Chemical Engineering Science 247 (2022) 117079

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Bio-fuel additive synthesized from levulinic acid using ionic

liquid-furfural based carbon catalyst: Kinetic, thermodynamic
and mechanism studies
Muzakkir Mohammad Zainol a, Mohd Asmadi b, Nor Aishah Saidina Amin b,⇑
a
School of Chemical Engineering, College of Engineering, Universiti Teknologi MARA, 40450 Shah Alam, Selangor, Malaysia
b
Chemical Reaction Engineering Group (CREG), School of Chemical and Energy Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81300 Johor Bahru, Johor, Malaysia

h i g h l i g h t s g r a p h i c a l a b s t r a c t

New carbon-based catalyst from ionic

liquid – furfural mixtures used for
reaction.

Kinetic models were developed to
describe the esterification reaction.

Pseudo first-order kinetic model
fitted-well with experimental data.
à

Thermodynamic parameters: Ea, DG ,
DSà., DHà were obtained for the
reaction system.

Combination of nucleophilic
substitution and Eley-Rideal as
postulated mechanism.

a r t i c l e i n f o a b s t r a c t

Article history: The ethyl levulinate is one of promising platform chemical from biomass and commonly involved the
Received 7 June 2021 esterification reaction of levulinic acid. The reactions are extensively focussed on the catalytic
Received in revised form 22 August 2021 performance by various catalysts and presented limited work on the kinetic, thermodynamic and mech-
Accepted 28 August 2021
anism study for heterogeneous catalyst reaction. To fill this gap, the reaction analysis over a new ionic
Available online 07 September 2021
liquid-furfural carbon catalyst has been investigated in this work. The mathematical equations were
derived to determine the kinetic-thermodynamic parameters, and proposed suitable mechanism for
Keywords:
the reaction. Pseudo-first order model presents high correlation coefficient and accuracy with the reac-
Kinetic and thermodynamic
Mechanism model
tion rate constant of 0.0037–0.0127 min1 and Ea = 17.3 kJ/mol. The reaction is endothermic and non-
Levulinic acid spontaneous with ordered system at transition state. The proposed combined nucleophilic substitution
Ethyl levulinate and Eley-Rideal mechanism is comprised of SN2 steps and heterogeneous catalytic reaction. The results
Esterification provide insights on the reaction for future designing and scaling-up the esterification process.
Ó 2021 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Chemical Reaction Engineering Group (CREG), School

of Chemical and Energy Engineering, Faculty of Engineering, Universiti Teknologi
Malaysia, 81310 Johor Bahru, Johor, Malaysia.
E-mail address: noraishah@cheme.utm.my (N.A.S. Amin).

https://doi.org/10.1016/j.ces.2021.117079
0009-2509/Ó 2021 Elsevier Ltd. All rights reserved.
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

1. Introduction usually determined on the basis of the activated complex theory.

Parameters such as enthalpy, entropy, and Gibbs free energy of
Bio-chemicals and bio-fuels derived from renewable resources activation were calculated using Eyring equation at which the acti-
such as lignocellulosic biomass and other bio-wastes are benefi- vated complex (intermediate structure between reactant and pro-
cial not only for the environment, but also for replenishing duct at transition state) is in quasi-equilibrium state with the
industrial demand due to the depletion of fossil fuel. Value- reactants (Borsato et al., 2014; Ong et al., 2013).
added chemicals from lignocellulosic biomass such as levulinic Studies pertinent to kinetic, thermodynamic and especially
acid has been used for synthesis of other chemicals including mechanism of esterification of levulinic acid with ethanol for
resins and polymers, plasticizers, and herbicides as well as phar- heterogeneous catalytic system are still limited. There are even
maceuticals and flavoring agents, succinic acid, solvents, and fuel limited studies on thermodynamic and kinetic studies for non-
additives (Kang et al., 2018). Ethyl levulinate is one of the levu- catalytic levulinic acid esterification system such as at supercrit-
linic acid-derived chemicals that have useful application which ical condition that was reported recently (Kothe et al., 2020) and
are synthesized via levulinic acid esterification with ethanol cat- also the kinetic and mass transfer studies for ethyl levulinate
alyzed by acid catalyst. Alkyl levulinate is suitable as fuel- synthesis from levulinic acid (Russo et al., 2020). Several
additives for fuel stability and engine performance improvement mechanism-based models have been applied such as nucle-
have been described as the most essential energy-added chemi- ophilic substitution, Langmuir-Hinshelwood and Eley-Rideal.
cals and there are more potential applications of levulinate com- The principle of nucleophilic substitution mechanism is well-
pound (Badgujar et al., 2020). known to adopt the esterification reaction of carboxylic acid
Homogeneous and heterogeneous acid catalysts have been and alcohol. However, the mechanism for the reaction using
used in ethyl levulinate synthesis from levulinic acid. However, the CCIL synthesized catalyst (heterogeneous catalyst) has yet
heterogeneous catalyst circumvents the drawbacks associated to be investigated. Preliminary mathematical modelling of sim-
with homogeneous catalyst. Heterogeneous catalyst can be easily ple second-order kinetics for the esterification reaction has been
removed during the separation step via filtration or centrifuga- developed based on the nucleophilic substitution mechanism for
tion for the next reaction cycle. Various solid acid catalysts for both homogeneous and heterogeneous catalyst (Lilja et al.,
the esterification reaction of levulinic acid such as desilicated 2002). In this work, the mathematical model has been improved
HZSM-5 (Nandiwale et al., 2014), SO-2 4 /ZrO2-membrane (Unlu by considering the adsorption of reactants and products for the
et al., 2016), Zr-containing metal–organic frameworks (MOF) case of heterogeneous catalyst. Explanation about nucleophilic
(Cirujano et al., 2015), Zr-phosphotungstic acid (Pasha et al., substitution in the reaction mechanism between the reactants
2019; Ramli et al., 2017), and sulfonic acid functionalized porous and products is more about the chemical elementary step on
organic polymer (Halder et al., 2020) have conferred high pro- the acid sites that lack physical adsorption step in the heteroge-
duct yield. Acidic carbon-based catalyst demonstrates a great neous catalyst mechanism (Van de Steene et al., 2012). The
potential with high ethyl levulinate yield in esterification reac- heterogeneous kinetics involving adsorption and desorption of
tion. Sulfonated carbon catalysts have been derived with higher one or more reactants and products, respectively, could be
performances during the esterification reaction of levulinic acid explained by other mechanism models. The Langmuir-
with ethanol (Ogino et al., 2018). Carbon catalysts derived from Hinshelwood and Eley-Rideal models are ubiquitous in kinetic
biomass such as sugarcane bagasse (Liu et al., 2019) and oil modelling. The Langmuir-Hinshelwood model basically considers
palm frond (Zainol et al., 2019c) have been applied for ethyl the adsorption of all reactants on the catalyst active sites
levulinate synthesis. In our previous work, ionic liquid and fur- (Delgado et al., 2007) while Eley-Rideal model assumes only
fural based carbon catalyst (CCIL) has exhibited good properties one reactant is adsorbed (Hoo and Abdullah, 2015) before reac-
as an acidic carbon catalyst for the levulinic acid conversion to tion takes place. The possible mechanism steps could be
ethyl levulinate (Zainol et al., 2019b). The CCIL synthesised from explained by verifying the postulated steps with experimental
the reaction between acidic ionic liquid with furfural could pro- data. Previously, the studied reaction of glycerol esterification
duce high acidity and large surface area of solid acid catalyst. with lauric acid has found that the reaction mechanism could
The CCIL as solid acid catalyst can be separated from the reac- be describe by a combination of both a nucleophilic substitution
tion mixture using filtration process and potential for reusability. mechanism and the Langmuir  Hinshelwood model (Hoo and
In other perspective, the synthesis of CCIL gives direction in uti- Abdullah, 2015).
lizing various ionic liquid to prepare various type and properties Herein, an extended investigation pertaining to kinetic, thermo-
of ionic liquid-based carbon catalysts. dynamic and mechanism for the performance of CCIL catalyst in
In kinetic study, pseudo-homogenous (PH) kinetic model is the esterification reaction has been conducted to fill the gap in this
usually applied by assuming a simple power-law. The PH model reaction study. The optimum condition of levulinic acid esterifica-
considers easy access of active sites by the reactants due to a com- tion reaction was determined through parametric studies. The
plete swelling of catalyst and solvent, and does not give a priority kinetic study was investigated based on pseudo-homogeneous
to the adsorption effect (Al-Jendeel et al., 2018). For heterogeneous kinetic model approach and the thermodynamic of activation
catalysis, the reaction rate constant and activation energy were parameters were quantified. In addition, the reaction mechanism
often determined by rapid investigation using simple power-law was studied by employing several mechanism-based models:
model (Hoo and Abdullah, 2015). Similar approach is used to nucleophilic substitution, Langmuir–Hinshelwood and Eley-
determine the kinetic parameter for levulinic acid esterification to Rideal. The postulated kinetic and mechanism models were identi-
ethyl levulinate in the presence of CCIL catalyst. Previously, fied based on high accuracy and good fitting of the experimental
pseudo-homogeneous kinetic model was applied for the esterifica- data. This work will provide the fundamental work of the reaction
tion of levulinic acid in ethanol using desilicated HZSM-5 study for esterification of levulinic acid in ethanol catalyzed by
(Nandiwale et al., 2014), DPTA/DH-ZSM-5 (Nandiwale et al., 2013), CCIL as heterogenous catalyst. The finding of this work from the
and SO-24 /ZrO2-membrane (Unlu et al., 2016). In the thermodynamic kinetic, thermodynamic and mechanism reaction could provide
study, the enthalpy, entropy and Gibbs energy are among the important insights for future reactor design which may involve
parameters to be determined (Teixeira et al., 2021). The the simulation and modelling of the reaction process and designing
thermodynamic parameter studies for the chemical reaction are for reactor material and sizing.

