.1 GENERAL
The study of the flow behaviour of superplasticized cements paste and its correlation
with concrete shows that the fluidity and flow retention depend on the type and
dosage of superplasticizers. The purpose of adding superplasticizers is to improve the
fresh state properties of concrete without adversely affecting the hardened state
properties. However, the literature shows that cement-superplasticizer incompatibility
results in an adverse effect on the rate of hydration and hydrated products (Jolicoeur
and Simard, 1998; Hanehara and Yamada, 2008; Agarwal et al., 2000; Prince et al.,
2002; Nkinamubanzi and Aitcin, 2004; Bedard and Mailvaganam, 2005).
The objective of this chapter is to have an understanding on the influence of type of
superplasticizers on the hydration behaviour and on the properties of hardened
concrete. Indirect techniques like differential thermal analysis/ thermogravimetric/
differential thermogravimetric (DTA/TG/DTG) and X-ray diffraction study (XRD)
give information about the average characteristics of the microstructure whereas
direct methods like scanning electron microscopy (SEM) provide information about
the way in which component phases are arranged in the microstructure (Scrivener,
1989; Ramachandran and Beaudoin, 2001).
The influence of superplasticizers on the hardened cement paste and more precisely,
on the hydration processes and the microstructural development is explained in this
Chapter. The cement paste phases are analyzed using XRD, micrographs, thermal
analysis and nuclear magnetic resonance (Si NMR) for pure cement paste as well as
for superplasticized pastes at saturation dosages. The complementary use of these

approaches is intended to provide a complete picture of the hydration processes, with
respect to the formation of new phases, the consumption of the anhydrous phases as
well as the formation of C-S-H and C-H.
5.2 X-RAY DIFFRACTION (XRD) STUDY
5.2.1 Background
X-ray diffraction study helps to understand the influence of the incorporation of the
superplasticizer on the hydration process of cement paste with respect to the
consumption of crystalline anhydrous phases, as well as through the identification of
crystalline products of hydration. The basic principle of XRD is that every crystalline
substance gives a specific pattern; the same substance always gives the same pattern
and in a mixture of substances each produces its pattern independently of others.
When a beam of X-rays falls on a crystalline material, it is diffracted by the incidence
angle given by Bragg’s law:
2 d sinθ = n λ (5.1)
where d is the distance between the crystallographic planes, λ is the wavelength of the
X-ray, n is an integer and θ is the diffraction angle.
XRD is used to identify the polycrystalline phases of hardened cement paste and
superplasticized cement paste through the recognition of the X-ray patterns that are
unique for each of the crystalline phases. So the technique allows the detection of
ettringite (AFt) and portlandite (C-H) along with the consumption of anhydrous
phases of the cement (gypsum, C3S, C2S, C3A, C4AF, SiO2 and CaCO3).
Superplasticizers generally affect the hydration by the reduced formation of Ca(OH)2
and ettringite in the initial ages of curing, which can be identified through XRD
121
(Singh et al., 1992; Mollah et al., 1995; Roncero, 2000; Prince et al., 2002; Roncero et
al., 2002; Prince et al., 2003).
5.2.2 Materials Used and Sample Preparation
Cement C1 of 53 grade and 6 types of superplasticizers with properties as given in
Section 3.2 along with a reference paste at several ages like half setting
(corresponding to 20 mm Vicat penetration), 3, 7 and 28 days were tested. These ages
are selected as reference ages so as to compare the effect of superplasticizers on the
rate of hydration. The pastes were prepared with the saturation dosage of
superplasticizers (determined from the Marsh cone test) and the mixing method
adopted was the same as selected from Section 3.4. The samples were cured under
water and the cement paste samples of 50mm×50mm×10mm were cut and kept in
acetone to prevent the hydration after the required age. Just before each test, the
samples were dried in an oven at 60°C and the surface was polished to remove
undulations. To account for the variations in the preparation of sample and the
instrument intensity, a standard sample of aluminium foil was placed over the
specimen and the normalized intensity was used for comparison purposes.
The XRD analysis was performed with a Bruker D8 Discover powder diffractometer.
In this apparatus, the Cu Kα (λ = 1.5418Å) radiation is generated in a Cu tube at
35kV, 25 mA. The tests were performed over a Bragg angle (2θ) range of 5-70° with
a scan speed of 1 sec per step on the samples. The ratio of the intensity of sample to
the intensity of aluminium at a particular Bragg angle is taken for the comparison of
crystalline phases of superplasticized pastes.
122
5.2.4 Results and discussions
The XRD pattern of cement paste and superplasticized pastes for half setting, 3, 7 and
28 days are shown in Figure 5.1 (a-g). The relative evolution with age of the identified
crystalline phases like C3S, gypsum, ettringite and portlandite can be compared with
the intensity. In the case of pure cement paste, amount of unhydrated phases decrease
with an increase in the age of curing and C-H increases with the age of curing, as
shown in the X-ray diffractograms. Ettringite is present at half setting and disappears
after 3 days as expected. The peak of calcite (CaCO3) indicates the occurrence of
carbonation in all the samples.
The XRD pattern of superplasticized paste shows some qualitative differences in the
hydration rate due to the incorporation of superplasticizers which is supported by the
reported results (Mollah et al., 1995, Roncero, 2000; Prince et al., 2003). This
qualitative difference varies with different superplasticizers. The delay of C3A
hydration in the presence of superplasticizers is obvious through the absence of
ettringite at half setting in the XRD pattern of superplasticized pastes. It affects the
setting and hardening due to the formation of electrical double layer along with “the diffuse
ion swarm” of superplasticizers on Ca2+ ions preventing the hydration (Mollah et al., 1995).
11 C-H
C-H
C-H C-H Al
10 C3S 28 days
9
8
C-H Calcite C-H Al
7 C3S C-H C-H
ISample/IAl
6 7 days
5
C-H Calcite C-H Al
4 C-H
3 days
3
2 C3S
C-H Al
C-H AFCalcite C-H C S Half setting
1 t 3
0
10 20 30 40 50 60 70 80
2θ (degrees)
(a) Cement paste with w/c 0.35
123
C-H
10 Al 28 days
8 C-H
7 days
Al
ISample/ I 6
C-H C-H 3 days
C3S Al
4
Gypsum
Calcite Calcite C3S
2 Half setting
0
10 20 30 40 50 60 70 80
2θ (degrees)
(b) LS-C1 based paste
C-H C-H Al
10 C-H
28 days
8
C-H Calcite C-H Al
C-H
7 days
ISample/IAl
Calcite
C-H C-H Al
4
3 days
C3S
Calcite C-H
2 Gypsum Al C3S
Al
Half setting
0
10 20 30 40 50 60 70 80
2θ (degrees)
(c) SNF-S1 based paste
124
10 C-H C-H
28 days
9
8
ISample/ IAl 7 7 days
6
5 C-H
4 Al
AFt
3 3 days
CS
2 Gypsum Calcite 3 C-H
C-H C-H C3S Half setting
1
0
10 20 30 40 50 60 70 80
2θ (degrees)
(d) SNF-D2 based paste
C-H C-H
10
28 days
8
ISample/IAl
6 C-H
Calcite
7 days
4
C-H SiO C-H
2 C3S
Al 3 days
2 AFt C-H C-H
Gypsum
C3S Half setting
0
10 20 30 40 50 60 70 80
2θ(degrees)
(e) SNF-C2 based paste
125
10 C-H C-H 28 days
9
8
C-H C-H
7 7 days
6
ISample/IAl
5
SiO C-H
4 AFt 2 3 days
3
2 Half setting
CS Al
1 Calcite 3 C-H C-H C3S
0
10 20 30 40 50 60 70 80
2θ(degrees)
(f) SMF-S1 based paste
C-H Al
10 28 days
8
C-H Al
7 days
ISample/IAl
C-H C-H Al
4 3 days
CalciteC3S
C-H
Gypsum Calcite C3S
2
Half setting
0
10 20 30 40 50 60 70 80
2θ
(g) PCE-D1 based paste
Fig. 5.1 (a-g) X-ray diffractogram of cement paste and superplasticized cement paste
The trends of the evolution of the crystalline phases of pure cement paste and
superplasticized cement paste at different ages are summarized in Table 5.1. The
arrows indicate the relative change in the amount of the crystalline phases observed.
The amount of unhydrated compounds like C3S/C2S decreases with an increase in the
126
age of curing. Even though the time of half setting is different for the studied pastes,
major qualitative difference could not be observed between the superplasticized
pastes. The presence of portlandite in the superplasticized cement pastes at half setting
indicates that the formation of portlandite is not affected by the superplasticizers.
Generally, the amount of portlandite increases with age of curing, as expected.
However, anomaly is observed with some types of superplasticizers.
Table 5.1 Relative evolution of the crystalline phases with age
Type of C3S,
Age SiO2 Gypsum Ettringite Portlandite
paste C2S
Half setting P P A P P
3 days P P A A P
Reference 7 days P P A A P
28 days P P A A P
Half setting P P P A P
3 days P P A P P
LS-C1 7 days P P P A P
28 days P P A A P
3 days P P A P P
SNF-S1 7 days P P A P P
28 days P P A P P
Half setting P A P A P
3 days P P A P P
SNF-D2 7 days P P A P P
28 days P P A P P
3 days P P A P P
SMF-S1 7 days P P A P P
28 days P P A P P
3 days P A A P P
SNF-C2 7 days P A A A P
28 days P P A P P
3 days P P A P P
PCE-D1 7 days P P P P P
28 days P P A P P
P - present; A – absent
Increase or decrease of the phase, respectively, with reference to the previous
age
127
The comparison of the diffractograms of pure cement paste with superplasticized
pastes at half setting indicates that superplasticizer mainly affects the hydration of
C3A and consequently the formation of ettringite; it can be observed that no gypsum
is present in the pure cement paste at half setting due to its conversion to ettringite by
reacting with C3A. However, in the superplasticized pastes, no traces of ettringite
were observed at half setting; instead, gypsum is present. This clearly shows that the
setting is retarded due to the addition of superplasticizers because it inhibits the
hydration reaction. The possible reasons for set retardation have been already
explained in Chapter 2. The results also support the data from the setting time tests
reported in Chapter 4.
Also, traces of ettringite are present upto 28 days in all the superplasticized pastes
except for the paste with LS-C1, in which it is present upto 3 days only. However, the
peak intensity of ettringite decreases between 3 days and 7 days and again increases
after 7 days in the case of SMF-S1 and PCE-D1. This is in accordance with the
reported results (Prince et al., 2003) that the addition of superplasticizer affects the
crystallization of primary ettringite and when no more superplasticizer is available to
block the natural growth of ettringite, it reappears after 28 days. The diffraction
pattern does not show any traces of AFm or secondary gypsum phases which indicate
that phase reversal (see Chapter 2) has not occurred. In short, from the diffractograms
observed in the particular cementitious system under study, it is seen that there is very
strong interaction between the superplasticizer and ettringite at early ages.
5.3 SCANNING ELECTRON MICROSCOPY (SEM)
5.3.1 Background
Backscattering electron (BSE) imaging is generally used to identify the different
phases of hardened cement paste and to understand the development of the

128
microstructure of cement paste. Scrivener (1989) has used the technique to distinguish
between the dense and porous phases in the microstructure of same paste. The basic
principle is that the electrons of the incident beam, rebound from the surface of the
specimen depending on the composition of the phases. The intensity of the BSE signal
is a function of the average atomic number of the sample and the contrast observed is
based on the atomic number of phases (Famy et al., 2002).
The unhydrated particles have higher molecular weight with higher back scattering
intensity than hydrated particles (Zhao and Darwin, 1992; Puertas et al., 2005). In
such studies it has been shown that C4AF has the highest intensity, C-H is brighter
than C-S-H and that the pores appear black.
A comparative study of the microstructure of different superplasticized cement pastes
and reference pastes is conducted here to understand the influence of different types
of superplasticizers on the hydration process and consequently on the development of
microstructure.
5.3.2 Sample Preparation
The quality of the BSE image depends on the sample surface. A highly polished
surface is required for BSE microscopy (Scrivener, 1989). The sample preparation
included the preparation of cement paste samples, curing for the prescribed ages,
cutting, grinding and polishing. The superplasticized cement paste cubes with four
families of superplasticizers along with the reference paste cubes were prepared as in
the case of the XRD study. The cubes were cured under water for 3, 7 and 28 days.
After the period of curing, the samples were put in acetone to remove the pore water
by the solvent replacement method to prevent further hydration as in Roncero (2000).
Samples of 10mm×10mm×10mm were cut and carefully ground by hand at moderate
129
pressure on a middle-speed lap wheel with p800, p1000, p1200, p1600 and p2000
sand papers. Afterwards, lap wheel with 6-12, 3-6 and ½ -1 μm diamond paste was
used to get a fine polished surface.
5.3.3 Comparison of the Microstructure of Superplasticized Paste and

Reference Paste
The microstructural study was performed on pastes with no admixture (i.e., reference
paste) and those containing saturation dosages of superplasticizers. The BSE
micrographs of reference paste and paste with LS-C1, SMF-S1, PCE-D1 and SNF-S1
based superplasticizer are shown in Figure 5.2 (a-i). The microstructure of cement
paste shows the bright grains of unhydrated cement, the intermediate gray of C-S-H
and slightly brighter gray (shown as red for easy identification) of calcium hydroxide,
along with dark pores. As the age of curing increases up to 28 days (see Figure 5.2 a
and b), the amount of unhydrated particles decreases and a uniform distribution of
hydrated particles is identified in the picture.
The BSE morphological analysis of superplasticized pastes showed no substantial
difference in either the form or the texture of the different hydration products in pastes
with and without admixtures. However, superplasticized pastes show more amount of
unhydrated particles at 3 days than reference paste confirming the retardation of
hydration in early ages due to the incorporation of superplasticizers as already
discussed in Chapters 2 and 3. Even though, the micrographs are obtained at the
saturation dosage of superplasticizers, the amount of unhydrated particles is different
for different pastes. In the case of LS based superplasticizer, more amount of C-S-H
can be observed. The amount of unhydrated products is more for the paste with the
SMF at 3 days compared to that with the PCE. More unhydrated particles than in the
pastes with SMF and PCE can be observed in the paste with SNF at 28 days. Also, the
130
pastes with the SMF micrograph shows more C-H at 28 days, compared to those of
the other superplasticizers.
A comparison of the microstructure of the pastes with SMF and PCE shows that a
more uniform microstructure is produced in the latter case. The number of pores in the
superplasticized paste is more than in the reference paste and it is higher in the paste
with PCE. The pores are smaller when PCE is added as reported by Sakai et al.
(2006). Generally, the increase in strength due to the addition of superplasticizer can
be attributed to making the microstructure more dense, uniform and compact
(Uchikawa, et al., 1995). The micrographs shown in Fig. 5.2 also support the
modification of the microstructure with uniform distribution of C-S-H and C-H
C-S-H
Unhydrated
C-H
(a) Reference paste, 3 days (b) Reference paste, 28 days
131
(c) LS-C1 based paste, 3 days (d) LS-C1 based paste, 28 days
(e) SMF-S1 based paste, 3 days (f) SMF-S1 based paste, 28 days
(g) PCE-D1 based paste, 3 days (h) PCE-D1 based paste, 28 days
132
(i) SNF-S1 based paste, 28 days
Fig. 5.2 (a-i) BSE (250X) micrograph of pure cement paste and
superplasticized cement paste
Energy dispersive X-ray analysis (EDAX) was also done for the selected samples to
find out the elemental composition of different phases. It is a technique used for
identifying the elemental composition of the specimen, or an area of interest thereof.
Each of the peaks in an EDAX spectrum is unique to an atom, and therefore
corresponds to a single element. The higher a peak in a spectrum, the more
concentrated the element is in the specimen. An EDAX spectrum not only identifies
the element corresponding to each of its peaks, but the type of X-ray to which it
corresponds as K-alpha peak or K-beta peak.
Figure 5.3 (a-c) shows an example of the BSEM micrograph along with the EDAX
spectrum of a C-S-H particle for the paste with the lignosulphonate. The elemental
composition of Ca and Si in the EDAX confirms the presence of C-S-H in the
micrograph. The Ca/Si ratio can also be determined for validating the identified
phases through the microstructure. A higher value of Ca/Si ratio for C-S-H in some
cases of superplasticized pastes shows that it is more amorphous than the pure cement
133
paste (Taylor, 1997). However, the summary of the results is not reported in this study
as the numbers of studied samples are limited.
Element Wt% At%

OK 40.83 61.13
AlK 01.76 01.56
SiK 12.88 10.99
KK 00.84 00.51
CaK 41.86 25.02
FeK 01.83 00.78
(a) EDAX spectrum of different elements (b) Elemental composition

in the selected spot corresponding to (Kα) radiation
(c ) Micrograph of C-S-H particle
Fig. 5.3 (a-c) BSE micrograph along with EDAX for a dark grey
(C-S-H) particle in the LS based paste
134
5.4 THERMAL (DTA/TG/ DTG) ANALYSIS
5.4.1 Background
Differential thermal analysis (DTA) along with thermogravimetric (TG) and
differential thermogravimetric analysis (DTG) can be used to understand the
hydration reactions of superplasticized cement paste (Prince et al., 2002). The basic
principle is that the material is subjected to continuous heating at a uniform rate and
the loss in mass is studied. In DTA, the difference in temperature (ΔT) between the
sample and reference material such as α-Al2O3 is recorded while both are subjected to
the same heating program (Ramachandran and Beaudoin, 2001).
TGA measures the loss in weight as the temperature of the substance is raised at a
uniform rate. The reactions occurring at the heating process are responsible for the
changes in weight. Knowing that certain reactions occur at specific temperatures
enables the identification of the constituents of the sample. TGA curves are usually
used to measure the weight loss corresponding to thermal decomposition whereas
DTG analysis helps locate the temperature peak due to the decomposition of different
phases in cement pastes.
5.4.2 Result and Discussions
The analyses were conducted with a NETZSCH STA 409 C/CD, in nitrogen
atmosphere, within a temperature range of 20-1400 ºC, at a uniform rate of 10 K/min.
The samples were obtained by powdering the prepared paste fragments and sieving
through 70 μm sieves.
Figure 5.4 (a-c) shows the DTA analysis of superplasticized paste and reference paste
after 3 days of curing. The small endothermic peak at 94.5ºC for pure cement paste
corresponds to dehydration of the non-evaporable water from the C-S-H gel (Prince et
135
al., 2002, 2003). In the case of pastes with SMF and PCE, an endothermic peak at
about 160ºC shows the decomposition of ettringite, gypsum and C-S-H gel (Prince et
al., 2003). The increase in the intensity of this effect with time is indicative of
increased formation of C-S-H with time. The endothermic peak in the range of 425-
550ºC shows the decomposition of C-H, in all the three cases (Tzouvalas et al., 2005;
Zhang, 2007).
o
1 peak at 94.5 C
exo 0
o
peak at 459.3 C
-1
DTA mw/mg
-2
-3
-4
-5
-6
-7
0 200 400 600 800 1000 1200
o
Temperature ( C)
(a) Reference paste, 3 days
0
0
Peak:164.1 C
-1 Peak:437.3 C
0
DTA (mw/mg)
-2
-3
-4
-5
-6
0 200 400 600 800 1000 1200
o
Temperature ( C)
(b) Paste with SMF-S1, 3 days
136
0.0 peak at 161.1
exo-0.5
-1.0 peak at 442
DTA mw/mg
-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
0 200 400 600 800 1000 1200
o
temperature ( C)
(c) Paste with PCE-D1, 3 days

Fig.5.4 DTA analysis of cement paste and superplasticized paste at 3 days
Figure 5.5 (a-d) shows the TG curves of the different pastes at 3 days. The total loss in
mass represents the degree of hydration of the pastes. The quantity of calcium
hydroxide is calculated from the weight loss at the temperature around 450-550ºC.
The loss in mass for superplasticized paste is less than that of pure cement paste at 3
days showing that the formation of C-H is slightly retarded at early ages due to the
presence of superplasticizers. This is in accordance with the reported results of
Ramachandran et al. (1995) and Puertas et al. (2005). The mass loss is small and
within the ranges of the reported results (Prince et al., 2002; Prince et al., 2003;
Puertas et al., 2005).
137
100
o
Inflection 78.8 C
98
Mass change -7.5%
mass %
96
Mass change -2.1%

94
o
Inflection 453.8 C
o
Inflection 646.4 C
92
0 200 400 600 800

o
Temperature ( C)
100
98 Inflection at 88.8 C
o
Mass change -7%

96
mass %
Total Mass change -10%

94
o Mass change -1.53%

92 Inflection 434.5 C
Mass change -1.4%

90
o
Inflection 645 C
88
0 200 400 600 800 1000 1200
o
Tempeature ( C)
138
100
o
96 Mass change -7.7%

Total mass change -14%
94
mass % 92
90 o
Mass change -2%
Inflection 438.3 C
88 Mass change -1%
86 o
Inflection 642.4 C
84
0 200 400 600 800 1000 1200 1400
o
Temperature ( C)
100
o
Inflection 100.7 C
98
Mass change -5.3%
96 Total Mass change -9.4%

mass %
94 o
Mass change -1.39%
Inflection 445.8 C
Inflection 685.2 C Mass change -2.1%

o
92
Mass change -0.057%
o
0 200 400 600 800 1000 1200 1400

o
Temperature ( C)
(d) Paste with SNF-C2, 3 days
Fig. 5.5 (a-d) TG and DTG analysis of pastes at 3 days
The results of the TG/DTG analysis for different pastes at 28 days are shown in
Figure 5.6 (a-d). The total mass loss increases in all cases with age of curing. The
mass loss corresponding to 600-800 ºC for the pastes at 3 and 28 days, is generally
139
interpreted as partly due to carbonation and partly to the final stages of dehydration of
C-S-H and the hydrated aluminate phases (Taylor, 1997). However, the XRD results
confirmed the presence of calcite; hence, this loss is mainly due to the dehydration of
C-S-H, hydrated aluminate phases and calcites.
100
o
Inflection 82.3 C
Mass change -8.7%

95
Total mass change -18%

mass %
Mass change -2%

90 o
Inflection 444.9 C
Mass change -6.4%

o
80
0 200 400 600 800 1000 1200 1400
o
Temperature ( C)
100
o
96 Mass change -9.3%

94
mass %

92
90
88 o
86 o
Mass change -2%
Inflection 695.3 C
Mass change -0.8%
84 o
Inflection 909.8 C
82
0 200 400 600 800 1000 1200 1400
o
Temperature ( C)
140
100
o
96
Mass change -10%
94
mass %
92
90
88 o Mass change -2.2%
Inflection 453.6 C
86 Mass change -2%
o
Inflection 691.6 C
Mass change -0.85%
84
o
Inflection 909.8 C
82
0 200 400 600 800 1000 1200 1400
o
Temperature ( C)
100
98 o
Inflection 97.6 C
96 Mass change -10%
94
mass %
92 Total Mass change -15%
90
o
88
Mass change -1.1%
86 o Mass change -0.58%
Inflection 681.1 C
o
Inflection 879.7 C
84
0 200 400 600 800 1000 1200 1400

o
Temperature ( C)
(d) Paste with SNF-C2, 28 days
Fig. 5.6 TG and DTG analysis of pastes at 28 days
Table 5.3 gives the summary of total mass loss, as well as the mass loss corresponding
to the decomposition of C-H at the temperature of around 450-550 ºC. It is observed
that the total mass loss increases with the age in all the cases whereas the mass loss
corresponding to C-H decomposition is comparable at 3 days and 28 days. This shows

141
that the superplasticizers retards the hydration mainly in the initial stage. However,
the percentage mass loss corresponding to the decomposition of C-H at 3 days and 28
days for different combinations seems to be too small for a detailed interpretation.
Table 5.2 Summary of mass loss in different pastes
Type of mix Mass loss corresponding to Total mass loss

the decomposition of C-H (%)
(%)
3 days 28 days 3 days 28 days
Pure Cement paste 2.1 2 * 18
Paste with SMF-S1 1.5 1.9 10 17
Paste with PCE-D1 2.0 2.2 14 17
Paste with SNF-C2 1.4 2.1 9.4 15
*Could not be determined
5.5 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY (Si NMR)
5.5.1 Background
Nuclear magnetic resonance spectroscopy is a technique used to characterize the state
of polymerization of silicates in the cement paste. NMR determines the energy
(resonance absorption) needed to invert the nuclear spin in Si atoms due to a powerful
alternating magnetic field. Magic angle spinning (MAS) NMR is a technique used on
solid phases, which consists of spinning the sample at high frequencies at the magic
angle of 54°44’ to the magnetic field (Roncero, 2000; Roncero et al., 2002; Puertas et
al., 2005).
The MAS NMR of siliceous materials provides useful information about the state of
the SiO44- tetrahedrons. The Si MAS NMR spectra exhibit peaks at 5 different
chemical shifts that correspond to the different electronic environments of the Si
atom, which are affected by the length of the Si-O bond, the Si-O-Si angle and the
nature of neighbouring atoms. The chemical shifts depend on the number of oxygen
atoms shared by the tetrahedral, known as the degree of condensation of the
142
tetrahedra. Each type of tetrahedron connectivity is denoted as Qn, where n denotes
the number of shared oxygen atoms. Thus, the five chemical shifts are Q0, Q1, Q2, Q3,
Q4 in which the Si tetrahedra share 0, 1, 2, 3 and 4 oxygen atoms, respectively, with
other Si atoms.
Table 5.4 shows the chemical shifts corresponding to the five electron environments
described above. The anhydrous silicate phases of the cement (C3S, C2S) correspond
to the monomer state (Q0), which undergoes polymerization as the hydration
proceeds. The hydrated cement pastes manifest the presence of Q1 and Q2
tetrahedrons. The Q1 and Q2 peaks correspond to the formation of the C-S-H gel, the
structure of which is formed mainly by 3n-1 long chains of tetrahedrons.
Table 5.3 Ranges of 29 Si chemical shifts of the Qn tetrahedrons
Ranges of chemical shifts (ppm)

0 1
Q Q Q2 Q3 Q4
-66 to -74 -75 to -82 -85 to -89 -95 to -100 -103 to -115
5.5.2 Sample Preparation
The preparation of the sample was the same as that in the case of the DTA/TG
analysis. Tests were conducted on selected pastes at half setting, 3 days and 28 days.
The powdered sample was kept in air tight containers till the date of testing.
5.5.3 Results and Discussions

