lable at ScienceDirect

Renewable Energy 130 (2019) 547e557

Contents lists avai

Renewable Energy

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

Kinetics and thermodynamic analysis of levulinic acid esterificationusing lignin-furfural carbon cryogel catalyst

Muzakkir Mohammad Zainol, Nor Aishah Saidina Amin*, Mohd AsmadiChemical Reaction Engineering Group (CREG), Faculty of Chemical and Energy Engineering, Universiti Teknologi Malaysia, 81310, UTM Johor Bahru, Johor,Malaysia

a r t i c l e i n f o

Article history:Received 3 February 2018Received in revised form18 April 2018Accepted 20 June 2018Available online 26 June 2018

Keywords:Ethyl levulinateEsterificationKinetic modelThermodynamicLevulinic acidCarbon cryogel

* Corresponding author. Chemical Reaction EngineeChemical and Energy Engineering, Universiti TeknologDarul Takzim, Malaysia.

E-mail address: noraishah@cheme.utm.my (N.A.S.

https://doi.org/10.1016/j.renene.2018.06.0850960-1481/© 2018 Elsevier Ltd. All rights reserved.

a b s t r a c t

The synthesis of ethyl levulinate, a fuel additive, by catalytic esterification of levulinic acid with ethanolover carbon cryogel has been investigated. The carbon cryogel catalyst, coupled with a large surface areaand strong acidity, has been identified as an effective carbon-based catalyst for obtaining high ethyllevulinate yield of 86.5mol%. The pseudo-homogeneous kinetic model is adopted to evaluate thedifferent reaction orders. The first-order pseudo-homogeneous model is considered most suitable(R2> 0.98) while the selection of kinetic model is also clarified and supported by the linearity of theparity plot. The activation energy of the esterification reaction is estimated to be 20.2 kJ/mol. Based onthe thermodynamic activation parameters, the reaction is classified as endergonic and more ordered. Theresults from this study could provide valuable information for reactor modeling and simulation purposesin the future.

© 2018 Elsevier Ltd. All rights reserved.

1. Introduction

The derivatives of bio-fuel from renewable sources are appro-priate for applications as transportation fuels and bio-basedchemicals. Biodiesel, derived from many sources including wasteoil, has been credited as a promising bio-fuel [1e3]. In addition,many processes have been conducted on lignocellulosic biomass toproduce bio-fuels and bio-chemicals via thermochemical conver-sion [4]. The catalytic hydrolysis of biomass has been conducted toderive chemicals such as reducing sugar and 5-hydroxymethylfurfural (HMF) for further conversion to levulinic acid [5]. Levu-linic acid is recognized as a building block and chemical platform[6]. Previous studies have reported the direct conversion of ligno-cellulosic biomass to levulinic acid using homogeneous and het-erogeneous catalysts [5,7e10].

As a platform chemical, levulinic acid can be used to synthesizelevulinate esters via esterification reactionwith alcohol. Esters suchas methyl, ethyl and n-butyl levulinate are produced using meth-anol, ethanol, and n-butanol, respectively as co-reactants [11e14].

ring Group (CREG), Faculty ofi Malaysia, 81310, UTM, Johor

Amin).

The production of alkyl levulinate from levulinic acid as biomass-derived chemical has attracted great attention in the green en-ergy sector. The reaction products could be separated and purifiedfrom the feedstock using separation methods such as reactivedistillation [15] and membrane technology [16,17]. Levulinate es-ters have extensive application mainly as additives in gasoline andbiodiesel, in g-valerolactone (GVL) production, and also in flavoringand fragrance industries [18e20].

The application of carbon-based catalyst has attracted signifi-cant attention owing to its large surface area, high activity andthermal stability. The catalyst is prepared by modification of acti-vated carbon [3,21,22] and sulfonation of carbonaceous material[23,24] mainly from biomass. The main motivation for synthesizinga new carbon cryogel is to tailor the properties of carbon-basedcatalyst more suited for esterification. Based on our previousstudies, carbon cryogel synthesized from lignin and furfural geldemonstrated good reactivity in esterification [25,26]. Lignin andfurfural, as major byproduct and derived chemicals from biomassprocessing, have advantages as green substitutes in gel synthesis.Previously, more toxic chemicals such as phenol or resorcinol andformaldehyde have been used for gel synthesis to produce carbongel [27e30]. Therefore, it is envisaged ligninefurfural based carboncryogel could substitute homogeneous acid catalysts, such as thestrong H2SO4 acid, for esterification reaction.

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557548

Modeling studies on kinetic and thermodynamic provideimportant parameters related to activity, reaction rate, and acti-vation energy in heterogeneous reaction. The models should fit theexperimental data at different reaction time and temperature andare very useful for scaling-up purposes [31]. Moreover, the kineticparameters are needed in chemical process analysis and optimi-zation of the reaction parameters for catalysis application in theindustry [32,33].

The kinetic model is commonly developed based on simplepower law for determining the reaction rate constant. Hoo andAbdullah [34] stated for rapid investigation on the reaction rateconstant and activation energy of developed heterogeneous cata-lysts, the simple power law model has frequently been used. Thepseudo-homogeneous approach is usually applied in determiningthe kinetic model [20,35e37]. The first-order kinetic model isfrequently adopted to fit the data based on the assumptions made[38e40]. Previous studies on the levulinic acid esterification re-ported first-order and second-order kinetic models [20,31,35].

Thermodynamic study involves parameters such as enthalpy,entropy, and Gibbs free energy. Based on the activated complextheory, Ong et al. [41] have calculated the Gibbs free energy withthe Eyring-Polanyi equation (generally known as Eyring equation).Similar equations has been used in previous studies for calculationof thermodynamic activation parameters [42e44].

In this paper, the kinetic and thermodynamic parameters oflevulinic acid esterification with ethanol by using lignin-furfuralcarbon cryogel have been determined. The effects of internal andexternal diffusions were studied before conducting the kineticanalysis on the reaction at various temperatures and reaction time.The experimental data were interpreted using the pseudo-homogeneous kinetic model with different reaction orders. Thekinetic parameters of the selected model were reported includingthe reaction rate constants, activation energy and the pre-exponential factor. In addition, the thermodynamic activation pa-rameters namely the enthalpy of activation, entropy of activationand Gibbs energy of activation were determined for the levulinicacid conversion. Important basic information will be obtainedthrough this study for process development of catalytic levulinicacid conversion to ethyl levulinate using carbon cryogel.

