Kinetic Study of Acetic Acid Esterification with Methanol Catalyzed by Gel and Macroporous Resins

Transcription

1 Vol25, No4 (2014) Article Kinetic Study of Acetic Acid Esterification with Methanol Catalyzed by Gel and Macroporous Resins Evelien VAN DE STEENE 1,2, Jeriffa DE CLERCQ 1 and Joris W THYBAUT 2,* 1 Ghent University, Department of Industrial Technology and Construction, Valentin Vaerwyckweg 1, B-9000 Gent, Belgium 2 Ghent University, Department of Chemical Engineering and Technical Chemistry, Laboratory for Chemical Technology, Technologiepark 914, B-9052 Gent, Belgium (Manuscript received June 30, 2014; Accepted July 31, 2014) Abstract The liquid-phase acetic acid esterification kinetics with methanol to methyl acetate and water catalyzed by gel (Lewatit K1221) and macroporous (Lewatit K2629) ion exchange resins have been investigated and compared with literature reported data acquired on Amberlyst 15 The effects of the resin swelling, the initial molar methanol to acetic acid ratio (1:1 10:1) and the temperature ( K) on the reaction kinetics were investigated The gel type resin exhibits a remarkably higher catalytic activity compared to the macroporous resins, despite its similar number of sulfonic acid groups This can be attributed to the differences in active site accessibility during reaction and, hence, in resin swelling, especially when it is in contact with polar components, such as water and methanol The differences in catalytic behavior between the considered resins have been assessed using an exchange based Eley-Rideal model in which it is assumed that (1) all the active sites are occupied, (2) the acid undergoes a proton exchange with the protonated methanol and (3) the reaction occurs according to an Eley-Rideal mechanism with the surface reaction between protonated acetic acid and methanol from the bulk as ratedetermining step This model, which implicitly accounts for resin swelling, could describe the observed differences between the various resins The activation energy was determined at 47 kj mol -1, irrespective of the resin used Keywords Kinetics, Esterification, Ion-exchange resin, Swelling, Eley-Rideal reaction mechanism 1 Introduction Organic esters are important fine chemicals widely used in the manufacturing of flavors, pharmaceuticals, plasticizers and polymerisation monomers They are also used as emulsifiers in the food and cosmetic industries (1-3) Several synthetic routes are available for obtaining organic esters, nevertheless, the most common used methodology for ester synthesis is direct esterification of carboxylic acids with alcohols in the presence of acid catalysts ( 1,4) Strong liquid mineral acids, such as H2SO 4, HCl and HI, are effective for the esterification of carboxylic acids These and other homogeneous acid catalysts have been the subjects of extensive studies (1,5-9) Despite a strong catalytic effect, the use of these catalysts suffers from some well-known drawbacks, such as the existence of side reactions with reactants/products, equipment corrosion, and having to deal with acid-containing waste ( 2,4) Many solid catalysts, eg, solid acids, ion-exchange resins, zeolites, and acidic clay catalysts, are used as heterogeneous catalysts in the esterification ( 4,10) Their weightbased activity show that ion-exchange resins are the most effective compared to the other solid acid catalysts (4) * Corresponding author: JorisThybaut@UGentbe 234

2 J ION EXCHANGE Additionally, ion-exchange resins are very commonly used heterogeneous catalysts and have already proven to be effective in liquid-phase esterification, transesterification and etherification reactions (4,11-19) The use of ion-exchange resins as catalysts presents some distinct advantages over homogeneous catalysis: (1) the purity of the products is higher, as side reactions can be completely or partially eliminated; (2) the catalyst can be easily removed from the reaction mixture by filtration; (3) the corrosive environment caused by the discharge of acid-containing waste is eliminated ( 2,18-22) The catalytic activity of ion exchange resins strongly depends on their swelling properties When a dry gel or macroporous resin is brought into contact with a liquid, the resin volume increases, ie, the resin swells, because of a portion of the liquid component is sorbed by the resin up to reaching equilibrium with the liquid phase (17,23) When, eg, water diffuses into the resin, it interacts with its sulfonic acid sites, forms protonated water molecules and primary solvation shells This solvation attracts additional water molecules and further extends the solvation shell This phenomenon is observed at the macroscale as resin swelling, leading to eg a 55% volume increase of the macroporous Amberlyst 15 in methanol ( 5,19,24,25) This