2
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

2. Materials and methods stirring (100 rpm) for 24 h. The mixture was filtered followed by
repeat washing with water and finally oven-dried at 110 °C, 24 h
2.1. Materials before being reused CCIL in the next cycle.
Another important criteria for kinetic study is to ensure the data
The chemicals used in the gel synthesis included 1-Butyl-3- collected are free from diffusional limitation. The internal diffusion
methylimidazolium hydrogen sulfate, [BMIM][HSO4] (Sigma- resistance as control factor was studied by conducting the reaction
Aldrich Co.), furfural, C5H4O2 (Merck), ethanol, C2H6O (95%, QRec), at 100 to 300 mm CCIL particle size while the effect of external dif-
and sulfuric acid, H2SO4 (95–97%, QRec). The other chemicals such fusion using CCIL was evaluated at reaction agitation speeds in
as levulinic acid, C5H8O5 (98%, Merck) and ethanol, C2H6O (99%, range 100 to 300 rpm.
Merck) were the main feedstocks for the esterification reaction The study on kinetic, thermodynamic and mechanism were
and the ethyl levulinate standard, C7H12O3 (99%, Merck) was used investigated at various reaction time (0–180 min) and reaction
for product analysis. All chemicals received were used directly for temperature (78–170 °C) with fixed loading of CCIL and molar ratio
the reaction. (ethanol to levulinic acid). The reproducibility of the data was
observed randomly by selecting and repeating the experiments.
All the product samples were vacuum filtered before conducting
2.2. Synthesis and characterization of CCIL
an analysis.

The method of gel synthesis via sol–gel polycondensation fol-

2.4. Analysis of reaction product
lows the process described elsewhere (Zainol et al., 2019b). The
mixing process was conducted in a flask consisted of 1.25 ratio
Gas chromatograph (Agilent Technology, GC-FID 7820A) was
of furfural to [BMIM][HSO4], 0.125 ratio of distilled water to
operated to identify the ethyl levulinate (EL) peak and determine
[BMIM][HSO4] and 0.5 mL of H2SO4 as the acid catalyst for the
its concentration. The sample is analyzed using HP-5 column
reaction. Ethanol was added to homogenize the mixture. The reac-
(length = 30.0 m; diameter = 320.0 lm) under the following condi-
tion mixture was heated at 100 °C, 1.0 h, and the gel produced after
tions: temperature of injector (270 °C); column ramp temperature
reaction was pre-frozen for 24 h in the refrigerator and further
(80 °C to 170 °C, 13 °C/min and 170 to 300 °C, 40 °C/min); N2 car-
dried at 60 °C, 24 h using freeze dryer to form ionic liquid-
rier gas (1.0 mL/min). High performance liquid chromatography
furfural cryogel (CIL). Finally, carbon cryogel (CCIL) was prepared
(Agilent Technology, HPLC 1260 Infinity) analysis was conducted
after calcined CIL in the furnace at 400 °C, 2 h.
to determine the concentration of levulinic acid using Hi-Plex H
The characterization of CCIL was conducted to evaluate the
column (length = 300 mm; diameter = 7.7 mm) and detector
important chemical and physical properties as catalyst. Nitrogen
(UV, 210 nm). The analysis conditions as follows: column temper-
sorption at 196 °C (Thermo Scientific, Surfer instrument) was
ature (60 °C); auto injection volume (20 lL) and mobile phase
conducted to evaluate the total surface area of CCIL via
(5 mM of H2SO4, 0.6 mL/min). The ethyl levulinate and levulinic
Brunauer-Emmett-Teller (BET) method, mesopores area via
acid standard concentrations were used to determine the respec-
Barrett-Joyner-Halenda (BJH) method and the value difference rep-
tive compounds concentration in the sample with the aid of stan-
resented the micropore area. Temperature programmed desorp-
dard calibration curves. Eqs. (1) and (2) were used to calculate
tion of ammonia (NH3-TPD) was handled by using analyzer from
product yield and feedstock conversion.
Micromeritics (AutoChem II 2920 V4.03) to measure the total acid-  
ity of CCIL. The pre-saturation step of 10% NH3/He (30 cm3/min) CE  V p
Ethyl levulinate yield ðwt:%Þ ¼ =W i  100 ð1Þ
was conducted on the sample for 30 min and the desorption pro- 1000
cess was carried out by heating up to 900 °C (10 °C/min) with
helium flow (30 cm3/min). Thermogravimetric analyzer (Perkin Levulinic acid conversion ðwt:%Þ
Elmer TGA 7) was used to analyze the derivative thermo-   
CL  V p
gravimetric (DTG) of CCIL through heating under nitrogen flow ¼ Wi  =W i  100 ð2Þ
1000
from 30 to 900 °C (10 °C/min). Field Emission Scanning Electron
Microscope (FESEM) by ZEISS Supra 35VP, GEMINI column with where the final concentration of ethyl levulinate and levulinic acid
energy dispersive X-ray (EDX) spectrometry reveal the morphol- (mg/L) are labeled as CE and CL, respectively, volume of product (L)
ogy and the elements in CCIL. The total organic carbon (TOC) anal- as Vp, initial weight levulinic acid (g) as Wi, and molecular weight of
ysis by Shimadzu SSM-5000A was applied to measure the total ethyl levulinate and levulinic acid (g/mol) as ME and ML,
carbon of CCIL. Total carbon of CCIL was determined by pushing respectively.
the sample in a ceramic boat into the 900 °C combustion tube.
As for the inorganic carbon (IC), the phosphoric acid (H3PO4) was 2.5. Mathematical equations for kinetic, thermodynamic and
injected in the sample before it was transferred into a 200 °C com- mechanism studies
bustion IC tube.
The reversible equation of the levulinic acid esterification reac-
tion with ethanol and the main products of ethyl levulinate and
2.3. Synthesis of ethyl levulinate
water is shown in Eq. (3). Based on the chemical reaction of levu-
linic acid esterification with ethanol, the rate of reaction can be
The levulinic acid esterification with ethanol over CCIL was con-
derived as in Eq. (4).
ducted in high pressure stirred reactor (100 mL) from Amar Equip-
ment Pvt. Ltd.. The reaction was investigated by manipulating the
following variables: CCIL weight (5–35 wt%), ethanol/levulinic acid
molar ratio (10–30), and reaction time (1–6 h) and temperature
(78–200 °C). The agitation speed was fixed at 200 rpm. At the ð3Þ
selected condition, the reusability of CCIL for the esterification
reaction was studied to access the catalyst life-time for remaining rA ¼ dC A =dt ¼ k1 C aA C bB  k2 C cC C dD ð4Þ
active. The CCIL was recovered by stirring the used CCIL in etha-
nol–water mixture (1:10) at 300 rpm for 2 h followed by slow
3
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

where the concentration of levulinic acid, ethanol, ethyl levulinate where the rate constant, k (s1), gas constant, R, transmission factor
and water are labeled as CA, CB, CC, and CD, respectively, and their (j = 1), Boltzmann’s constant, kB = 1.381  10-23 J/K, Planck’s con-
respective reaction order as a, b, c, and d, respectively. The kinetic stant, h = 6.626  10-34 Js, absolute temperature, T (K), and the state
constant for the forward and reverse reaction are labeled as k1 of activated complex notation, à, are used to represent the symbol in
and k2, respectively. equation.
Some assumptions were made in developing a kinetic model of Mechanism-based model equations are further derived based
the levulinic acid conversion. Excess ethanol was assumed for reac- on general reaction equation in Eq. (3) for the suggested reaction
tion (i.e. CB  other reactant or product concentration), which can mechanisms as conducted by Hoo and Abdullah (2015). The study
be treated as a constant in the reaction rate expression. Thus, the reaction mechanism models are nucleophilic substitution (NS)
equilibrium reaction could be pushed forward and considered as with the suggested reaction on the acid sites, and Langmuir–Hin-
irreversible reaction. A similar kinetic modelling approach was shelwood (LH) and Eley-Rideal (ER) models for the mechanism
adopted in the study on kinetic parameters of transesterification steps. The reaction mechanism step via LH model considers a
of oil in methanol (Andreo-Martínez et al., 2018). Accordingly, dual-site mechanism while ER model is related to the reaction of
Eqs. (5) and (6) are formed after simplifying the initial rate equa- the adsorbed molecule with a molecule from the bulk gas phase.
tion. The different orders of reaction equations were solved using The two-step NS reaction mechanism involved protonation of
integration method. The kinetic reaction order was determined levulinic acid by the acid site and the nucleophilic attack of ethanol
from best fitting of the experimental data on the model equation. as shown in Eqs. (12) and (13), respectively.

