High Conversion of Methyl Acetate Hydrolysis in a ... - ACS Publications

Article pubs.acs.org/IECR

High Conversion of Methyl Acetate Hydrolysis in a Reactive Dividing Wall Column by Weakening the Self-Catalyzed Esterification Reaction Xiaoda Wang, Hongxing Wang, Jinyi Chen, Weiyue Zheng, and Ting Qiu* Engineering Research Center of Reactive Distillation Technology (Fujian Province), School of Chemical Engineering, Fuzhou University, Fuzhou 350108, Fujian, China S Supporting Information *

ABSTRACT: We show the superiority of reactive dividing wall column (RDWC) to the single reactive distillation (RD) column in improving the conversion of reactant, taking the hydrolysis of methyl acetate (MA) as example. It is difficult to achieve above 99% conversion for MA in a traditional reactive distillation column (TRDC) due to the existence of selfcatalyzed methanol (MeOH)-acetic acid (HAc) esterification reaction in the column bottom. In this work, more than 99% conversion of MA hydrolysis was realized experimentally in an RDWC by separating MeOH from the hydrolysis mixture. The effects of several operation parameters on hydrolysis conversion were systematically investigated, including feedwater−MA mole ratio, heat duty, mole flow rate of feed MA, and vapor distribution. The simulation results by Aspen Plus showed that RDWC has several improvements in MA hydrolysis over TRDC, including lower energy consumption, lower water−MA mole ratio, and larger production capacity. With the increase in MA conversion, the superiorities became more obvious and contribute to the weaker self-catalyzed MeOH−HAc esterification reaction in the RDWC.

1. INTRODUCTION Reactive dividing wall columns (RDWC) can be understood as a combination of reactive distillation and dividing wall columns. During the last ten years, research on RDWC has shown its ability to reduce energy consumption and equipment cost.1−10 In the present work, we not only demonstrate its ability to save energy but also its superiority to the single RD column in improving conversion of the reactant. The hydrolysis of methyl acetate (MA) is taken as an example. MA is produced in large amounts as a byproduct in purified terephthalic acid and poly(vinyl alcohol) plants. Due to its low additional value as solvent, MA is always hydrolyzed to the more valuable chemicals methanol (MeOH) and acetic acid (HAc). The fixed-bed reactor (FBR) technology is the first industrialized technology for MA hydrolysis. The equilibrium conversion of MA is very low in the FBR because the MA hydrolysis is a reversible reaction with small equilibrium constant. It is impossible to separate unreacted MA from the reaction mixture via a distillation column due to the existence of the minimum azeotrope between MeOH and MA. The FBR for MA hydrolysis was followed by several distillation columns to separate and purify the reactants and products. Figure 1a shows the flow sheet for the FBR technology. Reactive distillation (RD) is a process where reaction and separation take place instantaneously in the same zone of a single distillation column. This technology has been considered to be an excellent alternative for improving the conversion of equilibrium-limited chemical reactions because the products © 2017 American Chemical Society

can be separated continuously from reactants as the reaction proceeds.11,12 The conventional multiunit reactor/column/ recycle system is also an alternative to equilibrium-limited chemical reactions. It has been proven that replacing the conventional multiunit reactor/column/recycle system with an RD column could reduce costs by a factor of 2−3 due to more effective energy utilization and less equipment investment.13 The liquid-phase reaction system where the reactant is the lowboiler is most unfavorable system for RD because the volatilization of reactant leads to its concentration reduction in liquid. Nevertheless, the saving of energy and equipment costs could even be realized for such an unfavorable reaction system.13 Fuchigami14 was the first to apply the RD technology to hydrolyze MA. It is impossible to conduct MA hydrolysis in a reactive distillation column with stoichiometric methyl acetate and water as feeds, methanol as overhead, and acetic acid as bottom because the reaction could not make the minimum MA-MeOH azeotrope disappear. Therefore, the RD column they adopted was operated at total reflux on the top with the products and unreacted reactants withdrawn from the bottom. It was reported that the conversion of MA could be remarkably improved, and the heat requirement was estimated to be about half that of the FBR technology. Such an RD column is always Received: Revised: Accepted: Published: 9177