2. Materials and methods

2.1. Materials

Ethanol (C2H6O, 95%, QRec), sulfuric acid (H2SO4, 95e97%,QRec), lignin (alkali, Sigma-Aldrich Co.), and furfural (C5H4O2,Merck) were used in gels synthesis. Levulinic acid (C5H8O5, 98%,Merck), ethanol (C2H6O, 99%, Merck) and ethyl levulinate standard(C7H12O3, 99%, Merck) were purchased with analytical grade andused as received for the esterification reaction.

2.2. Preparation and characterization of carbon cryogel

Lignin, furfural and distilled water were mixed with a ratio of1:1:1 for gel synthesis following the methods described by Zainolet al. [25] and [26]. Themixturewas homogenized and reactionwasconducted at 90 �C for 30min to synthesize the gel. Next, the gelwas immersed in t-butanol for solvent exchange, pre-frozen for24 h and freeze-dried for 8 h to form cryogel. Finally, the cryogelwas calcined in the furnace at ambient atmosphere and 500 �C for1 h to produce carbon cryogel.

The surface area of the carbon cryogel was evaluated by usingBrunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH)methods based on nitrogen adsorption and desorption at �196 �Cvia Surfer instrument by Thermo Scientific. The total acidity of the

carbon cryogel was measured by NH3-TPD (MicromeriticsAutoChem II 2920 V4.03) with thermal conductivity detector (TCD)and gas passing of 10% NH3/He up to 900 �C. The TGAwas analyzedby Perkin Elmer TGA 7 instrument at ramp temperature of 10 �C/min from 30 to 900 �C under nitrogen flow to evaluate the thermalstability.

2.3. Synthesis of ethyl levulinate

The esterification reaction of levulinic acid was conducted in a100mL high-pressure reactor. A series of carbon cryogel with par-ticle size in the range of 100e300 mm were utilized in the experi-ment to determine if internal diffusion resistance was a controllingfactor. Varying agitation speed from 100 to 300 rpmwas studied toevaluate the effect of external diffusion on levulinic acid conver-sion. For a typical experiment, the agitation speed was fixed at200 rpm and the reaction was conducted at 25wt% of catalystloading and 15M ratio of ethanol to levulinic acid. For determiningthe kinetic and thermodynamic parameters, the reaction time andtemperature range were from 0 to 180min and 78e150 �C,respectively. The experimental runs were randomly selected andrepeated to test the reproducibility of the data. The product sam-ples were separated using vacuum filtration with 0.45 mm nylonmembrane filter for further analysis.

2.4. Product analysis

A gas chromatograph system (GC-FID 7820A, Agilent Technol-ogy) was used to determine the number of moles of ethyl levulinate(EL) under the following conditions: HP-5 column (length, 30.0m;diameter, 320.0 mm); injector temperature, 270 �C; column tem-perature, set from 80 to 170 �C (13 �C/min) and the 170e300 �C(40 �C/min); carrier gas, nitrogen (1.0mL/min). The levulinic acidconcentration was determined for the calculation of conversionwith high performance liquid chromatography (HPLC) from AgilentTechnology (1260 Infinity) with the following conditions: Hi-Plex Hcolumn (length, 300mm; diameter, 7.7mm); UV detector(210 nm); column temperature of 60 �C; auto-injector (20 mL in-jection volume) and mobile phase of 5mMH2SO4 (0.6mL/min).

The standard concentration of ethyl levulinate and levulinic acidwere obtained from the standard calibration curves of the respec-tive compounds. The conversion and product yield were calculatedusing eqs. (1) and (2):

EL yield ðmol%Þ ¼ ��CE � Vp

��1000=ME

��ðWi=MLÞ � 100 (1)

Conversion ð%Þ ¼ �Wi �

��CL � Vp

��1000

� ��Wi � 100 (2)

where, CE and CL are the final product concentrations of ethyl lev-ulinate and levulinic acid (mg/L), respectively, Vp is the final volumeof product (L), Wi is the weight of levulinic acid (g), and ME and ML

are the molecular weight of ethyl levulinate and levulinic acid (g/mol), respectively.

2.5. Mathematical model for kinetic and thermodynamic studies

The esterification reaction of the levulinic acid and ethanol isshown in eq. (3), while the rate of the reaction is derived in eq. (4).

Levulinic acidCA

þ ЕthanolCB

%k1

k2Ethyl Levulinate

Cc

þ waterCD

(3)

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557 549

e rA ¼ e dCA=dt ¼ k1CaAC

bB e k2C

gCC

dD (4)

where CA, CB, CC, and CD are the concentrations of levulinic acid,ethanol, ethyl levulinate and water, respectively. a, b, g, and d aretheir respective reaction order. k1 is the kinetic constant for theforward reaction, while k2 is the kinetic constant for the reversereaction.

The kinetic model of levulinic acid conversion reaction wasbased on the following assumptions:

1) Excess ethanol was used (i.e. CB » CA, CC, CD), thus, the expressionfor concentration of ethanol (Cb

B) can be treated as a constant.2) The excess ethanol could push the equilibrium reaction forward

(i.e. k1 » k2) and is considered as irreversible reaction.

Thus, the kinetic model could be simplified to eqs. (5) and (6).

e dCA=dt ¼ k1CaAC

bB (5)

e dCA=dt ¼ k CaA (6)

where k ¼ k1CbB is a constant.

The equation was solved using integration method by studyingthe different reaction orders for levulinic acid. The kinetic order ofreaction was determined and justified based on the experimentalresults.

Based on the transition state theory, the Eyring equation isdeveloped to describe the relationship of the reaction rate withtemperature which could be also explained by Arrhenius equation.However, the Eyring equation is more appropriate to define thereaction in gas, condensed and mixed-phase. Thus, the Eyringequation and theory of the activated complex was applied toevaluate the thermodynamic activation parameters of the reaction[42e44] in accordance to eqs. (7) and (8).

DGz ¼ DHz � ΤDSz (7)

k ¼ kkBTh

þ e�DGz

RT(8)

The linearized Eyring equation (eq. (9)) has been used previ-ously to calculate the thermodynamic activation parameters[42e44]. The values of DHz, DSzwere determined from the plot of lnk/T versus 1/T and the Gibbs energy of activation was calculatedusing eq. (10).