swelling is of particular importance when using ion exchange resins as catalysts because the swelling capacity controls the reactants accessibility to the acid sites and, therefore, affects their overall reactivity ( 26) Ion exchange resin catalyzed esterification kinetics have been conceptualized and quantified by a pseudo-homogeneous model (PH), as well as by models including adsorption phenomena via an Eley-Rideal (ER) or a Langmuir-Hinshelwood (LH) mechanisms (5,17,27-30) Several authors found, however, that the classical adsorption-based models, such as ER and LH, were not sufficient to describe the reaction kinetics catalyzed by ion ( 17 exchange resins,19) Therefore, more advanced models, accounting for swelling, were developed ( 5,17,19) In the present work, the acetic acid esterification kinetics have been assessed on gel type and macroporous resins making use of a kinetic model according to an exchange based Eley- Rideal mechanism The latter model performs well in the simulation of transesterification kinetics data acquired on a gel type resin Lewatit K1221 and several macroporous resins such as Lewatit K2629, Lewatit K2640 and Amberlyst 15 (12) Kinetic measurements were performed to evaluate the effect of temperature and initial molar methanol to acetic acid ratio on the esterification rates catalyzed with Lewatit K1221 and Lewatit K2629 resins Moreover, a literature data set acquired on Amberlyst 15 ( 5) has been included in the comparison To independently quantify the swelling phenomenon, resin swelling tests with pure reactants and products have been performed, using the approach described by several authors ( 19,31-34) 2 Procedures 21 Materials The reagents, methanol (Fiers, purity > 9985 %) and acetic acid (Fiers, purity n-octane (Acros Organics, purity > 99%) was used as internal standard Methyl acetate (Acros, purity > 99%) and ethanol (Fiers, purity > 998) are used for calibration purposes The ion-exchange resins Lewatit K1221, Lewatit K2629 (Lanxess) and Amberlyst 15 (Rohm & Haas) are spherical polystyrene-based resin beads that are cross-linked with divinylbenzene, with sulfonic acid groups as active centres The resins properties are reported in Table 1 These resins were selected based on their good catalytic performance in the transesterification (11,12,21,35) These resins are heat-sensitive and experience an activity loss above 398 K Table 1 Chemical and physical properties of ionexchange resins Property K1221 K2629 Amberlyst 15 Appearance dark brown, translucent beige, opaque deep grey, beige opaque Type gel macroporous macroporous Acid site (eq/kg) % DVB 4 % 18 % 20 % Bead size (mm) BET area (23) < (m² g -1 ) The pretreatment of Lewatit K1221 and Lewatit K2629, consisted of 24 h freeze drying under vacuum at 233K, to completely remove any moisture, as advised by the suppliers The pretreatment of Amberlyst 15, was, in contrary to the latter, first washed with methanol followed by several wash steps with water to remove impurities Wash steps with water were repeated until the supernatant liquid became colorless Afterwards the washed Amberlyst 15 was dried under vacuum at 363 K until the mass remained constant, which usually took about 2 days (5) 21 Volumetric swelling experiments Resin swelling is quantified by measuring the swelling ratio S, cq the ratio of the volume of the swollen resin, in the presence of a solvent, to the volume of the dry resin A graduated cylinder of 25 ml was filled with 10 ml of dry resin and subsequently 25 ml of solvent was added The volume of the swollen resin was read, and the swelling ratio S, was calculated 22 Kinetic measurements Acetic acid esterification with methanol, was carried out in a three-necked glass flask of 180 ml capacity equipped with a 235

3 Vol25, No4 (2014) reflux condenser, a thermocouple and a sampling port The temperature in the reactor was maintained within 05 K from the set point with a thermostat (Lauda Proline RP845) equipped with a PID-controller The reaction mixture was stirred with a magnetic stirrer at a constant speed of 500 rpm throughout the experimentation which sufficed to establish complete mixing of the reaction mixture The reactor was first loaded with methanol and catalyst and subsequently heated to the reaction temperature When the methanol and catalyst reached the desired temperature, preheated acetic acid and n-octane were added through the sampling port in order to have an as well determined starting point of the esterification experiment as possible without disturbing the reactor temperature All experiments were performed at atmospheric pressure Each experiment was performed at least double if not quadruple with an experimental error below 5% The range of experimental conditions is given in Table 2 Samples of the