May 8, 2017 June 29, 2017 July 21, 2017 July 21, 2017 DOI: 10.1021/acs.iecr.7b01907 Ind. Eng. Chem. Res. 2017, 56, 9177−9187

Article

Industrial & Engineering Chemistry Research

RD column as traditional reactive distillation column (TRDC) and the flow sheet in Figure 1b as TRDC technology in the present work for the convenience of statement. Xiao et al.15 investigated in detail the effects of several operational parameters on the MA conversion, following the work of Fuchigami.14 Their experiment results showed that a high feed ratio of water to MA is necessary for high MA conversion.15 The work of Kim et al.16 showed that considerable MA conversion could also be achieved with the TRDC even though MA-MeOH azeotrope was used as feed instead of pure MA. Lee17 and Lin et al.18 used a fixed bed reactor as the reflux drum of the TRDC to reduce the total cost. Instead of operating the RD column at total reflux, Han et al.19 withdrew the products from both ends of the column. Pöpken et al.20 found that the equilibrium-stage concept is sufficient to model MA hydrolysis in such an RD column. In comparison with the RD column Han et al.19 applied, the TRDC has the advantage of improving the MA conversion but the disadvantage of losing the separating capacity.21 In addition, an RD process of MA hydrolysis intensified by MeOH dehydration was proposed to avoid separation of the hydrolysis mixture,22 but its industrial operability needs to be further examined. Among the MA hydrolysis technologies mentioned above, the one Fuchigami proposed14 attracted the most attention due to its prominent industrial applicability. However, there are two shortcomings for this configuration. The first one is that separation of the hydrolysis mixture remains an energyintensive process due to the total reflux operation and excessive supply of water. Several energy-saving technologies have been applied to TRDC, and considerable energy savings were achieved.23−26 The other shortcomings is that high MA conversion (>99%) is difficult to realize.14,15,27 The low MA conversion leads to the complex separation process of the hydrolysis mixture. Up to now, the effective solution to this shortcoming is limited. In the present work, the RD column Fuchigami14 designed was coupled with the MeOH distillation column following it, forming an RDWC to achieve high MA conversion (>99%). Figure 1c illustrates the schematic of this RDWC, which has a divided overhead section and common bottom section. Although Li et al.25 studied the issues of design and control for MA hydrolysis in such an RDWC by Aspen Plus, they did not pay any attention to its ability to realize high MA hydrolysis conversion. In this work, we first analyzed the reason for the difficulty in realizing high MA conversion (>99%) in TRDC and explained why the RDWC could be applied to overcome this difficulty. Then, the effects of several operation parameters on MA conversion were investigated experimentally to verify the feasibility of achieving more than 99% MA conversion in an RDWC. Finally, the superiority of RDWC to TRDC in MA hydrolysis was showed by the simulation results of Aspen Plus.

2. PROBLEM ANALYSIS AND SOLUTION The hydrolysis mixture withdrawn from the bottom of the TRDC at >95% conversion of MA mainly contains water, MeOH, and HAc.14 The hydrolysis mixture contains 20−30 wt % HAc at different feed mole ratios of water to MA with the temperature ranging from 80 to 90 °C.14 Because the MA hydrolysis to MeOH and HAc is an acid-catalyzed reversible reaction, the high concentration of HAc would inevitably catalyze the esterification reaction of MeOH with HAc in the hydrolysis mixture due to the high concentration of MeOH and HAc and the high temperature, resulting in the formation of

Figure 1. Development of the technologies for MA hydrolysis: (a) FBR, (b) TRDC, and (c) RDWC.

followed by a distillation column to separate MeOH from the hydrolysis mixture, as illustrated in Figure 1b. We defined this 9178

