Xiaocheng Lin,Youjie Huang,Ling Li,Changshen Ye,Jie Chen,2,,Ting Qiu,

1 Engineering Research Center of Reaction Distillation,Fujian Province University,College of Chemical Engineering,Fuzhou University,Fuzhou 350116,China

2 College of Environment and Resources,Fuzhou University,Fuzhou 350116,China

Keywords:Catalyst Ionic liquids Biodiesel Poly (ethylene imine)Esterification

ABSTRACT A series of polymeric ionic liquids (PILs) used as effective heterogeneous catalysts for biodiesel production via esterification of free fatty acids(FFAs)were effectively prepared by the reaction of poly(ethylene imine) (PEI) polymers with different molecular weight and 1,3-propanesultone,followed by the further acidification with differential effective acids,i.e.H2SO4,CF3SO3H,CH3SO3H or p-toluenesulfonic acid(p-TSA).Ultrahigh acidity and catalytic performance were achieved and could be fine-tuned by simply adjusting the molecular weight of PEI and by further treatment of acids.Specifically,under the optimal conditions (i.e.reaction temperature was 70 °C,reaction time was 2.0 h,catalyst dosage was 3.15%(mass),and alcohol/acid molar ratio was 14:1) acquired through the Box-BEHNKEN response surface methodology,a high oleic acid conversion of 98.42% could be obtained over the optimal PIL,PEI(70000)-PS-p-TSA.Additionally,our PILs also showed high generality for esterification of other FFAs,with general high conversion over 90% noted in each case even under much milder reaction conditions compared to other conventional catalysts.

1.Introduction

Due to rapidly growing energy demand and environmental concerns,the depletion of the world’s oil reserves has prompted interests to explore alternative sources of some petroleum-based fuels[1].In recent years,biodiesel has attracted wide attention because of its reproducibility and environmental friendliness compared with fossil diesel[2].Biodiesel can be generally prepared from vegetable oils and animal fats through the transesterification of triglycerides [3] or through the esterification of free fatty acids(FFAs) in the presence of a catalyst [4].

As an effective reaction for biodiesel production,transesterification can be catalyzed by both acidic and basic catalysts [5].However,feedstocks with low amount of FFAs were required in the base-catalyzed transesterification due to the easy saponification of FFAs.In addition,triglycerides and alcohol should be highly anhydrous to prevent undesirable saponification reactions [6].Compared to these,the esterification of FFAs shows much more promising advantages over transesterification in terms of cost and efficiency in the biodiesel production[7].As such,the research on acid-catalyzed biodiesel production through esterification has attracted great attention in recent years.However,homogeneous acid catalysts such as sulfuric acid have disadvantages of equipment corrosion,harmful waste emission and inconvenient recycling [8].Acidic ionic liquids (ILs) with non-corrosivity and high activity are therefore designed to replace traditional homogeneous catalysts for biodiesel production[9].Among the ILs,Br?nsted acid ionic liquids (BAILs) play an important role in esterification reaction application as catalysts[10].For the BAILs catalytic esterification,first,carbonyl oxygen on the oleic acid would interact with the protons released from the ionic liquids to form carbocation.The hydroxyl on methanol would then tend to perform a nucleophilic attack on the carbocation to form the intermediate,accompanied by the generation of ethyl oleate and water,while the protons would be released again to conduct the reaction.Thus,the high acid density (AD) of the BAILs promised themselves to be important catalysts for the esterification.However,their widespread utilizations are still limited by the difficulty in their recovery.For these reasons,heterogeneous catalysts [11],such as heteropoly solid acids [12] have attached considerable attention in recent years [13,14].Solid acids,as nonvolatile materials,are much less noxious than traditional liquid acids[15].However,they are still suffering from drawbacks of low active-site to molecular weight ratios and restricted accessibility of the matrix-bound acidic sites,resulting in the rapid deactivation from coking [16].