reaction mixture were withdrawn through the sampling port at regular time intervals The samples were analyzed by gas chromatography using a Focus GC, equipped with an AS3000 autosampler, a Stabilwax capillary column (30 m x 032 mm, 025 μm thickness) and a flame ionization detector Table 2 Range of experimental conditions for acetic acid esterification with methanol Catalyst Lewatit Lewatit Amberlyst K1221 K (5) Number of experimental points Number of experiments Temperature (K) , Pressure (MPa) MeOH:HOAc molar 07:1 ratio 10:1 1:1 10:1 1:1 20:1 Mass of catalyst ( kg) Experiment time (s) Batch time* (kg s) *Defined as product of experiment time multiplied by the mass of catalyst For Amberlyst 15, Pöpken et al (5) used an isothermal glass reactor with a volume of 500 ml A constant stirring at a speed of 250 rpm was imposed by a plate-type stirrer The kinetic experimental procedure was similar to the one described above, nevertheless, the experiments were performed until equilibrium was reached Only the experimental points before equilibrium was reached were considered here During each experiment between 15 and 45 samples were taken The acetic acid concentration of each sample was analyzed by potentiometric titration with sodium hydroxide ( 5) The number of experimental points, the number of experiments and also the experimental conditions of the three kinetics dataset are given in Table 2 Particular attention should be paid to the limited information obtained with respect to the temperature dependence of the acetic acid esterification kinetics on Lewatit K2629, ie, experiments are only available at and K, and to the amount of catalyst used in the experiments with Amberlyst 15 The latter exceeds that used in the case of Lewatit K1221 or Lewatit K2629 by at least a factor of 10 The absence of internal as well as external mass-transfer resistances in the kinetics observed on Lewatit K1221 and Lewatit K2629, was evaluated via quantitative criteria such as the Carberry number, for external mass transfer limitations and the Weisz modulus for internal diffusions The respectively values accounting to 13x10-2 and 80x10-5 indicate the absence of respectively external and internal mass transfer limitations Pöpken et al (5) have evaluated the mass transfer resistance for Amberlyst 15 on a qualitative manner, ie via experiments, and proved absence of internal and external mass transfer resistances 23 Thermodynamic calculations The reaction enthalpy was calculated at -439 kj mol -1, based on thermodynamic data obtained from the tabulated values of the NIST Chemistry Webbook At 298 K the values for the corresponding equilibrium coefficient K eq ranging from to 3016 were published (5,36-41) Due to the non-ideal character of the reaction mixture, activity coefficients for each component were calculated at temperatures ranging from 303 K to 333 K with the UNIQUAC method 24 Modelling and regression analysis The mass balance for species i in a batch reactor, which is considered to be spatially uniform in composition and temperature, is given by: with R i the specific net production rate of component i, W the catalyst mass in the reactor, i the stoichiometric coefficient of component i, r the reaction rate, n i the number of moles of component i in the reaction mixture, and t the time The net production rate is calculated according to equations derived from the various mechanisms that are proposed, ie, pseudohomogeneous (PH), Langmuir-Hinshelwood (LH) and Eley- Rideal (ER) The net production rate is a function of the rate coefficient (k), temperature, activities and adsorption or exchange equilibrium coefficients of component i (Ki) Within the limited temperature interval, no temperature dependence of the exchange equilibrium coefficients is considered The rate coefficient is described by the Arrhenius law, which is (1) 236

4 J ION EXCHANGE reparameterized according to Kittrell (42) in order to avoid strong binary correlation between the Arrhenius parameters: With T the temperature, T ref the reference temperature (331 K), E A the activation energy, R the universal gas constant and k Tref the reaction rate coefficient at the reference temperature The activation energy and the rate coefficient at the reference temperature as well as the adsorption or exchange equilibrium coefficient for the adsorption- or exchange-based mechanisms have been determined by regression The parameter estimates were determined by minimizing the sum of squared residuals between the observed and the calculated concentration of methyl acetate using the nonlinear least-squares technique and the Bayesian estimation technique applying a single response Levenberg-Marquardt algorithm, available in Athena Visual Studio (43) The global significance of the regression and the parameters