29
A Bruker BZH 300 solid state high resolution spectrometer was used for the Si
NMR of the pastes. Tetramethylsilane (TMS) was used as the reference material, and
the chemical shifts were determined with reference to that of the TMS. The details of
the Si NMR spectrum of reference paste, along with superplasticized pastes are given
in Figure 5.7 (a-g). The chemical shifts are represented along the horizontal axis and
the magnitude or intensity of the NMR resonance signal is displayed along the vertical
axis of the spectrum and it is proportional to the molar concentration of the sample. It
is observed that at half-setting, the peak corresponding to Q0 is identified in all the
143
pastes showing the anhydrous silicate phases of the cement (C3S, C2S) corresponding
to the monomer state. Two peaks corresponding to Q0 and Q1 can be identified in the
studied pastes for 3 days and 28 days. The peaks shows that the amount of dimer
gradually increases with a reduction in monomeric silicates for 3 to 28 days. Q2 units
were detected in the case of paste with SMF at 28 days compared to the paste with PCE.
This indicates that the retardation in hydration due to SMF is less than that with PCE.
Intensity of NMR signal
Chemical shifts
(a) Cement paste, half setting (b) Cement paste, 3 days (c) Cement paste, 28 days
144
-73.191
-82.091
-73.232
-84.669
-73.247
-86.299
-68.084
-76.155
-50 -100 -50 -100 -50 -100 -50 -100
(d) and (e) SMF based paste, 3 and 28days (f) and (g) PCE based paste, 3 and 28days
Fig. 5.7 Details of Si NMR spectrum of different pastes for different ages
The relative concentration of the Qn units can be determined by the integration of their
corresponding curves. Since the Q0 and Q1 peaks in the NMR spectra overlap, the
areas corresponding to the Q0 and Q1 peaks, denoted as A0 and A1, respectively, are
obtained for each spectrum as in Figure 5.8.
Q0
Q1
A0
A1
-60 -80 -100
Fig. 5.8 Determination of area of peaks of the Si NMR spectrum
145
The relative area of the Q1 peak (A1/A0) ratios for each paste is shown in Figure 5.9. It
can be observed that the superplasticizer has a significant effect on the polymerization
of the silicates during the hydration process. Its incorporation leads to the formation
of a lower amount of dimers at 3 days as well as at 28 days, reflected by the lower
relative areas of Q1 peak in all the superplasticized pastes compared to reference
paste.
2.5 2.436
Half setting
3 days
28 days
2.0 1.828
1.510
1.5
A1/A0
1.172
1.0
0.699 0.638
0.5
0.0
Reference SMF PCE
Type of mix
Fig. 5.9 Comparison of area of the Q1 peak (A1) with reference to the
area of the Q0 peak (A0) at different ages
It is observed that the silicate polymerization is more in the paste with SMF than in
the paste with PCE i.e., (A1/A0) is more for SMF). The results obtained from setting
tests also supports this result as the setting time for the paste with PCE was more than
that for the paste with SMF. Further to this, the change in the relative area of peaks
indicates a change in the C-S-H morphology in the presence of superplasticizers.
5.6 SUMMARY
A multiple technique based approach, instead of a single test method, is used for
studying the effect of the superplasticizer type on the hydration and the
146
microstructural development for a better understanding of the influence of the
incorporation of superplasticizers on the hardened properties. The comparison of X-
ray diffractograms of the reference paste and superplasticized paste shows that the
superplasticizers mainly influence the hydration of aluminates and consequently the
formation of ettringite rather than the formation of portlandites. However, the
quantitative analysis based on the thermal techniques shows that the amount of C-H
formed for superplasticized pastes is less than that of reference paste at 3 days
Also, the retardation in hydration in the presence of superplasticizers can be observed
through the BSE images, in which the amount of unhydrated particles in the
superplasticized paste is more than that in the reference paste. Also, a lower
polymerization of the silicates has been observed through the NMR study, in the
presence of the superplasticizer, which is reflected by lower proportion of dimers at
the age of 3 and 28 days. In addition, the silicate polymerization is also affected by
the type of superplasticizer; the paste with PCE has a lower amount of dimers
compared to the paste with the SMF. This has been observed in setting tests in which
PCE retarded the setting more than the SMF.
Even though the polymerization of C-S-H is lower in the superplasticized pastes, a
slight increase in compressive strength due the incorporation of superplasticizer was
observed (see Section 4.7); it appears that the mechanical behaviour is dependent
more on the dense and uniform microstructure developed in the superplasticized
pastes. Hence, it is confirmed through these studies that the superplasticizers not only
affect the fresh state properties but also the hardened state properties of the cement
paste.
147
CHAPTER 6
CEMENT-SUPERPLASTICIZER COMPATIBILITY
6.1 GENERAL
The influence of the type and dosage of superplasticizers on the flow behaviour of
paste and concrete, as well as on the microstructure of the cement paste, has been
discussed in the previous Chapters. In addition, the chemical composition of cement
also plays an important role in the cement-superplasticizer interaction and the
incompatibilities that can arise. An incompatibility can be defined as any undesirable
effect that may be produced in the properties of concrete due to the particular
combination of cement and superplasticizer. It is obvious that with a variety of
cements and variability of their chemical compounds incompatible situations could
arise. The literature reveals that cement-superplasticizer compatibility is affected by
the following parameters related to the cement: chemical and phase composition,
especially the C3A content, alkali content, amount and type of calcium sulphate
(gypsum, hemihydrates or anhydrite), cement fineness and free lime content
(Hanehara and Yamada., 1999; Jiang et al., 1999; Yamada et al., 2001; Tagnit-Hamou
et al., 2003).
The objective of this chapter is to understand the influence of the chemical
composition of the cement on the cement-superplasticizer interaction and to suggest
some guidelines for selecting the best combinations. Compatibility in terms of
rheology and setting time is investigated for different types of cements with different
superplasticizers. Also, an attempt is made through an adsorption study based on UV
absorbance to understand the reason for the loss of fluidity and to explain why a
higher dosage is required for some types of superplasticizers.

6.2 MATERIALS SELECTION AND CHARACTERISATION
6.2.1 Types of cements
To consider a reasonable number of material combinations, three commercial brands
of cements were identified. Cement that had created some incompatibility problems in
the laboratory trials was selected as the fourth one. The chemical analyses of these
cements were done in National Test House, Chennai, and the physical properties were
determined in the Construction Materials Laboratory of IITM. The chemical
properties of selected cements are given in Table 6.1, along with its Bogue
composition determined as per ASTM C 150 (2007a). The physical properties are
given in Table 6.2. The important aspects of the four cements are given below.
Cement 1 (C1)
This cement has C3S, C2S and oxide contents in the normal range and a low C3A
content. The alkali and SO3 contents are also within the allowable limits of 0.2-1.2%
and 2-3.5%, respectively, of ordinary portland cement.
Cement 2 (C2)
It contains high C2S and high insoluble residue but a low C3S content. It has the
highest SiO2 content of the cements studied whereas the other oxides are within
normal ranges. This cement has a low alkali content as the (Na2O)eq is 0.086%. Also,
its fineness is higher compared to the other cements.
Cement 3 (C3)
CaO content and alumina/iron oxide ratio is higher for this cement compared to other
cements. Other compounds are within the permissible limits.
Cement 4 (C4)
CaO and SO3 are least for this cement whereas Al2O3 and MgO are highest, compared
to the other cement studied. Loss on ignition is least for this cement whereas the
149
insoluble residue is high. It has the highest C3A content and also a high alkali content
compared to other cements.
Table 6.1. Chemical and Bogue compositions of different cements
Analyte Percentage of mass

C1 C2 C3 C4
CaO 60.81 61.14 62.13 59.44
SiO2 19.5 22.95 20.84 21.04
Al2O3 4.12 4.62 4.88 5.67
Fe2O3 4.72 4.1 4.29 4.99
MgO 1.52 0.72 0.85 1.78
SO3 2.48 2.48 1.69 1.39
L. O. I 3.41 3.61 3.47 1.97
Insoluble residue 1.51 2.98 1.41 3.3
Cl 0.01 0.02 0.03 0.03
Na2O 0.05 0.04 0.18 0.22
K2 O 0.28 0.07 0.52 0.54
(Na2O)eq 0.234 0.086 0.52 0.575
Lime saturation factor 0.93 0.82 0.91 0.86
Alumina/iron oxide ratio 0.68 1.13 1.14 1.136
C3S 53.99 28.21 48.72 30.78
C2S 15.18 44.51 22.99 37.10
C3A 2.94 5.31 5.673 6.58
C4AF 14.36 12.48 13.05 15.18
Table 6.2. Physical properties of different types of cements
Physical property Quantity

C1 C2 C3 C4
Water demand for normal consistency of
cement (%) 33 30 31 31
Initial setting time (minutes) 99 100 170 85
Final setting time (minutes) 184 220 290 215
Blaine specific surface area (m2/kg ) 316 371 355 301
6.2.2 Superplasticizers
Superplasticizers used in the present compatibility study with cements C2, C3 and C4
were selected from the work presented in Chapter 3 where only cement C1 was used.
Four superplasticizers (a lignosulphonate, a melamine, a naphthalene and a
150
polycarboxylate) all belonging to ASTM C 494 (2008a) type F and the fifth one
belonging to Type G were selected; see Table 3.4 for the properties. Another
naphthalene based superplasticizer (SNF-C2), which had created problems of slow
setting, was also used along with C3 and C4; the properties of this admixture are:
solid content of 40%, density of 1.2 kg/litre and pH of 7.5.
6.3 Experimental Procedure
The mixing procedure for the preparation of cement paste was the same as given in
Section 3.4. The pastes were tested at ambient temperature, in the range of 30-32°C.
The tests conducted to understand the flow behaviour were the Marsh cone, mini-
slump and viscometric tests. Marsh cone test was conducted to determine the
saturation dosage of superplasticizer in each case, which is supported by mini-slump
test results. The Bingham model was used to derive the yield stress and plastic
viscosity of different compositions from viscometric test results. The setting of
different combinations was studied with the Vicat apparatus.
6.4 Results and discussions
6.4.1 Influence of Cement Composition on the Flow Behaviour of

Superplasticized Pastes
The flow behaviour of cement C1 with different types of superplasticizers has been
discussed in detail already in Chapter 3; the Marsh cone flow time curves are
reproduced in Figure 6.1(a) to provide comparisons with the behaviour obtained with
other three cements that will be discussed here. The salient features of the behaviour
of the pastes with the cement C1 are: (1) the saturation dosage of LS-C1 is slightly
higher than those of the SNF and SMF based superplasticizers whereas the flow
behaviour is comparable; (2) at very low dosages, SMF-S1 seems to give a high flow
time, indicating the poorer dispersion of cement particles and lower fluidity; (3) the
lower saturation dosage of 0.07% for PCE-D1 indicates that the combined
151
mechanisms of electrostatic repulsion and steric hindrance makes it effective even at
very low dosages; and (4) the flow time curves corresponding to 60 minutes are
similar to those at 0-minutes for the saturation dosage and above (except for the case
of the SMF-S1), indicating a negligible loss in fluidity beyond the saturation dosage,
which reiterates the importance of using an appropriate dosage of superplasticizer for
maintaining the fluidity for the required period.
The Marsh cone flow time curves of 0 and 60 minutes for the other three cements C2,
C3 and C4 are given in Figure 6.1 (b-d). In the case of cement C2, the saturation
dosages are generally higher than those obtained with cement C1 and the loss of
fluidity is high only in the paste with LS-C1. With cement C3, though the initial
fluidity is in the same range as in the other cements for the studied combinations,
there is generally a higher loss of fluidity, especially at dosages less than the
saturation point. The loss of fluidity with cement C4 is even higher for all the studied
combinations and severe at low superplasticizer dosage. This can be attributed to a
high C3A content in cement C4 (compared to cement C1).
LS-C1 0
LS-C1 60
70 Cement C1 SNF-S1 0
SNF-S1 60
60 SNF-D1 0
SNF-D1 60
SNF-D2 0
50 SNF-D2 60
SNF-C1 0
40 SNF-C1 60
SMF-S1 0
SMF-S1 60
30
SMF-C1 0
SMF-C1 60
20 PCE-S1 0
PCE-S1 60
10 PCE-D1 0
PCE-D1 60
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
(a) Marsh cone flow time curves for Cement C1
152
60 Cement C2
LS-C1 0
LS-C1 60

50 SNF-S1 0
SNF-S1 60
SNF-D2 0
40 SNF-D2 60
SMF-S1 0
SMF-S1 60
30 PCE-D1 0
PCE-D1 60
20
10
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
(b) Marsh cone flow time curves for Cement C2
60
Cement C3
LS-C1 0
50 LS-C1 60
SNF-S1 0
SNF-S1 60
40 SNF-D2 0
SNF-D2 60
SNF-C2 0
30 SNF-C2 60
SMF-S1 0
SMF-S1 60
PCE-D1 0
20
PCE-D1 60
10
0.0 0.2 0.4 0.6 0.8

sp/c %
(c) Marsh cone flow time curves for Cement C3
153
70
LS-C1 0
Cement C4 LS-C1 60
60 SNF-S1 0
SNF-S1 60
Marsh cone flow time

50 SNF-D2 0
SNF-D2 60
SNF-C2 0
40 SNF-C2 60
SMF-S1 0
30 SMF-S1 60
PCE-D1 0
PCE-D1 60
20
10
0.0 0.1 0.2 0.3 0.4 0.5 0.6

sp/c %
(d) Marsh cone flow time curves for Cement C4
Fig. 6.1 Marsh cone flow times of different cement-superplasticizer combinations
The data from the Marsh cone, mini-slump and viscometer tests for the pastes with
different superplasticizers at 0 and 60-minutes and all the four cements are given in
Tables 6.3-6.6, respectively. The results show that the trends obtained are generally
the same; i.e., the fluidity of the paste increases with an increase in the dosage of
superplasticizer upto saturation dosage and remains practically constant after that.
The mini-slump spreads reflect the loss of fluidity at low superplasticizer dosages, as
already discussed in the context of the Marsh cone flow times. Beyond the saturation
dosage, the spread value at 60 minutes is almost the same as that obtained
immediately after mixing. It was seen that pastes with mini-slump spreads of more
than 165 mm exhibited bleeding. Also, some pastes that exhibited bleeding at 0
minutes improved in this aspect after 60 minutes due to the reduction in fluidity.
154
Table 6.3 Test data for pastes with different superplasticizers for cement C1
Flow time Mini slump spread Yield stress Plastic viscosity

Type of (sec) (mm) (Pa) (Pa s)
sp/c
sp
%
0 60 0 60 0 60 0 60
0.1 20 21 93 90 12.20 13.50 0.122 0.132

0.2 14 16 153 123 5.81 7.59 0.089 0.097
LS-C1 0.3 13 13 162 143 2.0 3.15 0.059 0.075
0.4 13 14 163 163 1.0 0.35 0.050 0.064
1 12 13 164 162 0.5 0.23 0.068 0.049
0.2 13 26 143 90 16.21 24.61 0.143 0.185
SNF-S1 0.3 11 11 151 150 1.80 5.91 0.055 0.066
0.4 11 11 165 163 0.0 0.21 0.033 0.054
1 11 11 170 178 0.09 0.10 0.031 0.025
0.1 45 * 82 * 18.31 20.23 0.197 0.241
0.2 20 70 96 * 12.13 16.61 0.139 0.178
SNF-D1 0.3 13 13 127 115 2.88 5.98 0.074 0.089
0.4 14 17 bleeding 141 0.08 0.09 0.049 0.083
0.5 14 18 bleeding 164 0.11 0 0.043 0.045
1 14 13 bleeding 163 0.09 0.12 0.033 0.035
SNF- D2 0.1 29 * 87 * 16.99 25.52 0.197 0.194

0.2 17 39 106 87 6.94 12.10 0.088 0.149
0.4 14 14 bleeding 148 0.55 0 0.042 0.043
0.6 13 13 bleeding 176 0.37 0.25 0.041 0.039
1 13 12 bleeding 160 0.13 0.07 0.045 0.038
SNF-C1 0.1 11 14 126 95 6.72 7.24 0.057 0.046

0.2 10 11 152 128 1.25 6.69 0.042 0.064
0.3 10 10 168 168 0.72 0.77 0.029 0.046
1 10 9 173 178 0.24 0.17 0.023 0.035
0.1 62 * * 20.39 * 0.177 *
SMF-S1 0.2 18 44 102 90 6.89 2.39 0.076 0.092
0.25 16 22 115 109 2.36 3.77 0.098 0.101
0.3 15 18 122 113 1.54 8.97 0.078 0.065
1 15 18 167 161 0.77 0 0.048 0.043
0.1 >120 * * * 33.14 * 0.199 *

SMF-C1 0.2 16 26 108 103 13.44 30.92 0.128 0.139
0.3 13 16 132 112 2.88 7.71 0.074 0.100
0.5 12 12 181 169 0 0 0.039 0.042
1 12 11 177 172 0 0 0.037 0.041
0.05 26 * 86.3 * 17.55 * 0.155 *

PCE-S1 0.1 16 * 110 * 11.07 23.67 0.089 0.162
0.2 12 14 bleeding 157 0.51 0.36 0.073 0.068
0.3 11 13 bleeding 179 0 0.08 0.034
0.5 11 14 bleeding 192 0 0 * 0.035
1 12 12 bleeding bleeding 0 0.034 0.034
0.05 18 * 97 * 24.45 * 0.214 *

PCE-D1 0.1 12 16 166 147 0.01 2.37 0.035 0.079
0.2 12 14 170 bleeding 0 0.09 0.035 0.043
0.4 12 14 199 bleeding 0 0 0.036 0.036
1 14 16 193 bleeding 0 0 0.047 0.054
155
From the data of the viscometer tests, it can be seen that in general the yield stress
decreases significantly with an increase in the dosage of superplasticizer down to a
value of practically zero beyond the saturation dosage. The plastic viscosity also
reduces with an increase in the superplasticizer dosage. It is observed that the yield
stress and plastic viscosity developed vary from 0-43 Pa and 0.034-0.188 Pa.s,
respectively. Both the rheological parameters increase with time as seen in the
comparison of the 0 and 60 minute data.
Type Flow time Mini slump spread Yield stress Plastic viscosity
of sp/c (sec) (mm) (Pa) (Pa s)
sp % 0 60 0 60 0 60 0 60
0.1 36 * 85 * 33.45 * 0.185 *
0.2 13 59 123 90 15.33 24.34 0.095 0.142
LS-C1 0.3 13 45 140 113 7.72 12.99 0.086 0.113
0.5 11 14 bleeding 143 0.26 3.61 0.047 0.090
1 14 23 bleeding 120 0.24 1.28 0.068 0.130
0.1 26 * 84 * 38.16 * 0.162 *
0.2 13 28 125 89 11.64 1.83 0.095 0.076
SNF- 0.3 11 15 bleeding 138 1.2 3.04 0.073 0.073
S1 0.5 12 11 bleeding bleeding 0.19 0.32 0.040 0.034
1 11 10 bleeding bleeding 0.31 0.13 0.043 0.043
0.1 31 * 85 * 22.56 * 0.170 *
0.2 12 19 130 113 6.78 26.27 0.074 0.181
SNF- 0.3 11 13 153 143 0.44 3.85 0.057 0.073
D2 0.5 10 10 bleeding bleeding 0 0.014 0.043 0.039
0.8 10 10 bleeding bleeding 0 0 0.038 0.046
0.1 * * * * * * * *
0.2 20 * 85 * 20.73 34.26 0.123 0.188
0.3 15 16 148 145 13.13 12.32 0.090 0.053
SMF- 0.4 15 15 149 150 1.65 3.72 0.083 0.078
S1 0.5 14 15 153 148 0.70 1.15 0.084 0.082
1 13 13 bleeding bleeding 0.04 0 0.052 0.052
0.05 47 * 75 * 42.97 * 0.165 *
0.1 18 40 115 85 13.39 23.98 0.110 0.143
PCE- 0.2 13 16 145 153 0.51 0.80 0.087 0.102
D1 0.3 13 14 155 163 0.42 0.36 0.074 0.074
1 16 18 158 148 0.32 0.56 0.086 0.092
156
Type Flow time Mini slump spread Yield stress Plastic viscosity
of sp/c (sec) (mm) (Pa) (Pa s)
sp % 0 60 0 60 0 60 0 60
0.05 22 * 83 * 14.99 * 0.163 *
0.1 15 * 102 74 13.98 17.33 0.119 0.112
0.2 13 * 128 77 11.39 21.82 0.099 0.147
LS-
C1 0.3 12 34 145 91 4.07 15.74 0.072 0.122
0.4 12 23 bleeding 108 0.29 12.70 0.056 0.100
0.6 12 21 bleeding 148 0.067 2.11 0.019 0.103
0.025 29 * 97 75 16.10 21.78 0.152 0.197
0.05 21 * 100 85 15.84 25.34 0.169 0.145
0.1 15 * 107 85 14.26 19.04 0.109 0.182
SNF-
S1 0.2 11 15 139 118 3.21 12.16 0.061 0.109
0.3 10 11 bleeding 154 0.59 0.53 0.043 0.054
0.6 10 11 bleeding 0 0.62 0.029 0.084
0.025 21 * 85 76 3.189 22.91 0.086 0.168
0.05 14 40 108 85 13.272 15.93 0.113 0.135
0.1 12 21 125 108 7.573 14.95 0.058 0.124
SNF-
D2 0.2 11 11 140 148 1.023 0.76 0.036 0.065
0.3 12 12 bleeding 165 0 1.61 0.048 0.289
0.6 10 11 bleeding 170 0 0 0.033 0.021
0.025 26 * 93 * 16.02 * 0.138 *
0.05 19 * - 78 15.68 18.39 0.145 0.182
0.1 16 44 - 79 9.61 18.65 0.079 0.176
SNF-
C2 0.2 15 27 108 83 7.75 16.05 0.073 0.137
0.3 15 15 144 108 3.22 5.02 0.048 0.065
0.4 12 12 bleeding 190 0 0 0.019 0.042
0.025 23 * 98 * 17.59 21.75 0.17 0.217
0.05 18 * 93 * 14.60 20.42 0.126 0.179
0.1 16 * 103 73 14.31 * 0.122 *
SMF-
S1 0.2 11 16 128 103 5.30 9.63 0.071 0.069
0.3 11 13 141 115 1.53 5.46 0.056 0.073
0.5 11 12 bleeding bleeding 0.72 0.15 0.043 0.039
0.025 12 * 130 * 6.78 * 0.077 *
0.05 11 55 135 85 9.45 17.02 0.056 0.116
0.075 10 13 bleeding 130 1.44 3.87 0.039 0.071
PCE-
D1 0.1 10 14 bleeding 120 1.12 4.38 0.035 0.094
0.2 10 10 * 168 0 - 0.030 -
0.3 10 9 bleeding 175 0 0.12 0.032 0.0096
157
Flow time Mini slump spread Yield stress Plastic viscosity

Type (sec) (mm) (Pa) (Pa s)
sp/c
of sp
% 0 60 0 60 0 60 0 60
0.1 36 * 78 * * * * *
0.2 14 * 105 * 22.83 * 0.144 *
LS-C1 0.3 13 * 125 * 10.36 * 0.088 *
0.4 12 18 165 100 2.59 * 0.073 *
0.6 14 16 bleeding bleeding 0.30 0.1 0.047 0.054
0.05 37 * 88 * 24.81 * 0.221 *
0.1 25 * 86 * 22.26 * 0.188 *
SNF- 0.2 13 23 130 106 5.42 17.40 0.072 0.137
S1 0.3 11 12 bleeding 146 0.15 1.07 0.038 0.044
0.5 11 10 bleeding 168 0 0 0.026 0.026
0.05 65 86 * 29.59 * 0.264 *
0.1 14 >84 118 83 10.95 22.60 0.095 0.238
SNF- 0.2 12 14 158 124 0.37 2.71 0.045 0.071
D2 0.3 11 11 178 160 0 0.0029 0.034 0.055
0.1 56 * 81 * 33.21 * 0.209 *
0.2 16 * 101 83 21.80 37.50 0.159 0.261
0.3 14 28 114 85 10.31 22.73 0.082 0.156
SNF- 0.4 11 15 bleeding 123 1.129 9.21 0.063 0.071
C2 0.5 11 12 bleeding bleeding 0 0.16 0.051 0.037
0.6 11 11 bleeding 165 0 0 0.043 0.037
0.1 40 * 87 * 35.41 * 0.304 *
0.2 16 * 107 * 20.69 * 0.185 *
0.3 12 21 139 91 3.99 22.03 0.072 0.182
SMF- 0.4 12 17 145 105 4.04 9.29 0.078 0.087
S1 0.5 10 11 bleeding 168 0.24 0.09 0.052 0.040
0.05 31 * 83 * * * * *
0.075 14 * 112 * 28.25 * 0.132 *
PCE- 0.1 13 * 123 * 13.33 * 0.079 *
D1 0.15 10 16 bleeding 111 0.21 7.48 0.044 0.102
0.2 11 13 bleeding bleeding 0.09 0 0.034 0.06
0.3 11 11 178 bleeding 0.03 0 0.025 0.035
0.6 10 10 188 184 0 0 0.026 -
* could not be determined
Table 6.7 gives a comparison of the saturation dosages for the four cements obtained
from Marsh cone flow curves at 0 minutes, using the same method as given in
Chapter 3. It is observed that the saturation dosages of the LS, SMF and PCE based
superplasticizers are generally the highest for the cement C2, mainly because it has
comparatively higher surface area and hence the amount of superplasticizer needed is
higher, which is in accordance with the reported results (Erdogdu, 2000). Also, it is a
158
low alkali cement and therefore is expected to adsorb more superplasticizer (Dodson
and Hayden, 1989; Jiang et al., 1999)
Table 6.7 Comparison of the saturation dosages at 0-minutes for different cements
Type of Saturation dosages%

superplasticizer C1 C2 C3 C4
LS-C1 0.25 0.40 0.25 0.4
SNF-S1 0.20 0.25 0.20 0.25
SNF-D1 0.24 * * *
SNF-D2 0.23 0.3 0.2 0.2
SNF-C1 0.16 * * *
SNF-C2 * * 0.2 0.4
SMF-S1 0.23 0.4 0.2 0.3
SMF-C1 0.20 * * *
PCE-S1 0.2 * * *
PCE-D1 0.07 0.2 0.08 0.15
Cement C3 gives good fluidity even at very low dosages and the saturation dosages
obtained are comparatively lower for the various superplasticizers. The yield stress
and plastic viscosity values are lower compared to those of pastes with other cements.
However, some pastes with cement C3 do show a loss of fluidity within 60 minutes.
All combinations with cement C4 had rapid slump losses within 60 minutes and the
saturation dosages of SNF-C2, PCE-D1 and LS-C1 were high. This may be due to the
high C3A content, which causes some of the superplasticizer to be absorbed by the
initial hydration products of C3A and hence less would be available in the solution for
further adsorption. Also, cement C4 has a low SO3 content and hence the SO42- ions in
the superplasticizer will compete with the sulphate ions of the cement, causing more
initial adsorption, resulting in the increase of the superplasticizer demand.
159
The results reported by Kumar (2006) on the compatibility of concrete using cement
C1 and C2 and different superplasticizers used here generally support the results of
the present study; he found that the superplasticizers SMF-S1, SNF-S1 and PCE-D1
in combination with cement C1 gave better slump retention in concrete, even at low
superplasticizer dosages, compared to the concretes with cement C2. As an extension
of the previous study, Das (2007) conducted tests to study the compatibility of
cements C3 and C4 and the same superplasticizers. His results also show that
concretes with cement C4 had more slump loss than those with cement C3, as
reported in the present study.
6.4.2 Influence of Cement Composition on the Setting of Superplasticized Pastes
The setting time is affected by both the chemical composition of cement, as well as by
the type of superplasticizer, as seen in Figure 6.2 (a-c). It is seen that the trend of
setting varies among the cements studied. It can be observed that for cement C1 and
C3, the lowest retardation is seen for SMF-S1 but not in C4. For cement C4, LS-C1
and SNF-D2 superplasticizers gave the least retardation compared to other
superplasticizers with the same cement. This is in sharp contrast to the behaviour of
LS-C1 and SNF-D2 in the other cements, where it leads to high retardation. The mix
with cement C4 and admixture SNF-C2 does not set even after 24 hours whereas the
paste with cement C3 and the same admixture sets in about 10 hours.
It can be seen that the setting behaviour can vary significantly depending on the
cement composition and the superplasticizer used. Some of the trends cannot be
explained on the basis of the literature or the data obtained in this work, and call for
further research to provide the understanding needed.
160
Cement C1 Control paste
40 LS-C1
SNF-S1
Initial setting SNF-D2