To overcome the above-mentioned shortcomings,it is proposed to use polymerized ionic liquids and immobilized acidic ionic liquids as heterogeneous catalysts for biodiesel production [9].Since then,some BAILs,including–SO3H-functionalized ionic liquids were immobilized on solid carriers to obtain efficient catalysts.For example,Wang et al.[17]prepared a reusable and efficient catalyst by modificating the ionic liquid on functionalized SBA-15 by thiol-ene click reaction,which showed high catalytic activity for various esterification.Miao et al.[18] reported that silica gel supported ILs[Silica-(CH2)3SO3-Im][HSO4],and used them for esterification of n-Butanol with acetic acid.A satisfying yield of over 99%under the optimal reaction condition was obtained.Liang et al.[19]found that the novel acidic polymeric ionic liquid (PIL),prepared by grafting the BAILs onto the copolymer of viny imidazole (VIm)and divinylbenzene (DVB),was an efficient catalyst for both the transesterification of triglycerides and esterification of free fatty acids with the overall yield of 99.0%.Even though the solidification from polymerization and immobilization can be favored for biodiesel conversion,the complex process for the synthesis and the harsh conditions for the esterification were always required due to the low number of active sites.Additionally,high catalytic conversion could be only achieved by using high catalyst dosage,which is still the main challenge to expand their practical application [9].

To address the challenges discussed above,in this paper,we report on a series of novel PILs with high number of catalytic sites and activity.These novel PILs could be easily obtained via a twostep method and be effectively applied in the biodiesel production under more gentle experimental conditions.Poly (ethylene imine)(PEI) polymers with different molecular weights were selected as starting materials to react with 1,3-propanesultone (PS) to obtain PEI-PS precursors,followed by the acidification with different acids(e.g.H2SO4,CF3SO3H,CH3SO3H,and p-toluenesulfonic acid(p-TSA))to obtain our novel PILs.The influences of the molecular weight of PEI and the type of the acid for acidification on the catalytic performance were investigated to propose a general design principle for such PEI-based PILs on the FFAs esterification.

2.Experimental

2.1.Materials

PEI polymers (99%) (with molecular weight of 600,1800,or 10,000 Da),PEI polymer solution (with molecular weight of 70,000 Da;50%(mass)in H2O),1,3-propanesultone(PS,99%),oleic acid(85%),H2SO4(99%),CF3SO3H(98%),CH3SO3H(98%)and p-TSA(98%)were purchased from Aladdin Chem.Co.,Ltd.,China.Methanol and ethanol of AR grade were purchased from Sinopharm Chemical Reagent Co.,Ltd.,China.PEI solution was thoroughly evaporated to remove water before use.

2.2.Preparation of PILs

PILs based on PEI were prepared by a two-step method including the PIL precursor preparation(Fig.1(a))and acidification(Fig.1(b)).Notes that due to the steric hindrance and different reactivity of different type of amines in the PEI,some of the amine groups could not be reacted with PS.First,PEI (1.0 g) was dissolved in ethanol(10 ml),and PS(5.0 g) was then slowly added to the solution.The solution was then stirred at 50 °C for 12 h to get a white solid,which was collected and washed for 3 times with ethanol to remove the unreacted materials.The solid was dried at 60°C under vacuum for 12 h to form PEI(X)-PS precursors,where X is the molecular weight of PEI(X=600,1800,10,000 or 70,000).Element analysis(EA)was carried out for the different PEI(X)-PS polymer to investigate their compositions (See Table S1 in Supplementary Material).

PEI(X)-PS polymers were then acidified with different acids including H2SO4,CF3SO3H,CH3SO3H or p-TSA.First,PEI(X)-PS(1.0 g) was added into a round bottom flask.Certain amount of the selected acid was then added,and the molar ratio of acid and the nitrogen element of PEI(X)-PS (measured by EA) was set as 1.2:1 molar to ensure the complete acidification (See Table S2).The mixture was heated gradually to 90 °C with a heating rate of 4°C.min-1and magnetically stirred for 6 h to form yellow viscous solids,which was then washed three times with ethyl acetate and dried at 60°C under vacuum for 12 h to get the final PEI(X)/PS-acid PILs (where acid=H2SO4,CF3SO3H,CH3SO3H or p-TSA).