was assessed with respectively the F and t test Apart from the statistical significance, the estimated parameters were evaluated on their physicochemical meaning (2) 32 Catalytic activity The catalytic activity of the considered resins is shown in Fig 1 Similar to transesterification (12), Lewatit K1221 exhibits a higher activity than the macroporous resins, despite the comparable active site concentration, see Table 1 The difference in catalytic activity between gel type and macroporous resins can be attributed to the accessibility of the active sites, which is related to the swelling capacity of the resin, see Table 3 The importance of resin swelling in esterification kinetics has also been demonstrated by Lotero et al (26) Resin swelling guarantees substrate accessibility to the active sites and, hence, affects its overall reactivity Upon swelling, the resins gel beads expand their polystyrene chains such that new/additional porous space and, hence, active sites become available 3 Experimental results 31 Volumetric swelling tests The swelling ratios of the investigated resins in contact with the pure reactants and products are reported in Table 3 Several factors potentially determine the swelling phenomenon (19), ie, (1) the nature of the solvent: polar or apolar; (2) the degree of cross-linking of the resin; (3) the nature of the functional groups in the resin and (4) the resins exchange capacity For the resins considered in Table 3, factors (1) and (2) can be invoked to ( 17 explain the observations,31,33,36,44-49) Gel type and macroporous resins swell more in polar than in less or apolar solvents, Swelling in water is more pronounced than in any of the solvents with a lower dielectric constant ( 17,23) It is also evident from Table 3 that a macroporous resin, cq, Lewatit K2629 or Amberlyst 15, swells to a lesser extent than a gel type resin Similar observations have been made when investigating the swelling with the components involved in ethyl acetate ( 12 transesterification with methanol ), and can be explained by the different crosslinking degrees, see Table 1 Table 3 Swelling ratio (S) of Lewatit K1221, Lewatit K2629 and Amberlyst 15 in water, methanol (MeOH), methyl acetate (MeOAc) and acetic acid (HOAc), and dielectric constant of these components Swellin g ratio K1221 K2629 Amberlyst 15 Dielectric constant water 256 ± ± ± MeOH 245 ± ± ± MeOAc 124 ± ± ± HOAc 118 ± ± ± Fig 1 Simulated (lines; ER_exchange model, Eq (3) with parameters as reported in Table 4) and experimental (symbols) methyl acetate yield versus batch time catalyzed by different ion exchange resins ( Lewatit K2629 and and C HOAc,0 = 950 M for Lewatit K1221 and Lewatit K2629; 323 K and C HOAc,0 = 402 M for Amberlyst 15; Initial molar ratio of MeOH:HOAc = 1:1) The evolution of the methyl acetate yield on both macroporous resins deserves some further, more specific attention Although the reaction with Amberlyst 15 has been performed at 323 K, compared to 333 K for Lewatit K2629, Amberlyst 15 exhibits at short batch times higher methyl acetate yields and, hence, a higher initial reaction rate, compared to Lewatit K2629 With increasing batch time, however, the methyl acetate yield on Amberlyst 15 increases more slowly than on Lewatit K2629 such that, ultimately, the highest methyl acetate yields are obtained with the latter resin Hence, whereas Amberlyst 15 initially, apparently, swells more than Lewatit K2629, the former exhibits more pronounced product adsorption and, hence, inhibition, than the latter Note that the swelling ratios for K2629 and Amberlyst 15 as reported in Table 3 are within each other s confidence interval and, hence, do not confirm of/nor form a contra-indication against the suggested 237

5 Vol25, No4 (2014) interpretation It should also be kept in mind that the pretreatment of Amberlyst 15 accounted for an additional wash step compared to Lewatit K1221 and K2629, see Section 2 This additional step, wherein impurities are removed can, to our knowledge, be the explanation for the higher initial reaction rate of Amberlyst Temperature and initial molar ratio effect The temperature effect and the effect of the initial molar methanol to acetic acid ratio were investigated by performing experiments between and K and 1:1 and 10:1 respectively As expected, a higher temperature results in a higher esterification rate and a correspondingly higher methyl acetate yield at the same batch time The equilibrium conversion is practically temperature independent, which is a logic consequence of the limited reaction enthalpy, ie, -439 kj mol -1, of the investigated reaction (5,29,36) Figure 2 shows that, under the investigated operating conditions, the reaction kinetics can be enhanced by using a methanol excess The esterification rate, however, will