SMF-S1
30 PCE-D1
20
10
0
200 400 600 800 1000 1200
Time (minutes)
(a) Vicat penetration test results of Cement C1
45 Control paste
Cement C3 LS-C1
40
SNF-S1
Initial setting SNF-D2
35
SNF-C2
30 SMF-S1
PCE-D1
25
20
15
10
5
0
200 400 600 800
Time (minutes)
(b) Vicat penetration test results of Cement C3
161
45
Cement C4 Control paste
40 LS-C1
Initial setting SNF-S1

35
SNF-D2
30 SNF-C2
SMF-S1
25 PCE-D1
20
15
10
5
0
400 800 1200 1600 2000
Time (minutes)
(c) Vicat penetration test results of Cement C4
Fig. 6.2 Influence of type of cement on setting
The results of Kumar (2006) on setting with cement C1 shows the same trend as
observed in the present paste study; the SNF-S1 gives lower setting time than PCE
and SNF-D2 based superplasticized paste. The results of Das (2007) on setting are
also comparable with the present results; he showed that concrete with the
superplasticizer SNF-S1 exhibited higher setting time with cement C4 whereas SNF-
D2 had a lower setting time, when compared with identical concretes with the cement
C3.
6.5 STUDY OF THE ADSORPTION OF SUPERPLASTICIZERS
6.5.1 Background
As discussed earlier, the superplasticizer is adsorbed on the cement surface; the rate
and amount of this adsorption varies with type of cement and superplasticizer, which
in turn affects the fresh state properties of concrete. Consequently, adsorption plays an
important role in the initial flow behaviour, the maintenance of fluidity as well as the
setting behaviour of pastes. Hence an understanding of adsorption of different
162
superplasticizers on different types of cement is required for answering several
questions like why some superplasticizers that have good initial fluidity can not
maintain it even for 60-minutes. Also, it could explain the variation in saturation
dosage of the same superplasticizer for different types of cements.
The increase in workability due to the addition of superplasticizers occurs through the
dispersion of agglomerated cement particles. However, the quantification of this
mechanism is a difficult task and is further complicated by the hydration reactions.
The adsorbed superplasticizer causes fluidity and the superplasticizer that is
remaining in solution is responsible for fluidity retention, as explained in Chapter 2.
UV absorbance techniques have been used to understand the amount of adsorption by
several researchers, who have concluded that the method is highly precise and the
results obtained are reproducible (Daimon and Roy, 1978, Andersen et al., 1987,
Krishna, 1996, Kim et al., 2000).
In the present study, the adsorption behaviour of four families of superplasticizers and
its relation to the fluidity of cement paste has been investigated at the saturation
dosage of superplasticizer using UV absorbance technique.
6.5.2 Materials used
Superplasticizers belonging to the four families of products, namely lignosulphonates
(LS), melamines (SMF), naphthalenes (SNF) and polycarboxylates (PCE), were used.
The details of the superplasticizers are given in Table 3.4. OPC 53 grade cements C1,
C2,C3 and C4, with the characteristics given in Table 6.1, were used in the study.
The concentration of polymer in the liquid phase was determined through a
quantification of the UV absorbance using a JASCO V-570 UV spectrophotometer in
163
the 190-400 nm region. According to the Beer-Lambert’s law, the absorbance is
directly proportional to the concentration of the absorbing material in solution. In
absorption spectroscopy, the intensity of light of a particular wavelength that is
absorbed by the sample is measured and used to determine the absorbance using the
following equations:
I0
A = log10 (6.1)
I1
where A is the absorbance, A = α l c (6.2)
α is the absorption coefficient, l is the distance that the light travels through the
material, c is the concentration of the absorbing species in the material, I0 is the
intensity of incident light, and I1 is the intensity after passing through the material.
UV spectrometry is used here for the measurement of the absorbance of different
superplasticizers. The test procedure is as follows.
Step 1: Determination of the characteristic peak wavelength (λmax) of the

superplasticizer
The first step in the method is to determine the wavelength corresponding to
maximum absorbance for a particular superplasticizer; absorbance measurements
made at that wavelength are expected to have the least errors. 0.1 ml of a solution of
water and superplasticizer, with a known superplasticizer concentration, is taken and
further diluted as 250X, 500X, 1000X, 10000X, etc. Water, used as the reference
liquid for the measurements, was placed in another couette along with the diluted
sample in the UV-visible spectrometer. Measurements are made for each of the
diluted solutions until an absorbance peak is obtained within the measurable range of
the instrument (maximum of 5 for the equipment used here). The UV spectra for the
different superplasticizers are shown in Figure 6.3, showing different peak absorbance
164
values and the corresponding wavelengths (denoted as λmax). Note that the dilutions
are different in all the cases so that absorbance values are not directly comparable. It
is observed that the λmax is 205 nm for LS-C1, 215 nm for SMF-S1, 225 nm for SNF-
S1 and 195nm for PCE-D1. The absorbance by the superplasticizer in the aqueous
solution was determined hereafter at the corresponding λmax.
3.5
205nm
3.0 LS-C1
SNF-S1
2.5 225nm SMF-S1
Absorbance
PCE-D1
2.0
1.5
1.0
0.5 215nm
195nm
0.0
200 250 300 350 400
Wavelength (nm)
Fig. 6.3 Absorbance spectra for different superplasticizers
Step 2: Determination of the absorbance of the superplasticizers in cement paste
The cement pastes were prepared with the saturation dosage of superplasticizer. The
aqueous solution was extracted by centrifuging the paste at 6000 rpm for 10 minutes.
The extracted solution was diluted with deionized water to obtain a solution with an
absorbance in the measureable range. The absorbance spectrum is obtained and the
peak absorbance was noted for the corresponding wavelength, as explained in step 1.
6.5.4 Results and discussion of absorbance tests
The absorbance spectrum of the aqueous solution extracted from the paste with
cement C1 and lignosulphonate based superplasticizer just after mixing is given in

165
Figure 6.4. The comparison of the same with Fig. 6.3 shows that peak absorbance for
this superplasticizer occurs at the same wavelength of 205 nm for both the solution in
water and the aqueous solution extracted from the cement paste. It is also seen that
there is no other peak in the spectrum, which indicates that there is no other species in
the aqueous solution of the paste with significant absorbance other than the
superplasticizer. The same is true for all the other superplasticizers, as seen in the
other plots of Figure. 6.4.
3.0
LS-C1 with cement C1
2.5 SNF-S1 with cement C1
SMF-S1 with cement C1
PCE-D1 with cement C1
2.0
Absorbance
Pure cement(C1) paste
1.5
1.0
0.5
0.0
200 250 300 350 400

Wavelength (nm)
Fig. 6.4 Absorbance spectrum of the diluted aqueous solutions

of superplasticized pastes extracted after mixing
The absorbance spectrum of the solution extracted from the pure cement paste is also
given in the same graph. Note that each of the curves is for a different dilution and are
not, therefore, directly comparable; the dilutions for the corresponding aqueous
solutions are: LS-C1: 250X (i.e., 0.1 ml of the aqueous solution is diluted in 25 ml of
water), SNF-S1: 500X, SMF-S1: 1000X and PCE-D1: 250X. The absorbance spectra
of the superplasticized cement pastes at 90 minutes were also obtained in the same
166
manner. Note that the age of the aqueous solution at the time of the measurements is
estimated as 15 minutes.
The absorbance values for tests at 15-minutes and 90-minutes for the different
superplasticizers in combination with cement C1 are summarized in Table 6.8. The
absorbance at the saturation dosage for different pastes depends on both the family of
superplasticizer and chemical composition of cement. The absorbance of SNF-S1 and
SMF-S1 is lower at 90-minutes compared to that at 15-minutes showing that the
adsorption of the superplasticizer has increased with time. In the other cases, the
absorbance is almost the same at both 15 and 90 minutes indicating that the
adsorption of those superplasticizers does not increase over this time range. It should
be noted that the presence of cement particles in the aqueous solution and the possible
desorption of the superplasticizer from these fine particles during subsequent dilution
process could distort the absorbance values by giving falsely high values that could be
attributed to the unadsorbed admixtures (Daimon and Roy, 1978). This is a major
limitation of the procedure and further studies are needed to get a better estimate of
the amount of adsorbed superplasticizer on the cement particles and its evolution with
time.
Table 6.8 Summary of absorbance test results

Dilution Absorbance
Combination 15 90 15 90
minutes minutes minutes minutes
Cement C1+LS-C1 250X 250X 2.7 2.9
Cement C1+SNF-D1 500X 500X 2.6 1.9
Cement C1+SMF-S1 1000X 1000X 1.1 0.8
Cement C1+PCE-D1 250X 250X 1.3 2.1
Cement C1 +PCE-D1 (0.05%) 250X 250X 1.1 1.2
Cement C2 +LS-C1 250X 250X 2.2 2.0
Cement C2+PCE-D1 250X 250X 0.4 0.4
Cement C4+PCE-D1 250X 250X 1.3 1.0
167
6.5.4.1 Variation in Superplasticizer Absorbance with Type of Cement
The absorbance for the pastes with cement C2 and LS-C1 is less than that with cement
C1 and LS-C1 for the same dilution (see Figure 6.5). This suggests that the initial
adsorption may be more in the case of the paste with cement C2 and LS-C1 than in
the paste with cement C1 and LS-C1, resulting in the higher loss of fluidity, as
described earlier. This could be attributed to the higher surface area of cement C2 and
its low alkali content, both of which can increase the adsorption. Similarly, for the
paste with cement C2 and the PCE, the absorbance at 15 and 90-minutes is low
compared to the C1-PCE-D1 combination, supporting the results shown earlier where
there is more loss of fluidity and higher superplasticizer demand in the former.
2.5
15 minutes
2.0 90 minutes
Absorbance
1.5
1.0
0.5
0.0
200 250 300 350 400
Wavelength (nm)
Fig. 6.5 Absorbance spectra for cement C2 and LS-C1
6.6 IDENTIFYING COMPATIBLE CEMENT-SUPERPLASTICIZER

COMBINATIONS
As indicated in the preceding chapters, a number of mechanisms and interactions can
contribute to the so-called incompatibility between the cement and superplasticizer.
All of these mechanisms are complex and interrelated. The results show that a single
168
test is not sufficient to determine a compatible cement-superplasticizer combination.
Consequently, a simple methodology that combines different test procedures is
developed based to select the compatible combinations.
A flow chart representing the methodology for the selection of an appropriate cement-
superplasticizer combination based on the present interaction study is shown in
Figure 6.6. The entire methodology is divided into two steps; the method of selection
of a compatible combination of cement and superplasticizer from the paste studies is
described in step 1. Step 2 describes the methodology for the final selection of
cement-superplasticizer combination based on workability, strength and cost by
conducting trial tests on concrete.
Step 1
• The water demand and setting time of cement, and the density, solid content
and pH of superplasticizer are to be determined for recording the basic
properties for the quality checks, as well as for comparing different batches of
products.
• The Hobart mixer or a similar intensive mixer is considered to be essential for
the preparation of paste for selecting a compatible combination based on the
paste studies as well as for correlating with concrete results.
• The first criterion in the selection process is based on the ability to obtain a
well-defined saturation dosage from the Marsh cone test that is within the
maximum dosage of the superplasticizer as recommended by the supplier. The
study of the flow behaviour of superplasticized paste has shown that results
from the mini-slump and viscometer tests can be correlated with those of the
Marsh cone test results and, therefore, need not be carried out since they
169
require more sophisticated equipment or may not be as sensitive. The
correlation between the flow behaviour of paste, mortar and concrete has
shown that the dosage obtained from the paste tests can be used as the
guideline for selecting the dosage of superplasticizer for concrete.
• As seen from the studies on the loss of flow and UV absorbance, the flow
behaviour can vary significantly with time in the fresh paste. Therefore, it is
recommended that the saturation dosage be obtained at the age that
corresponds to the actual initiation of the placing of concrete (e.g., age when
the concrete is expected to be delivered to the site, in the case of ready-mixed
concrete).
• The influence of superplasticizer on the setting behaviour of the paste is the
second criterion for compatibility. Here, the final setting time is limited to 16
hours as obtained in the Vicat penetration test, unless there is deliberate
retardation through the use of a retarder or a superplasticizer with retarding
properties. The results of the study suggest that if this criterion is satisfied, the
FST of concrete would be not more than 12 hours since the difference between
the FSTs of paste and concrete was found to be normally not more than 4
hours. Since the electrical resistivity measurements require more technical
expertise than the Vicat penetration apparatus and there is only limited
experience with such tests, it is deemed to be unnecessary for determining the
compatibility. In any case, the IST obtained with Vicat apparatus correlates
well with the dormant stage extension obtained from electrical resistivity
measurements.
• The microstructural analysis is not included in the methodology since the
studies previously discussed show that a paste with adequate workability and
normal setting behaviour generally has a uniform microstructure.
170
Step 2
• Final trials are done on concrete, within this selection procedure, to
complement the paste tests and, more importantly, to ensure a mix with the
required slump and compressive strength. It is assumed that the paste volume
and the aggregate skeleton are optimized independently, for example with a
method based on packing density or other, as in the methods of Toralles-
Carbonari et al. (1996) and Gomes et al. (2001).
• The study has shown that the fines content in the aggregates can increase the
superplasticizer demand to more than the saturation dosage of the
superplasticizer, which also supports the necessity for the final trials on
concrete. In such cases, a higher dosage than the saturation point may be used
for achieving the desired workability; however, it is recommended that the
dosage is not more than 150% of the saturation dosage in order to limit the
retardation and possible segregation, as well as cost. Obviously, the
superplasticizer dosage can be reduced if the slump obtained is more than the
desired value when the saturation dosage is used.
• From the point of view of productivity, a minimum value of 50% of the 28-
day compressive strength is recommended at 3 days. This criterion can be
waived when the early age strength is not critical.
• The final choice of the superplasticizer can be based on the minimum cost
considering the dosage of the superplasticizer for obtaining the desired slump
among the compatible combinations. Superplasticizers can also be short-listed
for further testing based on tests on pastes, for example, by comparing flow
times and saturation dosages.
171
• Determine density, pH and solid content of superplasticizer, and compare with previous
data and manufacturer’s specifications. If there is disagreement, check with another batch.
• Determine the consistency and setting time of cement and compare with previous data and
manufacturer’s specifications. If there is disagreement, check with another batch.
• In case of anomalies, reject corresponding batch.
Mixing of the paste with Hobart mixer or similar
Determination of saturation dosage of superplasticizer using Marsh cone test,

at the required age.
• Is there a well-defined saturation point?

• Is the saturation dosage within the No
Step 1 manufacturer’s recommended limit?
Yes
Incompatible
Determine the final setting time (FST) combination
using Vicat test at saturation dosage
No
FST < 16 hours
Compatible combination
Mix for concrete with saturation dosage of the

superplasticizer
Check for the required slump. If slump is higher than required, decrease the dosage. If slump
is lower than required, increase the dosage up to a maximum of 150% of the saturation
dosage; if the required slump is still not obtained, reject the superplasticizer.
Determine the compressive strength at 3 days.

Reject if the strength is less than 50% of the
Step 2 target 28 day strength.
• Determine the cost of the superplasticizer per m3 of

the concrete
• Select the superplasticizer with the minimum cost
for the dosage required for the desired slump
Fig. 6.6 Methodology to select the Compatible Cement-Superplasticizer Combination
172
The summary of the paste test results along with the compatibility criteria derived
from the methodology for different combinations at 0-minutes and 60-minutes are
given in Tables 6.9 and 6.10, respectively. The nomenclature used for the mixes is as
follows; the first letter represents the type of cement followed by the admixture code,
e.g.,C1/LS-C1 is the mix with cement C1 and superplasticizer, LS-C1.
Table 6.9 Summary of Superplasticized paste combinations at 0-minutes
Saturation dosage
within the Final
Mix
manufacturer’s Setting Compatible?
compositions
recommended ≤ 16 hr
limit at 0 min
C1/LS-C1 Yes Yes Yes
C1/SNF-S1 Yes Yes Yes
C1/SNF-D1 Yes Yes Yes
C1/SNF-C1 Yes Yes Yes
C1/SMF-S1 Yes Yes Yes
C1/SMF-C1 Yes Yes Yes
C1/PCE S1 Yes Yes Yes
C1/PCE-D1 Yes Yes Yes
C2/LS-C1 No * No
C2/SNF-S1 Yes * Yes
C2/SNF-D2 Yes * Yes
C2/SMF-S1 Yes * Yes
C2/PCE-D1 Yes * Yes
C4/LS-C1 No Yes No
C4/SNF-C2 No No No
*Could not be determined
The compatibility of different combinations are studied as per the criteria given in
step 1. Accordingly, C2/LS-C1 and C4/LS-C1 seem to be incompatible at 0-minutes
as the saturation dosages are more than the recommended limit. C4/SNF-C2 is also
173
incompatible at 0-minutes as per both criteria; the saturation dosage is above the
recommended limit and the setting time is more than 16 hours. The incompatible
combinations are marked with bold letters.
The test results at 60-minutes give more incompatible combinations due to the
significant loss of fluidity in some combinations (see Figure 6.1(a-d)). Accordingly,
C1/SMF-C1, C3/LS-C1 become incompatible as their saturation dosages are above
the manufacturer’s recommended limit whereas the absence of a well-defined
saturation dosage makes C2/LS-C1, C4/LS-C1 and C4/SMF-S1 incompatible. The
SNF-C2 is also incompatible due to the same reasons as mentioned for 0-minutes.
Table 6.10 Summary of Superplasticized paste combinations at 60-minutes
Saturation dosage within the Final

Mix
manufacturer’s recommended Setting Compatible?
Compositions
limit at 60 min ≤ 16 hr
C1/SNF-S1 Yes Yes yes
C1/SMF-C1 No Yes No
C1/PCE S1 Yes Yes Yes
C2/LS-C1 No * No
C2/SNF-S1 Yes * Yes
C2/SNF-D2 Yes * Yes
C2/SMF-S1 Yes * Yes
C2/PCED1 Yes * Yes
C3/LS-C1 No Yes No
C4/LS-C1 No Yes No
C4/SNF-D2 Yes Yes yes
C4/SNF-C2 No No No
C4/SMF-S1 No Yes No
*could not be determined
174
The literature gives other approaches to identify the compatible cement-
superplasticizer combinations. According to Aïtcin (1998), in a cement-
superplasticizer combination, if the 0 minutes Marsh cone flow time curve and 60
minutes flow time curve intersect or are close to each other, it can be considered as a
compatible combination, and the saturation dosage is taken as the dosage at the point
of intersection of the curves. According to Aïtcin's criterion, all the combinations
studied here are compatible, except those with the superplasticizer LS-C1 with the
cements C2, C3 and C4. The combinations that are incompatible according this
criterion are also incompatible as per the proposed selection criteria, which is
however, more stringent (by taking into consideration the setting of the paste) leading
to more incompatible combinations.
Furthermore, Nkinamubanzi et al. (2000) has classified the compatibility of different
cements through their chemical compositions. They suggested the use of the SO3/C3A
ratio and alkali content for selecting compatible combinations. Figure 6.7 attempts to
classify the tested cements accordingly. According to Nkinamubanzi et al. (2000), a
high SO3/C3A ratio and high alkali content are considered as the requirements for
compatibility. Consequently, the cements C1, C2 and C3 are classified as compatible
and cement C4 is less compatible due to a low SO3/C3A ratio. Note that the SO3
present is considered to be completely soluble, which may not be always the case.
According to the proposed method (see test results in Tables 6.9 and 6.10), the cement
C4 has more incompatible combinations, and therefore, can be considered as less
compatible than the other cements, which is in accordance with the classification of
Nkinamubanzi et al. (2000).
175
Fig. 6.7 Classification of cement based on its chemical composition
(based on Nkinamubanzi, et al., (2000)
6.7 SUMMARY
The influence of the chemical composition of cement and family of superplasticizers
on the flow and setting behaviour of different pastes are analysed in the present
chapter. It is confirmed that both the fresh and hardening properties vary for different
combinations. In order to understand this difference in nature, the absorbance of the
admixtures has been determined, as the rate and amount of adsorption of
superplasticizers on cement surfaces is responsible for the effective dispersion. Also,
the UV absorbance at different times helps understand the mechanisms that cause the
loss of fluidity of superplasticized pastes. This study confirms that a good correlation
exists between the loss of fluidity with initial adsorption. In addition to that, the λmax
is found to be a characteristic of a particular superplasticizer and hence UV
absorbance can be used as a simple test to identify the family of superplasticizer.
Finally, a simple methodology based on the flow and setting behaviour of pastes is
suggested to identify the compatible cement-superplasticizer combinations and is
validated with other approaches from the literature.
176
CHAPTER 7
FLOW BEHAVIOUR OF SUPERPLASTICIZED PASTES

INCORPORATING OTHER ADMIXTURES
7.1 GENERAL
An objective procedure for determining the saturation dosage of superplasticizer in a
cement paste using the Marsh cone test has been explained in Chapter 3 for
understanding the flow behaviour of the paste. The experimental and analytical
correlation between the Marsh cone flow times with rheological parameters has also
been presented. The flow behaviour of the paste was found to relate well to the fresh
state properties of normal concrete showing the relevance of conducting the tests on
paste. These procedures applied for understanding the behaviour of superplasticized
cement paste are extended in the present Chapter to pastes with other admixtures such
as metakaolin and viscosity modifying agents.
The incorporation of mineral admixtures, viscosity modifying admixtures (VMA) and
superplasticizers in the paste composition results in a balance between the cohesion
and fluidity of special cement-based materials such as cement grouts, shotcrete,
underwater concrete and self compacting concrete (SCC) (Khayat and Guizani, 1997;
Lachemi et al., 2004, 2007). Superplasticizers generally decrease the yield stress
whereas viscosity modifying admixtures tend to increase the viscosity at the same w/c
ratio (Ferraris et al., 2001; Sonebi, 2006). Moreover, the interaction between different
admixtures in a cementitious system sometimes results in delays in setting time and
rapid slump loss, which demand a compatible combination to achieve the best
rheological properties (Khayat and Yahia, 1997).

The objective of this chapter is to understand the flow behaviour of superplasticized
paste with components such as mineral admixtures and VMAs. The work is of
particular interest for the development of self compacting concrete (SCC) as it
generally incorporates both the superplasticizer and viscosity modifying agent.
The present chapter details the procedures and results of tests performed with the
Marsh cone, mini-slump and viscometer on superplasticized pastes with metakaolin or
a VMA. For correlating the results from the tests on pastes, an experimental study of
the fresh behaviour of SCC was done through its characteristic tests like slump flow,
V-Funnel, J-Ring in addition to compression tests. Based on the results, the
methodology of Gomes et al. (2001) and Gettu et al. (2004) for the mix design is
extended to SCC incorporating a VMA.
7.2 MATERIALS USED IN THE STUDY
53 grade ordinary portland cement (cement C1), PCE-D1 and SNF-D2
superplasticizers, metakaolin and four types of VMAs are used in the study. The
characteristics of cement and superplasticizers have already been reported in
Chapter 3.
The commercial availability of metakaolin in India has increased in the last few years
due to extensive kaolin sources across the country (Kumar and Kaushik, 2003).
Metakaolin is a primary product formed by the calcination of kaolin clay at a
temperature of 650-800°C. It belongs to the chemical family of China clay-
aluminium silicate (Al2O3 2SiO2). It contains 50-55% SiO2 and 40-45% Al2O3. Its
specific gravity is 2.6, bulk density is 300 kg/m3 and specific surface is 12000m2/kg.
Additional, C-S-H is produced due to its reaction with Ca(OH)2 as indicated in
Equation 7.1:
AS2 + 6CH + 9H Æ C4AH13 + 2C-S-H (7.1)
178
Four types of VMAs were used in the study with the properties given in Table 7.1.
VMA 4 is diutangum whereas the others are liquids based on polysaccharides.
Table 7.1 Properties of viscosity modifying agents
Density
Name of VMA Solid content (%) appearance
(gm/ml)
VMA 1 1.02 2 liquid
VMA 2 1.00 2 liquid
VMA 3 1.01 2 liquid
VMA 4 - 100 powder
7.3 FLOW BEHAVIOUR OF PASTES WITH METAKAOLIN
Metakaolin is said to increase the total volume of finer pores and decrease the volume
of coarser pores; however, the total volume of pores is reported to be less than in pure
cement paste (Khatib and Wild, 1996). Its density is lower than that of cement
whereas the surface area is higher with irregular shape particles. Consequently, the
flow behaviour is affected by the incorporation of metakaolin with higher loss of
fluidity and more shear thickening when compared with pure cement paste (Cyr et al.,
2000; Corinaldesi and Moriconi, 2003; Li and Ding, 2003).
It has also been reported that metakaolin increases the water demand and apparent
viscosity of paste due to an increase in the solid fraction of paste. The superplasticizer
type appears to significantly affect the workability and compressive strength of mortar
and concrete incorporating metakaolin as the rate of pozzolanic reaction and hydration
depends on the type of superplasticizer; Kim et al. (2003) found that SNF blends and
PCE cause slump retention whereas the SNF and SMF cause slump loss. It was also
reported that the PCE retards the hydration reaction and hence causes a reduction in
compressive strength.
179
7.3.1 Effect of Metakaolin in Paste incorporating PCE Based Superplasticizer
For the pastes, cement and metakaolin were taken in the ratio of 90:10 and the
water/powder ratio (w/p) was taken as 0.35. The Marsh cone, mini-slump and
viscometer tests were conducted immediately after mixing (0 minutes) and after 60
minutes. The trend observed in the Marsh cone flow time with metakaolin is same as
that of cement-PCE combination with a decrease in flow time with an increase in the
dosage of superplasticizer. All the combinations seem to exhibit well defined
saturation points. The addition of metakaolin increases the saturation superplasticizer
dosage from 0.1% to 0.2% (in terms of solid content of sp/c%), as shown in Figure
7.1, which may be due to the increase in solid fraction and surface area in the paste.
The loss of fluidity is also observed to be higher than in the cement-superplasticizer
combination, which confirms earlier results (Li and Ding, 2003). The level of flow at
the saturation dosages are, however, practically the same.
1.22
1.20 Cement-PCE-0-minutes
Cement-PCE-60-minutes
1.18 Cement-metakaolin-PCE-0-minutes
Cement-metakaolin-PCE-60-minutes
1.16
log (flow time,sec)
1.14
1.12
1.10
1.08
1.06
1.04
1.02
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 7.1 Marsh cone flow time curves for pastes with
metakaolin and PCE based superplasticizer
180
The mini-slump spread increases with an increase in the dosage of superplasticizer in
the cement-superplasticizer-metakaolin combination, as seen in Figure 7.2. At 0
minutes, both the pastes exhibit bleeding beyond the saturation dosage, and hence the
corresponding data are not plotted in the graph. The addition of metakaolin decreases
the mini-slump spread, as shown in Figure 7.2.
200
180
160
140
120 Cement-PCE-0-minutes
Cement-PCE-60-minutes
Cement-metakaolin-PCE-0-minutes
100 Cement-metakaolin-PCE-60-minutes
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 7.2 Variation of mini-slump spread with dosage of PCE for
metakaolin-PCE combination
The rheological study of the pastes showed that the addition of metakaolin results in
an increase in the yield stress at low dosages and remains constant beyond the
saturation dosage of superplasticizer, as seen in Figure 7.3. The plastic viscosity also
increases with the addition of metakaolin, especially at low dosages of
superplasticizer, as shown in Figure 7.4. With regards to both the parameters, the
addition of metakaolin does not produce any significant difference in the rheological
parameter beyond the saturation dosage. The observed trends are in disagreement with
the reported results of Ferraris et al. (2001), who did not find any significant increase
in the yield stress or viscosity due to the addition of metakaolin mix. However, the
181
present data are in accordance with the results of Cyr and Mouret (2003). It is possible
that the variations in the properties of the metakaolin lead to the differences in the
conclusions drawn from the different studies.
Cement-PCE 0
30 Cement-PCE 60
Cement-metakaolin-PCE 0
Yield stress (Pa)
20
10
0
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 7.3 Variation of yield stress with dosage of PCE for