2.3.Characterization

Fourier transform infrared(FTIR)spectra were obtained using a ALPHA FTIR spectrometer(Bruker,Germany)in the range from 400 to 4000 cm-1.UV–visible spectroscopy spectra were obtained with a B453 spectrophotometer (Agilent,USA).Elemental analysis (EA)was recorded on a Vario EL cube (Elementar,Germany).Thermogravimetric analysis (TGA) was investigated on a STA449F5 (Netzsch,Germany) under a nitrogen atmosphere and the temperature was increased from 50 to 600 °C with a heating rate of 10 °C.min-1.The acid density (AD) was determined according to the previously reported method [20].First,the fully dried sample was dissolved in a 1 mol.L-1NaCl solution for 24 h;the amount of the released H+in the sample was measured with the titration by using 0.05 mol.L-1Na2CO3aqueous solution,while the phenolphthalein was applied as the indicator.The acid density was calculated according to the following equation:

where V (L) is consumed volume of the Na2CO3solution used;CNa2CO3is the concentration of the Na2CO3solution used in titration(0.05 mol.L-1),and Wcatis dry weight of the catalyst sample(g).The Hammett acidity of our PILs was determined according to the literature [21–23].PIL samples and p-nitroaniline (pKa=0.99) as the basic indicator were dissolved in water to form the aqueous solutions with the concentrations of 5 mmol.L–1and 10 mg.L–1,respectively.The protonation extent of the indicator in terms of the [I]/[IH+] ratio in solution was calculated by through UV–vis spectrum(Carry 7000,Agilent,USA).The Hammett function(H0)is defined as:

where pK(I)aqis the pKavalue of the indicator referred to the aqueous solution,and [IH+] and [I] are the molar concentrations of the protonated and unprotonated forms of the indicator in the solvent PILs,respectively.Noted the maximum absorbance of the unprotonated form of 4-nitroaniline was observed at 380 nm in the deionized water.The [I]/[IH+] could be therefore acquired by using the absorbance change at 380 nm with a UV–vis spectroscopy.

2.4.Catalytic experiments

To study the catalytic performance of PILs in the biodiesel production via oleic acid esterification,different experimental conditions (i.e.reaction temperature of 50–90 °C;molar ratio of methanol to oleic acid of 6:1–14:1;catalyst amount based on oleic acid mass of 1.0%–4.0% (mass);reaction time of 0.5–2.5 h) were carefully applied to study their influence on the catalytic performance of PILs.After the reaction,PILs were separated by simple filtration,and then excess methanol was removed from the product by a rotary evaporator.According to the standard method of ASTM D974 oil acid value,we can determine the oleic acid conversion rate from the initial oleic acid value and the acid value in the biodiesel product through KOH titration according to our previous work as follow [24]:

Fig.1.Approach for the synthesis of PEI(X)-PS precursors (a) and PEI(X)/PS-acid PILs (b).

where AV0and AV1are the acid values of initial oleic acid and product biodiesel,respectively (mg KOH per g sample).

Statistical analysis for esterification of oleic acid with methanol was carried out using response surface methodology (RSM).The Box-BEHNKEN design method[25]was used to evaluate the interaction effects of the three main process variables,with which the optimal operating conditions for biodiesel could be predicted.The variables include the oleic acid conversion(Y),the molar ratio of methanol to oleic acid (X1),the dosage of catalyst (X2),and the reaction time (X3).The range and encoding level of the independent variable were encoded as (–1,1) intervals,and the center point of the variable was encoded as zero (0).The level,coded and actual values of the three operating variables were listed in Table S4.According to the design,17 experiments were performed with 12 factorial points and 5 center points.

The complete quadratic model is used to estimate the interaction between the operating variables of the oleic acid yield and is given by the following Eq.(4):

where Y is the predicted response;λ0is the intercept coefficient;λ1,λ2and λ3are the linear terms;λ11,λ22and λ33are the quadratic terms;λ12,λ13and λ23are the interaction terms.The experiments were carried out under the temperature of 70 °C because the catalytic performance showed no increase when the reaction temperature was fixed above 70 °C according to the single factor experiment.

