Optimization of Process Variables in the Synthesis of Tributyl Citrate Using a Polyvinylpolypyrrolidone-Supported Brønsted Acidic Ionic Liquid Catalyst

A polyvinylpolypyrrolidone(PVPP-) supported Brønsted acidic ionic liquid catalyst ([BsPVPP]HSO4) was synthesized by the reaction between SO3H-functionalized PVPP and H2SO4. The prepared catalyst was characterized by IR, XRD, FESEM, TG, and DSC. The catalytic activity of [BsPVPP]HSO4 in the preparation of tributyl citrate (TBC) by the esterification reaction between citric acid and n-butanol was investigated. Response surface methodology (RSM) was applied to optimize the process variables of the esterification reaction. The variables, including the reaction time, the n-butanol-to-citric acid mole ratio, the reaction temperature, and the catalyst amount, were optimized by a Box-Behnken design. Under optimized conditions, with a n-butanolto-citric acid mole ratio of 5.2 : 1 and a reaction temperature of 120C, the TBC yield reached 92.9% within 5.5 h in the presence of 6.6 wt% of catalyst; this result is in good agreement with the values predicted by the mathematical model. Moreover, the catalyst could be recycled four times with high catalytic activity.


Introduction
Tributyl citrate (TBC) is usually applied as a thermally stable, high-performance, and nontoxic plasticizer in toys, medical products, biodegradable polymers, and food additives [1][2][3][4].Moreover, TBC is a biorenewable and sustainable plasticizer.This is due to the fact that the raw materials applied in the preparation of TBC can be gotten from the fermentation processes using renewable resources [5,6].Thus, the preparation of TBC has drawn much attention.Traditionally, the catalyst applied in the preparation of TBC is sulfuric acid, a homogeneous catalyst that has some disadvantages such as high by-product generation, equipment corrosion, and environmental pollutions.To solve the above problems, the heterogeneous acid catalysts that offer easy separation, high catalytic activity, and environmental friendliness were developed for TBC preparation.Metal oxide and zeolitesupported catalysts have attracted more and more attention [7][8][9][10].Nevertheless, these heterogeneous catalysts have some disadvantages, including low product selectivity, high mass transfer resistance, and rapid deactivation during the esterification reaction.
Because of their excellent thermal stability, negligible volatility, and tunable acidity, acidic ionic liquids (ILs) have been utilized as environmental friendly reagents for organic reactions [11][12][13][14][15][16][17].However, ILs have been proved to have some disadvantages, including poor reusability, high viscosity, and large amount requirement for the reaction [18][19][20].Heterogeneous polymer-supported catalysts had been developed to solve these problems because of their environmentally friendly characteristics, such as excellent chemical stability, efficient recover ability, and recyclability [21][22][23][24].These properties make polymer-supported ILs suitable candidates for catalytic reactions.Polyvinylpolypyrrolidone (PVPP), a commercialized cross-linked polymer adsorbent, has been used as a polymer support since it has many advantages, including being nontoxic, high adsorption capacity, and being insoluble in all types of solvents [25,26].Moreover, it has been demonstrated that PVPP-supported catalysts have high catalytic ability and reusability in organic reactions [27][28][29].
Response surface methodology (RSM) is a frequently used mathematical method which can be applied to find the optimum reaction conditions by investigating the relationships between explanatory and response variables [30][31][32].Therefore, in the present work, a polymer-supported Brønsted acidic IL catalyst, [BsPVPP]HSO 4 , was applied in the preparation of TBC by the esterification of n-butanol with citric acid for the first time.The effect of reaction conditions on the citric acid conversion and TBC yield was investigated thoroughly (single factor).Based on the results of the single-factor investigation, the reaction conditions were further optimized to increase the TBC yield by applying RSM, and then the optimum reaction conditions were confirmed (multifactors).Additionally, the recyclability of [BsPVPP]HSO 4 was also discussed.[BsPVPP]HSO 4 was prepared as reported in previous publications with a little modification [28,33].In a typical preparation procedure, PVPP powder (10 g, monomer molar quantity was 0.09 mol) was put into toluene under vigorous stirring, and after a while, BS (6.8 g, 0.05 mol) was introduced into the mixture.The mixture was heated to 80 °C and stirred in a 100 mL flask for 24 h continuously.The resulting solid zwitterion was recovered by filtration and washed with ethyl acetate to completely remove the unreacted reactant and then dried at 60 °C under vacuum overnight.Then, sulfuric acid (5.0 g, 0.05 mol) was added to the respective zwitterion dispersed in ethyl acetate at 0 °C and stirred for 24 h.After filtration, the residue was first washed three times with methanol and then dried under vacuum overnight.