dC A =dt ¼ k1 C aA C bB ð5Þ

a
dC A =dt ¼ kC A ð6Þ

where k ¼ k1 C bB is a constant. ð12Þ

The rate equations are defined (Table 1) based on Eq. (6) in
accordance to the reaction order for pseudo-zero, first, and second k II
+
order kinetic model with a = 0, 1 and 2, respectively. Arrhenius Protonated intermediate + Ethanol Ethyl levulinate + H2O + H
equation (Eq. (7)) is used to determine the activation energy (Ea) CPI k –II
CB CC CD C H+
and exponential factor (A) of the reaction (Al-Rawi et al., 2020)
based on the selected kinetic model. ð13Þ

lnk ¼ lnA  Ea =RT ð7Þ The NS mechanism model equation was derived based on the
following assumptions: (i) ethanol (nucleophile) substitution is
The Eyring equation is used to further describe the reaction rate rate determining for ester and water formation, (ii) fast and rapid
and temperature relationship, as explained by Arrhenius equation for all other diffusion steps while the nucleophilic substitution is
(Andreo-Martínez et al., 2018). Generally, this equation has been likely to occur via SN2 mechanism, and (iii) the backward reaction
used to determine the thermodynamic parameters such as for step 2 is considered very small (i.e. kII  k-II). The initial rate
enthalpy, entropy and Gibbs energy (Teixeira et al., 2021). The the- equation of NS mechanism is expressed as Eqs. (14) and (15):
ory of the activated complex was applied in this equation to eval-
uate the thermodynamic activation parameters of the chemical rI ¼ kI C A C Hþ  kI C PI ð14Þ
reaction as shown in Eqs. (8) and (9). The parameters were calcu-
lated using linearized Eyring equation in Eq. (10) (Andreo-Martínez rII ¼ kII C B C PI  kII C C C D C Hþ ð15Þ
et al., 2018; Borsato et al., 2014). The plot of ln k/T versus 1/T was
Considering quasi-equilibrium theory in step 1 is at equilibrium
used to determine the values of enthalpy and entropy of activation
state (rI = 0), the assumption of very small k-II and the concentra-
(DHà and DSà, respectively), and Eq. (11) was applied to calculate
tion of acid site (CH+) is assumed to be constant due to fixed cata-
the Gibbs energy of activation (DGà).
lyst concentration, the equations are simplified as Eqs. (16) and
DGz ¼ DHz  T DSz ð8Þ (17):

kI C A C Hþ ¼ kI C PI ð16Þ
jkB T DGz
k¼ þ e RT ð9Þ
h
rII ¼ kII C B C PI ð17Þ
z z
k DH k B DS Substituting Eq. (16) in (17) where kI/k-I = KI and kIIKI = k’, the
ln ¼ þ ln þ ð10Þ
T RT h R rate equation expression of NS mechanism, shown in Table 2, is
  similar to second-order kinetics.
kh
DGz ¼ RTln ð11Þ Consider the esterification reaction of levulinic acid (reactant A)
kB T with ethanol (reactant B) as shown in Eq. (3). The LH mechanism
assumed that both reactants are adsorbed on the catalyst surface
for the reaction to occur. Therefore, the mechanism suggests that
the reaction occurs through the (i) external diffusion of both reac-
Table 1 tants from bulk to the external surface of the catalyst, followed by
Reaction rate expression and corresponding linear form of pseudo-homogeneous
kinetic model.
the (ii) internal diffusion to the internal surface of the catalyst and
(iii) adsorption of both reactants on the surface of the catalyst.
Model Reaction rate expression Linear form Then, (iv) the surface reaction between both adsorbed reactants
Pseudo-zero order rA ¼ k X A C AO ¼ kt on the catalyst surface occurred followed by the (v) desorption of
Pseudo-first order r A ¼ kC A lnð1  X A Þ ¼ kt both (C) and (D) products from the catalyst internal surface.
Pseudo-second order r A ¼ kC A 2 X A =ð1  X A Þ ¼ C AO kt
Finally, (vi) the internal diffusion of products (C) and (D) away
Where CAO and XA are the initial concentrations and conversion of levulinic acid, from internal surface of catalyst and (vii) external diffusion of both
and k is the kinetic constant. products from external catalyst surface into the bulk fluid. All
4
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Table 2
Reaction rate expression and corresponding linear form of mechanism-based models.

Model Reactant condition Reaction rate expression Linear form

0
NS Reactant A adsorbed on acid site, nucleophilic substitution of ethanol rA ¼ k C A C B 1 rX A
¼ k0 C AO t
r1 ln rð1X A Þ
LH-1 If both A & B are weakly adsorbed ðK A C A þ K B C B  1Þ r A ¼ k0 C A C B 1 rX A
¼ k0 C AO t
r1 ln rð1X A Þ
where k0 ¼ kT K A K 2B
LH-2 If only A is weakly adsorbed ð1 þ K A C A  K B C B Þ r A ¼ k0 CCAB ðr  1Þ ln 1
1XA þ XA ¼ k0 t
CAO

where k0 ¼ kT KA
KB
LH-3 If only A is strongly adsorbed ð1 þ K B C B  K A C A Þ r A ¼ k0 CCAB ð1  rÞln rX
r
þ X A ¼ CkAOt
0

A
0
where k ¼ kT KK AB
ER-1 At low concentration of adsorbed A ðK A C A  1Þ r A ¼ k0 C A C B 1 rX A
¼ k0 C AO t
r1 ln rð1X A Þ
where k0 ¼ kT K A
ER-2 At high concentration of adsorbed A ðK A C A  1Þ r A ¼ k0 C B where k0 ¼ kT  r12 ln rX
0
r ¼ C AO
A kt

Where CAO and XA are the initial concentrations and conversion of levulinic acid, respectively, r is the molar ratio of levulinic acid to ethanol, and k’ is the kinetic constant.

diffusion, adsorption and desorption steps were assumed relatively 3. Result and discussion
fast compared to the surface reaction of both adsorbed reactants.
Therefore, the mechanism step (iv) would be the rate- 3.1. CCIL production
determining step. Assuming the complete coverage of the catalyst
sites is by reactant (A) and (B) only, all steps are reversible but the The CCIL characteristics are shown in Fig. 1 and the result is
surface reaction in step (iv) is considered for forward reaction due summarized to describe the properties as catalyst for the esterifi-
to very small rate constant of reverse surface reaction as fast des- cation reaction. The CCIL with 560 m2/g of total surface area is con-
orption is taking placed. By considering all the assumptions made, sisting of 467 m2/g and 93 m2/g micropores and mesopores,
the rate equation in LH mechanism is expressed as Eq. (18) for the respectively. Fig. 1(a) attests the pore distribution of CCIL with
irreversible esterification reaction. Some cases of the reaction are pore radius > 1.5 nm relating to the mesoporous structure. High
considered and the simplified rate equation as shown in Table 2. distribution of mesopores is observed for pore radius in the range
of 1.5–3 nm. Even with high micropores, the presence of meso-
pores in the structure will provide good accessibility to facilitate
kT K A K B C A C B adsorption of reactants and enhance the reaction to yield ethyl
rA ¼ ð18Þ
ð1 þ K A C A þ K B C B Þ2 levulinate.
The catalyst active sites is represented by the acidity of CCIL,
evaluated by NH3-TPD graph in Fig. 1(b). The acidic sites of CCIL
where kT is the rate constant for the surface reaction, and KA and KB are contributed by the available sulfonate groups. Based on the
are the adsorption constants of A and B, respectively. desorption peaks area, CCIL registered total acidity of 20.0 mmol/
As for ER mechanism, similar reaction steps assumed in LH g attributed by strong acidic sites. High acidity of sulfonated
mechanism are used but with different steps in (iii) to (v). The carbon-based catalyst derived from palm kernel shell and bamboo
steps are (iii) the adsorption of reactant (A) only on the surface reported strong desorption peak between 500 and 700 °C (Farabi
of the catalyst followed by (iv) the surface reaction between et al., 2019). Luo et al. (2017) stated that desorption peak at above
adsorbed reactant (A) and bulk reactant (B) on the catalyst surface. 550 °C referred to superacidic sites. The NH3 desorption peak in
The mechanism is assumed that the reactant adsorbed is levulinic TPD analysis is related to the sulfonated group decomposition in
acid (A) while ethanol (B) is the reactant from the bulk. Then, (v) is the CCIL structure as observed in DTG curve. The DTG peak exhibits
the desorption of product ethyl levulinate (C) from inside the cat- major losses between 400 and 800 °C owing to the decomposition
alyst surface. Other external or internal diffusion of reactants and component structure of ionic liquid-furfural and sulfonated group
products are similar as described in LH mechanism. Several in CCIL. The decomposition of SO3H group has been reported to be
assumptions are made to derive the rate equation. Rapid steps of between 220 and 550 °C at which SO2 is released (Grishchenko
all diffusion and sorption were involved compared to the surface et al., 2020). Above 550 °C the remaining components of ionic
reaction. Thus, the surface reaction of adsorbed (A) and (B) in bulk liquid-furfural and the very strong sulfonate group decompose.
(step (iv)) is postulated as the rate-limiting step. It was also The surface acidity of CCIL is contributed by the strong Brønsted
assumed that the complete coverage of the catalyst sites by reac- acid sites (Zainol et al., 2019b). The sulfonated group in modified
tant (A) only and all steps are reversible and the surface reaction catalyst promoted the Brønsted acidic sites as reported by previous
is considered for forward reaction due to very small rate constant studies (Gupta and Kantam, 2019; Upare et al., 2019). The sul-
of reverse surface reaction and fast desorption step. The initial rate fonated catalyst provided good catalytic activity for fabricating
equation of the ER mechanism is shown in Eq. (19). Two models of ethyl levulinate via esterification of levulinic acid in ethanol (Li
ER mechanism have been considered for different reactant condi- et al., 2019; Liu et al., 2019).
tions as described in Table 2. The related elements in the CCIL structure including the SO3H
group is confirmed by the EDX analysis (Fig. 1(c)). The high carbon
content at 68.8 wt% is evident while O and S elements at 12.9 wt%
kT K A C A C B and 1.6 wt%, respectively, represent the SO3 functional group. The
rA ¼ ð19Þ
ð1 þ K A C A Þ carbon content is supported using TOC analysis with 67.7% of com-
The LH-1 and ER-1 models have similar second-order rate equa- position reported. The microspherical structures on CCIL is
tions as described by the NS mechanism model. The rate equation observed using FESEM at a higher magnification. Microspheres car-
in NS, LH and ER mechanism models are solved via the integration bon structure has been reported by previous studies using different
method and the models are justified based on experimental data feedstocks (Grishechko et al., 2016; Zainol et al., 2017). The
fitting.
5
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

(a) 0.40 0.020 (b) 0.20 3.0

Weight loss rate (wt.% min )

2.5

Micropore region, r <1.5 nm

0.32 0.016 0.16

Dv/Dr (cm /g/radius)

TCD signal (a.u.)