3.Results and Discussion

3.1.Optimization of the molecular weight of PEI

To obtain optimal PILs,based on PEI,for the biodiesel production,we firstly looked into the influence of molecular weight of PEI on the physical and catalytic properties of PILs and their precursors,while H2SO4was selected as the treated acid.

The FT-IR spectra of polymers PEI(70000),PEI(70000)-PS and PEI(70000)-PS-H2SO4PIL were shown in Fig.2.Taking PEI(70000)as an example,the bands at 2900–3000 cm-1were assigned to the vibrations of C-H;the wide double bands at 3000–3600 cm-1and the band at 1650 cm-1were assigned to the vibrations of N-H;the band at 1035 cm-1was assigned to the vibration of C-N[26–28].After modified with PS,the PEI(70000)-PS showed newly formed bands at 1020 cm-1[29]and 1135 cm-1[30],which were assigned to the sulfonate groups and sulfonic acid groups,respectively,showing the successful grafting of PS on the PEI.From PEI(70000)-PS to PEI(70000)-PS-H2SO4PIL,the intensity of the bands at 1020 cm-1[29] decreases while that at 1135 cm-1increase,suggesting the protonation of sulfonate groups to form sulfonic acid groups.Besides,the band at 867 cm-1can also confirm the presence of HSO4–[28,31,32],suggesting the successful preparation of PEI(70000)-PS-H2SO4PIL.These data fitted well with the results of1H-NMR shown and discussed in Fig.S1.The similar FTIR results were also observed in other PEI(X)-PS and their PILs,as shown in Fig.S2,confirming the effective fabrication of PEI(X)-PS-H2SO4PILs.

Fig.2.FTIR spectra of PEI(70000),PEI(70000)-PS and PEI(70000)-PS-H2SO4 PIL.

Acid density values of PEI(X)-PS-H2SO4PILs were conducted to evaluate their catalysis performance.As shown in Fig.3,from PEI(600)-PS-H2SO4to PEI(70000)-PS-H2SO4,the values of acid density(AD)increased from 4.65 to 7.50 mmolg-1.The increasing trend in AD should be due to the decreasing steric hindrance for the reaction between amine groups of PEI and PS with the increase of PEI molecular weight,resulting in the increasing amount of the grafted sulfonic acid groups onto PEI chain(more details please see Tables S1 and S2).We also studied their Hammett acidity function (H0),which is another important index of acidity evaluation of the acidic catalysts.The Hammett acidity determination of PEI(X)-PS-H2SO4PILs and H2SO4were conducted and the results were shown in Fig.4 and Table 1.As well known,the lower H0value can signify the stronger acidity of the acidic samples [21–23].Therefore,it can be seen that under our experimental conditions,the acidity order of the investigated samples was shown as PEI(70000)-PS-H2-SO4>PEI(10000)-PS-H2SO4≈PEI(1800)-PS-H2SO4>PEI(600)-PSCH3SO3H >H2SO4,which increased with the molecular weight of PEI,suggesting the acidity for PILs could be tuned through their molecular weight of PEI.Additionally,the acidities of our PILs were considerably higher than that of H2SO4,suggesting a more promising catalytic performance on the esterification of our PILs [33].

Driven by the tunable acidity shown above,we conducted a quick study on the catalytic performance of our PEI(X)-PS-H2SO4PILs,with different molecular weights of PEI,for biodiesel production via esterification of oleic acid with methanol.The reactions were initially conducted with a methanol/oleic acid molar ratio of 10:1 and a catalyst dosage of 2.5% (mass) at 50 °C for 0.5 h.It can be seen from the Fig.5 that the conversion of oleic acid was enhanced significantly with the presence of PEI(X)-PS-H2SO4PILs as the catalysts,compared to that without catalysts.Encouragingly,their catalytic activities also showed an order as follows:PEI(70000)-PS-H2SO4>PEI(10000)-PS-H2SO4>PEI(1800)-PS-H2SO4>PEI(600)-PS-H2SO4,which was in well accordance with their molecular weight,and can be attributed to the enhanced acid density and acidity from PEI(600)-PS-H2SO4to PEI(70000)-PS-H2SO4as mentioned above.