2.2.2.
Preparation of 1-Sulfobutyl-3-methylimidazolium p-Toluenesulfonate ([BsMIM]PTSA).Methylimidazole (9.1 g, 0.11 mol) and BS (13.6 g, 0.10 mol) were added into a 100 mL round-bottom flask, where ethyl acetate was added later as a solvent.This flask was magnetically stirred at 40 °C for 24 h until a solid zwitterion formed which was filtered, washed three times with diethyl ether to remove nonionic components, and then dried in vacuum.Then, p-toluenesulfonic acid (PTSA, 17.2 g, 0.1 mol) was introduced to the respective zwitterion dissolved in water and stirred at 80 °C for 8 h.Finally, the product was obtained by vacuum drying.

Catalyst Characterization.
The loading of the IL on the PVPP support was determined by elution experiment (supporting information).IR investigations were carried out on a Prestige-21 FTIR absorption spectrometer using KBr optics (Figure S1).Thermal decomposition of catalyst was investigated by a TA Q50 thermogravimetric instrument in N 2 with a heating rate of 10 °C/min (Figure S2).Morphologies of [BsPVPP]HSO 4 and PVPP were observed by a Hitachi SU8000 field emission scanning electron microscope equipped with energy-dispersive spectroscope (Figure S3).The element analysis of the catalyst was investigated via energy-dispersive spectroscope (Table S1), which has been proved to be a useful method to testify the chemical composition of the catalysts [34,35].The XRD measurement was conducted by a Shimadzu XRD-6000 apparatus with a scanning speed of 5 °•min −1 (Figure S4).The differential scanning calorimetry (DSC) examination was performed using a TA Q20 instrument with a heating rate of 5 °C/min in the temperature range of −80 °C to 200 °C as reported in literature [36] (Figure S5).NMR measurements were conducted on a Bruker AV 500 spectrometer (supporting information).

Esterification of Citric Acid with n-Butanol and Analysis.
Preparation of TBC by the esterification reaction of citric acid with n-butanol was performed in a 100 mL 3-necked flask equipped with a thermometer, a magnetic stirrer, and a segregator coupled with a reflux condenser at normal pressure (Scheme 1).
The typical reaction was applied as follows: citric acid (12.6 g, 0.06 mol), n-butanol (22.2 g, 0.3 mol), and [BsPVPP]HSO 4 (2.1 g, 6 wt% of raw material) were introduced into the flask.The reaction mixture was heated at 115 °C with vigorous stirring for 5 h.The conditions of the reaction system before and after the centrifugation separation were shown in Figure S6.After the reaction was completed, the product was investigated by gas chromatography (Agilent GC-7890A) using a Thermo TG-SQC capillary column (15 m × 0.25 mm × 0.25 μm), which equipped with a flame ionization detector (FID).Acid value was determined by titration according to the procedures described in the literature [29].Citric acid conversion and the TBC yield were calculated using the following equations [9]: actual amount of TBC analyzed by GC theoretical amount of TBC calculated f rom reactants × 100%, citric acid conversion = acid value bef ore reactionacid value af ter reaction acid value bef ore reaction × 100% 1 International Journal of Polymer Science 2.5.Response Surface Methodology.To optimize the reaction conditions for [BsPVPP]HSO 4 -catalyzed esterification reaction of n-butanol with citric acid, a three-level and four-factorial Box-Behnken design was designed to investigate the influences of the reaction conditions on the TBC yield.Esterification reactions using various molar ratios of n-butanol to citric acid (4.5-5.5),reaction temperatures (110-126 °C), catalyst amounts (4%-8%), and reaction times (4-6 h) were carefully investigated.Not only the variables and the levels of them but also the real and coded values are shown in Table 1.TBC yield was the response value of the designed experiment.
The relationship between the TBC yield and reaction factors was expressed using the following quadratic equation (2): where Y is the TBC yield; X 1 to X 4 are corresponding parameters shown in Table 1; the meaning of other parameters are clearly shown in literature [30].All of these parameters were obtained from the whole experimental results.The fitted equation was always expressed as response surface plots to demonstrate the relationship between the response and experimental levels of each variable.By this way, the optimal reaction conditions can be confirmed.Meanwhile, the coefficient of determination, R 2 , was usually applied to determine the general ability and accuracy of the multiple regression models.Stat-Ease Design Expert (Version 8. 0. 6, Stat-Ease Inc., USA) was used to establish the mathematical model, generate the three-dimensional response surface, and determine the optimal reaction conditions.