2.0
Volume (cm /g)

0.012 0.12
3

0.24
1.5

3
0.16 0.008 0.08
1.0
0.004
0.08 0.04
0.5

-1
0.000
0.00 0.00 0.0
1 2 3 4 5 6 7 8 9 10 200 400 600 800
o
Radius, r (nm) Temperature ( C)
(c)
20000 C
N
O
15000
S
Counts

10000

5000

0
0 1 2 3 4 5
Energy (keV)

Fig. 1. (a) Pore distribution, (b) NH3-TPD with DTG graph, and (c) FESEM image (2500X, 5.0 kV) and EDX spectrum of CCIL.

characteristic of CCIL is strong support of its ability to act as an sites, which could slow the reaction activity to produce ethyl
acidic catalyst for the levulinic acid esterification with ethanol. levulinate. The usage of more solvent could reduce the reactant
concentration (Pasha et al., 2019). The rate became limited by mass
3.2. Esterification reaction transfer decreasing the ethyl levulinate yield. Shu et al. (2010) sta-
ted that excess of alcohol might cause flooding of the active sites
CCIL is used as the catalyst in the synthesis of ethyl levulinate in and hinder the protonation of the limiting reactant.
the catalytic esterification of levulinic acid and ethanol. The opti- The effect of temperature is one of the important parameters
mum condition is determined by conducting a parameter study that significantly influences the reaction yield. The yield and con-
and the results of the effect of CCIL loading, ethanol to levulinic version in Fig. 2(c) have increased to 85.8 mol.% and 89.0%, respec-
acid molar ratio, reaction temperature and time on ethyl levulinate tively, when the temperature rises to 170 °C. However, the yield
yield is shown in Fig. 2. Fig. 2(a) depicts the effect of CCIL loading reduces beyond this temperature. At 200 °C, the yield decreases
from 5 wt% to 35 wt% on the reaction. The yield and conversion slightly to 79.7 mol.% even though the conversion is still high at
increase to 62.0 mol% and 64.5%, respectively, when CCIL loading 87.3%. This is possibly due to the ethyl levulinate being converted
increase up to 30 wt%. Small change in yield and conversion is to other products at higher reaction temperature. As reported by
observed when CCIL loading further increases to 35 wt%. The use Zhao et al. (2015), the conversion of ethyl levulinate to c-
35 wt% of loading in the reaction is still doable but it will not help valerolactone and 3-(1-ethoxyethoxy)-2-methylbutanal was
increase the yield and conversion. Thus, 30 wt% of CCIL loading is detected when the reaction temperature increased to 180 °C. The
selected as the optimum loading to obtain high product yield conversion has caused reduction of ethyl levulinate yield inferring
and reactant conversion. that high temperature is not favorable for increasing ethyl levuli-
The reaction study at different molar ratio (ethanol/levulinic nate yield.
acid) as in Fig. 2(b) renders increasing yield and conversion with Another important process parameter is the reaction time. Fig. 2
increasing molar ratio from 10 to 15. The yield and conversion (d) depicts the reaction performance of levulinic acid conversion to
increase up to 62.7 mol.% and 64.2%, respectively. However, the ethyl levulinate with various reaction time. In the first 2 h, the
yield and conversion decrease to 53.1 mol.% and 56%, respectively, reaction of levulinic acid conversion is fast, but slows as the reac-
when the molar ratio is further surged to 30. The reaction activity tion time prolongs. The yield and conversion increase to 85.8 mol.%
is affected due to the large amount of ethanol used. Excess ethanol and 89.0%, respectively, when the reaction is conducted up to 4 h.
is required to move the reaction forward for production of ethyl The reaction is extended up to 6 h, but the yield and conversion
levulinate. However, superfluous ethanol used in the reaction remained in equilibrium and showed only slight changes. Thus,
mixture might dilute the reactant and hinder the catalyst active the following reaction parameters have been selected: 30 wt% of

6
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

100
(a) Conversion (%) (b)
Yield (mol.%)
80
Yield / Conversion

60

40

20

0
5 10 15 20 25 30 35 10 15 20 25 30
Cat. loading (wt.%) Molar ratio (ethanol to LA)
100
(c) (d)
80
Yield / Conversion

60

40

20

0
78 100 120 150 170 200 1 2 3 4 5 6
O
Reaction temperature ( C) Reaction time (h)
Fig. 2. The yield of ethyl levulinate and levulinic acid conversion of esterification reaction catalyzed by CCIL at different (a) catalyst loading (78 °C, 4 h, 20:1), (b) molar ratio of
ethanol to levulinic acid (78 °C, 4 h, 30 wt%), (c) reaction temperature (4 h, 30 wt%, 15:1), and (d) reaction time (170 °C, 30 wt%, 15:1).

CCIL weight, 15 M ratio (ethanol/levulinic acid) at 170 °C and 4 h of high reaction temperature of 250 °C combined with short reaction
reaction temperature and time, respectively, to achieve the opti- time could give the optimum yield (Enumula et al., 2017) as the
mum ethyl levulinate yield. reaction temperature and time are important for determining the
The result of this study was compared with previous esterifica- optimum product yield.
tion results conducted at temperature above 100 °C. Propylsulfonic
acid-modified SBA-15 (Pr-SO3H-SBA-15) is a modified catalyst giv- 3.3. Reusability of carbon cryogel on esterification reaction
ing high yield and conversion of 96.0 mol.% at 117 °C, 2 h in ester-
ification reaction (Melero et al., 2013). Micro/meso HSZM-5 Catalyst reusability is important to be considered where cata-
catalyzed the reaction at 120 °C, 5 h to produce high product yield lyst activity is observed after being reused in a few reaction cycles
and levulinic acid conversion of 95.0 mol.% (Nandiwale and at the selected reaction condition. Fig. 3 exhibits the CCIL reusabil-
Bokade, 2015). Zirconium modified catalysts have been utilized ity testing results. The conversion of levulinic acid falls gradually
for the reaction such as 20-HPW/Zr with high product yield of from 89.0 to 78.2 % as the CCIL is reused up to the fourth run.
97.3 mol.% at 150 °C and 3 h (Ramli et al., 2017). ZrO2/TiO2 The activity of CCIL declines in the fifth run. The active sites deac-
obtained 81.2 mol.% of ethyl levulinate yield at 110 °C for 2 h (Li tivation by reactant or product and leaching from the catalyst pre-
et al., 2014), and ZrO2/SBA-15 conferred 50.0 mol.% of ethyl levuli- cursor possibly are the reasons of the levulinic acid conversion
nate yield with 100% conversion at 250 °C for 1 h (Siva Sankar et al., reduced. It has been reported that the reduction of yield observed
2016). Esterification reaction over WO3-SBA-16 at 250 °C, 1 h during the reusability testing with SO4/SnO catalyst is possibly
yielded 95.0 mol.% of ethyl levulinate (Enumula et al., 2017). caused by catalyst leaching, or active sites deactivation due to
Lignin-furfural carbon cryogel (CCLF) registered 86.5 mol.% and reactant/product adsorption (Fernandes et al., 2012). Similarly,
87.2% yield and conversion, respectively, at reaction temperature deactivation of catalyst sites by the reaction product possibly
of 150 °C (Zainol et al., 2019a). caused low catalyst activity (Casas et al., 2013).
The performance over CCIL is comparable where 85.0 mol.% of
yield and 89.0% of conversion are obtained at 170 °C in 4 h. Gradual 3.4. External and internal diffusion effect
increment of yield and conversion are observed for the reaction
with CCIL with the reaction temperature range 120 °C  170 °C. The internal and external diffusion limitations of CCIL as a
It is very likely high reaction temperature might have encouraged heterogeneous catalyst was first investigated before conducting
the side reaction reducing the ethyl levulinate yield. However, the the kinetic study. The external and internal resistances could result
7
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

100
(a) 180 - 300 m
100 - 180 m
80 80 < 100 m

Conversion (%)
60 60

40 40

20 20

0
0
100
1 2 3 4 5 (b) 100 rpm
200 rpm
Fig. 3. Reusability of CCIL on levulinic acid conversion.
80 300 rpm

Conversion (%)
in a limitation of the mass transfer between the reactants and the 60
CCIL catalyst and could affect the apparent reaction rate. Both dif-
fusion limitations are examined in this study via effect of speed of
agitations and catalyst particle sizes. Related studies to ensure the
40
reaction was free from external and internal resistances have been
previously reported (Nandiwale and Bokade, 2015; Ramli and
20
Amin, 2016). Reactions with different particle size were performed
to determine if internal diffusion resistance controlled the reaction
rate. No effect of particle size on levulinic acid conversion was 0
observed (Fig. 4(a)) inferring the internal resistance towards mass
1 2 3 4
transfer was negligible and should not be considered for the
reaction.
Reaction time (h)
As for external resistance influence, the effect of agitation speed Fig. 4. Effect of (a) CCIL particle sizes and (b) agitation speed on levulinic acid
on reaction system was performed. The increasing of the agitation conversion (78 °C, 30 wt%, 15:1).
speed might increase the contact between the reactant and CCIL,
and thus eliminate the reaction external resistance (Peng et al.,
2010; Ramli and Amin, 2016). The existence of external resistance
in the reaction system could contribute to the controlling factor in
the mass transfer. Only a small change on the conversion was
observed in Fig. 4(b), explaining an insignificant influence on the
reaction by stirring speed. Therefore, the reaction external mass
transfer resistance is negligible. Based on the two tests, the influ-
ence of internal and external diffusions are not the rate-
determining factors in the reaction with CCIL catalyst. The reaction
is not classified in the diffusion-limited regime and the approach of
pseudo-homogeneous model is applicable to represent the rate of
reaction in the kinetic study.