The results well confirmed that the catalysis-related performance of PEI(X)-PS-H2SO4PILs could be well varied and controlled through their molecular weight of PEI as the starting material,and finally,PIL based on PEI with the molecular weight of 70,000 showed the best performance and was selected for the further investigation.

Fig.3.Acid density values of PEI(X)-PS-H2SO4 PILs.

Fig.4.The UV absorption spectra of PEI(X)-PS-H2SO4 PILs.

3.2.Optimization of treated acid

We further used effective acids,including CH3SO3H,CF3SO3H,p-TSA and H2SO4to treat the PEI(70000)-PS to acquire PEI(70000)-PS-acid PILs for further enhancement on the catalytic performance.The structure of PEI(70000)-PS-acid PILs were also investigated by FTIR,as shown in Fig.6.It can be seen that the spectra of PEI(70000)-PS-CH3SO3H,PEI(70000)-PS-CF3SO3H and PEI(70000)-PSp-TSA were nearly identical to the spectrum of PEI(70000)-PSH2SO4due to the closely related chemical structures of these PEIs,and,however,some differences caused by the treating acids could also be observed.Specifically,for PEI(70000)-PS-p-TSA,the band at 675 cm-1was assigned to the in-plane-bending of aromatic CH on[p-TSA]–[34].For PEI(70000)-PS-H2SO4,the band at 867 cm-1,which was assigned to the vibration of [HSO4]–anions,could be clearly detected [31,35].For PEI(70000)-PS-CH3SO3,the peak intensity at 1342 cm-1,whichwas assignedto the-C-H bond,greatly increased,confirming an addition of -C-H on [CH3SO3]–[36].While for PEI(70000)-PS-CF3SO3,the band at 1143 cm-1,which was assigned to the vibrations of[CF3SO3]–anions,was also observed after acid treatment [37].The above results indicated that PEI(70000)-containing PILs could be acquired successfully by treating acids.

AD values of PEI(70000)-PS-acid PILs were then investigated and the results were shown in Fig.7,it can be seen that AD values of PEI(70000)-PS-CH3SO3H,PEI(70000)-PS-CF3SO3H and PEI(70000)-PS-p-TSA were in the range of 3.50–3.95 mmolg-1.By contract,the AD value of PEI(70000)-PS-H2SO4was 7.50 mmolg-1,which was nearly twice as high as those of the other PILs.The major difference was resulted from the presence of double acidic sites of H2SO4.Compared with the reported PILs,our PILs showed much higher AD (See Table S5).

We further conducted the Hammett acidity determination of PEI(70000)-PS-acid PILs to investigate the influence of the treated acid type for the acidification of PEI-PS on the acidity of PILs through the Hammett method,and the results were shown in Fig.8 and Table 2.It can be seen the order follows PEI(70000)-PS-H2SO4>PEI(70000)-PS-p-TSA ≈PEI(70000)-PS-CF3SO3H >PEI(70000)-PS-CH3SO3H >H2SO4,suggesting that PEI(70000)-PSH2SO4showed a much higher acidity than other acid treated PILs and even H2SO4.This could be explained by the higher ability to dissociate H+of H2SO4than other organic acids.

With the ultrahigh acidity in mind,we then investigated the influence of treated acid for the acidification of PEI-PS procedureon their catalytic activity.Reaction conditions of a methanol/oleic acid molar ratio of 10:1 and a low catalyst dosage of 2.5% (mass)at 50 °C for 0.5 h were applied.As shown in Fig.9,the sequence of catalytic activity was recorded as follows:PEI(70000)-PS-p-TSA >PEI(70000)-PS-CF3SO3H >PEI(70000)-PS-H2SO4>PEI(70000)-PS-CH3SO3H,which,however,did not well follow the orders of the acid density and acidity evaluated in Figs.7 and 8.These very different results demonstrated that the acid density and acidity were not the only two main factors influencing the esterification herein.After a careful review on the chemical properties of acids utilized in this study,we inferred that miscibility of acid in the reaction medium was another important factor needing to be taken account[38,39].The more catalysts remained in the organic phase;the higher conversion of oleic acid could be achieved.As an organic acids,a smaller amount of organic acids such as p-TSA,CF3SO3H or CH3SO3H was brought into the aqueous phase compared with that of H2SO4.As such,anion [p-TSA]–from the PEI(70000)-PS-p-TSA had higher miscibility in oleic acid than others,resulting in a higher activity than PEI(70000)-PS-H2SO4.Therefore,besides the molecular weight of the PEI matrix,the catalytic performance of our PILs should also be pre-determined through estimating the miscibility and acidity of the treated acids.We also compared our PILs to the commercially used Amberlys-15 resin(the chemical&physical characteristics of Amberlys-15 resin were shown in Table S6).Even though the Amberlys-15 resin showed a much higher AD (4.73 mmolg-1) compared to our PILs(the ADs were ranged from 3.50–3.95 mmolg-1),but much lower catalytic efficiency (29.1% vs.55.1% at reaction conditions of a methanol/oleic acid molar ratio of 10:1 and a low catalyst dosage of 2.5% (mass) at 50 °C for 0.5 h),confirming again the advantage of our PILs.