InkT¼ �DНz

RTþ In

kBh

þ DSz

R(9)

DGz ¼ �RT In�khkBT

�(10)

where, DGz is the Gibbs energy of activation, DHz is the enthalpy ofactivation and DSz is the entropy of activation, Kz is the thermo-dynamic equilibrium constant, k is rate constant (s�1), R is gasconstant, k is the transmission factor (assume to be equal to 1), kB isBoltzmann's constant (1.381� 10�23 J/K), h is Planck's constant(6.626� 10�34 Js), T is the absolute temperature (K), and notation z

refers to the state of activated complex.

3. Result and discussion

3.1. Carbon cryogel synthesis

Lignin and furfural were applied as alternative feedstocks in thesynthesis of gel to replace phenol and formaldehyde. The proposedreaction, shown in Fig. 1, involved condensation of the mixtureunder acidic condition. The acidic condition created high electrondensity on C2 and C6 of the lignin phenylpropanoid units for thelinkage between the lignin-lignin and lignin-furfural condensation[45]. Twomechanismswere reported to be involved in this reactionsystem: the induction effect of the alkyl group and the resonanceeffects of the electron pairs on the methoxy oxygen of the ligninstructure [45e47]. These mechanisms contributed to the linkage ofthe lignin-lignin and lignin-furfural through the electrophilic sub-stitution at the reactive site of phenylpropanoid unit. Li and Gel-lerstedt [48] and Samuel et al. [49] have also reported about thelignin-lignin condensation product in their study which showssimilar possible mechanism. Under acidic condition furfural asfuran monomer would undergo polymerization condensationincluding self-condensation [50]. The potential of furfural as cross-linking agent for condensation with lignin has been reported [45].As for lignin-furfural, the electrophilic substitution by lignin phe-nylpropanoid units will cause the condensation with the electro-philic aldehyde carbonyl carbon to form lignin-furfural linkage.Thus, the mechanism on gel synthesis reaction of the mixturepossibly produces a complex lignin and furfural condensationproduct.

Fig. 2(a) exhibits the N2 adsorption and desorption isotherms ofcarbon cryogel. The isotherm has type I influence as strongadsorption was evident at the initial step of the adsorption region(<0.05 relative pressure) due to the existence of microporesstructure. Around 0.05 to 0.3 relative pressures, monolayer fol-lowed by the first fewmultilayers formationwas likely based on thelow slope region in the middle of isotherm. Besides, the isothermalso presented the formation of small hysteresis loop (>0.5 relativepressure) that indicated capillary condensation. The carbon cryogelhas small mesoporous structure based on the small hysteresis loopin the figure. The total surface area was 426.5m2/g with the mi-cropores and mesopores areas were 364.6m2/g and 61.9m2/g,respectively. Large total surface area is desired for catalyst tofacilitate adsorption of reactants. The pore size distribution ofselected carbon cryogel presents a pore structure with mostly2.0 nm pore width attesting the micropores distribution of carboncryogel (Fig. 2(b)).

The carbon cryogel was also thermally stable based on thermaldegradation (TGA/DTG) study through a heating process in nitro-gen flow at high temperature up to 900 �C. The carbon cryogelregistered a small weight loss up to 500 �C with low massdecomposition about 8wt% [26]. The continuous decomposition upto 600 �C is attributed to low molecular weight organic compoundand the functional group present in furfural and lignin structureincluding the sulfonate group asmentioned by Zainol et al. [25] and[26]. Singare et al. [51] have stated about the decomposition ofsulfonate functional group at temperature up to 521 �C. Furtherdecomposition above 600 �C might be contributed by the heavyorganic material mainly from the lignin structures and the availablesulfonate functional group (the strong acid site presented).

The result is consistent with NH3-TPD thermograph in Fig. 2(c)that depicts the strong acidity of carbon cryogel, linked to thesulfonate group, is fully decomposed up to 800 �C. The decompo-sition of acid site, contributed by low and moderate acidity, weredecomposed between 100 to 300 �C and 300e600 �C, respectively.As for strong acidity, the decomposition occurred above 600 �C.

Fig. 1. Proposed reaction mechanism of lignin with furfural in acidic condition (adapted from Refs. [45e47,49,50]).

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557550

3.2. Esterification reaction

The synthesized carbon cryogel was applied in levulinic acidesterification reaction. Fig. 3 proposed a simplified reactionmechanism of levulinic acid with ethanol in the presence of lignin-furfural carbon cryogel for ethyl levulinate formation. The surfaceacidity of the carbon cryogel correlated to the Brønsted acid sitescontributed by sulfonate group (-SO3H). Modified catalyst with thesulfonate group has been reported to possess Brønsted acidic sites[52e54]. Therefore, the proposed mechanism considered the re-action of levulinic acid and ethanol on the Brønsted acid sites ofcarbon cryogel. The reaction mechanism was suggested accordingto the previous proposed mechanism [18,55e58]. Levulinic acidwas adsorbed over the Brønsted acid sites of carbon cryogel byjoining the oxygen of the carbonyl group and increased the elec-trophilicity of the adjoining carbon atom to form protonated lev-ulinic acid intermediate. Similarly, Khder et al. [59] have describedreaction mechanism between acetic acid and alcohol as acetic acidwas adsorbed on Brønsted acid sites to form protonated acetic acidintermediate. The protonated levulinic acid intermediate was thensubjected to nucleophilic attack by the oxygen of ethanol to formoxonium ion [18,58]. A proton transfer from the oxonium ion leadsto formation of a new oxonium ion. The final step of the reactioninvolves deprotonation of former carbonyl oxygen eliminatingwater and forming ethyl levulinate. The regeneration of the acidsites occurs from the deprotonation step [18,60].

3.3. Effect of external and internal diffusions

In the kinetic study of heterogeneous levulinic acid conversionthe influence of external and internal diffusions should be exam-ined. The external and internal resistance might lead to masstransfer limitation in the reaction between levulinic acid andethanol with carbon cryogel and could affect the apparent reactionrate [35,39,61]. Based on the experimental results in Fig. 4(a),particle size has no effect on levulinic acid conversion inferring thatthe internal mass transfer resistance is negligible.

In addition, agitation speed could improve the contact betweenthe levulinic acid and carbon cryogel, and thus eliminate theexternal mass transfer resistance in the reaction system [39,62]. Noclear influence of agitation speed on the levulinic acid conversionwas observed and the external mass transfer resistance was alsonegligible (Fig. 4(b)). Therefore, the internal and external diffusionswere not the rate limiting factors in levulinic acid conversioncatalyzed by carbon cryogel. Thus, the assumption of pseudo ho-mogeneous kinetic model is valid in this study.