3.5. Kinetic study

The levulinic acid concentration profile for the esterification

reaction using CCIL at 170 °C depict significant concentration
reduction in the first 2 h with the generation of ethyl levulinate
(Fig. 5). The concentration of levulinic acid continuously reduces Fig. 5. Concentration profile of levulinic acid consumption and ethyl levulinate
production at 170 °C (30 wt%, 15:1) in the presence of CCIL.
while ethyl levulinate increases with time. Thus, the reactant con-
sumption rate is inversely proportional to the product formation
rate. The development of pseudo-homogeneous model is based
on power-law model used for the investigation of reaction kinetic decomposition catalyzed by CCIL based on the best fitting of
in determining the reaction rate constant and activation energy. experimental data.
The rapid investigation of reaction kinetic study with heteroge- The reaction rate constants of each kinetic model are
neous catalyst has frequently used simple power law model (Hoo determined based on the respective plots from integrated rate
and Abdullah, 2015). The kinetic model with the assumption of equations. Table 1 shows the reaction rate expression and linear
pseudo-zero, first, and second order are evaluated for levulinic acid form of pseudo-zero, first and second order kinetic models. The
8
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

plot of XA, ln(1  XA) and XA/(1  XA) versus time for pseudo- significant deviation as the linearity obtained has a lower R2 value
zero, first and second order, respectively, are used to determine of 0.7828. Hence, the reaction is non-zero-order for the overall
the rate constant (k) from the slope value of linear plot. reaction and significant at certain condition. As stated by Pogliani
Fig. 6 exhibits the linearity of the experimental data with vari- (2008), a zero or pseudo-zero order reaction kinetics is consistent
ous pseudo-kinetic models. The regression coefficient obtained by within short reaction time. Hence, the reaction studied was not
the linear plot represents the precision of the experimental data accurately represented by pseudo-zero order kinetic model. There-
toward the kinetic model providing a good indication in the model fore, pseud-first and second order approaches are studied further.
selection. The parity plot representing the plotting of experimental Both plots of pseudo-first and second order models have shown
versus predicted data by the model is useful to validate the model good correlation of the experimental data (Fig. 6(c) and (e)). The cor-
selection. Hoo and Abdullah (2015) have verified the selection of relation coefficients of linear plot for both kinetic models are sum-
the kinetic model by observing the data variation and accuracy marized in Table 3. High value of R2 of the linear plots are
from the parity plot. In Fig. 6(a), the plot of pseudo-zero order indi- presented by pseudo-first-order kinetic model at different reaction
cates the reaction is independent of reactant concentration, which temperature. However, the linear plot of pseudo-second order
explained that the reaction rate is not affected by the change of kinetic models give better R2 for experimental data at higher reac-
conversion. The data fitting deviates from the linear plot model tion temperature such as at 150 °C and 170 °C. The parity plots of
at different reaction temperatures and only experimental data at both pseudo-first and pseudo-second order models are used to con-
78 °C fitted-well with the regression coefficient (R2) of 0.9818 firm the selection of the kinetic models. The parity plot of pseudo-
(Table 3). Also, the parity plot of pseudo-zero order model (Fig. 6 first order kinetic model (Fig. 6(d)) has R2 value of 0.9803. High R2
(b)) certifies the actual data disagrees with the predicted data with value is obtained as compared to pseudo-second order model

Fig. 6. Pseudo-homogeneous kinetic model and parity plots of pseudo-zero order (a) and (b), first order (c) and (d), and second order (e) and (f) reaction, respectively, for
levulinic acid (LA) conversion by CCIL. d78 °C, N100 °C, j120 °C, x 150 °C, +170 °C, fitted line (y = x) of experimental versus predicted LA conversion.

9
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Table 3
Kinetic parameters of pseudo-homogeneous model for levulinic acid esterification using CCIL.

Temp. (°C) Pseudo-zero order Pseudo-1st order Pseudo-2nd order Pseudo-1st order
R2 R2 R2 k (min1)
78 0.9818 0.9963 0.9806 0.0037
100 0.9013 0.9768 0.991 0.0053
120 0.8846 0.9862 0.9866 0.0070
150 0.8008 0.9715 0.9857 0.0102
170 0.7918 0.9741 0.9766 0.0127

(Fig. 6(f)) with R2 value of 0.9775. In addition, high accuracy of model (a)
is offered by pseudo-first order in reflecting the actual conversion of
-4.4
levulinic acid due to slope value (1.0199) being very near to unity.
The linearity of kinetic model and parity plots supported the notion
pseudo-first order is dependent on the conversion of levulinic acid.
Accordingly, this finding suggests the approach of using pseudo- -4.8
first order model is pertinent to the conversion of levulinic acid

ln k
experimental data using CCIL.
The reaction rate constant, k for levulinic acid reaction conver-
-5.2
sion to ethyl levulinic using CCIL at different reaction temperature
is determined for selected pseudo-first order model. The values of
rate constant (Table 3) infer the levulinic acid consumption rate
increases with the increasing temperature. As the ethyl levulinate -5.6
is the main esterification product from the levulinic acid conversion,
the levulinic acid consumption rate is proportional to the ethyl
levulinate formation rate, and the increase of reaction temperature 0.0022 0.0024 0.0026 0.0028 0.0030
forms ethyl levulinate more quickly as the rate of reaction increases 1/T (1/K)
as higher temperature enhances mass transfer effect and promotes (b) -11.7
the reaction process (Sahani and Sharma, 2018). Previously,
pseudo-first order kinetic model has been used to find the reaction
rate constant and activation energy for levulinic acid esterification -11.4
reaction (Yadav and Yadav, 2014) and other reaction such as trans-
esterification of oil (Andreo-Martínez et al., 2018; Malani et al.,
2019). The rate constant obtained at different reaction temperature -11.1
ln k/T

is used to determine the activation energy of the reaction.

Based on the Arrhenius plot in Fig. 7(a), the values of Ea and A for
the levulinic acid conversion are listed in Table 4. A larger Ea of -10.8
17.3 kJ/mol is registered for the conversion of levulinic acid to ethyl
levulinate in the presence of CCIL. A high value of Ea also explains
that the esterification reaction using CCIL is in the kinetic regime. -10.5
As stated by previous studies, a reaction is kinetically controlled if
the activation energy exceeds 15 kJ/mol (Bedard et al., 2012; Patel
and Brahmkhatri, 2013). Low activation energy in the range of 10– 0.0022 0.0024 0.0026 0.0028 0.0030
15 kJ/mol refers to diffusion-limited reaction as reported by 1/T (1/K)
(Nandiwale et al., 2014). The activation energy of previous studies
Fig. 7. (a) Arrhenius plot of ln k versus 1/T and (b) plot of ln k/T versus 1/T of
using different types of catalyst in their esterification reaction of pseudo-first order reaction for the levulinic esterification in the presence of CCIL.
levulinic acid are listed in Table 5. Above 15 kJ/mol of activation
energy has been reported for the conversion to ethyl levulinate with
various catalysts and kinetic models. Pseudo-first order derived Table 4
from Langmuir–Hinshelwood–Hougen–Watson model has been Activation energy and exponential factor of levulinic acid decomposition using CCIL.
used for the reaction with mesoporous super acidic zirconia Reaction condition Pseudo-1st order reaction
(UDCaT-5) (Yadav and Yadav, 2014). Pseudo-homogeneous second
R2 Ea (kJ/mol) A
order kinetic model was applied for the reaction with desilicated
78–170 °C 0.9998 17.3 1.399
HZSM-5 (Nandiwale et al., 2014), DPTA/DH-ZSM-5 (Nandiwale
CCIL
et al., 2013), and SO-24 /ZrO2-membrane (Unlu et al., 2016). Pseudo-
first order using pseudo-homogeneous kinetic model was for the
reaction with lignin-furfural carbon cryogel (CCLF) catalyst (Zainol
result showing positive value of DHà indicates the activation pro-
et al., 2019a). Ogino et al. (2018) have synthesized various sul-
cess requiring the heat or energy input from the external source.
fonated carbons for the esterification of levulinic acid in ethanol
This explains the energy required by reaction that brings the reac-
where the activation energy are all around 47 kJ/mol.
tants into the transition state (Nautiyal et al., 2014) or energy level
raising for reactants transformation to their transition state (Ong
3.6. Thermodynamic study et al., 2013) before products formation. The negative value of DSà
obtained indicates the system is less disordered due to more
The thermodynamic parameters are determined from the plot ordered structure of transition state. This explained that the
in Fig. 7(b) and the calculated data are tabulated in Table 6. The degrees of freedom were lost as the activated complex is formed
10
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Table 5 The NS mechanism model gives better fitting of the experimen-