not keep on increasing linearly with initial molar methanol to acetic acid (17,29,48) ratio It has already been demonstrated in the literature that the initial acetic acid esterification rate with methanol first increases with the initial molar methanol to acetic acid ratio before reaching a maximum and that it subsequently decreases The latter evolution can be relatively easily rationalized in terms of establishing optimum molar fractions within the resin for obtaining a maximum rate for a bimolecular reaction The higher methyl acetate yield as a function of the batch time at high initial molar methanol to acetic acid ratios for Amberlyst 15 and Lewatit K2629 are in line with what is obtained for acetic acid esterification on Lewatit K Kinetic modeling of gel and macroporous resin catalyzed esterification The acetic acid esterification with methanol has been simulated in literature using kinetic models based on several mechanisms, such as Eley-Rideal and Langmuir- Hinshelwood (2,5,8,13-15,17-19,27-31,36,37,50-57) A new acetic acid esterification kinetic model is constructed based on the exchange-based model as proposed for transesterification by Van de Steene et al ( 12,58) This model was selected making the best possible compromise between statistical significance and physicochemical meaning Because methanol and water strongly sorb in the resin, see Table 3, all the active sites of the resin are assumed to be occupied Acetic acid and methyl acetate undergo a proton exchange with protonated methanol on the active sites The surface reaction between protonated acetic acid and methanol from the bulk to form protonated methyl acetate and water is considered to be rate determining In contrast to transesterification the second reaction product in esterification, ie water, has a high affinity towards the sulfonic acid groups Hence, in the present esterification model the produced water is also allowed to exchange with protonated methanol in contrast to the ethanol produced in transesterification (12) Water therefore competes with methanol for the active sites and may lead to (significant) product inhibition ( 34) The corresponding rate expression (Eq (3)) depends on the rate coefficient (ksr), the temperature, the component activities a i, and the exchange equilibrium coefficients (K HOAc,K l ) k K a r a 1 SR HOAc HOAc 1 a Keq a a MeOAc H2O MeOH HOAc MeOAc H2O KHOAc Kl ameoh ameoh a a (3) Fig 2 Simulated (lines; ER_exchange model, Eq (3) with parameters as reported in Table 4) and experimental (symbols) methyl acetate yield versus batch time catalyzed by Amberlyst 15 (filled symbols), Lewatit K2629 (empty symbols) and Lewatit K1221 (crossed symbols) at different initial MeOH:HOAc molar ratios ( 83:1; 1:1) (Temperature = 323 K and C HOAc,0 = 086 M (83:1); C HOAc,0 = 402 M (1:1) for Amberlyst 15) (Temperature = 333 K and C HOAc,0 = 190 M (10:1); = 950 M (1:1) for Lewatit K2629 and Lewatit K1221) C HOAc,0 Because of the identical evolution of the water and methyl acetate concentration throughout the experimentation, it was impossible to determine separate values for the adsorption or exchange coefficients for water and methyl acetate (30,59-61) Independent exchange coefficient determination or incorporation of experimental measurements into the data set in which a specific amount of reaction product, ie, water or methyl acetate was added to the initial mixture may have allowed a separate determination of the adsorption/exchange coefficients for methyl acetate and water ( 61) Because such information is not contained in the kinetic data set acquired in the present work, ie, all experiments were performed in such conditions that only equal methyl acetate and water concentrations are part of the dataset, a single value was estimated for the methyl acetate and the water exchange 238

6 J ION EXCHANGE coefficient, indicated with K l 41 Parameter estimation for the exchange based model The parameter estimates and corresponding statistics for the three considered resins are reported in Table 4 All multiple correlation coefficients at least assume a value of 0990 All regressions are globally significant with F values exceeding 1000 if not The comparatively lower F value obtained for Amberlyst 15 stems from the lower number of data points In conclusion, these values indicate an appropriate representation of the experimental data by the ER-exchange model for all three catalysts The regressions resulted in comparable activation energies for acetic acid esterification catalyzed by Amberlyst 15 and K1221, as could be expected based on the results obtained for transesterification (12) Because the experimental data set acquired on Lewatit K2629 only contains experiments at and K, the activation energy estimate had an unacceptably wide