3.4. Effect of reaction temperature and time

Effects of temperature and time on ethyl levulinate yield andlevulinic acid conversion are shown in Fig. 5. The ethyl levulinateyield and levulinic acid conversion increased with time from 1 to4 h. Small changes in yield and conversionwere observed above 4 h

Fig. 2. (a) N2 sorption isotherm, (b) pore distribution and (c) NH3-TPD thermograph ofcarbon cryogel.

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557 551

of reaction time. Different reaction temperatures conferred asimilar trend and the reaction time of 4 h was the optimum. At78 �C, the ethyl levulinate yield and levulinic acid conversionincreased from 17.0 to 59.8mol% and 18.2e61.6%, respectively up to4 h of reaction time. The reaction time and temperature markedsignificant effect on the production of ethyl levulinate and alsolevulinic acid conversion. High yield and conversion were obtainedwith increasing temperature up to 150 �C. The yield increased from51.3 to 86.5mol% in 4 h and remained constant above this period.Similarly, the increment of levulinic acid conversion from 51.4 to87.2% was registered.

High ethyl levulinate yield attained at the selected condition

corresponded to the efficient catalyst activity. As reported by Fer-nandes et al. [63], the levulinic acid conversion was lower for blankreaction without any catalyst present. The application of catalysttogether with the reaction time and temperature is important todetermine the optimum ethyl levulinate yield. Therefore, high ethyllevulinate yield was obtained in less reaction time and moderatereaction temperature in the presence of carbon cryogel. Highselectivity (~100%) towards ethyl levulinate production and levu-linic acid conversion explained that the reactionwas not influencedby the side reaction.

3.5. Kinetic study

The concentration profile of levulinic acid consumption usingcarbon cryogel is displayed in Fig. 6. Significant reduction of thelevulinic acid concentration was observed at both reaction tem-peratures of 78 �C and 150 �C. The consumption of levulinic acidwas inversely proportional to the reaction rate of the ethyl levuli-nate production. Hoo and Abdullah [34] stated that the simplepower law model has frequently been used for the heterogeneouscatalyst for the rapid investigation in obtaining the rate constantand activation energy. In some cases, the mechanism-based modelis required to explain the detail of the mechanism for the reactionas the simple power law model is deficient to describe. However,the esterification of alcohol with a carboxylic acid is well-known tobe nucleophilic substitution which can be related to the reaction ofethanol and levulinic acid. Thus, themechanism-basedmodel is notconsidered in the present work. Hence, the power law model hasbeen adapted in this investigation to determine the rate constantand activation energy of the reaction.

Based on the mathematical model in eq. (6), the pseudo-zeroth-order, pseudo-first-order and pseudo-second-order kineticapproach were evaluated to study the reaction order for levulinicacid consumption. The integrated rate equation for pseudo-zero-order kinetic reaction (a¼ 0), pseudo-first order kinetic reaction(a¼ 1) and pseudo-second-order kinetic reaction (a¼ 2) aresimplified as in eqs. (11)e(13), respectively. CA0 and CB0 are theinitial concentrations of levulinic acid and ethanol, respectively, XAis levulinic acid conversion, and k is the kinetic constant.

XACA0 ¼ k t (11)

ln ð1 � XAÞ ¼ � k t (12)

XA= ð1 � XAÞ ¼ CA0 k t (13)

The reaction rate constants for three types of pseudo-order ki-netic studies are computed as follows:

(1) For pseudo-zero-order, the levulinic acid conversion rateconstant (k) was obtained from the slope of the plot of XAversus reaction time.

(2) As in pseudo-first-order, the levulinic acid conversion rateconstant (k) was calculated from the slope of the plot of ‒ln(1‒XA) versus reaction time.

(3) The plot of XA/(1‒XA) versus time determined the levulinicacid conversion rate constant (k) for pseudo-second-ordermodel approach.

The experimental data was subjected to the Arrhenius equation(eq. (14)) to calculate the activation energy (Ea) and exponentialfactor (A) of the reaction.

ln k ¼ ln A � Ea=RT (14)

Fig. 3. Proposed reaction mechanism of levulinic acid with ethanol in the presence of carbon cryogel (adapted from Refs. [18,55e59]).

Fig. 4. Effect of (a) carbon cryogel particle sizes and (b) agitation speed on levulinicacid conversion (78 �C, 25wt%, 15:1).

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557552

The linear plot of pseudo-zero-order in Fig. 7(a) exhibited therate was independent of the reactant concentration or conversion.Within the temperature range, the change of reactant concentra-tion or conversion with time did not affect the rate of reaction inaccordance with the zero-order kinetic. A good fit of experimentaldata was observed at the lower reaction temperature. However, asthe temperature increased, the data deviated from the linearmodel. Therefore, the reaction was not zero-order under all con-dition and only significant at certain condition. The situation couldalso be observed as the reaction time was further increased. Asshown in Fig. 5, the reaction above 78 �C slowed after 2 h of reac-tion, but the conversion increased which possibly involved transi-tion into different kinetic order. Thus, the approach of the zero-order model was insignificant since only valid for a limitedamount of time. The conversion of levulinic acid increased withtime until it reached a plateau inferring the reactionwas kineticallycontrolled by different reaction order. Therefore, pseudo-first-orderand pseudo-second-order model kinetic models were proposed.

Table 1 summarizes the correlation coefficients (R2) for allpseudo kinetic models. The coefficients at different reaction tem-peratures for pseudo-first-order model were all higher than thepseudo-second-order model (Fig. 7(b) and (c), respectively). A goodfit of the linear plot for the pseudo-first-order kinetic model wasobtained. The linearity inferred the pseudo-first-order wasdependent on the levulinic acid conversion. Therefore, the selectionof pseudo-first-order as the kinetic model for the catalytic reactionusing carbon cryogel was confirmed. The reaction rate constant oflevulinic acid conversion to ethyl levulinic using carbon cryogel wascalculated at different reaction temperature (Table 1). The rateconstant, k increased with temperature. The fast consumption oflevulinic acid at high temperature can be validated from the higherreaction rate of levulinic acid, compared to the reaction rate at lowtemperature. The yield of ethyl levulinate increased until it reacheda plateau. Hence, it can be claimed that levulinic acid was convertedinto ethyl levulinate as the final product of the esterification re-actions. This indicates that at high temperature, ethyl levulinateformation rate increases more quickly compared to reaction atlower temperatures.

