Study on the coupling of esterification with pervaporation

Study on the coupling of esterification with pervaporation

Journal of Membrane Science 182 (2001) 173–181 Study on the coupling of esterification with pervaporation Qinglin Liu a,∗ , Zhibing Zhang b , Hongfan...

190KB Sizes 2 Downloads 327 Views

Journal of Membrane Science 182 (2001) 173–181

Study on the coupling of esterification with pervaporation Qinglin Liu a,∗ , Zhibing Zhang b , Hongfang Chen b a b

Department of Chemical Engineering, Xiamen University, Xiamen 361005, China Department of Chemical Engineering, Tianjin University, Tianjin 300072, China

Received 6 April 2000; received in revised form 28 August 2000; accepted 29 August 2000

Abstract The separation characteristics of the crosslinked polyvinyl alcohol (PVA) membranes prepared in our laboratory were studied by pervaporation separation of the liquid mixtures of both water/acetic acid and water/acetic acid/n-butanol/butyl acetate. The permeation fluxes of water and acetic acid as a function of compositions were presented. The esterification of acetic acid with n-butanol catalyzed by Zr(SO4 )·4H2 O was carried out at a temperature range of 60–90◦ C. A kinetic model equation was developed for the esterification; then, it was taken as a model reaction to study the coupling of pervaporation with esterification. Experiments were conducted to investigate the effects of several operating parameters, such as reaction temperature, initial molar ratio of acetic acid to n-butanol, ratio of the membrane area to the reacting mixture volume and catalyst concentration, on the coupling process. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Composite membranes; Pervaporation; Organic separations; Pervaporation-aided esterification; Liquid permeability and separations

1. Introduction The use of membrane in a reactor for selective extraction of gaseous products resulting in favorable chemical equilibrium shift has been recognized. The removal by pervaporation of one reaction product to shift a liquid phase reaction has gained considerable interest in recent years [1–9]. Pervaporation esterification coupling has attracted, up to now, most attention. This can be explained by the availability of hydrophilic membranes which allows selective removal of the produced water, and by the fact that esterification is one of the most important chemical processes in organic chemical industry. There are at least four components in the liquid mixture during esterification, and the permeation flux of each component cannot be described by existing model equations ∗ Corresponding author. E-mail address: [email protected] (Q. Liu).

because of the strong interaction between components. In general, such a problem can be solved with some simplifications. For example, during the esterification of 1-propanol with propionic acid, there was only water permeating through the GFT composite membrane, and water permeation flux was proportional to the water content of less than 10 wt.% [10,11] in the reacting mixture. But, in most cases, both water and other components permeated simultaneously through the pervaporation membranes, and water permeation flux depends not only on water content but also on the concentration of other components in the system. In order to study the separate properties of the PVA/ceramic composite membranes prepared in our laboratory, the pervaporation separation of liquid mixtures of water/acetic acid and water/acetic acid/n-butanol/butyl acetate was carried out. Permeation flux expressions of water and acetic acid were thus developed. Then, the esterification of acetic acid with n-butanol in the presence of Zr(SO4 )2 coupling

0376-7388/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 6 - 7 3 8 8 ( 0 0 ) 0 0 5 6 8 - 8

174

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

Nomenclature denote acetic acid, n-butanol, butyl acetate and water, respectively Ci (i=A, B, E, W) compositions of component i. catalyst concentration Ccat C0 initial concentration of reactant (mol l−1 ) denote compositions of Ci0 Ci00 component i in feed liquid mixture and permeate (mol l−1 ) diffusion coefficient of Di component i in membrane permeation flux of component Ji i (mol m−2 h−1 ) k1 , k2 denote reaction rate constants of forward and backward reaction, respectively (mol−1 l−1 h−1 ) Pi permeate coefficient of component i in membrane (l m−2 h−1 ) constant as a function of PW , PA temperature S membrane area (m2 ) correction factors for water SW,B , SW,E permeation flux when there are n-butanol and butyl acetate in the liquid mixture correction factors for acetic SA,B , SA,E acid permeation flux when there are n-butanol and butyl acetate in the liquid mixture T temperature (K) V volume of the liquid mixture (ml)

acting mixture volume and the catalyst concentration, on the coupling process.