metakaolin-PCE combination
200
Cement-PCE 0
180 Cement-PCE 60
Plastic Viscosity (cP)
160 Cement-metakaolin-PCE 60
140
120
100
80
60
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 7.4. Variation of plastic viscosity with dosage of PCE for
Metakaolin-PCE combination
182
7.3.2 Effect of Metakaolin in Paste Incorporating SNF Based Superplasticizer
The flow curves at 0-minutes and 60-minutes for the cement-metakaolin-SNF based
paste are shown in Figure 7.5. The flow curves at 0-minutes shows that the addition of
metakaolin increases the saturation dosage of SNF. In spite of the retarding effect of
the superplasticizer SNF-D2, there is a significant loss of fluidity within 60-minutes in
the paste with metakaolin and a well defined saturation dosage cannot be obtained for
the flow curve corresponding to 60-minutes. This may be either due to incompatibility
between the metakaolin and SNF, as reported by Kim et al. (2003), caused by the
absorption of SNF by the metakaolin. The addition of metakaolin decreases the mini
slump spread at 0-minutes as well as for 60-minutes (see Figure 7.6) confirming the
results of Kim et al. (2003). The viscometer test results also shows the same trend,
with an increase in yield stress and plastic viscosity with an increase in the time after
mixing (see Figures 7.7 and 7.8).
The comparison of test results shows that SNF causes more loss of fluidity than PCE
based superplasticizer and is significantly less compatible with the metakaolin than
the PCE. The results also show that a compatible cement-superplasticizer combination
can give incompatibility related problems, such as loss of fluidity, when a mineral
admixture is incorporated.
183
1.5
Cement-SNF-0-minutes
1.4 Cement-metakaolin-SNF-0-minutes
Cement-metakaolin-SNF-60-minutes
log (flow time, sec) 1.3
1.2
1.1
1.0
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 7.5 Marsh cone flow time curves for pastes with metakaolin and SNF-D2
180
mini-slump spread (mm)
160 Cement-metakaolin-SNF-60-minutes
140
120
100
80
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 7.6 Variation of mini-slump spread for pastes with metakaolin and SNF-D2
184
30 Cement-SNF-0-minutes
Yield stress (Pa)
20
15
10
0.0 0.2 0.4 0.6 0.8 1.0

sp/c %
Fig. 7.7 Variation of yield stress for paste with metakaolin and SNF-D2
180
150
120
90
60
30
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
Fig. 7.8 Variation of plastic viscosity for paste with metakaolin and SNF-D2
7.4 INFLUENCE OF VMA ON THE FLOW BEHAVIOUR OF

SUPERPLASTICIZED PASTE
7.4.1 Characteristics of VMA
VMAs are water soluble polymers that increase the viscosity of the paste and enhance
the ability of the paste to retain its constituents in suspension (Khayat and Yahia,
1997; Lachemi et al., 2007). The main function of the VMA is to modify the
rheological properties of cement paste. As mentioned earlier, the rheological
behaviour of fluid cement pastes can be described by the yield stress and plastic
viscosity. The VMA changes the rheology by increasing the plastic viscosity but with
185
a small increase in yield shear stress. On the other hand, superplasticizers tend to
decrease the yield stress. In conjunction, the superplasticizer and VMA can be used to
optimize the yield shear stress. The test methods used are those explained earlier were
used here to understand the flow behaviour of paste incorporating VMA.
VMAs are either cellulose or polysaccharide based, and they are classified as natural,
semi-synthetic or synthetic polymers. Natural polymers include starches, guar gum,
locust bean gum, alginates, agar, gum Arabic, welan gum, xanthan gum, rhamsan
gum, and gellan gum, as well as plant protein. Semi-synthetic polymers include
decomposed starch and its derivatives, cellulose-ether derivatives such as
hydroxypropyl methyl cellulose (HPMC), hydroxyethyl cellulose (HEC) and carboxy
methyl cellulose (CMC), as well as electrolytes such as sodium alginate and
propylene glycol alginate. Synthetic polymers consist of polymers based on ethylene
such as polyethylene oxide, polyacrylamide, polyacrylate and those based on vinyl
such as polyvinyl alcohol (Khayat and Yahia, 1998). According to their physical
action, VMAs are classified into five categories (Rixom and Mailvaganam, 1999):
Class A: Water-soluble synthetic and natural organic polymers that increase the
viscosity of the mixing water. These types of materials include cellulose ethers,
polyethylene oxides, polyacrylamide, polyvinyl alcohol etc.
Class B: Organic water-soluble flocculants that become adsorbed on cement grains
and increase the viscosity due to enhanced interparticle attraction between cement
grains. It includes styrene copolymers with carboxyl groups, synthetic
polyelectrolytes and natural gums.
186
Class C: Emulsions of different organic materials that enhance interparticle attraction
and supply additional superfine particles in the cement paste. The materials belonging
to Class C are acrylic emulsions and aqueous clay dispersions.
Class D: Water-soluble inorganic materials having high surface area that increase the
water retaining capacity of the paste such as bentonites, silica fume, and milled
asbestos.
Class E: Inorganic materials having high surface area which increase the content of
fine particles in paste and hence the thixotropy. They include fly ash, hydrated lime,
kaolin, various rock dusts, and diatomaceous earth, etc.
7.4.2 Experimental Details
The methodology used for the mix design of SCC in the present study is an extension
of an approach proposed by Toralles-Carbonari et al. (1996) for high strength silica
fume concrete and later modified by Gomes et al. (2001) and Gettu et al. (2004) for
SCC. Within this approach, concrete is considered as a two phase material consisting
of a paste that provides fluidity and cohesion, and an aggregate skeleton that provides
the mechanical integrity. The flow behaviour of paste and concrete is studied in the
present work by varying the type and dosage of superplasticizer and VMA, while
keeping the other parameters such as temperature and cement type constant.
7.4.2.1 Study of the Interaction between Superplasticizer and VMA
The mixing method for the preparation of paste is same as that used in the preparation
of superplasticized cement paste, except that the VMA is added after three minutes
and it is further mixed for 2 minutes. The flow behaviour of the paste was determined
using Marsh cone, mini-slump and viscometer. Preliminary tests were conducted with
three different types of VMAs (VMA 1, VMA 2 and VMA 3) and a PCE based
superplasticizer at two w/c ratios (w/c = 0.5 and w/c = 0.35) to understand the
187
superplasticizer-VMA interaction. The saturation dosage of the superplasticizer was
determined as 0.2% and trials were conducted with a low dosage of 0.3% of different
VMAs. This study identified the VMA leading to the highest Marsh cone flow time
and the least mini-slump spread as VMA 1; see Figures 7.9 and 7.10 (VMA 0 denotes
the mix without VMA). This was considered as the most effective in terms of
achieving the cohesion compared to the other VMAs and hence it is used for further
studies. Also, it is observed that the effectiveness of the VMA increases with a
decrease in w/c as shown in Figures 7.9 and 7.10. This may be because, the molecules
in adjacent polymer chains can develop attractive forces at low w/c, further blocking
the movement of water, causing gel formation and a consequent increase in viscosity
(Khayat and Guizani, 1997). The plastic viscosity also shows the same trend as the
Marsh cone flow time and mini-slump spread and is given in Figure 7.11.
40
PCE- 0.2% VMA 0
35
0.3% VMA 1
30 0.3% VMA 2
0.3% VMA 3
25
20
15
10
0
w/c = 0.5 w/c = 0.35
Fig. 7.9 Marsh cone flow time for different viscosity modifying agents
188
180
PCE - 0.2% VMA 0
160
0.3% VMA 1
140
0.3% VMA 2
120 0.3% VMA 3
100
80
60
40
20
0
w/c = 0.5 w/c = 0.35
Fig. 7.10 Mini slump spread for different viscosity modifying agents
18 100
PCE - 0.2%
w/c = 0.35 90
16 yield stress
14 plastic viscosity 80
70 Plastic viscosity(cP)
yield stress(Pa)
12
w/c = 0.5 60
10
50
8
40
6
30
4 20
2 10
0 0
VMA 0 VMA 1 VMA 2 VMA 3 VMA 0 VMA 1 VMA 2 VMA 3
type of VMA
Fig. 7.11 Values of yield stress and plastic viscosity for

different viscosity modifying agents
189
With the w/c ratio set at 0.35, Marsh cone flow time curves were determined for low
and high dosages of VMA 1 with different dosages of superplasticizer (PCE-D1) The
test results are given in Figure 7.12; where it can be observed that Marsh cone flow
time increases with an increase in the dosage of the VMA. For comparison, results are
also reported for a Welan gum based VMA, namely VMA 4, at the dosages that are
normally used. It is seen that VMA 4 leads to higher flow times than VMA 1 at both
dosages.
2.0 0.3% VMA 1

1% VMA 1
1.9 0.01% VMA 4
0.03% VMA 4
log ( flow time, sec)
1.8
1.7
1.6
1.5
1.4
0.0 0.1 0.2 0.3 0.4 0.5 0.6
sp/c%
Figure 7.12 Determination of saturation dosage of superplasticizer
at different dosages of VMA
The mini-slump spread, the spread time to reach a diameter of 115 mm (t115) and
rheological parameters were determined for different dosages of VMA with the
corresponding saturation dosages of superplasticizers from paste tests. The results are
summarized in Table 7.2. Gettu et al. (2004) recommend a value of about 2 seconds
for t115 to obtain a paste with adequate cohesion for SCC, which is obtained in all the
present cases. However, their requirement of a minimum spread of about 170 mm is
not met, except in the case of VMA 4 at a dosage of 0.01%.
190
Table 7.2 Flow characteristics of pastes with different types and dosages of VMA
Saturation Type and Flow Mini- t 115 Yield Plastic

dosage of dosage of time slump (sec) stress viscosity
SP (sp/c VMA (%) (sec) spread (Pa) (cP)
%) (mm)
0.2 VMA 1 - 0.3 34 133 2 4 126
0.3 VMA 1 - 1.0 66 112 3 3 130
0.2 VMA 4 - 0.01 49 168 2 0 282
0.2 VMA 4 - 0.03 68 140 2 * *
* values could not be measured
7.4.2.2 Effect of Time on the Flow Behaviour of Paste Incorporating

Superplasticizer and VMA
The effect of VMA to increase as well as to maintain the viscosity depends on the
type of superplasticizer and type of VMA (Khayat and Yahia, 1997). In order to
understand the efficiency of VMA to maintain the cohesion for longer period, flow
tests were conducted at 0, 30, 60 and 90 minutes with PCE-D1 based superplasticizer
and VMA 1 and VMA 4 at a w/c of 0.35. The saturation dosage of superplasticizer for
different combinations were determined as explained in Section 7.4.2.1.
The rheological parameters were also investigated at low and high dosages of VMA.
The Marsh cone test results of different combinations are shown in Figure 7.13 and the
summary of flow test results are given in Table 7.3. The test results at different ages
show that only in the paste with PCE-D1 and VMA 4, the Marsh cone flow time
increases with time. It is observed that in other combinations, instead of increasing, the
Marsh cone flow time decreases with time. The mini-slump test result also supports the
Marsh cone results; the mini-slump spread decreases with time in PCE-D1-VMA 4
combination. The t115 value also increases with time in the same combination. This
shows that even though VMA 1 is effective in giving cohesion and stability to the paste
initially, it is not effective in maintaining the cohesion and fluidity even for a period of
60-minutes. This is because the yield stress of paste in this combination is very low and
the mix loses its cohesiveness as the plastic viscosity decreases with time.
191
Here, it is seen that VMA 4 is more effective in increasing, as well as maintaining the
viscosity of the paste than VMA 1. This may be due to the difference in the chemical
structure of the two VMAs.
90
VMA 1 0.3%
80 VMA 1 1%
VMA 4 0.01%
70 VMA 4 0.03%
60
50
40
30
20
10
0
0 30 60 90
Time after mixing (minutes)
Fig. 7.13 Effect of time on Marsh cone flow curves of SP-VMA mix
Table 7.3 Influence of time on SP-VMA mix
Type and Mini-

Type and Time after Flow Plastic
dosage of slump t115 Yield
dosage of mixing time viscosity
SP spread (sec) stress (Pa)
VMA (%) (min) (sec) (cP)
(sp/c%) (mm)
0 27 115 2 4.5 121
PCE- D1 VMA 1 30 18 125 1 3.8 82.4
0.2% 0.3% 60 16 122 1 3.6 85.7
90 18 140 1 3.3 84.4
0 43 147 3 0 125
PCE- D1 VMA 1 30 38 155 2 0 107
0.3% 1% 60 37 145 1.5 0 93
90 35 154 1 0 95
0 36 148 1.5 1.6 235
PCE-D1 VMA 4 30 43 149 2 2.9 263
0.2% 0.01% 60 51 135 3 5.4 260
90 89 109 6 8.6 285
0 64 138 2 * *
PCE-D1 VMA 4 30 75 128 3 * *
0.2% 0.03% 60 84 123 4 * *
90 87 115 5 * *
* data could not be determined
192
7.5 COMPARISON OF THE FLOW BEHAVIOUR OF PASTE AND SELF
COMPACTING CONCRETE
7.5.1 Mix design and testing of SCC
The procedure for the determination of the saturation dosage of superplasticizer in the
paste phase with VMA has been explained in Section 7.4, with the intention of
obtaining a paste with good cohesion and fluidity, which are essential for the
flowability and resistance to segregation of highly workable concretes. Since adequate
flow behaviour of paste is fundamental for achieving concrete with adequate fluidity
and viscosity, tests have been performed on SCC to validate the paste test results, as
done for normal concrete in Chapter 4. The mix proportioning of SCC is performed
by separately determining the saturation dosage of superplasticizer for prescribed
dosages of VMA in the paste phase, optimizing the aggregate proportion for minimum
void content, and then the paste content of the concrete to get the final composition,
following the method of Gomes et al. (2001).
As the first step in the mix design of SCC, the paste phase is optimized for high
fluidity and adequate cohesion. In the method of Gomes et al. (2001), a paste with the
saturation dosage of superplasticizer, as determined from the Marsh cone test, and
meeting the criteria of mini-slump spread of about 170 mm and a t115 of about 2
seconds (Gettu et al., 2004) is considered to be appropriate for SCC. The pastes
studied in the previous section with the superplasticizer and VMA dosages given in
Table 7.2, and with a w/c of 0.35, satisfied the above-mentioned criteria for pastes,
except for the requirement of a spread of about 170 mm that is met only in the case of
0.01% dosage of VMA 4. Nevertheless, SCC mixes were developed for all the four
pastes for further evaluation.
193
For the definition of the aggregate skeleton, the maximum aggregate size was limited
to 20 mm in order to have high fluidity without segregation. The optimum aggregate
combination is selected by considering the packing density and lowest void content
with the assumption that minimum void content leads to the minimum paste volume,
porosity and shrinkage. However, concrete requires more paste content than the
minimum void content to ensure cohesion and fluidity. The aggregate phase is
optimized here by measuring the dry uncompacted density of aggregate mixes and
choosing the mix with least void content (Gomes et al., 2001). The method is same as
that given in ASTM C29/C29M (2007), except that the uncompacted weight is taken,
which is more suitable for SCC.
In the present study, the sand, and coarse aggregates with maximum grain sizes of 20
mm and 10 mm were dry mixed manually in a tray and placed in a 15 litre container
without any compaction. The proportions of the 5-10 mm : 10-20 mm fractions were
fixed as 3:2. The density or unit weight was measured to determine the void content,
as shown in Figure 7.14. It was observed that the sand content that gave the minimum
void content was 45 %, with a void content of 28%.
46 2000
44 Void ratio
42 Bulk density 1900
40
Bulk density (kg/m )
3
1800
Void content (%)
38
36
1700
34
32 1600
30
28 1500
26
0 20 40 60 80 100
Sand (%)
Fig. 7.14 Determination of aggregate proportions for concrete
194
The SCC has to satisfy certain requirements in the fresh and hardened state .The
filling ability and stability of SCC in the fresh state can be defined by four key
characteristics, which can be determined by the following test methods recommended
by EFNARC Guide lines (2005), Italian UNI standards and ASTM C1621 (2006)
code.
Flowability Slump flow test
Viscosity and segregation resistance T 500 or V-Funnel test
Passing ability J-Ring test
The following are the acceptance criteria of SCC that were prescribed based on the
above test methods.
Table 7.4 Acceptance criteria for SCC
Test values
Slump flow 550-700 mm
Slump flow spread time (T 500) 2-5 sec
V-funnel flow time 10 ± 3 sec
Difference in V-funnel flow time between ≤ 3 sec
measurements made immediately and after 5
minutes
Difference between J ring Flow and slump < 50 mm
flow
28 days compressive strength >50 MPa
In the present study, the slump flow, V-funnel and J-ring tests have been conducted
for characterizing the self compacting concrete (EFNARC Guidelines for SCC, 2005).
In addition to that the compressive strength of concrete has been determined at 3 days,
7 days and 28 days.
Slump flow test is a simple method to evaluate the filling ability of SCC under its own
weight. The test method is shown in Figure 7.15. The slump cone is filled with
concrete without compaction and is lifted to allow the concrete to flow over a plain
195
surface. The time for 50cm spread (t50) is measured along with the final spread as the
average of two perpendicular diameters. The concrete is visually examined for the
segregation and bleeding.
Fig. 7.15 (a) Slump cone measurements (b) Slump Cone test
A combination of slump flow with a ring denoted as J-Ring has been used to study the
blockage due to reinforcement (see Figure 7.16). As in the case of slump flow test, the
slump cone kept inside the J-Ring is filled with concrete. The concrete flows through
the gaps as the cone is lifted. The final spread is measured in two perpendicular
directions which gives an idea about the passing ability of concrete.
(a)
196
(b)
Fig. 7.16 (a) J-Ring measurement (b) J-Ring test
The V-funnel test method consists of measuring the time taken for a certain volume
(approximately 10 litres) of concrete to flow through a funnel (see Figure 7.17). The
funnel is again filled with the same concrete and waited for 5 minutes. The outlet is
opened after 5 minutes and the time taken for it to flow is noted. The difference
between the initial and final time should be less than 5 seconds so as to satisfy the
characteristic of SCC. Thus the test gives a good indication of viscosity of the paste as
well as an indication of segregation.
Fig. 7.17 V-Funnel test

197
The trial mixes showed that adequate paste content is needed for achieving a
satisfactory SCC. Here, the paste volume was set as 38%, which is about 10% more
than the void content determined in the aggregate compaction tests. The mix
proportions used for the SCC tests are given in Table 7.5; the paste compositions are
the same those characterized in Section 7.4 except that the superplasticizer dosage has
been varied to evaluate the influence of the same. The aggregates used have been
described already in Chapter 5.
Table 7.5 Proportion of materials used in the SCC trials
Fine Coarse aggregate Superplasticizer –

Cement Water
aggregate PCE-D1 VMA (%)
Kg/m3 Kg/m3 10 mm 20 mm
Kg/m3 (sp/c %)
Kg/m3 Kg/m3
0.3%
0.1,0.2,0.25, 0.3
VMA 1
1%
0.25, 0.3, 0.35, 0.4
VMA 1
550 193 800 494 332
0.01%
0.1,0.2,0.25, 0.3
VMA 4
0.03%
0.2, 0.25, 0.3, 0.4
VMA 4
SCC was prepared with different dosages of superplasticizers for the different SP-
VMA combinations, and with low and high dosages of VMA in each combination.
The mixing procedure is same as that of normal concrete except that the total mixing
time is 7 minutes in which VMA was added after 4 minutes. The slump flow, V-
funnel and J-ring tests are used for studying the fresh state properties. In addition, 3, 7
and 28-day compressive strength were also determined.
The results are given in Table 7.6, with the mixes satisfying the acceptance criteria of
SCC denoted in bold letters. It is observed that in all the combinations, a dosage equal
198
to or higher than the saturation dosage is needed for obtaining an adequate slump
flow. Also, other acceptance criteria are not satisfied if the dosage is lower than the
saturation superplasticizer dosage. The superplasticizer dosage slightly higher than the
saturation dosage results in t50 value less than one second, showing that the mix is not
viscous enough for a stable concrete. Also the passing ability of concrete is not
improved with the extra superplasticizer. An overdosage of the PCE based
superplasticizer results in a reduced slump flow in some combinations as reported in
the paste study. This is attributed to the shear thickening and immediate stiffening of
the paste.
Table 7.6 Test results of SCC with various SP-VMA combinations
Saturation Slump V- V- J ring Compressive

Combi- sp/c T50
dosage of Flow Funnel Funnel spread strength (MPa)
nation (%) (Sec)
sp/c from (mm) Time0 Time5 (mm) 3 7 28
paste(%) (Sec) (Sec) days day days
0.1 290 * * * * 33 38 43
VMA 1 0.2 665 2.1 9 9.1 590 36 47 56
0.3% 0.2
0.25 675 1 4.3 5.2 610 36 39 49
0.3 715 <1 4.81 5.93 710 42 39 54
0.25 460 * 6.5 9.3 450 36 44 49
VMA 1 0.3 475 * 7.5 11 455 39 44 56
1% 0.3
0.35 580 2 5.1 5.6 575 38 43 59
0.4 565 1.41 4.9 5.7 555 37 43 52
0.1 * * 23.7 * * 27 37 47
VMA 4 0.2 600 3.2 7.1 9 570 38 44 52
0.01% 0.2
0.25 680 1.5 7 9.3 668 37 42 52
0.3 775 <1 6 6.2 650 39 44 53
0.2 545 2 12 14.3 460 39 48 54
VMA 4 0.25 645 2.2 15.4 15.6 605 40 50 58
0.03% 0.2
0.3 713 1.5 13 13.2 630 40 48 59
0.4 663 2.1 12 14 663 40 49 60
The best performance for each combination is observed at or slightly higher than the
saturation dosage of superplasticizer obtained from the paste studies; an increase in
dosage of 0.05% was required for the PCE at higher dosage of VMAs. The only
combination that satisfies all the criteria is with 0.01% of VMA 4, which was also the
199
paste that came closest to satisfying the criteria of Gettu et al. (2004). The
compressive strength results show that the strengths achieved are satisfactory in all
the mixes at the corresponding saturation dosages. The study shows that paste test
results can be used as guidelines for selecting the best combination of SP and VMA
with appropriate dosages for developing SCC.
7.6 SUMMARY
• The flow behaviour of paste composed of cement, superplasticizer, mineral
admixture and VMA was investigated using the Marsh cone, mini-slump and
viscometer tests. It is seen that Marsh cone test along with mini-slump spread
is required to understand the complete flow behaviour of the paste with these
compositions.
• The addition of mineral admixtures and VMA generally increases the flow
time with the corresponding change in rheological properties (i.e., the yield
stress and plastic viscosity increase).
• The study confirms that the interaction between other admixtures and
superplasticizer depends on the type and dosage of admixture and
superplasticizer, which in turn affects the fluidity and slump retention.
Generally, the paste with the PCE has good initial fluidity and maintains the
fluidity better than the paste with the SNF in all the tested compositions with
the mineral admixture. Among the different SP-VMA combinations studied, it
is seen that the PCE-VMA 4 combination is able to maintain the cohesion over
time unlike the other combinations. Consequently, it is emphasized that the
compatibility of the SP-VMA combination should be studied prior to use in
concrete since it can affect the time dependent behaviour of the fresh concrete
and result in loss of cohesion with time, in as early as 30-minutes. The slump
200
flow values gradually decrease and the viscosity increases as time passes, in
most of the cases.
• A mix design procedure for developing SCC with VMA has been presented
based on the flow tests of pastes. The study shows that the type and dosage of
superplasticizer and VMA together govern the viscosity of SCC. The
saturation dosage of superplasticizer obtained from paste studies were found to
give good results in SCC. Consequently, it can be confirmed that the
optimization of superplasticized cement paste can be used as a guideline in the
preliminary selection of chemical admixtures for SCC.
201
CHAPTER 8
CONCLUSIONS AND RECOMMENDATIONS FOR

FURTHER RESEARCH
8.1 GENERAL CONCLUSIONS
Overall, the study was able to improve the understanding of the mechanisms of
cement–superplasticizer interaction and their influence on concrete performance.
Moreover, the effectiveness of the Marsh cone test in characterising the fluidity of
cement paste and the factors to be considered while establishing the correlation
between flow behaviour of paste and concrete were brought out. The salient
conclusions that correspond to the fulfillment of the principal objectives of this work
are given below.
• The preparation of superplasticized cement paste with four mixing methods
and the comparison of the saturation dosages obtained from these mixing
methods with “ball milling” to simulate the mixing of concrete, demonstrate
the influence of mixing on the flow behaviour of the paste. The dependence of
the fluidity of the paste and the saturation dosage of the superplasticizer on the
intensity of mixing suggests that a paste prepared using Hobart with B-flat
beater is representative of the paste phase of concrete.
• The establishing of a correlation between the empirical methods with the
rheological parameters shows that the same trends are obtained in the Marsh
cone flow time, the yield stress and plastic viscosity when there is an increase
in the dosage of the superplasticizer. The prediction of Marsh cone flow time
using rheological parameters further confirms the correlation.