Fig. 5. Effect of temperature and time on the (a) levulinic acid conversion and (b) ethyllevulinate production (at 25wt%, 15:1).

Fig. 6. Concentration profile of levulinic acid decomposition and ethyl levulinateproduction at 78 �C and 150 �C (25wt%, 15:1).

Fig. 7. Plot of (a) pseudo-zeroth-order, (b) pseudo-first-order, and (c) pseudo-second-order kinetic model for levulinic acid conversion. C 78 �C, - 100 �C, : 120 �C, x150 �C.

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557 553

Fig. 8 compares the parity plot between experimental andpredicted data of all the kinetic models. Hoo and Abdullah [34] hasused the parity plot to verify and compare the variation and ac-curacy of the data for the selection of kinetic model. The pseudo-zero-order model could not fit the actual with the predicted dataas the slope and R2 of the plot (Fig. 8(a)) were lower compared tothe other twomodels. The pseudo-zeromodel attained a significantdeviation of the predicted levulinic acid conversion from the actualconversion data. The parity plots presented that both pseudo-firstand pseudo-second-order models (Fig. 8(b) and (c), respectively)have high R2 of 0.9869 and 0.9787, respectively. However, pseudo-

Table 1Kinetic parameters of levulinic acid decomposition using carbon cryogel.

Temp. (�C) Pseudo-Zero Order Pseudo-1st Order Pseudo-2nd Order Pseudo-1st Order

R2 R2 R2 k (min�1)

78 0.9967 0.9906 0.9566 0.0036100 0.9654 0.997 0.9584 0.0059120 0.9186 0.9917 0.9826 0.0095150 0.8529 0.9856 0.9766 0.0113

Fig. 8. Parity plot of (a) pseudo-zeroth-order, (b) pseudo-first-order, and (c) pseudo-second-order kinetic model.

Table 2Activation energy and exponential factor of levulinic acid decomposition usingcarbon cryogel.

Reaction condition Pseudo 1st Order

R2 Ea (kJ/mol) A (min�1)

78e150 �CCarbon Cryogel

0.9473 20.2 3.97

Fig. 9. Arrhenius plot of ln k versus 1/T of pseudo-first order reaction.

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557554

first-order model marked high accuracy in representing the actuallevulinic acid conversion as the slopes value of 0.994 was very near

to unity. Compared to pseudo-second-order, the model attainedhigh R2 value, but a small deviation from unity was observed be-tween the predicted and actual levulinic acid conversion. Thesefindings suggest that the pseudo-zero-order kinetic model is notaccurate to represent the reaction compared to the other twomodels. The pseudo-first-order model is pertinent in levulinic acidesterification in presence of carbon cryogel. Therefore, the reactionrate constant of the pseudo-first-order kinetic could be calculatedas the validity of the selected kinetic model was confirmed.

The activation energy (Ea) and pre-exponential factor (A) valuesfor levulinic acid conversion to ethyl levulinate using carbon cry-ogel are listed in Table 2. The Arrhenius plot in Fig. 9 is used todetermine both constants with the plot intercept and slope pro-vided the values of Arrhenius constant and the activation energy,respectively. The conversion of levulinic acid to ethyl levulinate hasa low activation energy of 20.2 kJ/mol, which indicates the reactionrequires low energy to initiate. Kinetically controlled reactions havebeen reported with activation energy above 14.0 kJ/mol, and forlower value diffusion limited reaction prevailed [35,64]. Therefore,the reaction conducted using carbon cryogel in the levulinic acidconversion process is kinetically controlled based on the Ea value.

From previous studies, levulinic acid esterification over various

Table 3Activation energy (Ea) of the esterification of levulinic acid by using a different typeof catalysts and solvents.

Catalyst Solvent Ea (kJ/mol) Ref.

UDCaT-5 Ethanol 37.7 [31]Desilicated HZSM-5 Ethanol 21.1 [35]DPTA/DH-ZSM-5 Ethanol 29.4 [20]micro/meso HZ-5 Ethanol 27.4 [61]micro/meso HZ-5 Methanol 51.6 [61]micro/meso HZ-5 n-Butanol 23.8 [14]micro/meso HZ-5 n-Octanol 17.1 [61]SO4

�2/ZrO2-membrane Ethanol 14.6 [65]Carbon Cryogel Ethanol 20.2 This study

Table 4Thermodynamic parameters of levulinic acid decomposition using carbon cryogel.

Reactioncondition

Thermodynamic parameters

R2 DHz

(kJ mol�1)DSz

(J mol�1 K�1)DGz

(kJ mol�1)

78e150 �CCarbon Cryogel

0.9473 17.0 �278.0 124.1

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557 555

types of catalysts and alcohol conferred different reaction rate andactivation energy values. As tabulated in Table 3, a high activationenergy of 37.7 kJ/mol was obtained in the presence of mesoporoussuper acidic zirconia (UDCaT-5) for the conversion of levulinic acidinto ethyl levulinate [31]. Besides, the Ea of 21.1 kJ/mol was obtainedby using desilicated HZSM-5 in synthesis ethyl levulinate [35]. Theethyl levulinate production using DPTA/DH-ZSM-5 presented Ea of29.4 kJ/mol [20] and micro/meso HZ-5was reported to have Ea of27.4 kJ/mol [61]. In addition, the conversion of levulinic acid usingdifferent types of alcohol was performed in the presence of micro/meso HZ-5 where the production of methyl, n-butyl, and n-octyllevulinate were reported to have an Ea of 51.6, 23.8 and 17.1 kJ/mol,respectively [14,61]. Low activation energy of 14.6 kJ/mol was ob-tained by SO4

�2/ZrO2-membrane for levulinic acid conversion toethyl levulinate [65]. The activation energy obtained in this study iscomparable with the other reported studies since low activationenergy is reported in the presence of carbon cryogel.

3.6. Thermodynamic study

Based on Fig. 10 and Table 4, a positive enthalpy of activation,DHz indicates that the activation process is endothermic. Ong et al.[41] have explained that positive enthalpy indicated that the en-ergy input from an external source which refers to heat, wasrequired to raise the energy level for reactants transformation totheir transition state. This is also agreed by Nautiyal et al. [66], aspositive enthalpy indicated that the heat input required by the

Fig. 10. Plot of ln k/T versus 1/T for the levulinic acid esterification in the presence ofcarbon cryogel.

reaction to bring the reactants to the transition state before formingthe product.