A, B, E and W

Greek letter δ effective depth of membrane (m)

with pervaporation was conducted to investigate the effect of the major operating parameters, such as, the temperature, the initial molar ratio of acetic acid to n-butanol, the ratio of the membrane area to the re-

2. Theoretical development According to Fick’s law in the frame of solution– diffusion mechanism, the permeation flux can be written dCi Ji = Di (1) dz If Di is assumed to be a constant, the above equation can be integrated into Ji = Di

Ci0 − Ci00 δ

(2)

Generally, Ci0  Ci00 , and the Eq. (2) can be simplified as Ci0 = Pi Ci0 (3) δ The variables involved in the above equations are defined in the nomenclature section. Water permeation flux can be calculated using Eq. (3) in the case of pervaporation of water/acetic acid mixture over water concentration of less than 10 wt.%. While during pervaporation of quaternary mixture of water/acetic acid/n-butanol/butyl acetate over the same water concentration range in the feed as that of water/acetic acid mixture, water permeation flux could not be directly obtained by Eq. (3) due to the interaction between water and n-butanol (or/and butyl acetate). As to the esterification, a general form can be written Ji = Di

A+B↔E+W

(4)

The rate of water production for the reaction can be written dCE (5) = k1 CA CB Ccat − k2 CE CW Ccat dt It can be concluded from the Eqs. (3) and (5) that the major influencing factors in the coupling of the esterification with pervaporation are (1) Temperature; (2) initial molar ratio of acetic acid to n-butanol; (3) ratio of the membrane area to reacting mixture volume; (4) catalyst concentration.

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

As for the coupling process, the ratio of rate of water removal to water production is found to be the key factor that represents a coupling factor being described as

175

(6)

into the flask immersed in a thermostatic water bath; and then, acetic acid was added to the reactor to start the reaction. Samples were withdrawn from the reactor at regular intervals and put into refrigerator immediately to prevent further reaction.

Here, F is a dimensionless parameter that stands for the interaction between water removal and water production during the coupling process, and when

3.2. Separation of liquid mixtures of water/acetic acid and water/acetic acid/n-butanol/butyl acetate by pervaporation

F =

JW × S/V dCE /dt

1. F = 1, the rate of water removal is equal to the water production rate, indicating that the conversion of the reaction could attain 100%; 2. F < 1, the rate of water removal is less than the water production rate, indicating that the conversion could be enhanced a little over the equilibrium conversion, and it is controlled by the water removal; 3. F > 1, the rate of water removal is larger than the water production rate, indicating that the conversion could attain 100%, and it is limited by the water production rate. Among the variables used in the above equations, the known quantities were: k1 and k2 which could be obtained from the esterification data at various temperatures. JW and dCE /dt were available from the pervaporation-aided esterification experiments. 3. Experimental 3.1. Heterogeneous catalytic reaction The esterification was carried out in a 250 ml round flask fitted with a reflux condenser. The experimental procedure was: n-butanol and Zr(SO4 )2 were placed

The schematic diagram of the pervaporation apparatus is shown in Fig. 1. The preparation of the PVA/ceramic composite membrane was in a similar way as that in reference [9,12]. The permeation cell was made of stainless steel and the PVA composite membrane with an effective area of 34 cm2 was placed on a porous stainless support plate. Pervaporation separations were performed with the mixtures of water/acetic acid and water/acetic acid/n-butanol/butyl acetate in a continuous mode with the permeate side at a reduced pressure of no more than 5 mm Hg. The samples taken from the permeate that was collected by turn in two liquid nitrogen traps were analyzed by gas chromatography (GC). 3.3. Pervaporation-aided esterification The schematic diagram of the experimental apparatus is the same as that in Section 3.2 and the procedure of the experiments was similar to that of the pervaporation process described above, while they were conducted in a batch mode to investigate the influence of the four operating parameters on F over varied compositions. Samples were withdrawn