• The correlation between the fluidity and setting of the fresh paste and that of
the mortar and concrete through standard tests shows that the trends are
similar in all cases. The comparison of the flow behaviour at 60-minutes
shows that the aggregates increase the loss of fluidity. Further, the presence of
fines in the coarse aggregate increases the superplasticizer demand for
adequate workability. However, it is confirmed that the optimization of the
paste phase is a preliminary step in the mix design of concrete and the
appropriate type and dosage of the superplasticizer can be selected on the basis
of tests on pastes.
• The use of multiple characterization techniques instead of a single method is
an effective approach for the analysis of the effect of superplasticizers on the
hydration processes as well as on the microstructural development. The
complementary use of XRD, SEM, DTA/TG/DTG and Si NMR has given an
insight into the properties of hardened cement paste. The retardation of the
setting of the superplasticized paste is attributed to the delay in the formation
of ettringite, as well as with the reduction in the silicate polymerization in the
presence of superplasticizers. The delay in hydration is also reflected by the
presence of more unhydrated particles in the micrographs of the
superplasticized pastes.
• The major characteristics of cement and superplasticizer affecting the cement-
superplasticizer interaction have been identified. The mechanisms of action of
the superplasticizer affect its effectiveness and consequently its compatibility
with cement. The major characteristics of the cement affecting the
compatibility are identified as SO3/C3A ratio, alkali content and fineness.
203
Finally, a methodology, based on the test results, is suggested as a guideline
for the selection of an appropriate cement-superplasticizer combination.
8.2 SPECIFIC CONCLUSIONS
The specific conclusions derived from the experimental and analytical investigations
on cement-superplasticizer interaction are classified under the following topics: (i)
selection of a suitable mixing method for the preparation of paste, (ii) selection of
type and dosage of superplasticizer based on Marsh cone test, (iii) rheological
characterization of superplasticized paste using viscometer and rheometer, (iv)
establishing experimental and analytical correlation between Marsh cone test with
rheological parameters, (v) validation of paste test results with normal concrete and
self compacting concrete, and (vi) influence of superplasticizers on hardening and
hardened state properties of concrete.
A number of tests on paste have been performed using the Marsh cone and
Viscometer to confirm the effectiveness of simple test methods in bringing out the
influence of superplasticizers on the flow behaviour of paste. The rheological
parameters like yield stress and plastic viscosity provided ample information about the
flow behaviour of paste immediately after mixing as well as at 60-minutes. The
complementary use of multiple techniques to understand the effect of
superplasticizers on hardened properties validates the test results even with a few
number of samples.
8.2.1 Influence of Mixing Method on the Flow Behaviour of Paste
• The type of mixing affects the fluidity as well as the saturation dosage of
superplasticizer. Hence, the selection of a proper mixing method for the
preparation of paste is required while correlating the superplasticized paste
204
flow behaviour with that of concrete.
• The study helps to classify the different mixing methods as less intensive,
intensive, and more intensive; less intensive mixing like hand mixing leads to
higher saturation dosage and low fluidity, intensive mixing using Hobart with
B-flat beater, reduce the superplasticizer demand in terms of saturation
dosage, more intensive mixing like blender mixing even though leads to
highest fluidity, and gives higher saturation dosage than the intensive mixing.
• The study confirmed that the paste prepared with the Hobart mixer with B-flat
beater simulates the paste phase of concrete. This clarifies the conflicting
views in the literature regarding the preparation of paste with blender and
Hobart mixer, for establishing a correlation between paste and concrete
behaviour.
8.2.2 Influence of Superplasticizers on the Flow Behaviour of Paste
• A detailed study of the influence of type and dosage of superplasticizers on the
flow properties of different compositions along with the determination of the
saturation dosage of superplasticizer has been conducted. The Marsh cone test,
a simple but effective method, has been used for measuring the fluidity. The
results of this test are found to have trends comparable to the rheological
parameters (i.e., yield stress and plastic viscosity) obtained from viscometric
tests.
• Even though the mini-slump test is also used to understand the flow behaviour,
the study showed that, in pastes with high fluidity, it is not useful for
identifying the influence of superplasticizers effectively.
205
• The saturation dosages of a polycarboxylate based superplasticizer with the
tested cements are confirmed to be less than lignosulphonates, melamines and
naphthalenes. However, the loss of slump in concrete with a polycarboxylate
based superplasticizer is more than in concrete with a naphthalene based
superplasticizer.
• The influence of the saturation dosage of superplasticizer on the loss of
fluidity, and overdosage on settling and bleeding throws light on the selection
of appropriate dosage of superplasticizer in any combination. The study on
loss of fluidity of paste shows that if the fluidity at a later stage is more critical
to the application than the immediate fluidity, then the saturation dosage
corresponding to 60-minutes or any other critical time is required for
maintaining the fluidity for that duration.
• The tests using a rheometer with parallel plate attachment showed that the
nature of the superplasticized paste changes from viscoelastic fluid-like
behaviour to linearly viscous flow with an increase in the dosage of the
superplasticizer.
• The addition of mineral admixtures and VMA makes the paste more cohesive,
and increases the yield stress and viscosity as their dosages increase.
Therefore, the saturation dosage of the studied superplasticizers were all
observed to increase with the addition of the admixtures.
8.2.3 Comparison between Marsh Cone Flow time and Rheological Parameters
• By applying the Bingham and Herschel Bulkley models to the flow behaviour
of superplasticized paste, it is seen that an increase in the dosage of the
206
superplasticizer leads to a significant decrease in the yield stress and plastic
viscosity up to saturation dosage, after which they remain practically constant.
• The prediction of Marsh cone flow time using rheological models shows that
the Marsh cone test results have a fundamental basis. Therefore, it is justified
that it is suitable for evaluating the relative fluidity of superplasticized paste.
8.2.4 Effect of Type and Dosage of superplasticizer on the Retardation of Setting
• The Vicat penetration test results confirmed that the superplasticizer retards
the setting of cement paste due to its effect on hydration. The present study
with saturation dosage of superplasticizer showed that amount of retardation
mainly depends on the type of superplasticizer. The PCE and LS based
superplasticizers retard more than other families.
• Electrical conductivity measurements on the cement paste have been used to
characterize the setting behaviour of cement paste. It has been observed that
the superplasticizer leads to retardation of the setting process, which is mainly
manifested as an increase in the duration of the dormant stage; the
superplasticizer that gives the highest initial setting time leads to the longest
dormant period.
• Even though the superplasticizer retards the initial setting, the duration
between the initial and final setting is less than in pure cement paste and is
comparable for different superplasticizers. This shows that superplasticizer
does not prolong the setting process once it is initiated.
8.2.5 Correlation between Paste, Mortar and Concrete
• It is found that the Marsh cone test is a satisfactory tool for comparing the
flow behaviour of different pastes and their components, and for defining the
207
saturation dosage of the superplasticizer. The saturation dosage of
superplasticizer immediately after mixing and after 60-minutes is observed to
be generally different. Hence, the saturation dosage corresponding to 60-
minutes flow time curves is more suitable to select a superplasticizer for
maintaining the fluidity for more than one hour.
• The mortar flow tests show that the incorporation of sand does not increase the
saturation dosage significantly in the context of the present study. However,
the loss of fluidity of mortar is higher than that of paste at low dosages and has
to be considered while comparing the mortar behaviour to that of paste phase.
• Flow tests on normal concrete and self compacting concrete show that a slight
increase in the saturation dosage of superplasticizer is required for adequate
workability in all combinations. This is attributed to the excess of fines in the
coarse aggregate used here.
• The comparison of the setting of paste and concrete showed that even though
the setting times are not directly related, the trends of the setting of paste and
concrete with superplasticizers are comparable. That is, SMF-S1 causes the
least retardation and PCE-D1 results in higher retardation than SNF-S1 in both
paste and concrete.
• The compressive strength increases with an increase in dosage of
superplasticizer and is maximum for the highest slump obtained without any
segregation or bleeding. This emphasizes the importance of good workability
in achieving adequate strength. The decrease in strength observed beyond
certain dosage is due to bleeding and segregation.
208
• The cost of concrete derived from the individual prices of superplasticizer,
cement and aggregate shows that low unit cost of superplasticizer need not
always result in cost effective concrete. For example, in the studied
combinations, even though the unit cost (cost/kg) of SMF-S1 is less than PCE-
D1, the former leads to more cost per m3 of concrete because the dosage of the
superplasticizer required for a particular slump is higher for SMF-S1.
8.2.6 Effect of Superplasticizer on the Hydration process and on the

Microstructural Development
• Instead of a single technique, multiple techniques (XRD, SEM, thermal
methods and Si NMR) have been used complementarily for understanding the
effect of superplasticizers on the hydration process and consequently on the
microstructure.
• The XRD study gives a qualitative evaluation of the effects of
superplasticizers on the hydration. It confirms that the superplasticizer affects
hydration and delays the formation of ettringite. In addition to that, it shows
that secondary ettringite is present even upto 28 days in some combinations.
The evaluation shows that the amount of C-H increases with the age of curing.
The presence of C-H peaks at half setting shows that the formation of C-H is
not delayed as the formation of ettringite in the presence of the
superplasticizers.
• The SEM study shows that the incorporation of a superplasticizer leads to more
dense and uniform structure than pure cement paste. However, the amount of
unhydrated particles at 3 days in the superplasticized paste is more than that of
pure cement paste confirming that the superplasticizer retards the hydration.
209
• The DTA/TG/DTG analysis also confirms the XRD results that the C-H
formation is less affected by the addition of superplasticizers. However, the
mass percentage of C-H is very low for a quantitative comparison.
• The NMR results that show that all the superplasticizers studied here decrease
the rate of silicate polymerization and lead to a lower amount of polymerized
silicates at 3 and 28-days which is indicative of the change in the morphology
of C-S-H gel in the presence of superplasticizers.
• As explained in the previous section, the addition of superplasticizer always
leads to a slightly higher strength in concrete. However, the physico-chemical
analyses indicate that polymerization of silicates is retarded and the amount of
C-H formed is lower at early ages. Hence, it can be concluded from the
microstructural studies that more uniform and dense structure of
superplasticized paste results in slightly higher compressive strength of
concrete.
8.2.7 Influence of Chemical Composition of Cement on Compatibility
• Based on this study it is recommended that specification for the alkali content
be included in IS code for the chemical composition of cement as the amount
of superplasticizer required varies with alkali content; low alkali cement
requires more amount of superplasticizer due to more initial adsorption.
• The absorbance study conducted using a UV spectrophotometer shows that it is a
good characterization technique for identifying the superplasticizers. This study
also suggests that the reason for the increased demand of certain superplasticizers
is that there is higher initial adsorption of these superplasticizers.
210
8.3 Guidelines for Selecting the Best Combination of Cement and
Superplasticizer
Based on the present study, a methodology is recommended for the selection of the
superplasticizer for concrete mix design. The methodology uses criteria based on the
Marsh cone and Vicat penetration tests to select a superplasticizer that is compatible
with the cement used. A compatible combination should have a well defined
saturation point and the dosage corresponding to the same should be within the
manufacturer’s recommended limit. Also, the final setting time of superplasticized
paste should be less than 16 hours, which is expected to lead to a concrete with a
setting time less than 12 hours. However, final trials have to be conducted on concrete
to select a cost-effective superplasticizer. The combinations studied here are used to
illustrate the evaluation for compatibility based on this methodology, and the results
are compared with other approaches in the literature.
8.4 Recommendations for Further Research
The study clarifies many of the conflicting views on the flow behaviour of paste,
effectiveness of the Marsh cone test in representing the flow behaviour of the
superplasticized paste, the correlation between the flow behaviour of paste and
concrete and the influence of superplasticizers on the hardened properties of concrete.
Continued research on the following aspects can give better understanding and be of
more use to the construction industry:
• The influence of the fines content in the aggregates on the saturation dosage as
well as the loss of fluidity with time.
211
• Regarding the microstructure and hydration processes, the hydration processes
could be studied further by incorporating the quantitative measurement of the
different phases involved in the hydration through XRD.
• The determination of the Ca/Si ratio of the C-S-H and image analysis of more
samples can give a clearer picture of the effect of superplasticizers on the
composition of C-S-H and pores in the microstructure. An extensive study on
the pore structure of the superplasticized paste is essential for understanding
the effect of superplasticizers on the shrinkage of concrete.
• The prediction of the Marsh cone time with the available rheological models is
suitable only for a certain range of fluidity. In the cases of pastes with low
flow, the models seem to be inappropriate. In addition to that, the viscoelastic
characterization shows the change in nature of the paste from non-linear
viscoelastic to viscous with variation in the type and dosage of
superplasticizer. Hence, an appropriate constitutive model is required for the
whole range of flow.
• The interaction of the superplasticizer with mineral admixtures and VMA
requires further study for the optimization of paste as well for understanding
the loss of fluidity. In addition, the influence of other admixtures like
shrinkage reducing and accelerators on the cement-superplasticizer interaction
need further attention.
• Absorbance based studies of superplasticizer adsorption can be extended for
understanding the influence of mineral admixtures and VMAs on the amount
and rate of adsorption of superplasticizers on cement surface which in turn
212
affects the loss of fluidity. The zeta potential of different pastes could be
measured for validating the adsorption study.
Finally, the number of tests on fresh cement paste are quite satisfactory for definite
conclusions to be made whereas in the studies performed on hardened cement paste,
the number of tests has been quite limited, which makes unambiguous conclusions
difficult. Therefore, more extensive tests can be performed on the properties of
hardened concrete for confirming the trends seen in paste.
213
APPENDIX A
Table A.1 Test results of three trials in the selection of mixing method
Optimum superplasticizer
Type of mixing dosage (sp/c%)
trial 1 trial 2 trial 3
Hobart with B-flat beater 0.10 0.13 0.09
Hobart with D-wire whip 0.19 0.16 0.12
Blender 0.13 0.14 0.16
Grinder 0.15 0.16 0.07
Hand 0.39 0.20 0.06
Ball mill mixing (steel balls) 0.07 0.16 0.11
Ball mill mixing (Glass marbles) 0.15 0.16 0.14
Table A.2 Comparison of superplasticizer dosages

(data from Marsh cone and viscometer tests)
Super- Dosage Dosage Optimum Dosage Dosage from

plasticizer for for dosage corresponding Aïtcin
constant constant corresponding to 140º (at 60 method using
yield plastic to 140º (at 0 minutes) flow-time
stress viscosity minutes) curves at 0
and 60
minutes
LS-C1 0.4 0.4 0.25 0.26 0.3
SNF-S1 0.3 0.3 0.20 0.34 0.3
SNF- D1 0.4 0.5 0.24 0.23 0.3
SNF-D2 0.4 0.4 0.23 0.46 0.4
SNF- C1 0.3 0.3 0.16 0.20 0.3
SMF-S1 0.3 0.3 0.23 0.30 0.5
SMF-C1 0.5 0.5 0.20 0.30 0.5
PCE-S1 0.3 0.3 0.20 0.24 0.4
PCE-D1 0.1 0.1 0.07 0.17 0.2
14
LS-C1 LS-C1 120
1.3 log (flow time, sec)
1.3 Marsh cone flow time 12
plastic viscosity (cP)

plastic viscosity

log (flow time, sec) yield stress 100
10
yield stress (Pa)

1.2 80
1.2 8
6 60
1.1
1.1 4
40
2
1.0 20
1.0 0 0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
sp/c%
(a) Correlation between Marsh cone flow time and rheological parameters for LS-C1
1.5
SNF-D2 1.5
SNF-D2 200
Marsh cone flow time 16
1.4
Yield stress Marsh cone flow time

1.4 160
yield stress (Pa)
log (flow time, sec) Plastic viscosity

12
1.3 120
1.3
8
1.2 80
4 1.2
40
1.1
0 1.1
0.0 0.2 0.4 0.6 0.8 1.0 0
sp/c % 0.0 0.2 0.4 0.6 0.8 1.0
sp/c %
(b) Correlation between Marsh cone flow time and rheological parameters for SNF-D2
35
SMF-C1 SMF-C1 200
1.20 30 1.20

25 Marsh cone flow time 160

yield stress (Pa)
yield stress
1.16 1.16 Plastic viscosity
20
120
15 1.12
1.12 80
10
5 1.08
40
1.08
0
1.04 0
0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0
sp/c % sp/c %
(c) Correlation between Marsh cone flow time and rheological parameters for SMF-C1
215
PCE-S1 PCE-S1 160
1.4 16 1.4
Marsh cone flow time 140


yield stress (Pa)
Yield stress Plastic viscosity
1.3 12 120
1.3
100
8
1.2 1.2 80
4 60
1.1 1.1
0 40
1.0 1.0 20
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
sp/c % sp/c%
(d) Correlation between Marsh cone flow time and rheological parameters for PCE-S1
Fig. A.1 Correlation between Marsh cone flow time and

rheological parameters of superplasticizers
35 0.25
30
0.20
25
Yield stress (Pa)
20 0.15
15
0.10
10
5 0.05
0 0.00
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
Flow time (sec) Flow time (sec)
(a) Correlation between Mash cone flow time, yield stress and
plastic viscosity for different combinations
35 0.25
30
0.20
25
Yield stress (Pa)
20 0.15
15
0.10
10
5 0.05
0
0.00
80 100 120 140 160 180 200 80 100 120 140 160 180 200
mini-slump spread (mm) mini-slump spread (mm)
(b) Correlation between Mini-slump spread, yield stress and

plastic viscosity for different combinations
Fig. A.2 Correlation between Marsh cone flow time and mini-slump spread
with rheological parameters for different compositions
216
APPENDIX B
Matlab programme to calculate flow time as per Nguyen model
%to calculate the velocity at the exit of the nozzle

%to calculate the time for falling the level from h1 to h2
clc
n=input('enter value of n = ');
R=0.004;
rho=input('enter value of rho = ');
g=9.81;
H=0.06;
tanphi=0.267;
lamda=0.96;
tuo0=input('enter value of tuo0 = ');
k=input('enter value of k = ');
alpha1= (1-(1.1*(tuo0/(rho*g*H*tanphi))));
alpha2=(1+(R/(3*n*H*tanphi)));
alpha=alpha1/alpha2;
syms x real
deltap=((alpha*rho*g)*(1+x/H)-rho*g)*H;
r0=2*tuo0/(rho*g+(deltap/H));
f1=1+((2*n)/(2*n+1))*(r0/R);
f2=((2*n^2)/((n+1)*(2*n+1)))*(r0/R)^2;
f3=(1-(r0/R))^((n+1)/n);
f=(f1+f2)*f3;
v1=n*R/(3*n+1);
v2=(alpha*rho*g*R)/(2*k);
v3=(H+x)/(H);
v=(v1*(v2*v3)^(1/n))*f;
Re=(rho*((v)^(2-n))*(2*R)^n)/k;
B1=(1/(2*lamda))*sqrt((alpha*R)/H);
B2=(n/(2*(3*n+1)))^(n/2);
B3=(1-(exp(-0.1*sqrt(Re))));
B4=(sqrt(Re)+(1.5/n)*(Re^(1/4)));
B=(B1*B2*B3*B4)+1;
vb= v/B;
t1=(R+(x*tanphi))^2;
217
t2=(vb)*R^2;
t3=t1/t2;
ii=linspace(0.224,0.165,100);
stepsize=ii(2)-ii(1);
for jj=1:length(ii)
fval(jj)=subs(t3,ii(jj));
end
cumsum=0;
for jj=2:length(ii)
cumsum=cumsum+(fval(jj)+fval(jj-1));
end
time=-((stepsize/2)*cumsum)
218
APPENDIX C
INFLUENCE OF SUPERPLASTICIZER ON THE NON-

NEWTONIAN CHARACTERISTICS OF CEMENT PASTE
C.1 GENERAL
The superplasticized cement paste usually exhibits non-Newtonian characteristics, which
depend on the type of superplasticizer and its dosage. The superplasticizer causes an
increase in the fluidity of the paste and shear thinning behaviour as explained in Chapter
3. However, in some instances of incompatibility, excessive amount of superplasticizer is
added, which can result in possible shear thickening. This shear thickening may be due to
high concentration of solids in nonflocculated suspensions as the hydration process is
accelerated during shearing (Struble and Sun 1995; Cyr et al. 2000). Also, an
agglomeration of excessive amount of superplasticizer available in the solution causes
more elasticity and possible shear thickening (Papo and Piani 2004).
Typically rotational viscometers have been used for characterizing the rheological
behaviour of cement paste as explained earlier (Roy and Asaga 1979; Masood and
Agarwal 1994; Ferraris et al. 2001; Park et al. 2005; Schwartzentruber et al. 2006). Most
of these investigations subject the material to a cycle consisting of shear rate increasing
and decreasing in a stepwise manner. Normally the cycles are repeated and the shear
stress-shear rate relationship in the down ramp of either the first, second or third cycle is
considered for further analysis (Papo and Piani 2004; Roussel and Roy 2005). Current
attempts in the cement paste rheology literature have mainly focused on reducing the data
from steady shear experiments to standard constitutive models, and exploring the
influence of superplasticizer dosage and type based on these model parameters as seen in
219
Chapter 3. A variety of models such as Bingham, Herschel- Bulkley, Power law models
etc., are used for characterizing the response of the material. The main intent of most of
these investigations have been to study the influence of these admixtures on some of
these model based parameters such as plastic viscosity, yield stress, consistency, flow
index etc. For instance, the influence of the dosage of superplasticizer on the Bingham
model parameters like yield stress and plastic viscosity shows that both of these
parameters decrease with increase in dosage of superplasticizer up to saturation dosage
and remains constant thereafter (Asaga and Roy 1980; Banfill 1981; Nehdi and Rahman
2004). Similar is the case with Herschel-Bulkley parameters like yield stress, consistency
and consistency index (Atzeni et al. 1985; Nehdi and Rahman 2004). There are also
contradictory observations reported in the literature for the same mixture of materials. For
instance, while the Bingham model reports negative yield stress for high dosage of
superplasticizers, Herschel-Bulkley model reports a non–negative yield stress and the
reason for this is normally ascribed to the so called non-linearity between shear stress and
shear rate of Herschel-Bulkley model (de Larrard et al. 1998).
As discussed earlier, the cement paste during the pre-induction period can exhibit shear
thinning when sheared in a coaxial viscometer. The rate at which the shear thinning
characteristic of the material changes depends on the time at which the material is
sheared. After pre-induction period, the same material can exhibit shear thickening
behaviour. The scenario is more complex for the superplasticizer-cement paste mixture.
Issues such as compatibility/incompatibility and low and high dosage change the
rheological behaviour considerably (Mailvaganam 1999). A compatible system at a low
dosage of superplasticizer normally will keep the cement particles in a state of dispersion
even up to 60 minutes (Agullo et al. 1999). This can be observed by the more or less
220
Newtonian behaviour in steady shear experiment, in a rotational viscometer. However if
the same material is sheared in a dynamic shear rheometer it is possible for it to exhibit
high normal stress differences thus conclusively proving the non–Newtonian response
characteristics (Mansoutre et al.1999; Min et al. 1994). Also, the same material can
exhibit creep and recovery and stress relaxation behaviour thus showing a viscoelastic
material response (Struble and Schultz 1993; Sun et al., 2006).
Significant insight into the nature of cement paste has been gained recently due to the
investigations of Van Damme et al. (2002). Essentially one can visualize the paste as a
dense assembly of deformable or rigid objects close to a jamming situation. These dense
assemblies are lubricated by the hydrodynamic or physical-chemical interactions of the
interstitial fluid, a mixture of water and superplasticizer. One important aspect of pastes
when viewed as an assemblage of granular particles is the property of dilatancy.
Essentially, pastes when subjected to shear strain will exhibit the tendency to expand in
the direction perpendicular to the shear plane. In a typical dynamic shear rheometer
experiment where the gap between the two plates is kept fixed, this essentially results in
the development of normal stresses. Van Damme et al. (2002) postulate that the two
important properties that give complete description of the mechanical behaviour of the
paste is the effect of dilatancy and inter-granular friction; the latter described very closely
by the ‘yield stress’. They used these concepts to study the rheological behaviour of
flocculated calcium silicate pastes and emphatically observe the need to measure normal
force in shear flow in addition to the ‘yield stress’ for a complete characterization of
superplasticized cement paste.
221
A proper constitutive model taking into account all the internal structural changes taking
place during the hydration process is lacking currently. This is not surprising as the
cement paste-superplasticizer mixture exhibits diverse and complex response
characteristics. However, even before one attempts in building a constitutive model, it is
imperative to understand through careful experimental investigations the transition in the
viscometric properties over a period of time. This information is necessary in building a
proper constitutive model. The intent of this investigation is precisely that.
It has been observed that the Marsh cone is extensively used in the optimization of high
performance concrete mixture design (see Chapter 3). Since the idea of using
superplasticizer is to impart linearly viscous behavior for a short period of time to the
paste and the mixture, the flow time measured in a Marsh cone has been linked to the
superplasticizer dosage beyond which there is no appreciable change in flow time is
normally taken as the optimum dosage. There are several interesting issues related to this
aspect. First, for a given cement paste – superplasticizer mixture, the optimum dosage
essentially relates the transition from a non-Newtonian to a Newtonian behavior. If one
assumes the Bingham model for the cement paste – superplasticizer mixture for dosages
less than the optimum dosages, the equations to be solved are the Buckingham-Reiner
equation and for the dosage equal to and greater than the optimum dosages, the classical
Hagen-Poiseuille equation are required to be solved. Also, discussions in the earlier
paragraphs clearly indicated that it is possible for the cement paste – superplasticizer
mixture to exhibit shear thickening, a typical non-Newtonian behavior for over dosage.
Hence, it is expected that the flow time might start increasing beyond the optimum
dosage as seen in the Chapter 3. It is also to be noted that this optimum dosage may not
be a sharp transition point, it is normally asymptotic in nature and hence it will be natural
222
that there can be several dosages near the optimum dosage which will result in the
transition from non-Newtonian to Newtonian. It will also be helpful if this transition
point is arrived at by appealing to kinematics of fluid flow rather than inferring it through
model dependent parameters.
In this investigation, several combinations of cement paste and superplasticizer are tested
in a Brookfield viscometer (details are given in Chapter 3) and Anton Paar dynamic shear
rheometer. The raw experimental data in terms of torque - shear stress – shear rate are
examined and detailed analysis is made on the various interesting manifestations of the
rheology of cement paste. The main thrust here is that the non-Newtonian behavior of
cement paste and superplasticizer mixture is characterized based on more fundamental
experimental investigations without recourse to any specific model. Steady shear
experiments, creep and recovery and stress relaxation experiments were conducted on a
variety of superplasticizer and cement paste combinations and the transition from non-
linear viscoelastic to non-Newtonian to linearly viscous fluid as the dosage rate is
increased is captured. Additionally the saturation dosages obtained from Marsh cone tests
were compared and it was found that dosage rates determined using Marsh cone were
closer to the dosage rates where the transition to the linearly viscous fluid was seen.
C.2 Materials used
The physical and chemical characteristics of cement and four types of superplasticizers
used for the study are same as given in Section 3.2
223
Table C.1–Mix details for steady shear experiments
Type of Dosage of Saturation dosage of

superplasticizer superplasticizer superplasticizer from
(sp/c %) Marsh cone test
(sp/c %)
SNF-S1 0.1, 0.2, 0.3, 0.4, 1 0.20
SNF-D2 0.1, 0.2, 0.4, 0.6, 1 0.23
SMF-S1 0.1, 0.2, 0.3, 1 0.23
PCE-D1 0.05, 0.1, 0.2, 0.4, 1 0.07
C.3 Experimental Procedure
The mixing procedure for the preparation of the paste is same as given in Section 3.5.The
Marsh cone test was used to evaluate the relative fluidity and the saturation
superplasticizer dosage. Table 1 gives The saturation superplasticizer dosages for the four
products used in this study are given in Table C.1. The details of the steady shear
experiments using Brookfield HA DV II +pro were explained in detail in Section 3.5. The
rheological experiments were also conducted with an Anton Paar dynamic shear
rheometer for studying the creep and recovery, and stress relaxation responses. The tests
were conducted with a parallel plate attachment with a gap of 1 mm (see Figure C.1). The
details of these tests are given in the respective sections that follow.
Fig. C.1(a) Anton Paar Rheometer (b) Shearing of cement paste with parallel plate
attachment
224
C.4 Results and Discussions
C.4.1 Marsh cone test on pastes
The saturation dosages obtained from Marsh cone test results for four superplasticized
pastes, already presented in Chapter 3, are summarized in Table C.1.
C.4.2 Steady shear tests with viscometer
Superplasticized cement paste generally shows shear thinning behaviour with an increase
in shear rate. This is due to the breakdown of the hydrated products formed around the
cement grain due to continuous mixing leading to the reduction in the apparent viscosity.
During unloading apparent viscosity remains constant during the down curve of the
hysteresis cycles for the PCE based product as shown in Figure C.2. It is observed that
the area within the loop decreases from the first cycle to the second cycle, because the
energy required for breaking down the hydration products decreases with time.
7 PCE-D1 0.1%
6
Shear stress (Pa)
3
First cycle
2 Second cycle
Area of first cycle = 86.49 J
1 Area of second cycle = 48.36 J
0
20 40 60 80 100 120 140 160 180
Shear rate (1/sec)
Fig. C.2- Loading-unloading cycles for the superplasticizer PCE-D1

at the dosage of 0.1%
Viscometric test results on paste shows that the behaviour changes from shear thinning
nature to Newtonian and shear thickening depending on the type and dosage of
superplasticizer (Lootens et al. 2004; Papo and Piani 2004; Chen et al. 2006). The
225
reported results on shear thickening are mainly based on oscillatory tests in terms of
storage modulus and viscous modulus (Chen et al., 2006; Chougnet et al. 2007). The
model parameters derived from viscometric tests and hysteresis cycles are used to capture
the change from shear thinning to shear thickening behaviour in Chapter 3.
The shear thinning and thickening nature of materials can also be classified based on the
area within the loading-unloading cycle as it reflects the energy required for breaking
down of the structure of the paste. Figures C.2 and C.3 show the work done in
deflocculating the particles for shear thinning and shear thickening compositions,
respectively. The area in the loop decreases from the first cycle to the second cycle for
shear thinning material because the energy required for breaking down of the particles
decreases. In the case of shear thickening material, the area within the loop increases.
30
PCE-D1 1%
25
Shear stress (Pa)
20
First cycle
15 Second cycle
Area of first cycle = 166.59 J