Table 1Hammett function parameters of PEI(X)-PS-H2SO4 PILs in water (5 mmolL–1) at 25 °C

Table 2Hammett function parameters of PEI(70000)-PS-acid PILs in water (5 mmolL–1) at 25 °C

Fig.5.Catalytic performance of PEI(X)-PS-H2SO4 PILs for the conversion of oleic acid with methanol.

Fig.6.FTIR spectra of PEI(70000)-PS-acid PILs.

Fig.7.AD values of PEI(70000)-PS-acid PILs.

Fig.8.The UV absorption spectra of PEI(70000)-PS-acid PILs.

Fig.9.Catalytic performance of PEI(70000)-PS-acid PILs for the conversion of oleic acid to methyl oleate.

3.3.Single factor optimization

After the careful evaluation above,the PEI(70000)-PS-p-TSA,which showed the best catalytic activity,was selected for further esterification optimization investigation.A series of influence-factor experiments were performed according to the single-factor design to obtain the optimal conditions for biodiesel production by the esterification of oleic acid.The effects of reaction time,reaction temperature,catalyst dosage and molar ratio of reactants on oleic acid conversion using PEI(70000)-PS-p-TSA as a catalyst were investigated and the results were displayed in Fig.10.

Specifically,the conversion increased initially with the reaction time(Fig.10(a)),and reached equilibrium at 1.5 h with a yield over 95.6% due to gradually occupation of reactive sites.An increase of conversion could be seen with an increase of reaction temperature(Fig.10(b)),but a further increase of the conversion could not be significantly acquired after the temperature reached 70 °C.The conversion was also controlled by the catalyst dosage (Fig.10(c)),which showed an increase with the dosage increase.Considering the cost and the reaction rate,2.5% (mass) of the dosage was applied as the optimal.Note that this amount of our PEI(70000)-PS-p-TSA,which was required by the high conversion,was considerably lower than that of other reported catalyst,which was due to the rich number of catalytic sites and the high acidity available on the PILs.As for the molar ratio of reactants(Fig.10(d)),the conversion also showed a gradual increase with the increase of methanol.This was because the excess of reactants could increase the rate of conversion and drive the reaction equilibrium towards the products.

Balancing the trade-off of cost and efficiency,10:1 was considered as the optimal ratio.Therefore,a reaction time of 1.5 h,a reaction temperature of 70 °C,a catalyst dosage of 2.5% (mass) and methanol/oleic acid(molar ratio)of 10:1 were selected as the optimal reaction conditions for as-prepared PEI(70000)-PS-p-TSA.Notes that due to the overloading of PS on the amine (with mole ratio of 3:1),the PIL precursors also showed Bronsted acidity,which could be applied as catalysts for the biodiesel production(ca.85.1% at condition of catalyst dosage of 5% (mass),methanol/oleic acid of 10:1,reaction temperature of 70 °C,reaction time of 1 h).However,a much higher conversion of ca.95.0% could be achieved by the PILs after p-TSA acid addition under a much milder condition of catalyst dosage of 2.5% (mass),methanol/oleic acid of 10:1,reaction temperature of 70°C,reaction time of 1.0 h,showing a much promising advantage of acid addition on the further improvement of the catalytic performance of PILs.Under these conditions,the conversion of oleic acid reached 95.6%,primarily showing the excellent catalytic potential as a promising esterification catalyst of PEI(70000)-PS-p-TSA.