Fig. 1. Diagram of esterification coupling with pervaporation: (1) reaction cell, (2) membrane, (3) porous support steel plate, (4) stirrer, (5) thermometer, (6) inlet of thermostatic water, (7) outlet of thermostatic water, (8) three way cock, (9) vent to atmosphere, (10) cold trap, (11) two way cock, (12) barometer, (13) buffer tank, (14) drying column, (15) vacuum pump.

176

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

simultaneously from both the reactor and the liquid nitrogen traps once for 1 h and put into a refrigerator immediately for analysis by GC.

 k2 = 4.376 × 106 exp

−7.09 T

 (8)

4.2. Pervaporation 4. Results and discussion 4.1. Heterogeneous catalytic reaction External diffusional resistance may affect the heterogeneous reaction greatly at certain circumstance but its negative effect on the reaction can be reduced or eliminated with stirring. The esterification reactions were conducted over stirrer speeds ranging from 200 to 500 rpm at the same reaction condition as follows: CA0 = 8.74 mol l−1 , CB0 = 5.47 mol l−1 , V = 150 ml, T = 90◦ C, 3.0 g of Zr(SO4 )2 . The conversion of n-butanol to water for 1 h was measured and the result for each run was shown in Table 1. It was concluded that the stirrer speed played an insignificant part in the conversion when it was above 400 rpm, indicating that external diffusional resistance could be neglected when the stirrer speed was higher than 400 rpm. It was also proved according to Weisz et al. criterion (see Darge et al. [13]) that pore diffusion resistance in the catalyst was negligible to the esterification (which is indicated in the Appendix A). The relation of the rate of reaction for the esterification to the compositions in the reactor was determined over the catalyst concentrations ranging from 4.5 to 13.3 g/l, and the way of catalyst concentration calculation is detailed in the Appendix B. It was concluded that water production rate took the form of the Eq. (5). Experiments were carried out at temperatures ranging from 60 to 90◦ C in order to obtain the kinetic model equation for the esterification, and the dependence of reaction rate constants on temperature was described as follows   −6.39 (7) k1 = 4.531 × 106 exp T Table 1 Effect of the stirrer speed on the conversion of n-butanol to water for the esterification Stirrer speed (rpm) Conversion (%)

200 41.8

300 45.3

400 47.6

500 47.8

Pervaporation-based separation of water/acetic acid mixture was conducted at various temperatures over water concentrations ranging from 0 to 10 wt.%. The experimental procedure was as follows: firstly, acetic acid was fed to perform pervaporation and the permeation flux was obtained as steady-state was reached; secondly, the feed composition was varied with water added to it to determine the time to reach the new steady-state and to determine the permeation flux in the new steady-state with the changed feed. It was found that the time taken from one steady-state to another with the changed feed was nearly 10 min, indicating that the membrane dynamics were sufficiently fast to guarantee that the apparent flux and selectivity during one of the sampling sequences were representative of at least a pseudo steady-state in the following pervaporation-aided esterification processes. And the permeation flux expressions of water and acetic acid are developed, in which the values of the parameters of PW , PA and m are shown in Table 2. JW = PW CW

(9)

JA = PA CW + m

(10)

Table 2 shows that the parameters changed much less at a lower temperature than at a higher temperature. Largely water, less acetic acid, much less n-butanol and actually no butyl acetate permeated through the membrane during pervaporation separation of the quaternary mixture of water/acetic acid/n-butanol/butyl acetate. The calculation of water permeation flux in this case should take the interaction between water and other components into consideration. Thus, neither the permeation flux of water obeyed Eq. (9) nor Table 2 Parameters in Eqs. (9) and (10) at different temperatures T (◦ C)