10 Area of second cycle = 188.21 J
5
20 40 60 80 100 120 140 160 180
Shear rate (1/sec)
Fig. C.3- Loading-unloading cycles for the superplasticizer PCE-D1 at the dosage of 1%
Table C.2 shows the areas within the loading-unloading loops for the different pastes
studied, along with the saturation dosages for different compositions and it can be seen
that more work is required at low dosages of superplasticizer to make the paste to flow,
and that the area of the loops decreases with an increase in the dosage of superplasticizer.
226
It can be observed that the area calculation is reliable upto the saturation dosage of
superplasticizer and the maximum percentage reduction occurs approximately at the
saturation dosage of superplasticizer. The paste with PCE-D1 shows shear thickening at
1% and hence work required for structural breakdown increases with time of shearing.
Thus, the reduction in the area of hysteresis cycles indicates the effectiveness of different
superplasticizers as it represents the amount of work required to deflocculate the
particles. From the perspective of rheology of cement paste-superplasticizer mixture, this
reduction in area is related to the transition from non-Newtonian to Newtonian regime.
However, the trend is not clear in other superplasticizers.
Table C.2 Work done in different loading-unloading cycles
Type of sp/c Saturation Loop area of Loop area of

SP % dosage first cycle second cycle
(sp/c%) (N/m2s) (N/m2s)
0.1 234 230
SNF-S1 0.2 108 92
0.3 0.20 28 3
0.4 85 65
1 44 7
0.1 667 356
SNF-D2 0.2 115 55
0.4 0.23 143 135
0.6 140 135
1 124 112
0.1 672 299
SMF-S1 0.2 0.23 81 41
0.3 118 58
1 130 85
0.05 1813 820
0.1 86 48
PCE-D1 0.2 0.07 91 22
0.4 37 24
1 167 188
227
C.4.2.1 Shear thinning index
Superplasticized paste exhibits shear thinning nature as the structural breakdown is faster
than structural build up. The degree of shear thinning of non Newtonian fluids for
different combinations can be represented by the shear thinning index (ASTM D 2196
(2005)); the shear thinning index is defined as the apparent viscosity at low speed to that
at high speed. Accordingly, in this case the shear thinning index is calculated based on
the apparent viscosity at 25 rpm and 175 rpm.
Table C.3 shows the influence of the type and dosage of superplasticizer on the shear
thinning index. Higher the shear thinning index, higher is the degree of shear thinning. It
is observed that generally the shear thinning index decreases with an increase in the
dosage of superplasticizer, which shows that the effect is predominant at low dosages.
PCE D1 shows considerable shear thinning even at low dosages and at overdosage,
causes some shear thickening (i.e., shear thinning index is less than 1). For all
compositions, shear thinning occurs upto the saturation dosage and after that the index
remains almost constant. The shear thinning index of less than one in all cases shows that
the apparent viscosity at high shear rate is slightly higher than that at low shear rate. This
generally occurs at overdosage of superplasticizer because of the excess amount of
superplasticizer available in the solution which forms a network. Hence, it is proposed
that the shear thinning index of one can be used as a limit for the dosage of the
superplasticizer. Interestingly, the dosage corresponding to this limit is practically
comparable with the saturation dosage obtained with Marsh cone and coaxial cylinder
viscometer (except SNF-D2). In the case of PCE based superplasticizer, a small increase
in dosage of 0.05% causes a large reduction in the shear thinning index, showing its
effectiveness at low dosages.
228
Table C.3 Shear thinning indices for different types and dosages of superplasticizers
Type of Dosage of SP Shear thinning

superplasticizer index
SNF-S1 0.1 4
0.2 3
0.3 2
0.4 1
1 1
0.1 3
0.2 3
SNF-D2 0.4 1
0.6 1
1 1
0.1 3
0.2 3
SMF-S1 0.3 2
1 1
0.05 3
0.1 1
PCE-D1 0.2 1
0.4 1
1 1
C.4.3 Creep and Recovery and Stress Relaxation Experiments
The behaviour of cement pastes under steady shear has been determined using different
procedures (Lapasin et al. 1979; Lapasin et al. 1983; Atzeni et al. 1995; Geiker 2002).
These studies explain that the material behaviour changes from solid-like to fluid-like
with the variation of shear stress with time. They consider the yield stress as a criterion
for the change of material behaviour. The concept of ‘yield stress’ is discussed in detail in
the excellent review article by Barnes (1999) which also discusses various issues related
to whether ‘yield stress’ is a model derived parameter or a true phenomenon. In the
present study, tests were conducted with an Anton Paar rheometer using five different
compositions - pure cement paste and cement paste with four types of superplasticizers to
229
understand the influence of different superplasticizers on the flow behaviour (see Table
C.4). The details of the testing methods are given in Table C.5.
Table C.4 Mix details for creep and recovery and stress relaxation experiments
Composition of paste w/c Dosage of superplasticizer

(sp/c%)
Pure cement paste 0.35 Nil
Cement paste with SNF-S1 0.35 0.05, 0.20
Cement paste with SNF-D2 0.35 0.05, 0.23
Cement paste with SMF-S1 0.35 0.05, 0.23
Cement paste with PCE-D1 0.35 0.01, 0.025, 0.05, 0.07, 0.1
Table C.5 Details of the different data acquisition frequencies
Details of test Number of Measuring point Testing details

data points duration
1. Creep and Recovery Measure Strain (Є) vs t
Creep (10 sec) 200 0.05 sec
Recovery (100 sec) 40 0.05 to 5 sec varied
linearly during 100 sec σ
10 sec t
100 sec
2. Stress relaxation Measure σ vs t

Ramping of strain 100 5×10-3 sec
(0.5 sec)
Stress relaxation 10 5×10-3 sec to 20 sec Є
(33 sec) varied logarithmically
during 33 sec
0.5 sec t
C.4.3.1 Creep and Recovery Test on Cement Paste
Most flow studies of paste and models used consider the material as viscous only. In this
section an attempt is made to observe whether the material shows any viscoelastic nature
and whether the addition of superplasticizer results in a transition to non-Newtonian
behaviour.
230
Struble and Leit (1995), and Struble and Schultz (1993) demonstrated that creep and
recovery behaviour of cement paste undergoes a transition from solid-like behaviour at
low stress to fluid like behaviour at high stress. However, at high stress, the material
behaves as a viscous liquid with a linear increase in strain throughout the duration of
stress with no recovery. This stress level associated such a transition agrees with the yield
stress estimated from flow curves and oscillatory tests in their study. Creep and recovery
tests have also been used to monitor the setting of cement paste based on the development
of yield stress with time (Struble and Leit 1995).
Generally, in a stress controlled test, when the material is subjected to instantaneous
stress followed by a constant stress, an instantaneous increase in strain occurs initially
and then there is a continuous increase in strain at a non-constant rate, i.e., an elastic
response followed by combination of elastic and viscous response. If the material is fluid-
like, the strain rate asymptotically approaches a constant value. While releasing the stress,
some instantaneous strain recovery followed by delayed recovery occurs (Wineman and
Rajagopal, 2000).
In the present study, creep and recovery tests were done on pure cement paste and
superplasticized cement pastes. Tests were conducted at different stress levels in order to
understand how the nature of the material changes at different stress levels for different
types of superplasticizers and dosage rates. Figure C.4 shows the test results for pure
cement paste, which shows a typical viscoelastic nature with some residual strain at
different stress levels. The stress levels used include 50 Pa, 75 Pa and 100 Pa and it can
be observed that the strain produced at higher stress levels is higher. Also, the recovery is
less for lower stress.
231
0.007
Cement paste with w/c 0.35
0.006
50 Pa
75 Pa
0.005 100 Pa
0.004
Strain 0.003
0.002
0.001
0.000
20 40 60 80 100 120
Time (sec)
Fig. C.4 Creep and recovery test for pure cement paste
Influence of type of superplasticizer on creep and recovery
Figure C.5 (a-d) shows the influence of types of superplasticizers on the rheology of
cement paste. Tests were conducted for four superplasticizers (see Table C.4) at constant
dosage of 0.05%. the pastes with the SNF and SMF based superplasticizers show the
same trend as the pure cement paste; however, strain recovery is less compared to pure
cement paste. Moreover, the cement paste with PCE based superplasticizer shows an
entirely different nature. It shows linear strain initially and no instantaneous strain or
recovery. This is because the paste becomes viscous at that dosage due to the better
dispersion produced by the combined action of electrostatic repulsion and steric
hindrance.
232
0.8 SNF-D2 - 0.05%
0.35 SNF-S1 - 0.05%
50 Pa
0.30 75 Pa 0.7
100 Pa
0.6 50 Pa
0.25
0.5 75 Pa
Strain
Strain
0.20 0.4
100 Pa
0.15 0.3
0.10 0.2
0.05 0.1
0.00 0.0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time (sec) Time (sec)
(a) Creep and recovery of SNF-S1 (b) Creep and recovery of SNF-D2
600
SMF-S1 - 0.05% 50 Pa PCE-D1 - 0.05%
0.4 75 Pa 500
100 Pa 50 Pa
0.3 400 75 Pa
100 Pa
Strain
Strain
300
0.2
200
0.1
100
0.0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
(c) Creep and recovery of SMF-S1 (d) Creep and recovery of PCE-D1
Fig. C.5 (a-d) Creep and recovery test result of cement paste and with
constant dosage of different superplasticizers
Influence of dosage of superplasticizer on creep and recovery
In order to understand the influence of dosage of superplasticizer on creep and recovery,
tests were conducted with the most effective superplasticizer PCE-D1 (determined from
Marsh cone test) at different dosages. The test results are given in Figure C.6 (a-c). It is
observed that the nature of the paste changes from viscoelastic to viscous as the dosage of
superplasticizer increases from low to saturation. Also, the strain/deformation of the
material increases with an increase in dosage of superplasticizer.
233
0.08 500
PCE-D1 - 0.025%
PCE-D1 - 0.01% 50 Pa
0.07 75 Pa
100 Pa
400
0.06
50 Pa
0.05 300
Strain
75 Pa
Strain
0.04 100 Pa
0.03 200
0.02
100
0.01
0.00 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
(a) Creep and recovery of PCE-D1, 0.01% (b) Creep and recovery of PCE-D1, 0.025%
600 600
PCE-D1 - 0.05% PCE-D1 - 0.1%
500 500
400 50 Pa 400 50 Pa
75 Pa 75 Pa
Strain
Strain
300 100 Pa 300 100 Pa
200 200
100 100
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
(c)Creep and recovery of PCE-D1, 0.05% (d) Creep and recovery of PCE-D1, 0.1%
Fig. C.6 (a-d) Variation in creep and recovery behaviour

with different dosages of PCE-D1
Influence of saturation dosage of superplasticizer on creep and recovery
A comparison of the flow behaviour of pastes was obtained by conducting creep tests at
the saturation dosage of superplasticizers (see Figure C.7 (a-d)). It is observed that all
pastes show a viscous nature with a linear increase in strain initially and no recovery
while releasing the stress at their saturation dosages. Thus a transition in the nature of
material occurs at saturation dosages, which is necessary to have a flowable concrete at
the saturation dosages.
234
SNF-S1 - 0.20% SNF-D2 - 0.23%
600 1200
500 1000
400 800
Strain
Strain
300 600
200 400
50 Pa 50 Pa
100 75 Pa 200 75 Pa
100 Pa 100 Pa
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
(a) SNF-S1, at saturation dosage, (b) SNF-D2, at saturation dosage ,
SMF-S1- 0.23% 1800 PCE-D1 0.07%

500 50 Pa
75 Pa
1500
400 100 Pa
1200
Strain
Strain
300
900
200 600
50 Pa
100 75 Pa
300 100 Pa
0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
(c) SMF-S1, at saturation dosage , (b) PCE-D1, at saturation dosage
Fig. C.7 (a-d) Creep test result at saturation dosage of superplasticizers
Check for linearity of response
The simplest modeling concept for viscoelastic materials for which the experimental data
can be characterized is called linearity of response (Wineman and Rajagopal, 2000). It
consists of two conditions (1) linear scaling and (2) superposition of separate responses.
If the response of a viscoelastic material displays both linear scaling and possibility
superposition of separate responses for all histories, the response is said to be linear. If
the stress (σ0) is changed by a factor α, and if the deformation is also changed
235
consequently by the same factor α, then the material shows linear scaling. In our current
investigation, the stress levels are 50 Pa, 75Pa and 100Pa and for checking linear
response, 50 Pa is multiplied by a factor of 1.5 and 2. If the material satisfies linear
scaling, the deformation response obtained by increasing the stress by these factors
should be same as that of the deformation produced by 75 Pa and 100 Pa respectively.
However, Figure C 8 (a and b) shows that the pure cement paste as well as that with
superplasticizer at a low dosage, violate the condition of linear scaling in the tested
compositions. The same behavior was found for all the combinations tested.
0.007
Cement paste with w/c 0.35 SMF-S1 - 0.05% Actual
0.4
0.006 Observed
Actual σ= 100 Pa Expected
0.005 Observed 0.3
Expected
Strain
0.004
Strain
0.003 0.2
σ=100Pa
0.002
0.1 σ=2*50 Pa
σ=2*50 Pa
0.001 σ=50 Pa
σ=50 Pa
0.000 0.0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Fig. C.8 Test for linear scaling
If the material follows the principle of superposition, then the instantaneous deformation
due to the applied load at time t = 0 and the instantaneous deformation recovered on the
release of the load should be same. The instantaneous deformation due to loading and
instantaneous deformation due to unloading as observed from above Figure C.8 and
Table C.6 at a dosage of 0.05%, are different. This shows that the cement paste tested in
this investigation also violates the condition of superposition of separate responses. These
two checks show that pure cement paste and superplasticized pastes have a non-linear
236
viscoelastic nature at dosages less than saturation. However, the recovery could not be
observed for superplasticizers at saturation dosages.
Table C.6 Superposition of separate responses
Composition Stress (Pa) Instantaneous Recovered

Strain on strain on
loading unloading
50 0.00037 0.000171
Cement paste 75 0.00038 0.00017
100 0.0026 0.00114
Cement paste 50 0.0107 0.0015
with SNF-S1 75 0.0119 0.003
100 0.0147 0.0045
Cement paste 50 0.0098 0.002
with SNF-D2 75 0.0141 0.003
100 0.0214 0.0035
Cement paste 50 0.0072 0.0014
with SMF-S1 75 0.0141 0.003
100 0.0171 0.0045
Cement paste 50 0.0127 0
with PCE-D1 75 0.0194 0
100 0.0254 0
C.4.3.2 Stress Relaxation
In the case of a viscoelastic material, when strain is instantaneously increased at t = 0 and
then maintained, the stress instantaneously increases and then gradually decreases with
time. Tests are conducted on different compositions as described in creep and recovery
tests. In the case of pure cement paste, the material shows viscoelastic nature as the stress
reduces asymptotically to zero (see Figure C.9).
237
1000 Cement paste with w/c 0.35
Shear stress (Pa)

Strain 0.25%
800 Strain 0.5%
Strain 0.75%
600
400
200
0
0 10 20 30
Time (sec)
Fig. C.9 Stress relaxation of pure cement paste
Influence of type of superplasticizer on stress relaxation
Figure C.10 (a-d) shows stress relaxation test results for different types of
superplasticizers at a constant dosage of 0.05%.The SNF and SMF based
superplasticizers show viscoelastic nature as in the case of pure cement paste. In the case
of PCE based paste, no external force is required to maintain the strain because the
dosage is high enough to ensure fluid-like behaviour. This is because of the better
dispersion produced due to PCE based superplasticizer.
35
SNF-S1- 0.05% SNF-D2 - 0.05%
30 0.25% 120 0.25%
Shear stress (Pa)
0.5% 0.5%
100
Shear stress (Pa)
25 0.75% 0.75%
20 80
15 60
10 40
5 20
0 0
0 5 10 15 20 25 30 35
0 5 10 15 20 25 30 35
(a) Stress relaxation of SNF-S1 (b) Stress relaxation of SNF-D2
238
10
50 SMF-S1 - 0.05% PCE-D1 - 0.05%
0.25%
Shear Stress (Pa) 0.25%
Shear stress (Pa)

0.5%
5 0.5%
40 0.75%
0.75%
30
0
20
-5
10
0 -10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(c) Stress relaxation of SMF-S1 (b) Stress relaxation of PCE-D1
Fig. C.10 (a-d) Stress relaxation test results for pastes with 0.05%
dosage for different superplasticizers
Influence of dosage of superplasticizer on stress relaxation
Tests were performed with different dosages of PCE-D1. It is seen that the shear stress
decrease with increase in dosage of PCE based superplasticizer, as in Figure C.11 (a-e).
Even though the paste shows stress relaxation at low dosages of 0.01% and 0.025%, the
behaviour is viscous at higher dosages.
120 PCE-D1 - 0.025%

PCE-D1 - 0.01% 50
Shear stress (Pa)
0.25%
Shear Stress (Pa)
100 0.25%
0.5% 0.5%
0.75% 40 0.75%
80
30
60
40 20
20 10
0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(a) Stress relaxation of PCE-D1, 0.01% (b) Stress relaxation of PCE-D1, 0.025%
239
10 10
PCE-D1 - 0.05% PCE-D1 - 0.07%
Shear stress (Pa) 0.25% 0.25%
Shear stress (Pa)

5 0.5% 5 0.5%
0.75% 0.75%
0 0
-5 -5
-10 -10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(c) Stress relaxation of PCE-D1, 0.025% (d) Stress relaxation of PCE-D1, 0.07%
10
PCE-D1 - 0.1%
Shear stress (Pa)
0.25%
5 0.5%
0.75%
-5
-10
0 5 10 15 20 25 30 35
Time (sec)
(e) Stress relaxation of PCE-D1, 0.1%
Fig. C.11 (a-e) Stress relaxation at different dosages of PCE-D1
Influence of saturation dosage of superplasticizer on stress relaxation
It is clear that the dosage of the superplasticizer affects the relaxation. From the tests on
pastes with PCE-D1, it appears that at dosages much lower than the saturation dosages,
there is appreciable relaxation with time whereas the behaviour is viscous at dosages
higher than saturation dosages. At the saturation dosage, the paste with PCE-D1 has a
practically viscous response. The stress relaxation shown by SNF-S1 and SMF-S1 at
saturation dosage is appreciable but small whereas there is zero stress relaxation for the
240
paste with SNF-D2 as in the case of PCE-D1 as shown in Figure C.12 (a-d). This results
further confirms the viscous nature of superplasticized pastes at and beyond saturation
dosages.
25 10
SNF-S1 - 0.20% SNF-D2 - 0.23%
20 0.25%
Shear stress (Pa)

Shear stress (Pa)
0.25%
5 0.5%
15 0.5%
0.75% 0.75%
10
5 0
0
-5 -5
-10
-15 -10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(a) SNF-S1, at saturation dosage (b) SNF-D2, at saturation dosage
25 10
SMF-S1 - 0.23% PCE-D1 - 0.07%
Shear stress (Pa)
0.25% 0.25%
Shear stress (Pa)
20 0.5% 0.5%
0.75% 0.75%
15
0
10
0 -10
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(c) SMF-S1, at saturation dosage (d) PCE-D1, at saturation dosage
Fig. C.12 (a-d) Stress relaxation of pastes at saturation dosage of superplasticizers
C.5 Comparison of rheological tests
The creep and recovery and stress relaxation tests shows that a transition from non-
linear viscoelastic to viscous nature occurs for pastes as the superplasticizer dosage
changes from low to saturation and beyond. The saturation dosages obtained from
Marsh cone tests are shown to yield viscous behaviour in all cases studied.
241
C.6. SUMMARY
In this investigation, steady shear, creep and recovery, and stress relaxation experiments
were conducted on cement pastes with different superplasticizers of varying dosage rates.
This helps develop an understanding of the transition of these mixtures from viscoelastic
fluid-like materials to non-Newtonian fluids to linearly viscous fluids, as the
superplasticizer dosage rate is increased. The work also increases the understanding of
shear thinning characteristics as a function of superplasticizer type and dosage rates. The
area of hysteresis cycles and shear thinning index were used for characterizing the shear
thinning nature. As superplasticizer dosage increases, in general, there is a transition in
the behaviour from viscoelastic to linearly viscous fluid behaviour. In the creep and
recovery experiment, this is reflected as a linear change of strain with zero recovery
during unloading. In the case of stress relaxation, the classical stress spike with zero
stress after application of constant strain was seen. The dosages at which the change
occurs compares well with the saturation dosages determined using the Marsh cone.
242
REFERENCES
ACI 116 R (2000) Cement and concrete terminology. ACI Publicaton, USA.
ACI E4 (2003) Eduction Bulletin, Chemical admixtures for concrete. ACI