DOI: 10.1021/acs.iecr.7b01907 Ind. Eng. Chem. Res. 2017, 56, 9177−9187

Article

Industrial & Engineering Chemistry Research MA.14,28 Although the self-catalyzed reaction of HAc and MeOH proceeds very slowly, the formation amount of MA could not be neglected due to the long residence time of the hydrolysis mixture in the bottom of the TRDC. Therefore, it is difficult to realize the complete conversion of MA in the TRDC due to the existence of the self-catalyzed MeOH−HAc esterification reaction. Fuchigami pointed out that 99.5% MA conversion could be obtained in a lab-scale RD column by greatly reducing the volume of the column bottom.14 However, it is impractical to greatly reduce the bottom volume for a plant-scale reactive distillation column because the small bottom volume is unfavorable for the stable liquid-level control. Another method to increase the MA conversion is to improve the feed mole ratio of water to MA.15 It could be deduced from the work of Xiao et al.15 that water−MA mole ratio higher than 7 is required to achieve more than 99% MA conversion at their operation conditions. However, high water−MA mole ratio would lead to the high energy consumption of the subsequent water−HAc separation process. Therefore, more exercisable and economical methods should be developed to suppress the self-catalyzed MeOH−HAc esterification reaction in TRDC to realize the high hydrolysis conversion (>99%) of MA. An effective method to suppress the self-catalyzed MeOH− HAc esterification reaction in TRDC is to remove MeOH or HAc from the hydrolysis mixture. There is no azeotrope in the ternary system of MeOH−water−HAc, as shown in Table 1.

catalyzed reaction, RDWC is a better alternative from the point of cost savings. First, RDWC can save column equipment cost due to the integration of two columns into one shell.8 Second, only one reboiler is required for the RDWC shown in Figure 1c, while the side-draw TRDC technology needs two reboilers, one for the side-draw TRDC and the other for the distillation column following the side-draw TRDC. In conclusion, the RDWC was applied for our further study.

3. EXPERIMENTAL SECTION 3.1. Chemicals and Analytical Method. MA (mass purity ≥98%), MeOH (mass purity ≥99.5%), and HAc (mass purity ≥99.5%) were purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd. Deionized water was prepared in our laboratory. C100E, a strong acid cation exchange resin functionalized with sulfonic groups, was purchased from Purolite Co., Ltd. (China) and used as catalyst for MA hydrolysis in the present work. The relevant characteristics of C100E resin are summarized in Table S1 in the Supporting Information. MA and MeOH were quantitatively analyzed by a gas chromatography (GC-2014, Shimadzu) equipped with a hydrogen flame ionization detector and a nonpolar capillary column (50 m × 0.32 mm × 0.5 μm). Nitrogen gas was used as the carrier gas at 0.1 MPa. Both the temperatures of the injector and the detector were set at 523.15 K. The column temperature was first kept at 313.15 K and then increased at 10 K·min−1 to 393.15 K. After, it was kept at 393.15 K for 1 min and then increased at 20 K·min−1 to 453.15 K. Finally, the column temperature was increased at 30 K·min−1 to 523.15 K and kept for 2 min. N,N-dimethylformamide was used as internal standard for the quantitative analysis of MA and MeOH. The concentration of HAc was detected by the method of acid−base titration with sodium hydroxide solution as standard solution and phenolphthalein as indicator. The water content was detected by a Karl Fischer titration (KLS-411, INESA Scientific Instrument Co.,Ltd.). 3.2. Apparatus and Procedure. The schematic of the labscale RDWC is illustrated in Figure 2. Two parallel glass columns representing the divided parts of RDWC were connected to a common stripping section through a Y-type junction. The two parallel columns were 1.5 m high with internal diameters of 24 mm. The left column named as main column was equipped with 1-m-high catalytic packing in the upper section and 0.5-m-high Θ ring packing in the lower section, playing the role of reactive distillation column. The catalytic packing was a series of catalyst bundles that consists of metal wave mesh and catalyst particles encapsulated with nylon cloth bags, as shown in Figure 2. The detailed structure of the catalyst capsules can be found in ref 29. The right column named as MeOH rectifier was filled with Θ ring packing for the separation of MeOH from the hydrolysis mixture. The common stripping section was 0.5 m high with an internal diameter of 35 mm and packed with 3 × 3 mm Θ ring. Two condensers equipped with reflux ratio controller were placed on the upper ends of these two parallel columns. The reflux volume of the main column was measured through a specially designed structure which is the combination of a valve and a scaled branch tubule.15 A three-necked round-bottom glass flask with a volume of 1200 mL was used as the column bottom. The middle neck of the flask was connected to the common stripping section. The other two necks were prepared for withdrawing sample and measuring reaction temperature,

Table 1. Boiling Temperatures of Pure Compounds and Azeotrope for the Quaternary System of MeOH−HAc− water−MA at 101.3 kPa compound

mass fraction (%)

temperature (K)