PW

PA

m

90 80 70 60

4.2 3.87 3.52 3.50

0.054 0.053 0.053 0.052

1.83 1.70 1.57 1.52

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

177

Table 3 Parameters in Eqs. (11) and (12) at different temperatures T (◦ C)

SA,B

SA,E

SW,B

SW,E

90 80 70 60

0.236 0.230 0.224 0.221

0.258 0.248 0.236 0.232

0.108 0.104 0.102 0.102

0.114 0.110 0.106 0.104

that of acetic acid obeyed Eq. (10). It was proved that the permeation flux of water as a function of both CB and CE was larger than that estimated from Eq. (9), while the permeation flux of acetic acid also depending on both CB and CE was less than that estimated from Eq. (10), indicating that both n-butanol and butyl acetate plasticize the membrane. There was largely acetic acid when water concentration was no more than 10 wt.% for the separation of water/acetic acid mixture, while it was different for the separation of the quaternary mixture because acetic acid concentration was decreased as n-butanol and butyl acetate were added to the liquid mixture. Then, acetic acid permeation flux was decreased in the latter case due to the decrease in its concentration, while water permeation flux was increased due to both invariability in water content and the interaction between water and the other components. The permeation flux expressions of water and acetic acid were thus modified as in the form of the Eqs. (11) and (12) in the case of pervaporation of the quaternary mixture. The values of the parameters in the equations obtained are presented in Table 3. JW = PW CW exp(SW,B CB + SW,E CE ) JA =

PA CW + m exp(SA,B CB + SA,E CE )

Fig. 2. Evolution of F as a function of time at various temperatures (R0 = 1.59, C cat = 8.9 g l−1 , S/V = 0.23 cm−1 ).

a time t during the reaction, F could be obtained by Eq. (6). The effects of several process variables on F during the esterification are reported in this section. Experiments were conducted to determine the effects of temperature, the initial molar ratio of acetic acid to n-butanol, the ratio of the membrane area to the volume of reacting mixture and the catalyst concentration on the coupling process. 4.3.1. Effect of temperature Variations of F and water concentration with time at different temperatures are shown in Figs. 2 and 3, respectively. Fig. 2 shows that F is increased with in-

(11) (12)

Table 3 shows that the values of SA,B , SA,E , SW,B and SW,E changed less with temperature and they changed much less at a lower temperature than at a higher temperature. 4.3. Coupling of pervaporation with esterification The water production rate could be enhanced during pervaporation-aided esterification due to water extraction by pervaporation and was larger than that of the esterification without water removal. With the rate of water production and water removal known at

Fig. 3. Evolution of water content as a function of time at various temperatures (R0 = 1.59, C cat = 8.9 g l−1 , S/V = 0.23 cm−1 ).

178

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

creasing temperature indicating that the acceleration of the rate of water extraction is faster than that of water production rate, and this can be explained by the fact that both the water production rate and permeability coefficient are higher at a higher temperature. It is known from Fig. 3 that the maximum in water concentration was increased with temperature but shifted to a shorter time due to the higher water production rate. Fig. 3 also shows that water concentration is changed less in nearly earlier 3 h than later at a higher temperature; while it is changed much in nearly earlier 5 h than later at a lower temperature. The reason may be that water production rate was changed less at a higher temperature than at a lower one, while the rate of water removal was changed much at a higher temperature than at a lower one. Before the water content went through its maximum value, the rate of water production was larger than that of water removal; after water content attained the maximum value, the rate of water production was less than that of water removal.

Fig. 5. Evolution of F as a function of time at various R0 (T = 90◦ C, C cat = 8.9 g l−1 , S/V = 0.23 cm−1 ).