Activation energy (Ea) of the esterification of levulinic acid by using a different type of tal data for the esterification reaction. The derivation of mathemat-
catalyst.
ical equation for NS mechanism is similar to the LH mechanism
Catalyst Solvent Ea (kJ/mol) Ref. model (LH-1) and Eley-Rideal (ER-1), confirming the possibility
UDCaT-5 Ethanol 37.7 (Yadav and Yadav, 2014) of combining the NS mechanism with either LH or ER mechanism.
Desilicated HZSM-5 Ethanol 21.08 (Nandiwale et al., 2014) The combination of both models could represent the overall reac-
DPTA/DH-ZSM-5 Ethanol 29.35 (Nandiwale et al., 2013) tion mechanism from the bulk to molecular interaction between
SO-2
4 /ZrO2-membrane Ethanol 14.61 (Unlu et al., 2016)
Sulfonated carbon Ethanol 47.0 (Ogino et al., 2018)
the reactant molecules and acid sites of the catalyst (Hoo and
Carbon Cryogel (CCLF) Ethanol 20.2 (Zainol et al., 2019a) Abdullah, 2015). The reaction mechanism of the reactants on the
CCIL Ethanol 17.3 This study acid site of catalyst and the chemical transition to end products
could be explained by NS mechanism (Fig. 9) while the mechanism
steps of adsorption, surface reaction and desorption on the hetero-
where the reacting species joined themselves and have a more geneous acid catalyst could be described through LH or ER mecha-
ordered state (Borsato et al., 2014; Ong et al., 2013). Similarly, nism (Fig. 10). The combination of NS with ER mechanism (ER-1)
the negative entropy has referred to the better degree of ordered model via second order reaction could be a better explanation for
alignment of transition state chemical compared to the ground the levulinic acid and ethanol esterification reaction on CCIL. High
state reactants (Nautiyal et al., 2014). The other thermodynamic accuracy is also shown by ER-2 model based on the parity plot,
parameter shows the positive DGà which indicates the endergonic even though the R2 value of data distribution is not high. The ER-
and non-spontaneous reaction (Nautiyal et al., 2014; Ong et al., 2 model considers high concentration of levulinic acid for the
2013), which contributes to a high degree of energy in the transi- model equation, but it is not suitable to represent the experimental
tion state compared with the ground (Ong et al., 2013). Similar data since high ethanol concentration is used for the reaction.
result on the thermodynamic parameter with negative and positive Therefore, the experimental data plot is not well represented by
values of DSà and DGà, respectively, has been reported by Badgujar the ER-2 model. The ER-1 model is more suitable to represent
and Bhanage (2015). the experimental data as this model equation considers low con-
centration of levulinic acid in the reaction system. Similar studies
3.7. Mechanism study have considered ER mechanism for esterification with the assump-
tion that propanoic acid was adsorbed on the catalyst and alcohol
Mathematical equations on the mechanism-based model are from the bulk fluid (Dakshinamurty et al., 1984; Liu and Tan, 2001).
developed to study the mechanism of the reaction, which is not In another study, the ER mechanism has been reported where the
described by pseudo-homogeneous kinetic model via power-law rate-limiting step is the surface reaction of adsorbed oleic acid
model. As stated by Hoo and Abdullah (2015), the reaction mech- with methanol from the liquid phase (Chaemchuen et al., 2020).
anism could be explained by the mechanism-based model due to As for LH mechanism, all models have validated that levulinic
limitation of the simple power-law model to describe it. Based acid is weakly adsorbed on CCIL surface including the LH-3 model.
on the derivation of NS, LH and ER mechanism model equations, Levulinic acid is weakly adsorbed on CCIL surface with the ionic
the integrated rate equation is summarized in Table 2 with all bond becoming weak as the reaction temperature increases. LH-1
the assumptions. model, assuming that both levulinic acid and ethanol are weakly
In Fig. 8(a), the linearization of the experimental data for the NS adsorbed on catalyst surface, has similar model equation with NS
second order mechanism model at different reaction temperatures mechanism and presents a better fitting on the experimental data.
versus time gives good plot fitting with high R2 value (Table 7). The Since the LH-2 mechanism model also gives high accuracy and data
high value of R2 infers that this range of reaction temperature is fitting of the experimental data, both LH-1 and LH2 models seem to
possible for the esterification of levulinic acid to produce ethyl conflict with each other. Both LH-1 and LH-2 models infer different
levulinate catalyzed by CCIL. The parity plot of the NS mechanism conditions of ethanol adsorption at the catalyst surface, weakly
model (Fig. 8(b)) also reveals high accuracy of the actual levulinic and strongly adsorbed, respectively. Thus, the combination of NS
acid conversion with the predicted conversion since the slope with LH mechanism is not considered to represent the levulinic
(1.018) is very close to unity. The linearity of the parity plot has acid reaction with ethanol using CCIL.
the highest R2 value as compared to the other models. The linear For the combination of both NS and ER-1 mechanism models,
plot of LH model (LH-2) with assumption of weak adsorption of the assumptions from both models could be considered. The over-
levulinic acid (Fig. 8(c)) also exhibits good fit with high R2 value. all steps involved in the reaction are suggested to concur with the
Meanwhile, the linearity of the LH-2 model parity plot in Fig. 8 proposed ER-1 mechanism in Fig. 10(a). The description of the
(d) confers slope and R2 values of 1.0225 and 0.9772, respectively. steps involved has been explained earlier for all the diffusion steps.
In contrast, large deviation of the experimental data is registered The adsorption of levulinic acid on the acid site at CCIL surface
by LH-3 model where the R2 value reduces to 0.4950 at high reac- (step (iii)) will activate levulinic acid to form the protonated levu-
tion temperature (Fig. 8(e)). The deviation of linear plot is sup- linic acid intermediate as described by the postulated NS mecha-
ported by the parity plot (Fig. 8(f)) representing low accuracy nism in Fig. 9. The surface reaction occurs between adsorbed
and high deviation with slope of 1.1004 and R2 value of 0.4585. levulinic acid and ethanol from bulk for formation of ethyl levuli-
As for the ER-2 model (Fig. 8(g)), the linear plot deviation is large nate and water (step (iv)). The protonated intermediate increases
especially at high reaction temperature with the R2 parity the electrophilicity of the carbonyl group of levulinic acid for
plot = 0.8392. nucleophilic attack by ethanol from bulk (Khder et al., 2008;

Table 6
Thermodynamic parameters of levulinic acid decomposition using CCIL.

Reaction condition 78–170 °C Thermodynamic parameters

R2 DHà (kJ mol1) DSà (J mol1 K1) DGà (kJ mol1)
Pseudo-1st order reaction 0.9998 14.0 252.8 114.3

11
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Mechanism-based Model Parity Plot

2 1
(a) (b) y = 1.018x

Experimental LA conversion
ln [(r - X A) / r (1 - X A)] 1.6 0.8 R2 = 0.9824

1.2 0.6
NS

0.8 0.4

0.4 0.2

0 0
0 20 40 60 80 100 120 140 160 180 0 0.2 0.4 0.6 0.8 1
Time (min) Predicted LA conversion
30 1
(c) (d)

Experimental LA conversion
(r - 1) ln [1/(1 - X A)] + X A

24 y = 1.0225x
0.8
R2 = 0.9772
18 0.6
LH-2

12 0.4

6 0.2

0 0
0 20 40 60 80 100 120 140 160 180 0 0.2 0.4 0.6 0.8 1
Time (min) Predicted LA conversion
0.05 1
(e) (f) y = 1.1004x
Experimental LA conversion
(1 - r) ln [r/(r - X A)] + X A

0.04 0.8 R2 = 0.4585

0.03 0.6
LH-3

0.02 0.4

0.01 0.2

0 0
0 20 40 60 80 100 120 140 160 180 0 0.2 0.4 0.6 0.8 1
Time (min) Predicted LA conversion
4E-4 1
(g) (h)
Experimental LA conversion

y = 0.9968x
0.8 R2 = 0.8392
-1/r2 ln [(r-X A)/r]

3E-4
0.6
ER-2

2E-4
0.4

1E-4
0.2

0E+0 0
0 20 40 60 80 100 120 140 160 180 0 0.2 0.4 0.6 0.8 1
Time (min) Predicted LA conversion

Fig. 8. Mechanism-based models and parity plots of levulinic acid (LA) conversion in the presence of CCIL via nucleophilic substitution (a) and (b), Langmuir  Hinshelwood
for weak adsorption (LH-2) (c) and (d) and strong adsorption (LH-3) (e) and (f) of levulinic acid, and Eley-Rideal (ER-2) at high concentration of levulinic acid (g) and (h).
d78 °C, N100 °C, j120 °C, x 150 °C, +170 °C, fitted line (y = x) of experimental versus predicted LA conversion.

Table 7
R2 value of data plot using different mechanism-based models for esterification of levulinic acid using CCIL.

Temp. (°C) R2 value

Nucleophilic substitution (NS) Langmuir  Hinshelwood (LH), LH-2 Langmuir  Hinshelwood (LH), LH-3 Eley-Rideal (ER), ER-2
78 0.9962 0.9963 0.9395 0.9833
100 0.9787 0.9747 0.7778 0.9062
120 0.9880 0.9842 0.6968 0.8911
150 0.9745 0.9683 0.5275 0.8100
170 0.9768 0.9711 0.4950 0.8015

12
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Ethanol H
O H H
O O
O O
O +
+ OH
OH + O
H
+ O
O H
H
Levulinic acid Acid site

O
O H H
+
O
H2O O O
O + O O
Ethyl Levulinate
O O
+
H+ H H

Fig. 9. Proposed SN2 nucleophilic substitution mechanism of levulinic acid with ethanol on CCIL for producing ethyl levulinate.

Fig. 10. Proposed (a) Eley-Rideal and (b) Langmuir–Hinshelwood mechanism models for the reaction of levulinic acid with ethanol on CCIL. (i, vii) external diffusion, (ii, vi)
internal diffusion, (iii) adsorption, (iv) surface reaction, and (v) desorption.