MeOH−MA MA−water MA MeOH water HAc

33.56−66.44 91.55−8.45 100 100 100 100

326.72 329.81 330.95 337.85 373.15 391.05

Additionally, MeOH has the minimum boiling point in this ternary system. Therefore, the RDWC with common bottom section is a good alternative to separate MeOH from the hydrolysis mixture. The configuration of the RDWC we used is illustrated in Figure 1c. A main column and a rectifier are integrated in one column, as shown in Figure 1c. The catalyst packing was loaded in the main column. It should be operated at total reflux and divided from the rectifier on the top to avoid drawing the azeotrope of MeOH−MA from the RDWC. The high conversion of MA can be achieved in the main column. The rectifier is used to momentarily separate the product MeOH from the hydrolysis mixture to avoid the occurrence of the self-catalyzed MeOH−HAc esterification reaction as far as possible. The rectifier is operated at partial reflux with highly purified MeOH continuously taken from the top. Except for RDWC, another solution to suppress the selfcatalyzed reaction in TRDC is to draw a vapor stream below the catalytic section of TRDC to remove MeOH from the bottom. The drawn stream mainly contains MeOH and water, which could be separated by a distillation column. Such a method has been applied by Gao et al.24 to reduce the energy consumption of the TRDC technology. Although both TRDC with side-draw and RDWC have the ability to suppress the self9179


Article


Figure 2. Schematic of the experimental setup.

respectively. The temperatures of the reflux liquid into the columns, the liquid at the bottom of the catalytic packing, and the liquid in the three-necked glass flask were measured by calibrated PT-100 thermocouples. The liquid in the flask was heated by an electric heating jacket to produce vapor phase for the RDWC. The distribution of vapor phase in main column and rectifier was adjusted by the valve in the Y-junction. The main column was operated at total reflux, while the rectifier was operated at partial reflux. The reactants, water and MA, were separately supplied into the main column by two peristaltic pumps from the top and bottom of the reaction zone, respectively. The products were withdrawn from the top of the rectifier and the bottom of the RDWC. All of the experiments were conducted at atmospheric pressure.

All of the model parameters were taken from Aspen Plus 7.3 data bank. 4.2. Reaction Kinetics. The reaction kinetics for MA hydrolysis on catalyst CE-100 was determined experimentally in our laboratory. The experimental data were used to fit the parameters of the activity-based pseudohomogeneous (PH) kinetic model. The determination method, experimental results, and model fitting are given in the Supporting Information. The PH model is written as rMA = −

1 dCMA = k+αMAαWater − k−αMeOHαHAc W dt

(1)

with

4. MODELING All simulations were performed with the RADFRAC module by the commercial simulation software Aspen Plus 7.2. This module is based on the equilibrium-stage model for solving the mass balance, phase equilibrium, summation, and energy balance (MESH) equations. The theoretical plates are numbered from top to bottom with the total condenser being the first stage and the reboiler being the last stage. 4.1. Vapor−Liquid Equilibrium. Accurate phase equilibrium data is necessary for the simulation of reactive distillation process. The UNIQUAC30 and Hayden−O’Connell (HOC)31 models were used to describe the liquid nonideality and dimerization of acetic acid in vapor phase, respectively.18,19,25,26

⎛ −70.767 × 103 ⎞ k+ = 1.644 × 108⎜ ⎟ RT ⎝ ⎠

(2)

⎛ −51.782 × 103 ⎞ k− = 3.083 × 106⎜ ⎟ RT ⎝ ⎠

(3)

where rMA is the reaction rate of MA catalyzed by ion-exchange resin with the unit of mol/g(cat)/min, W is the catalyst concentration with unit of g/m 3 , C MA is the molar concentration with unit of mol/m3, k+ is the rate constant of MA hydrolysis with units of mol/g(cat)/s, k− is the rate constant of MeOH−HAc esterification with units of mol/ g(cat)/s, α is the liquid activity, R (8.314 J/mol/K) is the ideal gas constant, and T is the reaction temperature in Kelvin. The reaction rate on native catalyst particle is different from that in 9180


Article


Figure 3. Effect of process parameters on MA hydrolysis in RDWC. (a) Feed water−MA molar ratio, Rm. (b) Heat duty, He. (c) Feed mole flow rate of MA, Sv. (d) Mass ratio of side product to feed MA, Sm.