3.4.Orthogonal optimization

In the single factor analysis,only one influence factor could be considered,thus the optimized results may not be the most optimized ones when the results were influenced by multi factors.While multi factors could be considered comprehensively in the Box-BEHNKEN design method to acquire results with higher efficiency and accuracy.Thus,the single factor analysis method was used at first to have a quick scan on the influence of the factors on the conversion to determine which factors,with higher impact,need to be considered in the Box-BEHNKEN design.With which,the Box-BEHNKEN design method was then used to further improve the efficiency and accuracy of the optimized results.Therefore,based on the results of single-factor experiments,an orthogonal experiment was used to further optimize the reaction conditions using PEI(70000)-PS-p-TSA as a catalyst for the best conditions for the biodiesel production,and the relationship between the variables and the response was studied.The coded values,test values and predicted values of oleic acid conversion at each test point were shown in Table 3,and the standard deviation and relative deviation of each set of experimental data.There was no significant difference between the experimental data and the predicted results.According to the experimental design results,a second-order multiple regression model of oleic acid conversion is given in Eq.(5):where Y is oleic acid conversion rate;X1,X2,and X3are the coded values of variables of the reaction time,methanol to oleic acid molar ratio,and catalyst amount.

Fig.10.Influence of reaction conditions including reaction time (a;70 °C,catalyst dosage of 2.5%(mass),molar ratio of methanol/oleic acid of 10:1),temperature (b;1.5 h,catalyst dosage of 2.5% (mass),molar ratio of methanol/oleic acid of 10:1),catalyst dosage (c;1.5 h,70 °C,molar ratio of methanol/oleic acid of 10:1) and molar ratio of methanol/oleic acid (d;1.5 h,70 °C,catalyst dosage of 2.5% (mass)) on the conversion of oleic acid using PEI(70000)-PS-p-TSA as the catalyst.

Table 3Experimental and predicted response values by Box-BEHNKEN design

The second-order response surface model and linear regression equations were fitted by analysis of variance(ANOVA).The results are shown in Table S7.The F-value of model was 524.14 which demonstrates the model was significant.And the p-values of model were less than 0.0500,which illustrated the model terms were significant.It can be seen from the analysis of ANOVA that,since p-value was less than 0.0001,the linear terms X1,X2,X3and the quadratic termswere the key terms that affected the conversion of oleic acid.The “Pre R2” of 0.986 was in reasonable consistency with the“Adj R2”of 0.997,which indicated that the model was significant.The p-values of “l(fā)ack of fit” was greater than 0.05,which illustrated the model agrees with the fitted results and the absence of a fit was not significant.

It can be seen from the above simulation results that the model was suitable for the calculation and optimization of oleic acid conversion in the study variable range.The 3D response diagram generated by the model was shown in Fig.11.Fig.11(a) showed the effect of the molar ratio of methanol to oleic acid and the amount of catalyst on the yield of biodiesel when the reaction time was 1.5 h.When the enough large amount of catalyst was applied,the yield of biodiesel gradually increased and then decreased with the increased of the molar ratio of methanol to oleic acid.This is because increasing the molar ratio of methanol to oleic acid could provide more reactants to drive the reaction to the right side.However,excessive methanol would decrease the yield of biodiesel due to the dilution of catalysts.The interaction of the molar ratio of methanol to oleic acid and reaction time was displayed in Fig.11(b).It could be found that the conversion of oleic acid increased with increasing reaction time,and reached an equilibrium afterwards.Fig.11(c)showed the relationship among the amount of catalyst,the reaction time and oleic acid conversion.With the increase of the reaction time and the amount of catalyst,the conversion of oleic acid showed an increase trend because more catalyst dosages could provide more activated sites for the catalytic reaction.

Fig.11.Response surface graphs for conversion of oleic acid under the influence of(a) molar ratio and catalyst dosage;(b) molar ratio and reaction time;(c) catalyst dosage and reaction time.