Comparison of the Catalytic Activities of Different
Catalysts.The catalytic performance of different catalysts for the esterification reaction of n-butanol with citric acid is presented in Table 2.
Based on the proton acid catalytic mechanism (Scheme S1), supported IL catalysts can catalyze the esterification reactions as efficient as inorganic acid and acidic IL [8,37,38].The catalytic ability of inorganic acid, acidic IL, and [BsPVPP]HSO 4 was compared.It can be observed from Table 2 that, without using catalysts, the TBC yield only reached 12.5% (entry 1).When H 2 SO 4 and PTSA were used as catalysts in esterification, homogeneous reaction systems were formed and all exhibited a high citric acid conversion, above 96%, and TBC yield, above 91% (entries 2 and 3).However, the homogeneous reaction systems made the separation of H 2 SO 4 and PTSA from the product difficult.When [BsMIM]PTSA was used as a catalyst, a liquid-liquid heterogeneous system was formed after the reaction.The citric acid conversion was 96.1%, and the TBC yield was 91.9% (entry 4).However, [BsMIM]PTSA was easily attached to the bottom of the flask like a gelatinous liquid, which can greatly reduce its reusability.[BsPVPP]HSO 4 is a heterogeneous catalyst, when it was used, a liquid-solid reaction system formed and a high citric acid conversion, 96.9%, and good TBC yield, 92.1%, were obtained (entry 5).Moreover, [BsPVPP]HSO 4 could be easily separated by vacuum filtration and then reused.Thus, [BsPVPP]HSO 4 was used as the catalyst to optimize various reaction parameters in the esterification of n-butanol with citric acid.3 International Journal of Polymer Science TBC yield and the citric acid conversion.Figure 1 shows the effect of the mole ratio of n-butanol to citric acid on the TBC yield and the citric acid conversion.It can be observed that the TBC yield and the citric acid conversion demonstrated a marked improvement with the increase of the molar ratio of n-butanol to citric acid from 4 : 1 to 5 : 1, and the maximum conversion of 96.9% and yield of 92.1% were achieved with a 5 : 1 molar ratio.This means that the excess nbutanol led to the shifting of the equilibrium towards the product side.Whereas, a further increase of the n-butanolto-citric acid mole ratio did not lead to a remarkable increase in the yield or conversion, most likely because the excessive n-butanol can dilute the concentration of citric acid and [BsPVPP]HSO 4 too much.This phenomenon was in good agreement with other reported literature [9].Taking the TBC yield and the cost of n-butanol into account, the optimal molar ratio of n-butanol to citric acid was chosen to be 5 : 1.