Publication, USA.
Agarwal, S.K., I. Masood, and S.K. Malhotra (2000) Compatibility of

superplasticizers with different cements. Construction and Building Materials, 14,
253-259.
Agullo, L., B.T. Carbonari, R. Gettu, and A. Aguado (1999) Fluidity of cement
pastes with mineral admixtures and superplasticizer - a study based on the Marsh cone
test. Materials and Structures, 32, 479-485.
Aїtcin, P.C., S.L. Sarka, M. Regourd, and D. Volant (1987) Retardation effect of
superplasticizers on different cement fractions. Cement and Concrete Research, 17,
995-999.
Aїtcin, P.C., C. Jolicoeur, and J.G. Macgregor (1994) Superplasticizers: How they
work and why they occasionally don’t. Concrete International, 16, 45-52.
Aïtcin, P.C. High performance concrete. E& FN Spon, London, 1998.
Aїtcin, P.C., S. Jiang, B.G. Kim, P.C. Nkinamubanzi, and N. Petrov (2001)
Cement/ superplasticizer interaction: The case of polysulphonates. Bulletin Des
Laboratories Ponts et Chausses, July-August, 89-99.
Akman, M.S and Z. Cavdar, Efflorescence problem in superplasticized concretes

pp. 316-329. J.G. Cabrera, and R.R. Villarreal (eds) Proceedings of International
Symposium on the Role of Admixtures in High Performance Concrete, RILEM,
Mexico, 1999.
Andersen, P.J., D.M. Roy, and J.M. Gaidis (1987) The effects of adsorption of
superplasticizers on the surface of cement. Cement and Concrete Research, 17, 805-
813.
Asaga, K. and D.M. Roy (1980) Rheological properties of cement mixes: 1V. Effects
of superplasticizers on viscosity and yield stress. Cement and Concrete Research, 10,
287-295.
Assaad, J., K.H. Khayat, and H. Mesbah (2003) Assessment of thixotropy of

flowable and self-Consolidating Concrete. ACI Materials Journal, 100, 99-107.
ASTM C 29/29M (2007) Standard test method for bulk density and voids in
aggregate. Annual book of ASTM standards, American Society of Testing and
Materials, USA.
ASTM C 109/109M (1998) Standard test method for compressive strength of

hydraulic cement mortars. Annual book of ASTM standards, American Society of
Testing and Materials, USA.
ASTM C 150 (2007) Standard specification for Portland cement. Annual book of
ASTM standards, American Society of Testing and Materials, USA.
ASTM C 187 (2004) Standard test method for normal consistency of hydraulic
cement. Annual book of ASTM standards, American Society of Testing and Materials,
USA.
ASTM C 191 (2001a) Standard test method for time of setting of hydraulic cement by
Vicat needle. Annual book of ASTM standards, American Society of Testing and
Materials, USA.
ASTM C 204 (2007) Standard test method for fineness of hydraulic cement by air
permeability apparatus. Annual book of ASTM standards, American Society of
ASTM C 305 (1989) Standard practice for mechanical mixing of hydraulic cement
pastes and mortars of plastic consistency. American Society for Testing and Materials,
USA.
ASTM C 494/C 494M (2008a) Standard specification for chemical admixtures for
concrete. Annual book of ASTM standards, American Society for Testing and
Materials, USA.
ASTM C 939 (1987) Standard test method for flow of grout for preplaced-aggregate
concrete (flow cone method), Annual book of ASTM standards, American Society for
ASTM C 1621/C 1621M (2006) Standard test method for passing ability of self
consolidating concrete by J-ring Annual book of ASTM standards, American Society
for Testing and Materials, USA.
ASTM D 2196 (2005) Standard test methods for rheological properties of non-
Newtonian materials by rotational (Brookfield Type) viscometer, American Society
for Testing and Materials, USA.
ASTM D 4643 (1987) Standard test method for determination of water (moisture)
content of soil by the microwave oven method, Annual book of ASTM standards,
American Society for Testing and Materials, USA.
Atzeni, C., L. Massida, and U. Sanna (1985) Comparison between rheological

models for portland cement pastes. Cement and Concrete Research, 15, 511-519.
Banfill, P.F.G. (1981) A viscometric study of cement pastes containing

superplasticizers with a note on experimental techniques. Magazine of Concrete
Research, 33, 37-47.
Banfill, P.F.G. and D.C. Saunders (1981) On the viscometric examination of cement
pastes. Cement and Concrete Research, 11, 363-370.
Banfill, P.F.G. (2003) The rheology of fresh cement and concrete - A review.
Proceedings of 11th International congress on Cement Chemistry, Durban, South
Africa, 50-62.
244
Barnes, H.A. (1999) The Yield stress – a review, Journal of Non-Newtonian Fluid
Mechanics 81, 133-178.
Basile, F., S. Biagini, G. Ferrari, and M. Collepardi, Influence of different

sulphonated naphthalene polymers on the fluidity of cement paste. pp. 209-220, V.M
Malhotra (ed) Third International Conference on Superplasticizers and Other
Chemical Admixtures in Concrete, American Concrete Institute, Detroit, USA, 1989.
Basu, P.C. and S. Choudhury (2005) Influence of minor constituents of portland

cement on rheology of mortar for self compacting concrete. The Structural
Engineering Convention (SEC 2005), Indian Institute of Science, Bangalore, India,
December, 209-214.
Basu, P.C. and S. Choudhury (2007) Impact of fine aggregate particle size on
mortar rheology for SCC. The Indian concrete Journal, January, 7-14.
Baum, H. and M. Ben-Bassat Effect of superplasticizers on some properties of fresh

and hardened concrete mixes. pp. 482-492. J.G. Cabrera, and R.R. Villarreal (eds)
Proceedings of International Symposium on the Role of Admixtures in High
Performance Concrete, RILEM, Mexico, 1999.
Beaupre, D. and S. Mindess, Rheology of fresh concrete: Principles, measurement

and applications. pp. 149-190. J. Skalny and S. Mindess (eds) Materials Science of
Concrete V, The American Ceramic Society Publication, USA, 1998.
Bedard, C.P.E. and N.P. Mailvaganam (2005) The use of chemical admixtures in
concrete: Part1: Admixture –Cement compatibility. ASCE Journal of Performance of
Constructed Facilities, 19, 263-266.
Bedard, C.P.E. and N.P. Mailvaganam (2006) The use of chemical admixtures in
concrete: Part 11: Admixture-admixture compatibility and practical problems. ASCE
Journal of Performance of Constructed Facilities, 20, 2-5.
Bentur, A. (2002) Cementitious materials-Nine millennia and a new century: Past,

present and future. Journal of Materials in Civil Engineering, January/February, 2-22.
Bhatty, J.I. and P.F.G. Banfill (1982) Sedimentation behaviour in cement pastes
subjected to continuous shear in rotational viscometers. Cement and Concrete
Billberg, P., The effect of mineral and chemical admixtures on fine mortar rheology.
pp. 301-320. V.M. Malhotra (ed) Proceedings of Fifth CANMET/ACI International
Conference on superplasticizers and other chemical admixtures in concrete,
American Concrete Institute, USA, 1997.
Bonen, D. and S.L. Sarkar (1995) The superplasticizers adsorption capacity of

cement pastes, pore solution composition, and parameters affecting flow loss. Cement
and Concrete Research, 25, 1423-1434.
Brooks, J.J., M.A.M. Johari, and M. Mazloom (2000) Effect of admixtures on

the setting times of high-strength concrete. Cement and Concrete Composites, 22,
293-301
245
Calleja, J. (1952a) Determination of setting and hardening time of high-alumina
cements by electrical resistance techniques. Journal of American Concrete Institute, 4,
329-330.
Calleja, J. (1952b) New techniques in the study of setting and hardening of hydraulic
materials. Journal of American Concrete Institute, 23, 525-536.
Chang, P.K. and Y.N. Peng (2001) Influence of mixing techniques on properties of
high performance concrete. Cement and Concrete Research, 31, 87-95.
Chandra, S. and J. Bjornstrom (2002a) Influence of cement and superplasticizers

type and dosage on the fluidity of cement mortars- Part 1. Cement and Concrete
Research, 32, 1605-1611.
Chandra, S. and J. Bjornstrom (2002b) Influence of superplasticizers type and

dosage on the slump loss of Portland cement mortars- Part 11. Cement and Concrete
Research, 32, 1613-1619.
Chen, C.T., L.J. Struble, and H. Zhang (2006) Using dynamic rheology to measure
cement-admixture interactions. Journal of ASTM International, 3, 1-13.
Chougnet A., A. Audibert, and M. Moan (2007) Linear and non-linear rheological
behaviour of cement and silica suspensions - Effect of polymer addition. Rheologica
Acta, 46, 793-802.
Claisse, P.A., P. Lorimer, and M.H. Al-Omari, The effect of changes in cement on
the properties of cement grouts with superplasticizing admixtures, pp. 19-33. J.G.
Cabrera, and R.R. Villarreal (eds) Proceedings of International Symposium on the
Role of Admixtures in High Performance Concrete, Mexico, 1999.
Collepardi, M. (1998) Admixtures used to enhance placing characteristics of

concrete. Cement and Concrete Composites, 20, 103-112.
Collepardi, M., Admixtures enhancing concrete performance. pp. 217-230. R.K.

Dhir, P.C. Hewlett and M.D Newlands (eds) Proceedings of International
Conference on Admixtures-Enhancing Concrete Performance, Thomas Telford,
Dundee, UK, 2005.
Corinaldesi,V. and G. Moriconi, The influence of mineral additions on the rheology

of self compacting concrete. pp. 227-240. V.M Malhotra (ed) Seventh CANMET/
ACI International Conference on Superplasticizers and Other Chemical Admixtures in
Concrete, Berlin, Germany, 2003
Corradi, M., R. Khurana, and Magarotto, New superplasticizers for the total
control of performances of fresh and hardened concrete. pp. 1-10. R.K. Dhir, P.C.
Hewlett and M.D. Newlands (eds),Proceedings of International Conference on
Admixtures-Enhancing Concrete Performance, Thomas Telford, London, 2005.
Cyr, M., C. Legrand, and M. Mouret (2000) Study of the shear thickening effect of
superplasticizers on the rheological behaviour of cement pastes containing or not
mineral additives. Cement and Concrete Research, 30, 1477-1483.
246
Cyr, M. and M. Mouret, Rheological characterization of superplasticized cement
pastes containing mineral admixtures: Consequences on self compacting concrete
design pp. 241-256. V.M Malhotra (ed) Seventh CANMET/ ACI International
Conference on Superplasticizers and Other Chemical Admixtures in Concrete, Berlin,
Germany, 2003
Daimon, M. and D. M. Roy (1978) Rheological properties of mixes: I. methods,

preliminary experiments, and adsorption studies. Cement and Concrete Research, 8,
753-764.
Das, A.C. (2007) Admixture-cement Compatibility, M.Tech Project report, Indian

Institute of Technology, Madras, India.
de Larrard, F (1990) A method for proportioning high strength concrete mixtures.

Cement, Concrete and Aggregates, 12, 47-52.
de Larrard, F., C.F. Ferraris, and T. Sedran (1998) Fresh concrete: A Herschel-
Bulkley material. Materials and Structures, 31, 494-498.
de Larrard, F.(1999) Why rheology matters? Concrete International, 21, 79-81.
Dhir, R.K. and A.W.F. Yap (1983) Superplasticized high workability concrete:
some properties in the fresh and hardened states. Magazine of Concrete Research, 35,
215-228.
Diamond, S. (2005) The patch microstructure in concrete: effect of mixing time.

Cement and Concrete Research, 35, 1014-1016.
Diop, M.B., A.A. Ndiaya, S. Ndiaye, B. Lo, O. Sakho, P.D. Tall, A.C. Beye, and
W. Soboyejo, Evaluation of phosphogypsum as an alternative of gypsum in the
fabrication of CEM II/A-M 32.5 R Portland cement. R.K. Dhir, T.A. Harrison and
M.D. Newlands (eds) Proceedings of International conference on Cement
Combinations for Durable Concrete, University of Dundee, UK, 2005.
Dodson, V.H. and T.D. Hayden (1989) Another look at the Portland cement
/chemical admixture incompatibility problem. Cement Concrete and Aggregates, 11,
52-56.
Domone, P. (2006) Mortar tests for self consolidating concrete- A contribution from
ACI committee 236, Concrete International, April, 39-45.
EFNARC (2005) The European Guidelines for Self Compacting Concrete:

Specification, Production and Use. The European Federation of Specialist
Construction Chemicals and Concrete Systems. www.efnarc.org visited on 25/7/2007.
Emoto, T. and T.A. Bier (2007) Rheological behaviour as influenced by plasticizers

and hydration kinetics. Cement and Concrete Research, 37: 647-654.
Erdogdu, S. (2000) Compatibility of superplastcizers with cements different in

composition. Cement and Concrete Research, 30, 767-773.
Falikman, V.R. Y.V Sorokin, A.Y.Vainer, and N.F Bashlykov, New high
performance polycarboxylate superplasticizers based on derivative copolymers of
247
maleinic acid. pp. 41-46. R.K. Dhir, P.C. Hewlett and M.D Newlands (eds)
Proceedings of International Conference on Admixtures-Enhancing Concrete
Performance. Thomas Telford, Dundee, UK, 2005.
Famy, C., K.L. Scrivener, A. Atkinson, and A.R. Brough (2002) Effects of an early
or late heat treatment on the microstructure and composition of inner C-S-H products
of portland cement mortars. Cement and Concrete Research, 32, 269-278.
Faroug, F., J. Szwabowski, and S. Wild (1999) Influence of superplasticizers on

workability of concrete, Journal of materials in civil engineering, May, 151-157.
Felekoglu, B., B. Baradan, H. Sarikahya, and L. Yuceer, Compatibility of a new

generation polycarboxylate-type superplasticizer with different set accelerators. pp.
89-96. R.K. Dhir, P.C. Hewlett and M.D Newlands (eds) Proceedings of
International Conference on Admixtures-Enhancing Concrete Performance. Thomas
Telford, Dundee, UK, 2005.
Felekoglu, B. and H. Sarikahya (2007) Effect of chemical structure of

polycarboxylate based superplasticizers on workability retention of self compacting
concrete. Construction and Building Materials (Article in press).
Fernandez-Altable, V. and I. Casanova (2006) Influence of mixing sequence and

superplasticizer dosage on the rheological response of cement pastes at different
temperatures. Cement and Concrete Research, 36, 1222-1230.
Fernon, V., A. Vichot, N. Le Goanvic, P. Colombet, F. Corazza, and U. Costa,

Interaction between Portland cement hydrates and polynaphthalene sulfonates. pp.
225-248. V.M. Malhotra (ed) Proceedings of 5th Canmet/ACI International
Conference on Superplasticizers and Other Chemical Admixtures in Concrete,
American Concrete Institute, Detroit, 1997.
Ferraris, C.F. and J.M Gaidis (1992) Connection between the rheology of concrete
and rheology of cement paste. ACI Materials Journal, 88, 388- 393.
Ferraris, C.F. (1999) Measurement of rheological properties of HPC: State of art

report. Journal of research of the National Institute of Standards and Technology,
104, 461-478.
Ferraris, C.F. (2001) Concrete mixing methods and concrete mixers: State of the art
report. Journal of Research of the National Institute of Standards and Technology,
106, 391-399.
Ferraris, C.F., K.H. Obla, and R. Hill (2001) The influence of mineral admixtures
on the rheology of cement paste and concrete. Cement and Concrete Research, 31,
245-255.
Flatt, R.J., Y.F. Houst, P. Bowen, H. Hofmann, J. Widmer, U. Sulser, U. Maeder,

and T.A. Bürge, Interaction of superplasticizers with model powders in a highly
alkaline medium. pp. 743-762. V.M. Malhotra (ed) Proceedings of 5th
CANMET/ACI International Conference on Superplasticizers and Other Chemical
Admixtures in Concrete, American Concrete Institute, Detroit, 1997.
248
Flatt, R.J., Y.F. Houst, P. Bowen, H. Hofmann, J. Widmer, U. Sulser, U. Maeder,
and T.A. Bürge (1998) Analysis of superplasticizers used in concrete. Analusis
Magazine, 26, 28-34.
Flatt R.J. and Y.F. Houst (2001) A simplified view on chemical effects perturbing
the action of superplasticizers. Cement and Concrete Research, 31, 1169-1176.
Flatt, R. J. (2004) Dispersion forces in cement suspensions. Cement and Concrete

Research, 34, 399- 408.
Gaidis J.M. and E.M. Gartner, Hydration Mechanisms II. pp. 9-39. J. Skalny and
S. Mindess (eds), Materials Science of concrete II, The American Ceramic Society,
USA, 1991.
Geiker M.R., B. Mari, L. Nyholm, and N. Lauge (2002) On the effect of coarse
aggregate fraction and shape on the rheological properties of self compacting
concrete. Cement, Concrete and Aggregates, 24, 3-6.
Gettu, R., A. Aguado, L. Agullo, B. Toralles-Carbonari, and J. Roncero,

Characterization of cement pastes with silica fume and superplasticizer as components
of high performance concretes. pp. 331-344. P.K Mehta (ed) Mario Collepardi
Symposium on Advances in Concrete Science and Technology, 1997.
Gettu, R., J. Roncero, and M.A. Martin (2002) Long term behaviour of concrete
incorporating a shrinkage-reducing admixture. The Indian Concrete Journal,
February, 1-7.
Gettu, R., P.C.C. Gomes, L. Agullo, and A. Josa, High strength self compacting
concrete with fly ash: Development and utilization. pp. 507-522. V.M. Malhotra (ed)
Eighth CANMET/ACI International Conference on Flyash, Silica Fume, Slag and
Natural Pozzolans in Concrete, ACI International, USA, 2004.
Gettu, R., B. Barragan, T. Garcia, J. Ortiz, and R. Justa (2006) Fiber Concrete
tunnel lining. Concrete International, August, 63-69.
Giaccio, G. and R. Zerbino (2002) Optimum superplasticizer dosage for systems

with different cementitious materials. The Indian Concrete Journal, September, 553-
557.
Goisis M., A. Buscema, and T. De Marco, Characterization of the rheological

properties of cement paste prepared with polycarboxylate type superplasticizer. pp. 1-
15. V.M Malhotra (ed) Seventh CANMET/ ACI International Conference on
Superplasticizers and Other Chemical Admixtures in Concrete, Berlin, Germany,
2003.
Golaszewski J. and J. Szwabowski (2004), Influence of superplasticizers on

rheological behaviour of fresh cement mortars. Cement and Concrete Research, 34,
235-248.
Gomes, P.C.C., R. Gettu, L. Agullo, and C. Bernad, Experimental optimization of

high strength self compacting concrete. pp. 377-386. K. Ozawa, and M. Ouchi (eds)
Proceedings of Second International Symposium on SCC, Kochi, Japan, 2001.
249
Gomes, P.C.C. (2002) Optimization and characterization of high strength self
compacting concrete. Doctoral thesis, Universitat Politecnica De Catalunya,
Barcelona, Spain.
Gopinath, R. and R. Gettu, Influence of shrinkage reducing admixtures on early age

shrinkage. pp.1132-1142. P.J. Rao, V. Ramakrishnan, I. Patnaikuni, and V.S.
Parameswaran (eds) Proceedings of Second International Conference on Advances
in Concrete and Construction, Hyderabad, India, 2008.
Grabiec, A.M. and Z. Piasta (2004) Study on compatibility of cement –

superplasticizer assisted by multicriteria statistical optimization. Journal of Materials
Processing Technology, 152, 197-203.
Griesser, A. (2002) Cement-superplasticizer interactions at ambient temperatures,

Doctoral thesis, Swiss Federal Institute of Technology, Zurich.
Grzeszczyk, S., and M. Sudol, The influence of temperature on the fluidity of fresh
self compacting concrete. pp. 165-170. R.K. Dhir, P.C. Hewlett and M.D Newlands
(eds) Proceedings of the International Conference on Admixtures-Enhancing
Concrete Performance, Thomas Telford, Dundee, UK, 2005.
Hanna, E., M. Ostiguy, K. Khalifé, O. Stoica, B.-G. Kim, C. Bédard, M. Saric-

Coric, M. Baalbaki, S. Jiang, P.C. Nkinamubanzi, P.C. Aїtcin, and N. Petrov
(2000) The importance of superplasticizers in modern concrete technology.
Superplasticizers conference, Nice, 2-18.
Hanehara, S. and K. Yamada (1999) Interaction between cement and chemical

admixture from the point of cement hydration, absorption behaviour of admixture and
paste rheology. Cement and Concrete Research, 29, 1159-1165.
Hanehara, S. and K. Yamada (2008) Rheology and early age properties of

cement systems. Cement and Concrete Research, 21, 175-195.
Hewlett P.C. (ed) (2004) Lea’s chemistry of cement and concrete. Elsevier Ltd,
Burlington, UK.
Hidalgo, J., C.T. Chen, and L.J. Struble (2008) Correlation between paste and
concrete flow behaviour, ACI Materials Journal, 105, 281-288.
Houst, Y.F., R.J. Flatt, H. Hofmann, U. Mader, J. Widmer, U. Sulser, and T.A.
Burge, Influence of superplasticizer adsorption on the rheology of cement paste,
pp.71-78, J.G. Cabrera and R.R. Villarreal (eds) Proceedings of International
Symposium on the Role of Admixtures in High Performance Concrete, Mexico, 1999a.
Houst,Y.F., R.J. Flatt, P. Bowen, and H. Hofmann, Optimization of

superplasticizers: From research to application, pp. 121-134, J.G. Cabrera, and R.R.
Villarreal (eds) Proceedings of International Symposium on the Role of Admixtures in
High Performance Concrete, Mexico, 1999b.
Houst, Y.F., P. Bowen, and F. Perche, Towards tailored superplasticizers. pp. 11-20.
R.K. Dhir, P.C. Hewlett and M.D. Newlands (eds) Proceedings of International
250
Conference on Admixtures-Enhancing Concrete Performance, , Thomas Telford,
London, 2005.
Hsu, K.C., J.J. Chiu, S.D. Chen, and Y.C. Tseng (1999) Effect of addition time of a
superplasticizer on cement adsorption and on concrete workability. Cement and
Concrete Composites, 21, 425-430.
Hui, W.U., G. Huiling, L. Jiaheng, Z.Rongguo, and L Yong (2007) Research on

synthesis and action mechanism of polycarboxylate superplasticizer. Front. Chem.
China, 2, 322-325.
IS 383 (2002) Indian standard specification for coarse and fine aggregates from
natural sources for concrete. Bureau of Indian Standards, New Delhi.
IS 1199 (2004) Indian standard specification for methods of sampling and analysis of
concrete. Bureau of Indian Standards, New Delhi.
IS 2386 (2007) Indian standard specification for methods of test for aggregates of
concrete. Bureau of Indian Standards, New Delhi.
IS 4031 – Parts 2, 4 and 5 (2005) Indian standard specification for method of

physical tests for hydraulic cement. Bureau of Indian Standards, New Delhi.
IS 8142 (2002) Indian standard specification for method of test for determining
setting time of concrete by penetration resistance. Bureau of Indian Standards, New
Delhi.
IS 9103 (2004) Indian standard specification for Concrete Admixtures. Bureau of

Indian Standards, New Delhi.
IS 10262 (2004) Recommended guidelines for concrete mix design, Bureau of Indian
Standards, New Delhi.
IS 12269 (2004) Indian standard specification for 53 grade ordinary Portland cement.
Bureau of Indian Standards, New Delhi.
Jiang, S., B.G. Kim, and P.C. Aїtcin (1999) Importance of adequate soluble alkali
content to ensure cement/superplasticizer compatibility. Cement and Concrete
Johnston, C.D. (1987) Admixture-cement incompatibility: A case history. Concrete

International, 9, 51-60.
Jolicoeur, C. and M.A. Simard (1998) Chemical admixture–cement interactions:

Phenomenology and physico-chemical concepts. Cement and Concrete Composites,
20, 87-101.
Justnes, H., M.V. Dooren, and D.V. Gemert, Reasons for workability loss in
cementitious binders. pp 53-65. V.M. Malhotra (ed) Seventh CANMET/ACI
International Conference on Superplasticizers and Other Chemical Admixtures in
Concrete, Berlin, Germany, 2003
251
Kantro, D.L. (1980) Influence of water-reducing admixtures on properties of cement
paste-A miniature slump test. Cement, Concrete and Aggregates, 2,
95- 102.
Kapelko, A. (2006) The possibility of adjusting concrete mixtures’ fluidity by means

of superplasticizer SNF. Archives of Civil and Mechanical Engineering, VI, Wroclaw
University of technology, 3, 37- 53.
Khayat, K. H. and Z. Guizani (1997) Use of viscosity modifying admixture to

enhance stability of fluid concrete. ACI Materials journal, 94, 332-340.
Khayat, K.H. and A. Yahia (1997) Effect of welangum-high range water reducer
combinations on rheology of cement grout. ACI Materials journal, 94, 365-372.
Khayat, K.H. and A. Yahia (1998) Simple field tests to characterize fluidity and
washout resistance of structural cement grout. Cement, Concrete and Aggregates, 20,
145-156.
Khayat, K.H., M. Saric-Coric, and F. Liotta (2002) Influence of thixotropy

onstability characteristics of cement grout and concrete. ACI Materials Journal, 99,
234-241.
Khayat, K.H., A. Yahia, and M. Sayed (2008) Effect of supplementary cementitious

materials on rheological properties, bleeding and strength of structural grout. ACI
Materials Journal, 105, 585-593.
Khatib, J.M. and S. Wild (1996) Pore size distribution of metakaolin paste. Cement
Khatib, J.M. and P.S. Mangat (1999) Influence of superplasticizer and curing on
porosity and pore structure of cement paste. Cement and Concrete Composites, 21,
431-437.
Kim, B.G., S. Jiang, and P.C. Aїtcin, Influence of molecular weight of PNS
superplasticizers on the properties of cement pastes containing different alkali
contents. J.G. Cabrera, and R.R. Villarreal (eds) Proceedings of International
Symposium on the Role of Admixtures in High Performance Concrete, Mexico, 1999.
Kim, B.G., S. Jiang, C. Jolicoeur, and P.C. Aїtcin (2000) The adsorption behaviour
of PNS superplasticizer and its relation to fluidity of cement paste. Cement and
Concrete Research, 30, 887-893.
Kim, B.G. and P.C. Aїtcin, Dispersing mechanism of PNS superplasticizers in high
performance concrete. G. Grieve and G. Owens (eds) Proceedings of 11th
international congress on the chemistry of cement, South Africa, 11-16 May, Durban,
South Africa, 2003.
Kim, B.G., T.H Ahn,. B.G. Kang and Y.T. Kim, Effect of superplasticizer type on
the properties of high performance concrete incorporating metakaolin,. pp. 83-
97.Supplementary papers compiled by P.C. Nkinamubanzi, Seventh CANMET/ACI
International Conference on Superplasticizers and Other Chemical Admixtures in
Concrete, Berlin, Germany, 2003.
252
Kirca, O., L. Turanli, and T.Y. Erdogan (2002) Effects of retempering on
consistency and compressive strength of concrete subjected to prolonged mixing.
Kjellsen, K.O., R.J. Detwiler, and O.E. Gjorv (1991) Backscattered Electron image
analysis of cement paste specimens: specimen preparation and analytical methods.
Krishna, R.N. (1993) Compatibility and strength characteristics of superplasticized

high strength concrete. Doctoral thesis, Anna University, Madras, India.
Krishna, R.N. (1996) Dispersing action of a superplasticizer with different grades of

cements and flyash. ACI materials Journal, 93, 351-355.
Kumar, P. and S.K. Kaushik (2003) Some trends in the use of concrete: Indian
Scenario. The Indian Concrete Journal, December, 1503-1508.
Kumar, J.V. (2006) Study of cement-superplasticizer compatibility in concrete.

M.tech Project Report, Indian Institute of Technology, Madras, India.
Lachemi, M., K.M.A Hossain, V. Lambros, P.C. Nkinamubanzi, and N.