Based on analysis above,the optimal conversion efficiency for oleic acid of 99.05% could be reached under the reaction time of 2.0 h,the methanol/oleic acid molar ratio of 14:1,and the catalyst amount of 3.15%(mass)according to the Eq.(5).Accordingly,three identification experiments were performed under these recommended conditions,and the experiment showed that the ultrahigh oleic acid conversion of 98.42%could be acquired with the convincible standard deviation of 0.72.

3.5.Catalytic performance comparison

The results of the conversion of different catalysts for the biodiesel production via the esterification of oleic acid by methanol were shown in Table S8.Compared with conventional homogeneous acidic catalysts (e.g.H2SO4[40] and [BMIM] [HSO4] [7]),PEI(70000)-PS-p-TSA showed a very competent catalytic activity even under much milder conditions (i.e.much lower temperature,smaller molar ratio and shorter reaction time).More notably,PEI(70000)-PS-p-TSA,as a heterogeneous catalyst,had a good recyclability during use.Compared with heterogeneous catalysts reported in the literature[20,37,41–44],PEI(70000)-PS-p-TSA showed much better catalytic performance under milder conditions.Therefore,PEI(70000)-PS-p-TSA,showing low cost and excellent catalytic activity,can be utilized as an excellent heterogeneous catalyst for biological production.

3.6.Catalytic application generality

To verify the universal applicability of PEI(70000)-PS-p-TSA in biodiesel production,more raw materials such as stearic acid,palmitic acid,n-decanoic acid,lauric acid,and n-octanoic acid were applied under the same experimental conditions as above.The results are shown in Table 4.The experimental results showed that PEI(70000)-PS-p-TSA had good catalytic activities in the esterification of FFAs with different chain lengths to biodiesel,and the conversion of free fatty acids were all greater than 90%.At the same time,the catalytic activity of PEI(70000)-PS-p-TSA for esterification of different FFAs was very competitive to those of homogeneous catalysts.It was also considerably higher than that of heterogeneous catalysts,even under milder conditions (i.e.much lower temperature,smaller catalyst dosage and shorter reaction time).It is worth noting that compared with homogeneous catalysts,our PILs showed higher thermostability (see Fig.S3) and can beeasily separated from the reaction medium(Fig.S4).Thus,we proposed that the PILs we designed in this study has good application prospects from both an economic perspective and an innovation perspective.

Table 4Conversion comparison of different catalysts

4.Conclusions

In summary,this study reported a series of PIL catalysts for biodiesel production.This novel series of PIL catalysts,with high density of active sites,could be acquired in a facile and low-cost manner by using a radical polymerization of PEI and PS,followed by treating with acids.The PILs showed tunable acidity activity,in which both the acidity and catalytic activities could be enhanced with the increase of the molecular weight of the PEI matrix.Owing to the high acidity and appropriate miscibility of treated acid with reactants,PEI(70000)-PS-p-TSA showed the most excellent catalytic activity for the esterification between oleic acid and methanol among these PILs.An ultrahigh biodiesel yield of 98.42% was thus reached under the optimal conditions of reaction time of 2.0 h,catalyst dosage of 3.15%(mass),and alcohol/acid molar ratio of 14:1.Moreover,high generality as an efficient catalyst for esterification of other FFAs was observed by our PEI(70000)-PS-p-TSA,with general high conversion over 90% for each case even under much milder conditions compared to other conventional catalysts.As such,our multiple acidic-site modified PEI(70000)-PS-p-TSA exhibited novel features of low-cost,easy tuning,efficient catalysis,and high generality,which could be exploited as a very promising catalyst for biodiesel production in next generation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We acknowledged the National Natural Science Foundation of China (21878054),Project on the Integration of Industry and Education of Fujian Province (2018Y4008),Science and Technology Project of Fujian Educational Committee (JAT190051),Fuzhou University Testing Fund of precious apparatus (2020T008) and Research Initiation Funding of Fuzhou University (GXRC-19051).X.Lin also acknowledged the Award Program for Minjiang Scholar Professorship.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.04.006.