13
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Pasquale et al., 2012). Deprotonation leads to levulinate ester and Declaration of Competing Interest
water formation. Ethyl levulinate is desorbed from the internal
surface of the CCIL (step (v)) with the regeneration of vacant acid The authors declare that they have no known competing finan-
site. All diffusion steps in the reaction mechanism are assumed cial interests or personal relationships that could have appeared
to occur very fast. to influence the work reported in this paper.
The ER mechanism models first consider levulinic acid to be
adsorbed on CCIL. If ethanol is adsorbed on the catalyst, the ER
mechanism model will give a second-order rate equation Acknowledgement
(first-order in levulinic acid and ethanol, respectively) at low con-
centration of ethanol, while the first-order reaction rate with The authors would like to express their sincere gratitude to the
respect to levulinic acid component is obtained at high concentra- Ministry of Higher Education (MOHE), Malaysia and Universiti
tion of ethanol. Both reaction rates give similar model equations as Teknologi Malaysia (UTM) for the financial support via Research
presented by the NS mechanism and pseudo-first order models, University Grant (Q.J130000.2546.19H95 and Q.J130000.2651.16J62)
respectively, and have shown good data fitting. However, the dif- and Professional Development Research University (PDRU) Post-
ference in data fitting between low and high ethanol concentration Doctoral Fellowship (R.J130000.7113.04E78).
in the ER model could not be clearly observed for model selection
to represent the data. Thus, the ER mechanism model equation
with ethanol adsorbed is not considered in this work. Previous References
studies have reported ER model with the reaction between
Al-Jendeel, H.A., Noori, W.O., Al-Jubouri, S., Alhassani, M.H., 2018. Esterification
adsorbed 1-propanol (Sharma et al., 2016), methanol (Lamba reaction kinetics using ion exchange resin catalyst by pseudo-homogenous and
et al., 2018), n-butanol (Bhusari et al., 2020), and ethanol Eley-Ridel models. Int. J. Eng. 31, 1172–1179.
(Merchant et al., 2013) with unadsorbed respective carboxylic Al-Rawi, U.A., Sher, F., Hazafa, A., Rasheed, T., Al-Shara, N.K., Lima, E.C., Shanshool, J.,
2020. Catalytic activity of Pt loaded zeolites for hydroisomerization of n-hexane
acids for the reaction. using supercritical CO2. Ind. Eng. Chem. Res. 59, 22092–22106.
For this work, more experimental data is required to confirm Andreo-Martínez, P., García-Martínez, N., del Mar Durán-del-Amor, M., Quesada-
the adsorption of the constituent species (reactants and products) Medina, J., 2018. Advances on kinetics and thermodynamics of non-catalytic
supercritical methanol transesterification of some vegetable oils to biodiesel.
by considering the adsorption constant to describe the mechanism Energy Convers. Manage. 173, 187–196.
in detail. If the situation of ethanol adsorbed for ER mechanism is Badgujar, K.C., Badgujar, V.C., Bhanage, B.M., 2020. A review on catalytic synthesis of
the best to represent the reaction, the NS mechanism (Fig. 9) is not energy rich fuel additive levulinate compounds from biomass derived levulinic
acid. Fuel Process. Technol. 197, 106213.
appropriate to represent the mechanism at surface reaction for ER
Badgujar, K.C., Bhanage, B.M., 2015. Thermo-chemical energy assessment for
mechanism. Thus, another plausible mechanism and model dis- production of energy-rich fuel additive compounds by using levulinic acid
crimination technique should be recommended for the reaction and immobilized lipase. Fuel Process. Technol. 138, 139–146.
on the catalyst surface between adsorbed ethanol and levulinic Bedard, J., Chiang, H., Bhan, A., 2012. Kinetics and mechanism of acetic acid
esterification with ethanol on zeolites. J. Catal. 290, 210–219.
acid from the bulk fluid, which contradicts with the NS Bhusari, A.A., Mazumdar, B., Rathod, A.P., Khonde, R.D., 2020. Kinetics of catalyzed
mechanism. esterification of acetic acid with n-butanol using carbonized agro-waste. Int. J.
Chem. Kinet. 52, 450–462.
Borsato, D., Galvan, D., Pereira, J.L., Orives, J.R., Angilelli, K.G., Coppo, R.L., 2014.
Kinetic and thermodynamic parameters of biodiesel oxidation with synthetic
4. Conclusion antioxidants: simplex centroid mixture design. J. Braz. Chem. Soc. 25, 1984–
1992.
Casas, A., Ramos, M.J., Rodríguez, J.F., Pérez, Á., 2013. Tin compounds as Lewis acid
CCIL catalyst obtained renders good reaction performance on catalysts for esterification and transesterification of acid vegetable oils. Fuel
levulinic acid esterification to ethyl levulinate. The parametric Process. Technol. 106, 321–325.
studies of the esterification reaction at the optimum condition con- Chaemchuen, S., Heynderickx, P.M., Verpoort, F., 2020. Kinetic modeling of oleic
acid esterification with UiO-66: from intrinsic experimental data to kinetics via
fers high ethyl levulinate yield and levulinic acid conversion of elementary reaction steps. Chem. Eng. J. 394, 124816.
85.0 mol.% and 89.0%, respectively. The pseudo-first order reaction Cirujano, F.G., Corma, A., 2015. Llabrés i Xamena, F.X., Conversion of levulinic acid
kinetic model with 17.3 kJ/mol of activation energy also reveals into chemicals: Synthesis of biomass derived levulinate esters over Zr-
containing MOFs. Chem. Eng. Sci. 124, 52–60.
that the reaction conversion to ethyl levulinate is an endergonic,
Dakshinamurty, P., Ramarao, M., Ramachandramurty, C.V., 1984. Kinetics of
not spontaneous with more ordered transition state. The investiga- catalytic esterification of propan-1-ol with propanoic acid using cation-
tion of the heterogeneous reaction mechanism provides important exchange resin. J. Chem. Technol. Biotechnol. Chem. Technol. 34, 257–261.
insights for the esterification reaction of levulinic acid to ethyl Delgado, P., Sanz, M.T., Beltrán, S., 2007. Kinetic study for esterification of lactic acid
with ethanol and hydrolysis of ethyl lactate using an ion-exchange resin
levulinate. The plausible mechanism for esterification of levulinic catalyst. Chem. Eng. J. 126, 111–118.
acid and ethanol is postulated to conform to the combination of Enumula, S.S., Gurram, V.R.B., Chada, R.R., Burri, D.R., Kamaraju, S.R.R., 2017. Clean
nucleophilic substitution and Eley-Rideal mechanisms. The results synthesis of alkyl levulinates from levulinic acid over one pot synthesized WO3-
SBA-16 catalyst. J. Mol. Catal. A: Chem. 426, 30–38.
of this study are pertinent for future scaling-up and designing pro- Farabi, M.S.A., Ibrahim, M.L., Rashid, U., Taufiq-Yap, Y.H., 2019. Esterification of
cess of ethyl levulinate production using CCIL involving simulation, palm fatty acid distillate using sulfonated carbon-based catalyst derived from
modelling and reactor design. In addition, the results from this palm kernel shell and bamboo. Energy Convers. Manage. 181, 562–570.
Fernandes, D., Rocha, A., Mai, E., Mota, C.J., da Silva, V.T., 2012. Levulinic acid
study could also be used for isolation of ethyl levulinate to obtain esterification with ethanol to ethyl levulinate production over solid acid
high product purity yield. catalysts. Appl. Catal. A 425, 199–204.
Grishchenko, L.M., Diyuk, V.E., Mariychuk, R.T., Vakaliuk, A.V., Radkevich, V.Z.,
Khaminets, S.G., Mischanchuk, O.V., Lisnyak, V.V., 2020. Surface reactivity of
nanoporous carbons: preparation and physicochemical characterization of
CRediT authorship contribution statement sulfonated activated carbon fibers. Appl. Nanosci. 10, 2923–2939.
Grishechko, L., Amaral-Labat, G., Fierro, V., Szczurek, A., Kuznetsov, B., Celzard, A.,
2016. Biosourced, highly porous, carbon xerogel microspheres. RSC Adv. 6,
Muzakkir Mohammad Zainol: Conceptualization, Writing –
65698–65708.
original draft, Methodology, Investigation, Data curation, Visual- Gupta, S.S.R., Kantam, M.L., 2019. Catalytic conversion of furfuryl alcohol or
ization. Nor Aishah Saidina Amin: Conceptualization, Writing – levulinic acid into alkyl levulinates using a sulfonic acid-functionalized
original draft, Supervision, Resources, Funding acquisition. Mohd hafnium-based MOF. Catal. Commun. 124, 62–66.
Halder, M., Bhanja, P., Islam, M.M., Chatterjee, S., Khan, A., Bhaumik, A., Islam, S.M.,
Asmadi: Conceptualization, Writing – review & editing, Visualiza- 2020. Porous organic polymer as an efficient organocatalyst for the synthesis of
tion, Funding acquisition. biofuel ethyl levulinate. Mol. Catal. 494, 111119.