4.3.2. Effect of initial molar ratio (R0 ) of acetic acid to n-butanol R0 was increased by keeping the reacting mixture volume fixed, so both n-butanol and acetic acid were changed. The relation of the conversion to R0 is represented in Fig. 4, and it is known that the conversion was increased with R0 increasing. Evaluations of F and water content as a function of time at various R0 are displayed in Figs. 5 and 6, respectively. Water

production rate is decreased with the increase of R0 , the reason is illustrated in the Appendix C, and consequently, the rate of water removal is decreased, while the former decreasing was much faster than that of the latter. So water content in liquid mixture was less at a higher R0 than at a lower R0 . The maximum in water content in the mixture had a higher amplitude at a lower R0 , but its position was practically the same. The value of F increased with the increase of R0 . One can also conclude from Figs. 5 and 6 that F and water content change less at a higher R0 than at a lower R0 reflecting that R0 affect the rates of water production and water removal less when R0 increasing.

Fig. 4. Effect of R0 on conversion (T = 90◦ C, C cat = 8.9 g l−1 , S/V = 0.23 cm−1 ).

Fig. 6. Evolution of water content as a function of time at various R0 (T = 90◦ C, C cat = 8.9 g l−1 , S/V = 0.23 cm−1 ).

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

179

Fig. 7. Evolution of F as a function of time at various S/V (T = 90◦ C, R0 = 1.59, C cat = 8.9 g l−1 ).

Fig. 9. Evolution of F as a function of time at various catalyst concentrations (T = 90◦ C, R0 = 1.59, S/V = 0.23 cm−1 ).

4.3.3. Effect of the ratio (S/V) of the membrane area to the volume of reacting liquid Water production rate was the same at various S/V, but the rate of water removal was reduced with the decrease of S/V. As a result, water concentration in the liquid had higher amplitude at a lower S/V, but F was in reverse, those are shown in Figs. 7 and 8, respectively. When the reaction comes to a certain extent, water production rate is decreased due to the high value of water content, then F is increased with the S/V decrease.

4.3.4. Effect of catalyst concentration Water production rate was proportional to and increased with the increase of catalyst concentration, and thus resulting in the increase in water permeation flux. And the acceleration of the latter was faster than the former and thus induce F increasing, that is indicated in Fig. 9. When the catalyst concentration was increased, the maximum in water content in the mixture was increased and shifted to shorter time and the final water content decreased, that was reflected in Fig. 10.

Fig. 8. Evolution of water content as a function of time at various S/V (T = 90◦ C, R0 = 1.59, C cat = 8.9 g l−1 ).

Fig. 10. Evolution of water content as a function of time at various catalyst concentrations (T = 90◦ C, R0 = 1.59, S/V = 0.23 cm−1 ).

180

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181

5. Conclusions The performance of the membrane was studied with pervaporation separation of the liquid mixtures of water/acetic acid and water/acetic acid/n-butanol/butyl acetate, and the permeation flux expressions of water and acetic acid were presented. The kinetic model equation of the esterification of n-butanol with acetic acid catalyzed by Zr(SO4 )2 was developed. Pervaporation-aided esterification was conducted, and the major operating parameters on the coupling process were studied. The ratio of the rate of water removal to water production was presented as the coupling factor F. The variation of F with reaction time at different temperatures, different molar ratios of acetic acid to n-butanol, different ratios of the membrane area to reacting mixture volume and different catalyst concentrations were studied. It was concluded from the experiments that water content in the mixture increased earlier during the reaction and then decreased when it reached to the maximum amplitude. Before water content passed through the maximum amplitude, it increased and F was less than 1, and after water content reached to the maximum amplitude, it decreased and F was larger than 1. S/V had a different effect on F from the other cases in that F increased earlier and then decreased with the increase of S/V. The temperature or the catalyst concentration had a different influence on water content in liquid mixture from R0 and S/V. The increase of temperature or catalyst content resulted in water content increase earlier and then decrease later; while the decrease of R0 or S/V resulted in the water content increase. Fig. 3 or Fig. 10 shows that the concentration has a sharper decrease at a higher temperature or at a higher Ccat after the concentration achieves its summit, while there is not the same in Fig. 6 or Fig. 8.