confidence interval Hence, an average value from those obtained for Lewatit K1221 and Amberlyst 15 was used, ie, 47 kj mol -1 for this resin Several authors published similar values ranging from 3813 to 5188 kj mol -1 for acetic acid esterification with methanol catalyzed with Amberlyst 15 ( 36,37,57) Table 4 Statistical information and parameter estimation of the ER-exchange model for the esterification catalyzed by Lewatit K1221, Lewatit K2629 and Amberlyst 15 Catalyst K1221 K2629 Amberlyst 15 N RSSQ R² F value k SR Tref (mol kg -1 cat s ± ± ± 0124 ) E A (kj mol -1 ) 4600 ± b 4758 ± 044 KHOAc 4580 ± ± ± 0657 Kl 3611 ± ± ± 1478 b This parameter was fixed, as there was not enough temperature variation in the experimental dataset ( Table 2) 1 In section 32, the difference in catalytic activity was attributed to the accessibility of the active sites, which is related to the swelling of the resin in the reaction mixture The swelling ratio depends on the crosslinking degree and, hence, the surface reaction rate coefficient for K1221 is expected to be higher than the ones for Lewatit K2629 and Amberlyst 15 A similar surface reaction rate coefficient is expected for Lewatit K2629 and Amberlyst 15 It is evident from Table 4 that the expected order in surface reaction rate coefficients is not respected by the obtained parameter estimates This may be the result of correlation between this surface reaction rate coefficient and the acetic acid exchange coefficient When looking at the product of both also denoted as the composite rate coefficient, as it occurs in the numerator of the rate expression, see Eq (3), the expected order is indeed found: 0985 (K1221) > 0331 (K2629) (Amberlyst 15) The ratio between the composite rate coefficient on gel type versus macroporous resins amounts to about three, which is in to the range of 3- to 4-fold as reported for transesterification (12) Whereas the products exchange coefficients, Kl, are relatively similar in a range from 2 (K2629) to 5 (Amberlyst 15), slightly more pronounced differences are obtained for the acetic acid exchange coefficients, K HOAc, that range from 1 (Amberlyst 15) to 6 (K2629) Particularly noteworthy is that the ratio of the acetic acid and products exchange coefficients exceeds one for Lewatit K1221 and Lewatit K2629 while it is lower than one for Amberlyst 15 The higher acetic acid exchange coefficient compared to that of the products can be attributed to the lack of distinction between the bulk liquid and resin phase in addition to the protonated species Apparently, additional phenomena are at stake on Amberlyst 15 It was already evident from Figure 1 that, at higher batch times, the methyl acetate yield evolves less steep on Amberlyst 15 than on Lewatit K2629 at higher batch times, which can be interpreted in terms of more pronounced product inhibition The latter can, indeed, become evident via a comparatively higher product exchange coefficient It should also, again, be pointed out that the acetic acid esterification data on Amberlyst 15 were taken from the literature, while those on Lewatit K1221 and Lewatit K2629 have been measured as part of the present work As a result, the differences observed in the estimates for the exchange coefficients may also be related to procedures that have been followed, such as the resin s pretreatment Firstly, the Amberlyst 15 resin was washed with methanol and water to remove the impurities, and afterwards dried under vacuum for 2 days at 363 K This pretreatment is distinct from the 24h freeze drying at 233 K, which has been performed for Lewatit K1221 and Lewatit K2629 Another difference in the experimental data obtained on Lewatit K1221 and Lewatit K2629, compared to Amberlyst 15 is that the ratio of the reactant concentration to that of the active sites is 20 to 60 times higher on the former than on the latter Implicitly, there will be another reactant-product composition at the active site, leading to a higher value for the product exchange coefficient and a lower value for the acetic acid exchange coefficient It is evident from the above discussion that the results obtained so far for acetic acid esterification indicate the need for some further refinement, in particular with respect to the thermodynamics within the resin More specifically the presence of water as reaction product in acetic acid esterification results in much more pronounced variations in polarization of the reaction mixture compared to transesterification As a result, the resin swelling and component sorption in an esterification 239