14
M.M. Zainol, M. Asmadi and N.A.S. Amin Chemical Engineering Science 247 (2022) 117079

Hoo, P., Abdullah, A.Z., 2015. Kinetics modeling and mechanism study for selective Pasquale, G., Vázquez, P., Romanelli, G., Baronetti, G., 2012. Catalytic upgrading of
esterification of glycerol with lauric acid using 12-tungstophosphoric acid post- levulinic acid to ethyl levulinate using reusable silica-included Wells-Dawson
impregnated SBA-15. Ind. Eng. Chem. Res. 54, 7852–7858. heteropolyacid as catalyst. Catal. Commun. 18, 115–120.
Kang, S., Fu, J., Zhang, G., 2018. From lignocellulosic biomass to levulinic acid: A Patel, A., Brahmkhatri, V., 2013. Kinetic study of oleic acid esterification over 12-
review on acid-catalyzed hydrolysis. Renew. Sustain. Energy Rev. 94, 340–362. tungstophosphoric acid catalyst anchored to different mesoporous silica
Khder, A., El-Sharkawy, E., El-Hakam, S., Ahmed, A., 2008. Surface characterization supports. Fuel Process. Technol. 113, 141–149.
and catalytic activity of sulfated tin oxide catalyst. Catal. Commun. 9, 769–777. Peng, L., Lin, L., Zhang, J., Zhuang, J., Zhang, B., Gong, Y., 2010. Catalytic conversion of
Kothe, V., Melfi, D.T., dos Santos, K.C., Corazza, M.L., Ramos, L.P., 2020. cellulose to levulinic acid by metal chlorides. Molecules 15, 5258–5272.
Thermodynamic analysis, experimental and kinetic modeling of levulinic acid Pogliani, L., 2008. Pseudo-zero-order reactions. React. Kinet. Catal. Lett. 93, 187–
esterification with ethanol at supercritical conditions. Fuel 260, 116376. 191.
Lamba, R., Kumar, S., Sarkar, S., 2018. Esterification of decanoic acid with methanol Ramli, N.A.S., Amin, N.A.S., 2016. Kinetic study of glucose conversion to levulinic
using Amberlyst 15: Reaction kinetics. Chem. Eng. Commun. 205, 281–294. acid over Fe/HY zeolite catalyst. Chem. Eng. J. 283, 150–159.
Li, N., Wang, Q., Ullah, S., Zheng, X.-C., Peng, Z.-K., Zheng, G.-P., 2019. Esterification Ramli, N.A.S., Sivasubramaniam, D., Amin, N.A.S., 2017. Esterification of levulinic
of levulinic acid in the production of fuel additives catalyzed by porous acid using ZrO2-supported phosphotungstic acid catalyst for ethyl levulinate
sulfonated carbon derived from pine needle. Catal. Commun. 129, 105755. production. Bioenergy Res. 10, 1105–1116.
Li, Z., Liu, Y., Kwapinski, W., Leahy, J.J., 2014. ZrO2-modified TiO2 nanorod Russo, V., Rossano, C., Salucci, E., Tesser, R., Salmi, T., Di Serio, M., 2020. Intraparticle
composite: Hydrothermal synthesis, characterization and application in diffusion model to determine the intrinsic kinetics of ethyl levulinate synthesis
esterification of organic acid. Mater. Chem. Phys. 145, 82–89. promoted by Amberlyst-15. Chem. Eng. Sci. 228, 115974.
Lilja, J., Murzin, D.Y., Salmi, T., Aumo, J., Mäki-Arvela, P., Sundell, M., 2002. Sahani, S., Sharma, Y.C., 2018. Economically viable production of biodiesel using a
Esterification of different acids over heterogeneous and homogeneous catalysts novel heterogeneous catalyst: Kinetic and thermodynamic investigations.
and correlation with the Taft equation. J. Mol. Catal. A: Chem. 182–183, 555– Energy Convers. Manage. 171, 969–983.
563. Sharma, M., Toor, A.P., Wanchoo, R.K., 2016. Esterification of pentanoic acid with 1-
Liu, C., Zhang, K., Liu, Y., Wu, S., 2019. Esterification of Levulinic Acid into Ethyl propanol by sulfonated cation exchange resin: Experimental and kinetic
Levulinate Catalyzed by Sulfonated Bagasse-carbonized Solid Acid. studies. Chem. Eng. Commun. 203, 801–808.
BioResources 14, 2186–2196. Shu, Q., Gao, J., Nawaz, Z., Liao, Y., Wang, D., Wang, J., 2010. Synthesis of biodiesel
Liu, W.-T., Tan, C.-S., 2001. Liquid-phase esterification of propionic acid with n- from waste vegetable oil with large amounts of free fatty acids using a carbon-
butanol. Ind. Eng. Chem. Res. 40, 3281–3286. based solid acid catalyst. Appl. Energy 87, 2589–2596.
Luo, Y., Mei, Z., Liu, N., Wang, H., Han, C., He, S., 2017. Synthesis of mesoporous Siva Sankar, E., Mohan, V., Suresh, M., Saidulu, G., David Raju, B., Rama Rao, K.S.,
sulfated zirconia nanoparticles with high surface area and their applies for 2016. Vapor phase esterification of levulinic acid over ZrO2/SBA-15 catalyst.
biodiesel production as effective catalysts. Catal. Today 298, 99–108. Catal. Commun. 75, 1–5.
Malani, R.S., Umriwad, S.B., Kumar, K., Goyal, A., Moholkar, V.S., 2019. Ultrasound– Teixeira, R.A., Lima, E.C., Benetti, A.D., Thue, P.S., Cunha, M.R., Cimirro, N.F., Sher, F.,
assisted enzymatic biodiesel production using blended feedstock of non–edible Dehghani, M.H., dos Reis, G.S., Dotto, G.L., 2021. Preparation of hybrids of wood
oils: Kinetic analysis. Energy Convers. Manage. 188, 142–150. sawdust with 3-aminopropyl-triethoxysilane. Application as an adsorbent to
Melero, J.A., Morales, G., Iglesias, J., Paniagua, M., Hernández, B., Penedo, S., 2013. remove Reactive Blue 4 dye from wastewater effluents. J. Taiwan Inst. Chem.
Efficient conversion of levulinic acid into alkyl levulinates catalyzed by sulfonic Eng. 125, 141–152.
mesostructured silicas. Appl. Catal. A 466, 116–122. Unlu, D., Ilgen, O., Hilmioglu, N.D., 2016. Biodiesel additive ethyl levulinate
Merchant, S.Q., Almohammad, K.A., Al Bassam, A.A., Ali, S.H., 2013. Biofuels and synthesis by catalytic membrane: SO2 4 /ZrO2 loaded hydroxyethyl cellulose.
additives: Comparative kinetic study of Amberlite IR 120-catalyzed Chem. Eng. J. 302, 260–268.
esterification of ethanol with acetic, propanoic and pentanoic acids to Upare, P.P., Hong, D.-Y., Kwak, J., Lee, M., Chitale, S.K., Chang, J.-S., Hwang, D.W.,
produce eco-ethyl-esters. Fuel 111, 140–147. Hwang, Y.K., 2019. Direct chemical conversion of xylan into furfural over
Nandiwale, K.Y., Bokade, V.V., 2015. Environmentally benign catalytic process for sulfonated graphene oxide. Catal. Today 324, 66–72.
esterification of renewable levulinic acid to various alkyl levulinates biodiesel. Van de Steene, E., De Clercq, J., Thybaut, J.W., 2012. Adsorption and reaction in the
Environ. Prog. Sustainable Energy 34, 795–801. transesterification of ethyl acetate with methanol on Lewatit K1221. J. Mol.
Nandiwale, K.Y., Niphadkar, P.S., Deshpande, S.S., Bokade, V.V., 2014. Esterification Catal. A: Chem. 359, 57–68.
of renewable levulinic acid to ethyl levulinate biodiesel catalyzed by highly Yadav, G.D., Yadav, A.R., 2014. Synthesis of ethyl levulinate as fuel additives using
active and reusable desilicated H-ZSM-5. J. Chem. Technol. Biotechnol. 89, heterogeneous solid superacidic catalysts: Efficacy and kinetic modeling. Chem.
1507–1515. Eng. J. 243, 556–563.
Nandiwale, K.Y., Sonar, S.K., Niphadkar, P.S., Joshi, P.N., Deshpande, S.S., Patil, V.S., Zainol, M.M., Amin, N.A.S., Asmadi, M., 2017. Effects of thermal treatment on carbon
Bokade, V.V., 2013. Catalytic upgrading of renewable levulinic acid to ethyl cryogel preparation for catalytic esterification of levulinic acid to ethyl
levulinate biodiesel using dodecatungstophosphoric acid supported on levulinate. Fuel Process. Technol. 167, 431–441.
desilicated H-ZSM-5 as catalyst. Appl. Catal. A 460, 90–98. Zainol, M.M., Amin, N.A.S., Asmadi, M., 2019a. Kinetics and thermodynamic analysis
Nautiyal, P., Subramanian, K., Dastidar, M., 2014. Kinetic and thermodynamic of levulinic acid esterification using lignin-furfural carbon cryogel catalyst.
studies on biodiesel production from Spirulina platensis algae biomass using Renew. Energy 130, 547–557.
single stage extraction–transesterification process. Fuel 135, 228–234. Zainol, M.M., Amin, N.A.S., Asmadi, M., 2019b. Synthesis and characterization of
Ogino, I., Suzuki, Y., Mukai, S.R., 2018. Esterification of levulinic acid with ethanol porous microspherical ionic liquid carbon cryogel catalyst for ethyl levulinate
catalyzed by sulfonated carbon catalysts: Promotional effects of additional production. Diam. Relat. Mater. 95, 154–165.
functional groups. Catal. Today 314, 62–69. Zainol, M.M., Amin, N.A.S., Asmadi, M., Ramli, N.A.S., 2019c. Esterification of
Ong, L., Kurniawan, A., Suwandi, A., Lin, C., Zhao, X., Ismadji, S., 2013. levulinic acid to ethyl levulinate using liquefied oil palm frond-based carbon
Transesterification of leather tanning waste to biodiesel at supercritical cryogel catalyst. Bioenergy Res. 12, 1–11.
condition: Kinetics and thermodynamics studies. J. Supercritical Fluids 75, Zhao, S., Xu, G., Chang, J., Chang, C., Bai, J., Fang, S., Liu, Z., 2015. Direct production of
11–20. ethyl levulinate from carbohydrates catalyzed by H-ZSM-5 supported
Pasha, N., Lingaiah, N., Shiva, R., 2019. Zirconium exchanged phosphotungstic acid phosphotungstic acid. BioResources 10, 2223–2234.
catalysts for esterification of levulinic acid to ethyl levulinate. Catal. Lett. 149,
2500–2507.

15