Considering the value of entropy of activation, DSz, the negativeDSz value predicted the system is less disordered. Negative entropyinfers no degrees of freedom due to the activated complex forma-tion. This means the reacting species formed the transition stateduring the reaction, thereby having a more ordered state than thereactants in the ground state structure [43]. Ong et al. [41]mentioned that the negative value of DSz might result in an asso-ciative mechanism confirming that reactants interacted with eachother to form transition state along the reaction pathway. Nautiyalet al. [66] also stated that negative entropy indicated the degree ofthe ordered alignment of transition state was better than the re-actants in the ground state.

As for the Gibbs free energy of activation, DGz, the positivevalues of DGz inferred that the reaction is endergonic and notspontaneous in nature [41,66]. Besides, the positive DGz wasattributed to the higher energy level in the transition statecompared to the reactant species in the ground state [41]. Based onthe study performed by Badgujar and Bhanage [64], negative DSz

and positive DGz were obtained for the levulinic acid conversion toalkyl levulinate.

4. Conclusion

Large surface area and high acidity carbon cryogel provides agood surface chemistry to enhance esterification reaction. Based onthe external and internal diffusion studies, the esterification reac-tion using carbon cryogel has been confirmed to be in the kineticregime. The developed kinetic models are applied to investigate thereaction rate of levulinic acid conversion to ethyl levulinate. Amongthe pseudo-order kinetic models (zero, first and second-order), thereaction fits with the pseudo-first-order kinetic model. Theselected kinetic model is in good agreement with the experimentaldata as observed through the parity plot. The activation energy ofthe esterification reaction, performed in the presence of carboncryogel, is reported as 20.2 kJ/mol. The thermo activation parame-ters (DHz, DSz, DGz) are determined for this reaction using Eyringequation for energy assessment. The reaction is endergonic, notspontaneous and the system is more ordered at the transition stateof ethyl levulinate production. The proposed kinetic model andthermo activation parameters could be applied for modeling andsimulation studies in industrial reactor applications.

Acknowledgement

The authors would like to express their gratitudes to the Min-istry of Higher Education (MOHE), Malaysia and Universiti Tekno-logi Malaysia (UTM) for financial assistance under the ResearchUniversity Grant (vote 04H69), Fundamental Research GrantSchemes (vote 4F160) and Professional Development ResearchUniversity Post-Doctoral Fellowship (vote 03E82).

References

[1] Z. Ullah, M.A. Bustam, Z. Man, A.S. Khan, N. Muhammad, A. Sarwono,

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557556

Preparation and kinetics study of biodiesel production from waste cooking oilusing new functionalized ionic liquids as catalysts, Renew. Energy 114 (Part B)(2017) 755e765.

[2] Y.H. Tan, M.O. Abdullah, C. Nolasco-Hipolito, N.S. Ahmad Zauzi, Application ofRSM and Taguchi methods for optimizing the transesterification of wastecooking oil catalyzed by solid ostrich and chicken-eggshell derived CaO,Renew. Energy 114 (Part B) (2017) 437e447.

[3] M.A. Ahmad Farid, M.A. Hassan, Y.H. Taufiq-Yap, M.L. Ibrahim, M.R. Othman,A.A.M. Ali, Y. Shirai, Production of methyl esters from waste cooking oil usinga heterogeneous biomass-based catalyst, Renew. Energy 114 (Part B) (2017)638e643.

[4] W. Won, C.T. Maravelias, Thermal fractionation and catalytic upgrading oflignocellulosic biomass to biofuels: process synthesis and analysis, Renew.Energy 114 (Part B) (2017) 357e366.

[5] N.A.S. Ramli, N.A.S. Amin, Catalytic hydrolysis of cellulose and oil palmbiomass in ionic liquid to reducing sugar for levulinic acid production, FuelProcess. Technol. 128 (2014) 490e498.

[6] S. Choi, C.W. Song, J.H. Shin, S.Y. Lee, Biorefineries for the production of topbuilding block chemicals and their derivatives, Metab. Eng. 28 (2015)223e239.

[7] D.W. Rackemann, W.O. Doherty, The conversion of lignocellulosics to levulinicacid, Biofuels, Bioproducts and Biorefining 5 (2) (2011) 198e214.

[8] N.A.S. Ramli, N.A.S. Amin, Optimization of biomass conversion to levulinic acidin acidic ionic liquid and upgrading of levulinic acid to ethyl levulinate, Bio-Energy Res. (2016) 1e14.

[9] A.S. Khan, Z. Man, M.A. Bustam, A. Nasrullah, Z. Ullah, A. Sarwono, F.U. Shah,N. Muhammad, Efficient conversion of lignocellulosic biomass to levulinic acidusing acidic ionic liquids, Carbohydr. Polym. 181 (2018) 208e214.

[10] L. Liu, Z. Li, W. Hou, H. Shen, Direct conversion of lignocellulose to levulinicacid catalyzed by ionic liquid, Carbohydr. Polym. 181 (2018) 778e784.

[11] F. Su, L. Ma, D. Song, X. Zhang, Y. Guo, Design of a highly ordered mesoporousH 3 PW12O40/ZrO2eSi(Ph)Si hybrid catalyst for methyl levulinate synthesis,Green Chem. 15 (4) (2013) 885e890.

[12] K. Yan, G. Wu, J. Wen, A. Chen, One-step synthesis of mesoporousH4SiW12O40-SiO2 catalysts for the production of methyl and ethyl levulinatebiodiesel, Catal. Commun. 34 (2013) 58e63.

[13] K.C. Maheria, J. Kozinski, A. Dalai, Esterification of levulinic acid to n-butyllevulinate over various acidic zeolites, Catal. Lett. 143 (11) (2013) 1220e1225.

[14] K.Y. Nandiwale, V.V. Bokade, Esterification of renewable levulinic acid to n-butyl levulinate over modified H-ZSM-5, Chem. Eng. Technol. 38 (2) (2015)246e252.

[15] Y.-H. Chung, T.-H. Peng, H.-Y. Lee, C.-L. Chen, I.L. Chien, Design and control ofreactive distillation system for esterification of levulinic acid and n-butanol,Ind. Eng. Chem. Res. 54 (13) (2015) 3341e3354.

[16] D. Unlu, O. Ilgen, N.D. Hilmioglu, Reactive separation system for effectiveupgrade of levulinic acid into ethyl levulinate, Chem. Eng. Res. Des. 118 (2017)248e258.