Temperature (◦ C)

0 DAB (cm2 /s)

0 DBA (cm2 /s)

90 80 70

3.24 × 10−5 2.41 × 10−5 2.03 × 10−5

2.63 × 10−5 2.30 × 10−5 2.03 × 10−5

where, α = [∂ln a/∂x]T,P , A denotes acetic acid and 0 is the diffusivity of B butanol, x is mole fraction, DAB acetic acid in infinite dilution, which can be obtained by 0 = 7.4 × 10−8 (φMB )1/2 × DAB

T µB VA0.6

(A.2)

here, µB is the viscosity of butanol (cp), MB is molecular weight of butanol (g/mol), VA is the mole volume at normal boiling point (ml/mol), φ = 1 for acetic 0 can also be calculated in a acid/butanol mixture. DBA 0 , D 0 at temsimilar way. In the end, the values of DAB BA ◦ peratures of 70, 80 and 90 C are available in Table 4. α in the Eq. (A.1) can be estimated by Margules activity equation and α = 1.581. The value of De at 90◦ C was taken as an example to be calculated 0 , D0 due to DAB BA being changed not much in the range of experimental temperature and D e = 4.53 × 10−5 was obtained. The initial concentrations of acetic acid and butanol were CA0 = 8.74 mol/l and CB0 = 5.47 mol/l, respectively. Radius of catalyst particles was r = 4.04 × 10−2 cm. The initial reaction rate was the largest during the esterification and its value of 4.55 mol/(l h) was determined experimentally. Then, dC 1 r 2 = 5.21 × 10−3  1 dt C0 De so it may be reasonable to conclude that the pore diffusion resistance is negligible.

Appendix A According to Weisz et al. criterion, if (dC/dt) (1/C0 )(r 2 /De ) < 1, pore diffusion resistance in the catalyst could be neglected. Here, r denotes the radius of catalyst particles, (dC/dt) gives the rate of chemical reaction and De is the effective diffusivity of acetic acid in butanol that can be estimated by 0 0 0 − DAB ) + DAB ]α De = [xA (DBA

Table A.1 Diffusivities of acetic acid and butanol in infinite dilution at various temperatures

(A.1)

Appendix B The use of catalyst concentration is related to homogeneous reaction. The activity of an insoluble catalyst is evaluated by its activity surface area in a heterogeneous reaction. Generally speaking, the activity surface area of the catalyst is proportional to its amount

Q. Liu et al. / Journal of Membrane Science 182 (2001) 173–181 Table B.1 Amount of catalyst via n-butanol conversion to water for the esterification Amount of catalyst (g) Conversion (%)

2.0 47.6

2.5 56.4

3.0 63.7

3.5 65.6

4.0 67.8

5.0 68.6

in a certain range. Accordingly, reactions were done to test the activity via the amount for the catalyst. The catalyst activity was represented by the n-butanol conversion, and the results were presented in Table 5, each run lasted for 2 h at the operating condition as follows: CA0 = 8.74 mol/l, CB0 = 5.47 mol/l, T = 90◦ C, V = 150 ml The catalytic activity is almost unchanged when the amount of catalyst were more than 3.0 g (20 g/l). And in the range of the amount of catalyst less than 3.0 g, the water production rate was in linearship with the catalyst concentration that was calculated by Wcat /V. Here, Wcat is the amount of catalyst in the reacting mixture, V is the volume of reacting mixture. Appendix C Knowing that the density of acetic acid is ρ A = 1.05 g/ml, and that of n-butanol is ρ B = 0.813 g/ml; the molecular weight of acetic acid is M A = 60.06 g/mol, and that of n-butanol is M B = 74.12 g/mol. R0 was increased by keeping the reacting mixture volume fixed at 150 ml. Firstly, the volume of acetic acid (VA ) was set equal to that of n-butanol (VB ), say V0 . Then the mole number of acetic acid is N A = V A ρ A /M A = 0.0175VA , and that of n-butanol is N B = 0.011V B . Therefore, we get NA NB = 1.925 × 10−4 V02 Secondly, VA was changed to V0 + 1V , and VB to V0 –1V , then, N A = 0.0175(V0 + 1V ), N B = 0.011(V0 − 1V ), consequently, we obtain the following relationships N A N B = 1.925 × 10−4 (V02 − (1V )2 ) CA CB =