7 Vol25, No4 (2014) mixture, with changing polarization with methyl acetate yield, is much more complex Mazzotti et al (17) found that the esterification equilibrium that could be obtained in a batch reactor depended on the amount of Amberlyst 15 used By increasing the resin to reactants ratio, not only the esterification rate was enhanced but also the equilibrium composition shifted towards the reaction products This effect could be attributed to the significant swelling of the resin due to solvent sorption and, hence, due to the non-negligible volume of the sorbed phase compared to the bulk phase Such effects are less likely or will even not occur in non-swelling solid catalysts ( 17) This was also concluded by Sainio et al ( 33) and Tesser et al ( 19) They attributed the effect of the amount of resin used in a batch reactor on the equilibrium composition to the simultaneous chemical and phase equilibrium between liquid bulk and resin Sainio et al ( 33) also mentioned that a significant solvent uptake by the resins occurred in a batch reactor when using a low reactants to resin ratio, such as for the Amberlyst 15 Given the above described indications that a further model enhancement is required with respect to the description of the thermodynamics within the resin, it can be understood that kinetics measurements performed at significantly different reactant to catalyst/resin ratio lead to different estimates for the exchange coefficients and/or different trends in these values In summary, several potential causes have been identified for the apparent discrepancies between the exchange coefficients obtained based on the literature reported Amberlyst 15 data on the one hand and the data measured on Lewatit K1221 and Lewatit K2629 on the other hand Mainly differences in pretreatment may have resulted in a more pronounced product inhibition, which became evident via a higher product to acetic acid exchange coefficient ratio On the other hand, the reactant concentration to active sites ratio in the Amberlyst 15 data was 20 up to 60 times smaller compared to the other data This range of reactant concentration to active sites ratio may be wider than can be adequately described by the model for the exchange of components between the bulk and the resin with a unique set of parameter values, indicating, again, that an enhanced model, eg, based on Flory-Huggins theory will be required 5 Conclusions Gel and macroporous acid ion exchange resins can be used as catalysts for acetic acid esterification with methanol to methyl acetate and water Due to the chemical similarity between the considered resins, differences in observed reaction rates were mainly related to the accessibility of the active sites, which are inversely proportional with the amount of crosslinking by divinylbenzene due to the corresponding swelling behavior of the resins An exchange based kinetic model, which implicitly accounts for resin swelling, allowed interpreting the observed differences between the various resins, although the estimates obtained for the acetic acid and products exchange coefficients indicated the need for a more elaborate description of the thermodynamics in the resin The composite rate coefficient, defined as the product of the surface reaction rate coefficient and the acetic acid exchange coefficient, was found to be proportional to the accessibility of the active sites and, hence, inversely proportional to the crosslinking degree The composite rate coefficient on the gel type resin, Lewatit K1221, was 3-fold higher than on the macroporous resins The composite activation energy amounts to 46 kj/mol independently of the considered resin, which is an indication of their chemical similarity Acknowledgements The Research Fund of the University College Ghent and the faculty of Engineering and Architecture of the University of Ghent are acknowledged for their financial support Lanxess is acknowledged for providing the ion-exchange resin References 1) Liu, Y J; Lotero, E; Goodwin, J G Journal of Catalysis, 242, 278 (2006) 2) Teo, H T R; Saha, B Journal of Catalysis,228, 174(2004) 3) Dossin, T F; Reyniers, M F; Marin, G B Applied Catalysis B-Environmental, 62, 35 (2006) 4) Peters, T A; Benes, N E; Holmen, A; Keurentjes, J T F Applied Catalysis a-general, 297, 182 (2006) 5) Popken, T; Gotze, L; Gmehling, J Industrial & Engineering Chemistry Research, 39, 2601 (2000) 6) Aranda, D A G; Santos, R T P; Tapanes, N C O; Ramos, A L D; Antunes, O A C Catalysis Letters, 122, 20 (2008) 7) Duque-Bernal, M; Quintero-Arias, J D; Osorio-Viana, W; Dobrosz-Gomez, I; Fontalvo, J; Gomez-Garcia, M A International Journal of Chemical Kinetics, 45, 10 (2013) 8) Lilja, J; Murzin, D Y; Salmi, T; Aumo, J; Arvela, P M; Sundell, M Journal of Molecular Catalysis a-chemical, 182, 555 (2002) 9) Ronnback, R; Salmi, T; Vuori, A; Haario, H; Lehtonen, J; Sundqvist, A; Tirronen, E Chemical Engineering Science, 52, 3369 (1997) 10) Kiss, A A; Dimian, A C; Rothenberg, G Advanced Synthesis & Catalysis, 348, 75 (2006) 11) Van de Steene, E; De Clercq, J; Thybaut, J Journal of Molecular Catalysis a-chemical, 359, 57 (2012) 12) Van de Steene, E; De Clercq, J; Thybaut, J Chemical Engineering Journal, 242, 170 (2014) 13) Altiokka, M R; Citak, A Applied Catalysis a-general, 239, 141 (2003) 14) Bart, H J; Kaltenbrunner, W; Landschutzer, H International Journal of Chemical Kinetics, 28, 649 (1996) 15) Yadav, G D; Thathagar, M B Reactive & Functional Polymers, 52, 99 (2002) 16) Lee, M J; Wu, H T; Lin, H M Industrial & Engineering Chemistry Research, 39, 4094 (2000) 240