[17] I.M. Atadashi, M.K. Aroua, A.A. Aziz, Biodiesel separation and purification: areview, Renew. Energy 36 (2) (2011) 437e443.

[18] G. Pasquale, P. V�azquez, G. Romanelli, G. Baronetti, Catalytic upgrading oflevulinic acid to ethyl levulinate using reusable silica-included Wells-Dawsonheteropolyacid as catalyst, Catal. Commun. 18 (2012) 115e120.

[19] G.M.G. Maldonado, R.S. Assary, J.A. Dumesic, L.A. Curtiss, Acid-catalyzedconversion of furfuryl alcohol to ethyl levulinate in liquid ethanol, EnergyEnviron. Sci. 5 (10) (2012) 8990e8997.

[20] K.Y. Nandiwale, S.K. Sonar, P.S. Niphadkar, P.N. Joshi, S.S. Deshpande, V.S. Patil,V.V. Bokade, Catalytic upgrading of renewable levulinic acid to ethyl levuli-nate biodiesel using dodecatungstophosphoric acid supported on desilicatedH-ZSM-5 as catalyst, Appl. Catal. Gen. 460 (2013) 90e98.

[21] A.S. Badday, A.Z. Abdullah, K.-T. Lee, Transesterification of crude Jatropha oilby activated carbon-supported heteropolyacid catalyst in an ultrasound-assisted reactor system, Renew. Energy 62 (2014) 10e17.

[22] S. Baroutian, M.K. Aroua, A.A.A. Raman, N.M.N. Sulaiman, Potassium hydrox-ide catalyst supported on palm shell activated carbon for transesterification ofpalm oil, Fuel Process. Technol. 91 (11) (2010) 1378e1385.

[23] T. Liu, Z. Li, W. Li, C. Shi, Y. Wanga, Preparation and characterization ofbiomass carbon-based solid acid catalyst for the esterification of oleic acidwith methanol, Bioresour. Technol. 133 (2013) 618e621.

[24] F.D. Pileidis, M. Tabassum, S. Coutts, M.-M. Titirici, Esterification of levulinicacid into ethyl levulinate catalysed by sulfonated hydrothermal carbons, Chin.J. Catal. 35 (6) (2014) 929e936.

[25] M.M. Zainol, N.A.S. Amin, M. Asmadi, Synthesis and characterization of carboncryogel microspheres from lignin-furfural mixtures for biodiesel production,Bioresour. Technol. 190 (2015) 44e50.

[26] M.M. Zainol, N.A.S. Amin, M. Asmadi, Effects of thermal treatment on carboncryogel preparation for catalytic esterification of levulinic acid to ethyl levu-linate, Fuel Process. Technol. 167 (2017) 431e441.

[27] Z. Zapata-Benabithe, F. Carrasco-Marín, J. de Vicente, C. Moreno-Castilla,Carbon xerogel microspheres and monoliths from resorcinoleformaldehydemixtures with varying dilution ratios: preparation, surface characteristics,and electrochemical double-layer capacitances, Langmuir 29 (20) (2013)6166e6173.

[28] A. Moreno, A. Arenillas, E. Calvo, J. Bermúdez, J. Men�endez, Carbonisation ofresorcinoleformaldehyde organic xerogels: effect of temperature, particle sizeand heating rate on the porosity of carbon xerogels, J. Anal. Appl. Pyrol. 100

(2013) 111e116.[29] C. Scherdel, G. Reichenauer, Carbon xerogels synthesized via

phenoleformaldehyde gels, Microporous Mesoporous Mater. 126 (1e2)(2009) 133e142.

[30] T. Yamamoto, T. Sugimoto, T. Suzuki, S.R. Mukai, H. Tamon, Preparation andcharacterization of carbon cryogel microspheres, Carbon 40 (8) (2002)1345e1351.

[31] G.D. Yadav, A.R. Yadav, Synthesis of ethyl levulinate as fuel additives usingheterogeneous solid superacidic catalysts: efficacy and kinetic modeling,Chem. Eng. J. 243 (2014) 556e563.

[32] L. Wie, X. Li, L. Yanxun, L. Tingliang, L. Guoji, Process Variables of the esteri-fication reaction of 3-pentadecylphenol and acryloyl chloride catalyzed by anionic liquid, Chem. Eng. Technol. 36 (4) (2013) 559e566.

[33] W. Shi, B. He, Y. Cao, J. Li, F. Yan, Z. Cui, Z. Zou, S. Guo, X. Qian, Continuousesterification to produce biodiesel by SPES/PES/NWF composite catalyticmembrane in flow-through membrane reactor: experimental and kineticstudies, Bioresour. Technol. 129 (2013) 100e107.

[34] P. Hoo, A.Z. Abdullah, Kinetics modeling and mechanism study for selectiveesterification of glycerol with lauric acid using 12-tungstophosphoric acidpost-impregnated SBA-15, Ind. Eng. Chem. Res. 54 (32) (2015) 7852e7858.

[35] K.Y. Nandiwale, P.S. Niphadkar, S.S. Deshpande, V.V. Bokade, Esterification ofrenewable levulinic acid to ethyl levulinate biodiesel catalyzed by highlyactive and reusable desilicated H-ZSM-5, J. Chem. Technol. Biotechnol. 89 (10)(2014) 1507e1515.

[36] H. Zhang, J. Ding, Y. Qiu, Z. Zhao, Kinetics of esterification of acidified oil withdifferent alcohols by a cation ion-exchange resin/polyethersulfone hybridcatalytic membrane, Bioresour. Technol. 112 (2012) 28e33.

[37] J. Bedard, H. Chiang, A. Bhan, Kinetics and mechanism of acetic acid esterifi-cation with ethanol on zeolites, J. Catal. 290 (2012) 210e219.

[38] A.H. Mohammad Fauzi, N.A.S. Amin, R. Mat, Esterification of oleic acid tobiodiesel using magnetic ionic liquid: multi-objective optimization and ki-netic study, Appl. Energy 114 (2014) 809e818.

[39] N.A.S. Ramli, N.A.S. Amin, Kinetic study of glucose conversion to levulinic acidover Fe/HY zeolite catalyst, Chem. Eng. J. 283 (2016) 150e159.

[40] G. Wang, Z. Li, C. Li, H. Wang, Kinetic and thermodynamic studies on one-stepsynthesis of methyl acrylate promoted by generated ionic liquid at mildtemperature, Chem. Eng. J. 319 (2017) 297e306.