NA NB V2

181

Here, V is the volume of the reacting mixture, CA , CB are concentrations of acetic acid and n-butanol, respectively. It is obviously that V02 is greater than V02 − (1V )2 , so it may be reasonable to conclude that water production rate was decreased with the increase of R0 . References [1] I.K. Song, W.Y. Lee, Methyl t-butyl ether decomposition in an inert membrane reactor composed of 12-tungstophosphoric acid catalyst and polyphenylene oxide membrane, Appl. Catal. A: General 96 (1993) 53–63. [2] H. Kita, K. Tanaka, K. Okamoto, M. Yamamoto, The esterification of oleic acid with ethanol accompanied by membrane separation, Chem. Lett. (1987) 2053–2056. [3] M.O. David, Q.T. Nguyen, J. Neel, Pervaporation membranes endowed with catalytic properties based on polymer blends, J. Membr. Sci. 73 (1992) 129–141. [4] K. Okamoto, M. Yamamoto, Y. Otoshi, T. Semoto, Pervaporation-aided esterification of oleic acid, J. Chem. Eng. Jpn. 26 (5) (1993) 475–481. [5] H. Kita, S. Sasaki, K. Tanaka, K.I. Okamoto, M. Yamoto, Esterification of carboxylic acid with ethanol accompanied by pervaporation, Chem. Lett. (1988) 2025–2028. [6] J.F.F. Keurentjes, G.H.R. Janssen, J.J. Gorissen, The esterification of tartaric acid with ethanol; kinetics and shifting the equilibrium by means of pervaporation, Chem. Eng. Sci. 49 (1994) 4681–4689. [7] X. Feng, R.Y.M. Huang, Studies of a membrane reactor: esterification facilitated by pervaporation, Chem. Eng. Sci. 51 (1996) 4673–4679. [8] L. Bagnell, K. Cavell, A.M. Hodges, A.W.H. Mau, A.J. Seen, The use of catalytically active pervaporation membranes in esterification reactions to simultaneously increase product yield, membrane permselectivity and flux, J. Membr. Sci. 85 (1993) 291–300. [9] Y.S. Zhu, H.F. Chen, Pervaporation separation and pervaporation–esterification coupling using crosslinked PVA composite catalytic membranes on porous ceramic plate, J. Membr. Sci. 138 (1998) 123–134. [10] M.O. David, R. Gref, Q.T. Nguyen, J. Neel, Pervaporation– esterification coupling, I. Basic kinetic model, Trans. Inst. Chem. Eng. 69 (1991) 335–340. [11] M.O. David, R. Gref, Q.T. Nguyen, J. Neel, Pervaporationesterification coupling, II. Modeling of the influence of different operation parameters, Trans. Inst. Chem. Eng. 69 (1991) 341–346. [12] Q.L. Liu, Y.S. Zhu, H.F. Chen, Esterification accompanied by pervaporation using crosslinked PVA composite membrane, J. Chem. Eng. Chin. Universities 11 (2) (1997) 172–176. [13] O. Darge, F.C. Thyrion, Kinetics of the Liquid phase esterification of acrylic acid with butanol catalyzed by cation exchange resin, J. Chem. Tech. Biotechnol. 58 (1993) 351– 355.