8 J ION EXCHANGE 17) Mazzotti, M; Neri, B; Gelosa, D; Kruglov, A; Morbidelli, M Industrial & Engineering Chemistry Research, 36, 3 (1997) 18) Yadav, G D; Rahuman, M Org Process Res Dev, 6, 706 (2002) 19) Tesser, R; Casale, L; Verde, D; Di Serio, M; Santacesaria, E Chemical Engineering Journal, 157, 539 (2010) 20) Russbueldt, B M E; Hoelderich, W F Applied Catalysis a-general, 362, 47 (2009) 21) Zielinska-Nadolska, I; Warmuzinski, K; Richter, J Catalysis Today, 114, 226 (2006) 22) Ozbay, N; Oktar, N; Tapan, N A Fuel, 87, 1789 (2008) 23) Helfferich, F G Ion Exchange; Dover Publications, ) Polyanskii, N G; Sapozhnikov, V K Russian Chemical Reviews, 46, 445 (1977) 25) Gusler, G M; Browne, T E; Cohen, Y Industrial & Engineering Chemistry Research, 32, 2727 (1993) 26) Lotero, E; Liu, Y J; Lopez, D E; Suwannakarn, K; Bruce, D A; Goodwin, J G Industrial & Engineering Chemistry Research, 44, 5353 (2005) 27) Sanz, M T; Murga, R; Beltran, S; Cabezas, J L; Coca, J Industrial & Engineering Chemistry Research, 41, 512 (2002) 28) Ali, S H International Journal of Chemical Kinetics, 41, 432 (2009) 29) Ali, S H; Merchant, S Q International Journal of Chemical Kinetics, 38, 593 (2006) 30) Song, W; Venimadhavan, G; Manning, J M; Malone, M F; Doherty, M F Industrial & Engineering Chemistry Research, 37, 1917 (1998) 31) Ali, S H; Al-Rashed, O; Azeez, F A; Merchant, S Q Bioresour Technol, 102, (2011) 32) Lilja, J; Murzina, E; Grenman, H; Vainio, H; Salmi, T; Murzin, D Y Reactive & Functional Polymers, 64, 111 (2005) 33) Sainio, T; Laatikainen, M; Paatero, E Fluid Phase Equilib, 218, 269 (2004) 34) Tesser, R; Di Serio, M; Casale, L; Carotenuto, G; Santacesaria, E Canadian Journal of Chemical Engineering, 88, 1044 (2010) 35) Zielinska-Nadolska, I; Warmuzinski, K; Richter, J Inz Chemiczna Proces, 25, 1861 (2004) 36) Tsai, Y T; Lin, H M; Lee, M J Chemical Engineering Journal, 171, 1367 (2011) 37) JagadeeshBabu, P E; Sandesh, K; Saidutta, M B Industrial & Engineering Chemistry Research, 50, 7155 (2011) 38) Wyczesany, A Chem Process Eng, 30, 243 (2009) 39) Wyczesany, A Industrial & Engineering Chemistry Research, 46, 5437 (2007) 40) Bematova, S; Aim, K; Wichterle, I Fluid Phase Equilib, 247, 96 (2006) 41) Poling, B E; Prausnitz, J M; Connell, J O The Properties of Gases and Liquids; McGraw-Hill, ) Schwaab, M; Pinto, J C Chemical Engineering Science, 62, 2750 (2007) 43) Stewart, W E; Caracotsios, M Computer-Aided Modeling of Reactive Systems; Wiley, ) Rehfinger, A; Hoffmann, U Chemical Engineering Science, 45, 1605 (1990) 45) Corain, B; Zecca, M; Canton, P; Centomo, P Philos Trans R Soc A-Math Phys Eng Sci, 368, 1495 (2010) 46) Coutinho, F M B; Rezende, S M; Soares, B G Journal of Applied Polymer Science, 102, 3616 (2006) 47) Bozek-Winkler, E; Gmehling, J Industrial & Engineering Chemistry Research, 45, 6648 (2006) 48) Mazzotti, M; Kruglov, A; Neri, B; Gelosa, D; Morbidelli, M Chemical Engineering Science, 51, 1827 (1996) 49) Zimehl, R Thermochimica Acta, 310, 207 (1998) 50) Zhang, Y; Ma, L; Yang, J C Reactive & Functional Polymers, 61, 101 (2004) 51) Delgado, P; Sanz, M T; Beltran, S Chemical Engineering Journal, 126, 111 (2007) 52) Qu, Y X; Peng, S J; Wang, S; Zhang, Z Q; Wang, J D Chinese Journal of Chemical Engineering, 17, 773 (2009) 53) Santacesaria, E; Tesser, R; Di Serio, M; Guida, M; Gaetano, D; Agreda, A G Industrial & Engineering Chemistry Research, 46, 5113 (2007) 54)Pasias, S; Barakos, N; Alexopoulos, C; Papayannakos, N Chemical Engineering & Technology, 29, 1365 (2006) 55) Pasias, S A; Barakos, N K; Papayannakos, N G Industrial & Engineering Chemistry Research, 48, 4266 (2009) 56) Xu, Y; Dou, W J; Zhao, Y J; Huang, G X; Ma, X B Industrial & Engineering Chemistry Research, 51, (2012) 57) Yu, W F; Hidajat, K; Ray, A K Applied Catalysis a- General, 260, 191 (2004) 58) Van de Steene, E; De Clercq, J; Thybaut, J In American Institute of Chemical Engineers Annual Meeting, Abstracts ) Gangadwala, J; Mankar, S; Mahajani, S; Kienle, A; Stein, E Industrial & Engineering Chemistry Research, 42, 2146 (2003) 60) Parra, D; Tejero, J; Cunill, F; Iborra, M; Izquierdo, J F Chemical Engineering Science, 49, 4563 (1994) 61) Huang, Y S; Sundmacher, K International Journal of Chemical Kinetics, 39, 245 (2007) 241