Bouzouba (2004) Self compacting concrete incorporating new Viscosity modifying
admixtures. Cement and Concrete Research, 34, 917-926.
Lachemi, M., K.M.A. Hossain, R. Patel, M. Shehata, and N. Bouzoubaa (2007)

Influence of paste/mortar rheology on the flow characteristics of high volume fly ash
self consolidating concrete. Magazine of Concrete Research, 59, 517- 528.
Lapasin, R., A. Papo, S. Rajgelj (1983) The phenomenological description of the

thixotropic behaviour of fresh cement pastes. Rheologica Acta, 22, 410-416.
Laskar, A.I. and S. Talukdar (2007) Correlation between compressive strength and
rheological parameters of high performance concrete. Research Letters in Materials
Science, Article ID 45869, 1-4.
Li, Z., and Z. Ding (2003) Property improvement of portland cement by

incorporating with metakaolin and slag. Cement and Concrete Research, 33, 579-584.
Li, G., A. Tagnit-Hamou, and P.C. Aïtcin, Improving cement-superplasticizer

compatibility by using soluble alkalis as a chemical additive in concrete. pp.655-665.
G. Grieve and G. Owens (eds) Proceedings of 11th international congress on the
chemistry of cement, South Africa, 11-16 May, Durban, South Africa, 2003.
Liao, T.S., C.L. Hwang, Y.S Ye, K.C. Hsu (2006) Effects of a carboxylic acid/
sulfonic acid copolymer on the material properties of cementitious materials. Cement
Lim, G.G., S.S. Hong, D.S. Kim, B.J. Lee, and J.S. Rho (1999) Slump loss control
of cement paste by adding polycarboxylic type slump- releasing dispersant. Cement
and Concrete Research, 29, 223-229
253
Lootens D., P. Hebraud, E. Lecolier, and H.V. Damme (2004) Gelation, shear-
thinning and shear-thickening in cement slurries. Oil and Gas Science andTechnology
– Rev. IFP, 59, 31-40.
Maeder, U. and I. Schober, Performance of blends of polycarboxylate polymers in

different cements. G. Grieve and G. Owens (ed) Proceedings of 11th international
congress on the chemistry of cement, 11-16 May, Durban, South Africa, 2003.
Magarotto, R., I. Torresan, and N. Zeminian, Influence of molecular weight of

polycarboxylate ether superplasticizers on the rheological properties of fresh cement
pastes, mortar and concrete. pp. 514-526. G. Grieve and G. Owens (eds) Proceedings
of 11th international congress on the chemistry of cement. 11-16 May, Durban, South
Africa, 2003.
Mailvaganam, N.P. (1999) Chemical admixtures in concrete- side effects and

compatibility problems, The Indian Concrete Journal, 73, 367-374.
Mailvaganam, N. P. (2001) How chemical admixtures produce their effects in

concrete? Indian Concrete Journal, 75, 331-334.
Mansoutre S., P. Colombet, H. Van Damme (1999) Water retention and granular
rheological behaviour of fresh C3S paste as a function of concentration. Cement and
Concrete Research, 29, 1441-1453
Masood, I. and Agarwal (1994) Effect of various superplasticizers on rheological

properties of cement paste and mortars. Cement and Concrete Research, 24, 291- 302.
McCarter, W.J. and P.N. Curran (1984) The electrical response characteristics of
setting of cement paste. Magazine of Concrete Research, 36, 42-49.
McCarter, W. J (1987) Gel formation during early hydration. Cement and Concrete
McCarter, W.J. and D. Tran (1996) Monitoring pozzolanic activity by direct

activation with calcium hydroxide. Construction and Building materials, 10, 179-184.
McCarter, W.J., T.M. Chrisp, and G. Starrs (1999) The early hydration of alkali-
activated slag: Developments in monitoring techniques. Cement and Concrete
Composites, 21, 277-283.
McCarter, W.J., T.M. Chrisp, G. Starrs, and J. Blewett (2003) Characterisation

and monitoring of cement based systems using intrinsic electrical property
measurements. Cement and Concrete Research, 33, 197-206.
Mehta, P.K. and P.J.M. Monteiro, Concrete-Microstructure, properties and

materials, Indian Concrete Institute, Chennai, India, 2005.
Meyer, L.M. and W.F. Perenchio (1979) Theory of concrete slump loss as related to
the use of chemical admixtures, Concrete International, 1, 36-43.
Min B.H., L. Erwin, and H.M. Jennings (1994) Rheological behaviour of fresh
cement paste as measured by squeeze flow, Journals of Materials Science, 29, 1374-
1381.
254
Mollah, M.Y.A., P. Palta, and T.R. Hess, R.K. Vempati, and D. L. Cocke (1995)
Chemical and physical effects of sodium lignosulphonate superplasticizer on the
hydration of portland cement and solidification/stabilization consequences. Cement
Monte R., and A.D. Figueiredo, Comparative evaluation of test methods for
determination of superplasticizers saturation dosages in cement pastes. pp 147-159.
V.M. Malhotra (ed) Seventh CANMET/ACI International Conference on
Superplasticizers and Other Chemical Admixtures in Concrete, Berlin, Germany,
2003.
Mork, J.H. and O.E Gjoerv (1997) Effect of gypsum-hemihydrate ratio in cement
on rheological properties of fresh concrete. ACI Materials Journal, March-April, 142-
146.
Mullick, A.K. (2008) Cement-superplasticizer compatibility and method of

evaluation, The Indian Concrete Journal, 82, 8-16.
Nagataki, S., E. Sakai, and T. Takeuchi (1984) The fluidity of flyash-cement paste
with superplasticizer. Cement and Concrete Research, 14, 631-638.
Nakajima, Y. and K. Yamada (2004) The effect of the kind of calcium sulphate in
cements on the dispersing ability of poly β naphthalene sulphonate condensate
superplasticizer. Cement and Concrete Research, 34, 839-844.
Nawa, T., H. Ichiboji, and M. Kinoshita, Influence of temperature on fluidity of

cement paste containing superplasticizer with polyethylene oxide graft chains. Pp.
195-210. V.M Malhotra (ed) Sixth CANMET/ ACI International Conference on
Superplasticizers and Other Chemical Admixtures in Concrete, Ottawa, Canada, 2000
Nehdi, M and M.A. Rahman (2004) Estimating rheological properties of cement

pastes using various rheological models for different test geometry, gap and surface
friction. Cement and Concrete Research, 34, 1993-2007.
Neville, A.M. Properties of concrete, ELBS with Longman, Harlow, UK, 2000.
Nguyen, V.H., S. Remond, J.L. Gallias, J.P. Bigas, and P. Muller (2006) Flow of
Herschel- Bulkley fluids through the Marsh cone, Journal of Non Newtonian Fluid
mechanics, 139, 128- 134.
Nkinamubanzi, P.C., B.G. Kim, and P.C. Aїtcin, Some key cement factors that
control the compatibility between naphthalene-based superplasticizers and ordinary
Portland cements. pp.33-54 V.M. Malhotra (ed) 6th CANMET/ACI International
Conference on superplasticizers and other chemical admixtures in concrete, Paris,
2000.
Nkinamubanzi, P.C. and P.C. Aїtcin (2004) Cement and superplasticizer

combinations: Compatibility and Robustness. Cement, Concrete and Aggregates, 26,
1-8.
Ohta, A., T. Sugiyama, and T. Uomoto, Study of dispersing effects of

polycarboxylate based dispersant on fine particles, pp. 211-227. V.M Malhotra (ed)
255
Sixth CANMET/ ACI International Conference on Superplasticizers and Other
Chemical Admixtures in Concrete, Ottawa, Canada, 2000
P 18-358 (1985) Adjuvants pour bètons, mortiers et coulis. Coulis courants

d’injection pour prècontrainte. Measure de la fluiditè et da la reduction d’eau,
Association francaise de normalization (afnor), Juillet (French Code).
Page, M., P.C Nkinamubanzi, and P.C. Aїtcin, The cement/superplasticizer

compatibility: A headache for superplasticizer manufacturers. J.G. Cabrera, and
R.R. Villarreal (eds) Proceedings of International Symposium on the Role of
Admixtures in High Performance Concrete, Mexico, 1999.
Papayianni, I., G. Tsohos, N. Oikonomou, and P. Mavria (2005) Influence of

superplasticizer type and mix design parameters on the performance of them in
concrete mixtures. Cement and Concrete Composites, 27, 217-222.
Papo, A., and L. Piani (2004) Effect of various superplasticizers on the rheological
properties of Portland cement pastes. Cement and Concrete Research, 34, 2097-2101.
Park C.K., M.H. Noh, and T.H. Park (2005) Rheological properties of cementitious
materials containing mineral admixtures. Cement and Concrete Research, 35, 842-
849.
Phan, T.H., M. Chaouche, and M. Moranville (2006) Influence of organic

admixtures on the rheological behaviour of cement pastes. Cement and Concrete
Research, 36, 1807-1813.
Plank, J. and C. Hirsch (2007) Impact of zeta potential of early cement hydration
phases on superplasticizer adsorption. Cement and Concrete Research, 37, 537-542.
Porter, M.R. Handbook of surfactants. Blackie Academic and Professional, United

Kingdom, 1994.
Prince, W., M. Edwards-Lajnef, and P.C. Aïtcin (2002) Interaction between

ettringite and polynaphthalene sulphonate superplasticizer in cementitious paste.
Prince, W., M. Espagne, and P.C. Aїtcin (2003) Ettringite formation: A crucial step
in cement superplasticizer compatibility. Cement and Concrete Research, 33, 635-
641.
Puertas, F., H. Santos, M. Palacios, and S. Martinez-Ramirez (2005)

Polycarboxylate Superplasticizer admixtures: effects on hydration, microstructure and
rheological behaviour in cement pastes. Advances in Cement Research, 17,
77- 89.
Rahman, M.A and M. Nehdi (2005) Empirical correlations between rheological

properties of cement pastes from various models. The Indian Concrete Journal,
October, 52- 60.
Ramachandran, V S., M.S Lowery, V. Malhotra, (1995) Behaviour of ASTM Type

V cement hydrated in the presence of sulfonated melamine formaldehyde. Materials
and Structures, 28, 133-138.
256
Ramachandran, V.S., V.M Malhotra, C. Jolicoeur, and N. Spiratos,
Superplasticizers: Properties and Applications in Concrete. Materials Technology
Laboratory, CANMET, Canada, 1998.
Ramachandran, V.S. and J.J. Beaudoin, Handbook of analytical techniques in

concrete science and technology. Noyes and William Andrew Publications, USA,
2001.
Ramachandran, V.S. (2002) Concrete admixtures handbook. Standard Publishers,

New Delhi, India.
Ramezanianpour, A.A., V. Sivasundaram, and V.M. Malhotra (1995)

Superplasticizers: Their effect on the strength properties of concrete. Concrete
International, April, 30-35.
Ravina,D. and I. Soroka (1994) Slump loss and compressive strength of concrete
Made with WRR and HRWR admixtures and subjected to prolonged mixing. Cement
Rixom R and N. Mailvaganam, Chemical admixtures for concrete. E&FNSpon,

London, UK, 1999.
Roberts, L.R. (1995) Dealing with cement-admixture interactions. 23 rd Annual

convention of the Institute of concrete technology ,11th April, Telford. p.16
Roncero, J., R. Gettu, P.C.C Gomes, and, L. Agullo (2000) Study of flow
behaviour of superplasticized cement paste systems and its influence on properties of
fresh concrete. pp. 273-294. High Performance Concrete: Research to practice,
American Concrete Institute, USA.
Roncero, J. (2000) Effect of superplasticizers on the behaviour of concrete in the

fresh and hardened states: Implications for high performance concretes, Doctoral
thesis, Universitat Politecnica de Catalunya, Barcelona, Spain.
Roncero, J., R. Gettu, E. Vazquez, and J.M. Torrents, Effect of superplasticizer

content and temperature on the fluidity and setting of cement pastes. pp. 343-356.
J.G. Cabrera, and R.R. Villarreal (eds) Proceedings of International Symposium on
the Role of Admixtures in High Performance Concrete, Mexico, 1999.
Roncero, J., S. Valls, and R.Gettu (2002) Study of the influence of superplasticizers
on the hydration of cement paste using nuclear magnetic resonance and X-ray
diffraction techniques. Cement and concrete Research, 22, 103-108.
Roussel, N. (2005) Steady and transient flow behaviour of fresh cement pastes.
Cement and concrete Research, 35, 1656-1664.
Roussel, N. and R.L. Roy (2005) The Marsh cone: a test or a rheological apparatus?
Roussel, N and P. Coussot (2005) “Fifty cent rheometer” for yield stress
measurements: From slump to spreading flow. Journal of Rheology, 49, 705-718.
257
Roussel, N. (2006) Correlation between yield stress and slump: Comparison between
numerical simulations and concrete rheometers results. Materials and Structures, 39,
501-509.
Roy, R.L. and N. Roussel (2005) The Marsh cone as a viscometer: theoretical
analysis and practical limits. Materials and Structures, 38, 25-30.
Roy, D.M. and K. Asaga (1980) Rheological properties of cement mixes: V. The
effects of time on viscometric properties of mixes containing superplasticizers;
conclusions. Cement and Concrete Research, 10, 389-394.
Sakai, E. and M. Daimon, Mechanisms of superplastification. pp.91-111. J. Skalny

and S. Mindess (eds) The American Ceramic Society Publication, USA, 1995.
Sakai, E., K. Yamada, and A. Ohta (2003) Molecular structure and dispersion-
adsorption mechanisms of comb-type superplasticizers used in Japan. Journal of
Advanced Concrete Technology, 1, 16-25.
Sakai, E., T. Kasuga, T. Sugiyama, K. Asaga, and M. Daimon (2006) Influence of

superplasticizers on hydration of cement and pore structure of hardened cement.
Schmidt, G. and E. Schlegel (2002) Rheological characterization of C-S-H phases-

water suspensions. Cement and Concrete Research, 32, 593-599.
Schultz, M.A. and L.J. Struble (1993) Use of oscillatory shear to study flow
behaviour of fresh cement paste. Cement and Concrete Research, 23, 273- 282.
Schwartzentruber, L.D., R. Roy, and J. Cordin (2006) Rheological behaviour of

fresh cement pasts formulated from a self compacting concrete (SCC). Cement and
Scrivener, K.L., The microstructure of Concrete. pp. 127-157. J.P. Skalny (ed)
Materials science of concrete I, The American ceramic society publication,
Westerville, 1989.
Seung-Bum, P. (1999) The effects of superplasticizers on the engineering properties

of plain concrete. KCI Concrete Journal, 11, 29-43.
Shah, S. (2006) Gypsum, a valuable input for agriculture, John Harrison Publications
(visited on 8.11.08)
(www.allotment.org.uk/articles2/Gypsum_Valuable_Input_for_Agriculture.php)
Simard, M.A., P.C. Nkinambanzi, and C. Jolicoeur (1993) Calorimetry Rheology

and compressive strength of superplasticized cement pastes, Cement and Concrete
Research, 23, 939-950.
Singh, N.B., R. Sarvahi, and N.P Singh (1992) Effect of superplasticizers on the
hydration of cement. Cement and Concrete Research, 22, 725-735.
Sonebi, M. (2006) Rheological properties of grouts with viscosity modifying agents

as diutan gum and welan gum incorporating pulverized fly ash. Cement and Concrete
Research, 36, 1609-1618.
258
Soroka I., Portland Cement Paste and Concrete. The Macmillan Press Limited,
London, 1979.
Spiratos, N., M. Page, N.P. Mailvaganam, V.M. Malhotra, and C. Jolicoer,

(2003) Superplasticizers for Concrete- Fundamentals, Technology and Practice,
Supplementary Cementing materials for Sustainable Development, Ottawa, Canada.
Spiratos, N. and C. Jolicoeur (2006) Concrete chemical admixtures: Perspective and

challenges. Indian Concrete Institute Journal, 6, 23-32.
Srinath, B. (2007) Use of non destructive techniques to analyze fresh and hardened
state properties of concrete. Master of Science Thesis, Indian Institute of Technology,
Madras, India.
Struble L.J. and M.A. Schultz (1993) Using creep and recovery to study flow
behaviour of fresh cement paste. Cement and Concrete Research, 23, 1369- 1379.
Struble, L. and G.K. Sun (1995) Viscosity of Portland cement paste as a function of
concentration. Advanced cement based materials, 2, 62-69.
Struble L.J. and Leit (1995) Rheological changes associated with setting of cement
Paste. Advanced cement based materials, 2, 224-230.
Struble, L., R. Szecsy, W.G. Lei and G.K. Sun (1998) Rheology of cement paste
and concrete, Cement, Concrete and Aggregates, 20, 269-277.
Struble L.,T.Y. Kim, and H.Zhang (2001) Setting of cement and concrete. Cement,
Concrete and Aggregates, 23, 88-93.
Sugiyama, T., A. Ohta and T. Uomota, The dispersive Mechanism and applications
of polycarboxylate based superplasticizers. pp. 560-568. G. Grieve and G. Owens
(eds) Proceedings of 11 th international congress on the chemistry of cement, Durban,
South Africa, 2003.
Sun Z., T. Voigt, and S.P. Shah (2006) Rheometric and ultrasonic investigations of
viscoelastic properties of fresh Portland cement pastes. Cement and Concrete
Research, 36, 278-287
Swamy R.N. (1991) Mineral admixtures for high strength concrete. The Indian
Concrete Journal, 65, 265-271.
Tagmit–Hamou, G., A. Li, and P.C Aїtcin, (2003) Improving cement

superplasticizer compatibility by using soluble alkalis as a chemical additive in
concrete. Proceedings of 11 th international congress on the chemistry of cement,
Durban, South Africa.
Tanaka, Y., A. Ohta, and T. Sugiyama (1999) Polycarboxylate based advanced

superplasticizers for high performance concrete. pp. 135-142. J.G. Cabrera, and
R.R. Villarreal (eds) Proceedings of the International Symposium on the Role of
Admixtures in High Performance Concrete, Mexico, 1999.
259
Tang, C.W., T. Yen, C.S. Chang, and K.H. Chen (2001) Optimizing mixture
proportions for flowable high performance concrete via rheology tests. ACI Materials
Journal, 98, 493-502.
Tattersall G.H., P.F.G Banfill, The rheology of fresh concrete. Pitman Books
Limited, London, 1983.
Taylor, H.F.W., Cement Chemistry, Thomas Telford, London UK, 1997.
Toralles-Carbonari B., R. Gettu, L. Agullo, A. Aguado, and V.A. Acena (1996) A

Synthetic approach for the experimental optimization of high strength concrete. pp.
161-168. F. de Larrard and R. Lacroix (eds) Proceedings of fourth International
Symposium on Utilization of High Strength/ High Performance Concrete, Paris, 1996.
Torrents, J.M., J. Roncero, and R. Gettu (1998) Utilization of impedence

spectroscopy for studying the retarding effect of a superplasticizer on the setting of
the cement. Cement and Concrete Research, 28, 1325-1333.
Torresan, I. and R. Khurana, New superplasticizers based on modified melamine

polymer. pp. 235-253, V.M Malhotra, (ed) Recent Advances in Concrete Technology,
Proceedings Fourth CANMET/ACI/JCI International Conference, ACI International,
Tokushima, Japan, 1998.
Tregger, N., L. Ferrara, and S.P. Shah (2008) Identifying viscosity of cement paste
from mini-slump flow test. ACI Materials Journal, 105, 558-566.
Tzouvalas, G., A. Papageorgiou, S. Tsimas, and D. Papageorgio (2005) Study of

early hydration of cement pastes containing alternative calcium sulphate bearing
materials, Proceeding of International Conference on Use of Foamed Concrete in
Construction, Dundee, UK, 41-51.
Uchikawa, H., D. Sawaki, and S. Hanehara (1995) Influence of kind and added
timing of of organic admixture on the composition, Structure, and property of fresh
cement paste. Cement and Concrete Research, 25, 353-364.
Uchikawa, H., S. Hanehara, and D. Sawaki (1997) The role of steric repulsive force
in the dispersion of cement particles in fresh paste prepared with organic admixture.
Van Damme H., S. Mansoutre, P. Colombet, C. Lesaffre, and D. Picart (2002)

Pastes: Lubricated and cohesive granular media. C.R Physique, 3, 229-238
Vandanjon, P.O., F. de Larrard, B. Dehouse, G. Villain, R. Maillot, and P.

Laplante (2003) Homogenisation of concrete in a batch plant: The influence of
mixing time and method on the introduction of mineral admixtures. Magazine of
Vikan, H., H. Justnes, F. Winnefeld, and R. Figi (2007) Correlating cement

characteristics with rheology of paste. Cement and Concrete Research, 37, 1502-
1511.
Wallevik, J.E. (2005) Thixotropic investigation on cement paste: Experimental and

numerical approach. Journal of Non-Newtonian Fluid Mechanics, 132, 86-99.
260
Williams, D.A., A.W. Saak, and H.M. Jennings (1999) The influence of mixing on
the rheology of fresh cement paste. Cement and Concrete Research, 29, 1491-1496.
Wineman, A.S. and K.R. Rajagopal, Mechanical response of polymers. Cambridge

University Press, United Kingdom, 2000.
Winnefeld, F., S. Becker, J. Pakusch and T. Gotz (2007) Effects of the molecular
architecture of comb shaped superplasticizers on their performance in cementitious
system. Cement and Concrete Composites, 29, 251-262.
Xiao, L.Z., Z.J. Li and X.S. Wei (2007) Selection of superplasticizer in concrete mix
design by measuring the early electrical resistivities of pastes. Cement and Concrete
Composities, 29, 350-356.
Yamada, K., T. Takahashi, S. Hanehara, and M. Matsuhisa (2000) Effects of the

chemical structure on the properties of polycarboxylate type superplasticizer. Cement
Yamada, K., S. Ogawa, and Hanehara (2001) Controlling of the adsorption and
dispersing force of polycarboylate-type superplasticizer by sulfate ion concentration
in aqueous phase. Cement and Concrete Research, 31, 375-383.
Yamada, K. and S. Hanehara, Working mechanism of polycarboxylate

superplasticizer considering the chemical structure and cement characteristics. pp.
538-549. G. Grieve and G. Owens (eds) Proceedings of 11 th international congress
on the chemistry of cement, Durban, South Africa, 2003.
Yamada, K., T. Sugamata, and H. Nakanishi (2006) Fluidity performance

evaluation of cement and superplasticizers. Journal of Advanced Concrete
Technology, 4, 241-249.
Yang, M., and H.M. Jennings (1995) Influences of mixing methods on the
microstructure and rheological behaviour of cement paste. Advanced Cement Based
Materials, 2, 70-78.
Yen, T., C.W. Tang, C.S. Chang, and K.H. Chen (1999) Flow behaviour of high
strength high performance concrete. Cement and Concrete Composites, 21, 413-424.
Yoshioka, K., E. Tazawa, K. Kawai, and T. Enohata (2002) Adsorption

characteristics of superplasticizers on cement component minerals. Cement and
Concrete Research, 32, 1507-1513
Young, J.F. (2008) Looking ahead from the past: The heritage of cement chemistry.
Yuan, C.Z., and W.J Guo (1987) Bond between marble and cement paste. Cement
Zhang, T., S. Shang, F. Yin, A. Aishah, A. Salmiah and T.L. Ooi (2001)
Adsorptive behaviour of surfactants on surface of portland cement. Cement and
261
Zhang, X. (2007) Quantitative microstructural characterization of concrete cured
under realistic temperature conditions. Doctoral thesis, Ecole Polytechnique Federale
de Lausanne, Zwitzerland.
Zhao, H. and D. Darwin (1992) Quantitative backscattered electron analysis of

cement paste. Cement and Concrete Research, 22, 695-706.
Zhor, J. and T.W. Bremner, Advances in evaluation of lignosulphonates as concrete

admixtures. pp. 1011-1042. V.M Malhotra (ed) Recent Advances in Concrete
Technology, Proceedings of Fourth CANMET/ACI/JCI International Conference, ACI
International, Tokushima, Japan, 1998.
Zingg, A., L. Holzer, A. Kaech, F. Winnefeld, J. Pakusch, S. Becker, L. Gauckler.

(2008) The microstructure of dispersed and non-dispersed fresh cement pastes – New
insight by cryo-microscopy. Cement and Concrete Research, 38, 522-529.
262
LIST OF PAPERS SUBMITTED ON THE BASIS
OF THIS THESIS
I REFERRED JOURNALS
1. C Jayasree and Ravindra Gettu (2008) Experimental Study of the Flow

Behaviour of Superplasticized Cement Paste. Materials and Structures,
RILEM, 41, 1581-1593.
2. C Jayasree and Ravindra Gettu. Correlations between the Fresh and

Hardened Properties of Superplasticized Paste, Mortar and Concrete, Indian
Concrete Journal (Under review)
II PRESENTATIONS IN CONFERENCES
1. C Jayasree and Ravindra Gettu Influence of Mixing Method on the Fluidity

of Superplasticized Cement Paste. Proceedings of the Fifth Asian Symposium
on Polymers in Concrete, Chennai, India, 11-12 September 2006, 665-670.
2. C Jayasree and Ravindra Gettu. Optimization of Superplasticized Cement

Paste and its Correlation with Fresh Concrete Behaviour. Proceedings of the
International Conference on Advances in Concrete and Construction,
Hyderabad, India, 7-9 February 2008, 1042-1052.
3. C Jayasree and Ravindra Gettu. Mix Design of Self Compacting Concrete

with Viscosity Modifying admixtures, International Conference on Innovative
World of Concrete, 11-14 December 2008, New Delhi, India.
263

THESIS - C. JAYASREE - Jayasree Chakkamalayath

Uploaded by

Copyright:

Available Formats

THESIS - C. JAYASREE - Jayasree Chakkamalayath

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

THESIS - C. JAYASREE - Jayasree Chakkamalayath

Uploaded by

Copyright:

Available Formats

STUDY OF CEMENT-SUPERPLASTICIZER INTERACTION AND

ITS IMPLICATIONS FOR CONCRETE PERFORMANCE

for the award of the degree

BUILDING TECHNOLOGY AND CONSTRUCTION MANAGEMENT DIVISION

This is to certify that the thesis entitled “STUDY OF CEMENT-

SUPERPLASTICIZER INTERACTION AND ITS IMPLICATIONS FOR

CONCRETE PERFORMANCE” submitted by C. Jayasree to the Indian Institute

of Technology Madras for the award of the degree of Doctor of Philosophy is a

Institute or University for the award of any degree or diploma.

Chennai 600 036 Prof. Ravindra Gettu

I am deeply indebted to my research guide Prof. Ravindra Gettu, Department of Civil

I also wish to place on record my sincere gratitude to Prof. K. Ramamurthy,

I am extremely thankful to Dr. J. Muralikrishnan, Assistant Professor, Department of

I express my sincere thanks to the present Head of Civil Engineering Department,

I am pleased to record my gratitude to the Management of Amrita Vishwa Vidya

I would like to thank the chemical admixture companies, BASF Construction

I am deeply indebted to my in-laws who have encouraged me in fulfilling my dream

Chemical admixtures, particularly superplasticizers have become indispensable

compatibility between the cement and admixture, which is influenced by the

the benefits of incorporating a superplasticizer. It is, therefore, essential to understand

the mechanisms that influence the cement-superplasticizer interaction for selecting an

appropriate type of admixture at an optimum dosage. The current study addresses

chemical and mineral admixtures, especially superplasticizers.

prepared will be representative of that of concrete. In order to evaluate the influence

of superplasticizers and cement characteristics, studies are conducted with four

rheological models to understand its effectiveness in capturing the relative fluidity.

of retardation varies with the type of superplasticizer.

hydration processes and on the development of microstructure of hydrated cement

techniques (X-ray diffraction, scanning electron microscopy, thermal analysis and Si

superplasticizers in delaying the hydration. The superplasticizer delays the formation

of ettringite as well as the silicate polymerization; however, it produces a dense and

uniform microstructure, enhancing the strength development.

cement and superplasticizer characteristics on the absorbance of the superplasticizers

cement-superplasticizer interaction shows that the addition of metakaolin and

viscosity modifying admixtures increases the saturation dosage of the superplasticizer

and the viscosity of the paste.

Finally, based on the test results, a simple methodology is recommended to select a

compatible cement-superplasticizer combination is recommended. The use of Marsh

strength and cost evaluation are suggested as part of this methodology.

Keywords: Superplasticizer; Marsh cone; mini-slump; rheology; loss of

Title Page No.

1.1 General ...........................................................................................................1

CHAPTER 2 STATE-OF THE-ART REPORT ON CEMENT-

2.1 General ...........................................................................................................7

2.3.6.1 Cement superplasticizer Interaction in the Presence of Mineral

2.4 Need for the Present Study ...........................................................................44

CHAPTER 3 STUDY OF THE FLOW BEHAVIOUR OF

3.1 General ................................................................................................ 47

3.5.5 Change of Flow Behaviour with Time .........................................................82

CHAPTER 4 CORRELATIONS BETWEEN THE FRESH AND HARDENED

4.1 General .........................................................................................................87

CHAPTER 5 EFFECT OF SUPERPLASTICIZER ON PROPERTIESS OF

5.1 General .......................................................................................................120

CHAPTER 6 CEMENT-SUPERPLASTICIZER COMPATIBILITY

6.1 General .......................................................................................................148

6.5 Study of the Adsorption of superplasticizers .............................................162

CHAPTER 7 FLOW BEHAVIOUR OF SUPERPLASTICIZED PASTES

7.1 General .......................................................................................................177

CHAPTER 8 CONCLUSIONS AND RECOMMENDATIONS FOR