Effect of the Amount of [BsPVPP]HSO 4 .
The effect of the catalyst amount of [BsPVPP]HSO 4 on its catalytic activity is illustrated in Figure 2. It can be observed that both the TBC yield and the citric acid conversion changed with increase in the amount of [BsPVPP]HSO 4 ; a citric acid conversion rate of 96.9% and a TBC yield of 92.1% were obtained in the presence of 6% of [BsPVPP]HSO 4 within 5 h of the reaction.However, citric acid conversion and the TBC yield changed slightly when the catalyst amount was further increased.This phenomenon might be caused by the limitation of the chemical reaction equilibrium, as reported in literature [31].Considering citric acid conversion, the TBC yield, Figure 1: Effect of the n-butanol to citric acid molar ratio on the esterification of citric acid with n-butanol.Reaction conditions: citric acid, 0.06 mol; n-butanol, 0.3 mol; catalyst, 6 wt%; temperature, 115 °C; time, 5 h.3.4.Effect of the Reaction Temperature.Since high reaction temperature always accelerates the reaction rate and enhances the conversion of reagents, temperature is an important parameter for the esterification reaction and the effect of temperature must be investigated.Figure 3 depicts the changes of the TBC yield and the citric acid conversion at different temperatures.The results indicate that both the citric acid conversion and the TBC yield increased when the temperature increased from 105 °C to 115 °C.The relatively high citric acid conversion of 96.9% and TBC yield of 92.1% were achieved at 115 °C.However, when the reaction temperature was higher than 115 °C, the TBC yield did not increase obviously.Thus, the optimum reaction temperature for the esterification reaction of n-butanol with citric acid was 115 °C.
3.5.Effect of the Reaction Time.The effect of the reaction time on the TBC yield and the citric acid conversion is presented in Figure 4.It can be observed that the TBC yield increased with the reaction time, that is, when the reaction time increased from 3 h to 5 h, the TBC yield increased from less than 85% to 92.1%.However, when the reaction time was more than 5 h, the TBC yield decreased slightly.Therefore, the optimum reaction time was 5 h for this experiment.5 International Journal of Polymer Science conditions, fully considering multifactor interactions.Using the independent variables and the corresponding levels for the Box-Behnken design shown in Table 1, the experimental results are presented in Table 3.
The following regression equation was generated from the application of RSM.It is an empirical relationship between the response variable, TBC yield, and the tested variables: R adj 2 = 0 97 5 The regression result also shows that the value of the regression parameter, R 2 , is close to 1.This result indicates that there is a good fitness between the predicted values and experimental data.Thus, using this model to predict the TBC yield is reliable.
The significant test for each regression parameter involved in (3) and the variable analysis are presented in Table 4.The detailed three-dimensional response diagrams obtained from the predicted models are shown in Figure 5.As seen from Table 4, the P value (<0.0001) of the model is very low, which indicates that the regression is statistically significant.Additionally, the P values of X 1 (P < 0 0001), X 2 (P < 0 0001), X 3 (P = 0 0030), and X 4 (P = 0 0087) were all lower than 0.05, indicating the high significance of (3).This result also showed that the n-butanol-to-citric acid molar ratio and reaction temperature have a highly significant effect on the TBC yield.Moreover, the interactions between the factors of X 1 X 2 (P = 0 0009) and X 3 X 4 (P = 0 0176) were significant, that is, the interactions, including the molar ratio of n-butanol to citric acid-reaction temperature and amount of catalyst-reaction time, had a noticeable influence on the TBC yield.
Optimum reaction conditions can also be deduced from (3) by means of analyzing the multiparameter regression.Optimum parameters for the esterification reaction of nbutanol with citric acid were as follows: a n-butanol to citric acid mole ratio of 5.19 : 1, reaction temperature of 121.00 °C, catalyst amount of 6.60 wt%, and reaction time of 5.32 h.A maximum TBC yield of 93.2% was obtained with this set of optimum parameters.Considering the predicted value and actual experimental conditions, a n-butanol to citric acid

Conclusion
A polymer-supported Brønsted acidic IL catalyst, [BsPVPP]HSO 4 , was synthesized and used for the esterification of n-butanol with citric acid.To optimize and study the effect of the process parameters, including the reactant mole ratio, reaction temperature, catalyst amount, and reaction time, on the TBC yield and the citric acid conversion, RSM was used based on single-factor experiments.According to the optimization results obtained with RSM, the optimum esterification conditions of n-butanol with citric acid are the n-butanol-to-citric acid mole ratio of 5.2 : 1, reaction time of 5.5 h, reaction temperature of 120 °C, and catalyst amount of 6.6 wt%.Under the optimized conditions, citric acid conversion and the TBC yield reached 97.5% and 92.9%, respectively.Moreover, [BsPVPP]HSO 4 exhibited excellent reusability.After four cycles of recycling, both citric acid conversion and the TBC yield decreased by no more than 3%.Thus, [BsPVPP]HSO 4 is an attractive candidate for industrial TBC synthesis.

Figure 5 :
Figure 5: Response surface plots showing the predicted values of the TBC yield: effect of the (a) n-butanol-to-citric acid molar ratio and temperature, (b) n-butanol-to-citric acid molar ratio and amount of catalyst, (c) n-butanol-to-citric acid molar ratio and reaction time, (d) temperature and amount of catalyst, (e) temperature and reaction time, and (f) amount of catalyst and reaction time on the TBC yield.Other variables were kept at their central level.

Table 1 :
Coded values of the variables used in the experimental design.

Table 2 :
Esterification of citric acid with n-butanol using different catalysts.

Table 3 :
Experimental design and the results of the Box-Behnken design for the optimization process.

Table 4 :
Significance of the coefficient and the corresponding statistical analysis.
4was washed with diethyl ether and vacuum heated for 2 h at 80 °C.As seen in Figure6, the citric acid conversion changed from 97.5% to 95.3% and the TBC yield changed from 92.9% to 90.1% after four cycles of recycling.Thus, the good recyclability reveals that [BsPVPP]HSO 4 is very stable for the esterification reaction of n-butanol with citric acid.