[41] L. Ong, A. Kurniawan, A. Suwandi, C. Lin, X. Zhao, S. Ismadji, Transesterificationof leather tanning waste to biodiesel at supercritical condition: kinetics andthermodynamics studies, J. Supercrit. Fluids 75 (2013) 11e20.

[42] P.D. Patil, S. Deng, Optimization of biodiesel production from edible and non-edible vegetable oils, Fuel 88 (7) (2009) 1302e1306.

[43] D. Borsato, D. Galvan, J.L. Pereira, J.R. Orives, K.G. Angilelli, R.L. Coppo, Kineticand thermodynamic parameters of biodiesel oxidation with synthetic anti-oxidants: simplex centroid mixture design, J. Braz. Chem. Soc. 25 (11) (2014)1984e1992.

[44] T. Sismanoglu, A. Ercag, S. Pura, E. Ercag, Kinetics and isotherms of dazometadsorption on natural adsorbents, J. Braz. Chem. Soc. 15 (5) (2004) 669e675.

[45] P. Dongre, M. Driscoll, T. Amidon, B. Bujanovic, Lignin-furfural based adhe-sives, Energies 8 (8) (2015) 7897e7914.

[46] G.H. van der Klashorst, Modification of Lignin at the 2-and 6-positions of thePhenylpropanoid Nuclei, ACS Publications, 1989.

[47] L. Bu, Y. Tang, Y. Gao, H. Jian, J. Jiang, Comparative characterization of milledwood lignin from furfural residues and corncob, Chem. Eng. J. 175 (2011)176e184.

[48] J. Li, G. Gellerstedt, Improved lignin properties and reactivity by modificationsin the autohydrolysis process of aspen wood, Ind. Crop. Prod. 27 (2) (2008)175e181.

[49] R. Samuel, S. Cao, B.K. Das, F. Hu, Y. Pu, A.J. Ragauskas, Investigation of the fateof poplar lignin during autohydrolysis pretreatment to understand thebiomass recalcitrance, RSC Adv. 3 (16) (2013) 5305e5309.

[50] J.K. Fink, 7-Furan Resins, Reactive Polymers Fundamentals and Applications,William Andrew Publishing, Norwich, NY, 2005, pp. 307e320.

[51] P.U. Singare, R.S. Lokhande, R.S. Madyal, Thermal degradation studies of somestrongly acidic ca-tion exchange resins, Open J. Phys. Chem. 1 (2) (2011) 45.

[52] K. Nakajima, M. Hara, Amorphous carbon with SO3H Groups as a solidBrønsted acid catalyst, ACS Catal. 2 (7) (2012) 1296e1304.

[53] M. Hara, Biodiesel production by amorphous carbon bearing SO3H, COOH andphenolic OH groups, a solid Brønsted acid catalyst, Topic Catalysis 53 (11e12)(2010) 805e810.

[54] Y.-X. Zhou, Y.-Z. Chen, Y. Hu, G. Huang, S.-H. Yu, H.-L. Jiang, MIL-101-SO3H: ahighly efficient Brønsted acid catalyst for heterogeneous alcoholysis of ep-oxides under ambient conditions, Chem. Eur J. 20 (46) (2014) 14976e14980.

[55] F. Su, Q. Wu, D. Song, X. Zhang, M. Wang, Y. Guo, Pore morphology-controlledpreparation of ZrO2-based hybrid catalysts functionalized by both organosilicamoieties and Keggin-type heteropoly acid for the synthesis of levulinate es-ters, J. Mater. Chem. 1 (42) (2013) 13209e13221.

[56] S.S. Enumula, V.R.B. Gurram, R.R. Chada, D.R. Burri, S.R.R. Kamaraju, Cleansynthesis of alkyl levulinates from levulinic acid over one pot synthesizedWO3-SBA-16 catalyst, J. Mol. Catal. Chem. 426 (Part A) (2017) 30e38.

[57] E.S. Sankar, V. Mohan, M. Suresh, G. Saidulu, B.D. Raju, K.R. Rao, Vapor phaseesterification of levulinic acid over ZrO2/SBA-15 catalyst, Catal. Commun. 75(2016) 1e5.

[58] N.A.S. Ramli, D. Sivasubramaniam, N.A.S. Amin, Esterification of levulinic acidusing ZrO2-supported phosphotungstic acid catalyst for ethyl levulinate

M.M. Zainol et al. / Renewable Energy 130 (2019) 547e557 557

production, BioEnergy Research 10 (4) (2017) 1105e1116.[59] A.S. Khder, E.A. El-Sharkawy, S.A. El-Hakam, A.I. Ahmed, Surface character-

ization and catalytic activity of sulfated tin oxide catalyst, Catal. Commun. 9(5) (2008) 769e777.

[60] D. Song, S. An, B. Lu, Y. Guo, J. Leng, Arylsulfonic acid functionalized hollowmesoporous carbon spheres for efficient conversion of levulinic acid or fur-furyl alcohol to ethyl levulinate, Appl. Catal. B Environ. 179 (2015) 445e457.

[61] K.Y. Nandiwale, V.V. Bokade, Environmentally benign catalytic process foresterification of renewable levulinic acid to various alkyl levulinates biodiesel,Environ. Prog. Sustain. Energy 34 (3) (2015) 795e801.

[62] L. Peng, L. Lin, J. Zhang, J. Zhuang, B. Zhang, Y. Gong, Catalytic conversion ofcellulose to levulinic acid by metal chlorides, Molecules 15 (8) (2010)5258e5272.

[63] D. Fernandes, A. Rocha, E. Mai, C.J. Mota, V.T. da Silva, Levulinic acid esteri-fication with ethanol to ethyl levulinate production over solid acid catalysts,Appl. Catal. Gen. 425 (2012) 199e204.

[64] K.C. Badgujar, B.M. Bhanage, Thermo-chemical energy assessment for pro-duction of energy-rich fuel additive compounds by using levulinic acid andimmobilized lipase, Fuel Process. Technol. 138 (2015) 139e146.

[65] D. Unlu, O. Ilgen, N.D. Hilmioglu, Biodiesel additive ethyl levulinate synthesisby catalytic membrane: SO4

�2/ZrO2 loaded hydroxyethyl cellulose, Chem. Eng.J. 302 (2016) 260e268.

[66] P. Nautiyal, K. Subramanian, M. Dastidar, Kinetic and thermodynamic studieson biodiesel production from Spirulina platensis algae biomass using singlestage extractionetransesterification process, Fuel 135 (2014) 228e234.