this kinetic model was set in the theoretical plates where the catalyst capsule was loaded. The kinetic model for the self-catalyzed esterification reaction of MeOH with HAc was given by Pöpken et al:28

catalyst capsule due to the concentration gradient between inside and outside of the capsule. An effectiveness factor η is defined to describe the difference:32 η=

reaction rate in catalyst capsule (rcap) reaction rate in batch reactor (rMA)

(4)

rMA,self =

The magnitude of this effectiveness factor depends on the mass transfer driving force and resistance between the inside and outside of the capsule. In a special RD loaded with catalytic capsule, the mass transfer driving force and resistance are related to the reactant concentrations outside the capsule and the liquid film thickness on the capsule, respectively. These two factors could be varied by the molar ratio of water to MA and the flux of liquid phase, respectively. The dependence of effectiveness factor on water−MA molar ratio and liquid flux has been correlated for MA hydrolysis in catalytic capsule:24 η = 0.3431 + 311.3034L − 0.1573ln R m

dxMA + − = αHAc(kself αMAαwater − kself αMAαHAc) dt (6)

with ⎛ −6.35 × 104 ⎞ + = 5.11 × 105exp⎜ kself ⎟ RT ⎝ ⎠

(7)

⎛ −8.02 × 104 ⎞ − kself = 9.83 × 106exp⎜ ⎟ RT ⎝ ⎠

(8)

where xMA is the mole fraction of MA and k+self and k−self are the rate constants of forward and backward reactions for the selfcatalyzed MeOH−HAc esterification reaction with unit of 1/s, respectively. In the simulation, the self-catalyzed kinetic model was set in the theoretical plates where the Θ ring packing was loaded. 4.3. Liquid Holdup and Separation Efficiency. The liquid holdup in the catalytic capsule should be specified by using the RADFRAC module to simulate the reactive dividing wall distillation process because the liquid holdup has an

(5)

where L is the flux of liquid phase with unit of m3/m2/h and Rm is the feed mole ratio of water−MA. Therefore, the actual reaction rate of MA in catalyst capsule could be calculated by the reaction kinetics model with effectiveness factor rcap = ηrMA. Because the reaction kinetics model in Aspen Plus does not contain an effectiveness factor, the reaction kinetics model with effectiveness factor was incorporated into the process simulator using an additional FORTRAN subroutine. In the simulation, 9181


Article

Industrial & Engineering Chemistry Research influence on the extent of the reaction. The liquid holdup of the catalytic capsule was previously measured by Xu et al.:33 hd = 0.0336u 0.0109L0.429

uneconomical to further increase the heat duty as the required MA conversion is achieved. 5.1.3. Feed MA Mole Flow Rate, Sv. The effect of Sv on MA conversion in RDWC is illustrated in Figure 3c. It can be seen that the MA conversion reduces from 99.6 to 98.5% with Sv rising from 0.6757 to 1.0811 mol·h−1 respectively. The similar variation tendency was also observed in other configurations of the RD column for MA hydrolysis.5 The declining MA conversion with increasing Sv should be contributed to the shorter residence time of reactants in the reaction zone of the RDWC at larger Sv. Besides, larger Sv means more feed throughput, inevitably inducing the reduction of the separation efficiency of the main column. As a result, more hydrolysis products stay in the reaction zone, inhibiting the reaction of water with MA. Both of these two effects lead to the reduction of MA conversion with rising Sv. Although rising Sv is adverse to high MA conversion, the column equipment is always expected to be operated at higher Sv to achieve larger production capacity and create more economic benefits. Therefore, as long as the MA conversion is satisfied, larger Sv is expected. 5.1.4. Mass Ratio of Vapor Entering into MeOH Rectifier to Feed MA, Sm. Because it is difficult to experimentally determine the vapor distribution ratio in RDWC, we replaced this parameter by the mass ratio of vapor entering into MeOH rectifier to feed MA, Sm. The mass flow rate of the vapor entering into the rectifier was calculated through multiplying the mass flow rate of the liquid drawn from the top of the MeOH rectifier with the reflux ratio of the MeOH rectifier. The effect of Sm on MA conversion is displayed in Figure 3d. It can be seen that the MA conversion grows from 94.5 to 99.3% with Sm decreased from 3.9 to 1.6, respectively. The explanations for the effect of Sm on MA conversion are similar to those for the effect of heat duty on MA conversion. The results in Figure 3d indicate that the vapor distributor of the RDWC should be carefully designed to bring more vapors into the main column to improve the MA conversion. However, it is inadvisable to operate the RDWC at too small of an Sm. With less vapor entering into the rectifier, the separation performance of the rectifier is likely to reduce. This will result in the incomplete separation of MeOH from the hydrolysis mixture. The coexistence of MeOH and HAc in the RDWC is adverse to realizing complete conversion of MA. Therefore, it is advised to operate the RDWC for MA hydrolysis at small Sm only if the separation requirement of the rectifier is satisfied. 5.1.5. Response Surface Methodology (RSM) Analysis. In this section, RSM analysis was applied to optimize the operation conditions based on a three-factor three-level Box− Behnken design (BBD). Four manipulated variables, including Rm, He, Sv, and Sm, were studied in Section 5.1. Because Sv impacts only the column sizing in industrial applications, it was ignored in the RSM analysis to produce a three-variable test matrix. The results of BBD are listed in Table 2. The MA conversion obtained by Aspen Plus was fitted with a second order polynomial equation:

(9)

where hd is the liquid holdup with unit of m3/m3 and u is the gas superficial velocity with unit of m/s. The liquid holdup is in the stripping section of main column; the public stripping section, the MeOH rectifier, and the reboiler were also specified because the self-catalyzed esterification reaction of MeOH with HAc occurs there. The liquid volume in the reboiler was 900 mL in our experiments. The liquid holdup for the Θ ring packing was specified to be 8% of its superficial volume.34 The height equivalent to a theoretical plate (HETP) of the catalytic capsule was set as 0.13 m according to our experimental measurement. The HETP of the 3 × 3 mm Θ ring packing was specified as 0.06 m according to ref 35.

5. RESULTS AND DISCUSSION 5.1. Effects of Process Parameters on MA Hydrolysis in RDWC. 5.1.1. Feed Water−MA Molar Ratio, Rm. The effect of Rm on MA conversion in RDWC is shown in Figure 3a. It indicates that a high feed ratio leads to a high MA conversion. Similar results were also reported for the MA hydrolysis in TRDC.15 The increase in conversion with Rm should be attributed to the fact that the MA hydrolysis is a reversible reaction. The increase in Rm not only facilitates the MA hydrolysis in the region packed catalyst but also suppresses the self-catalyzed MeOH−HAc esterification reaction in the region where Θ ring packing is loaded. In the TRDC for MA hydrolysis, Rm is the most important operating variable to improve the MA conversion.15,24 In the RDWC, Rm also plays a significant role. As shown in Figure 3a, the mole ratio of water− MA rises merely from 3 to 4, while the MA conversion increases by about 10% (from 90.2 to 98.8%). With the further increase in Rm, 99.5% MA conversion was achieved. Although increasing the water−MA mole ratio is favorable for improving the MA conversion, more energy consumption is required to separate water from HAc. Therefore, a suitable feed mole ratio of water−MA is expected to simultaneously realize high MA conversion and low energy consumption. 5.1.2. Heat Duty, He. The effect of He on MA conversion is given in Figure 3b. It can be seen that the MA conversion can be improved from 88.0 to 99.5% with the heat duty increased from 0.06 to 0.09 kW, respectively. Because the main column of the RDWC was operated at total reflux, increasing the heat duty leads to the increase in reflux volume. Increasing the reflux volume improves the separation performance of the main column, which means that more hydrolysis products can be removed from the reaction zone to increase the MA conversion. The increase in the reflux liquid will also introduce more MA into the reaction zone because the reflux liquid is an azeotrope in which the mass concentration of MA is about 90%. With more reflux liquid introduced into the reaction zone, the concentration of MA is increased, while the concentrations of the MeOH and HAc are decreased, promoting the growth of MA conversion. Moreover, the increase in heat duty can improve the separation performance of the rectifier so that MeOH can be more thoroughly separated from the hydrolysis mixture in the column bottom to inhibit the occurrence of the self-catalyzed esterification reaction. All of these effects together give rise to the increase in MA conversion. Of course, it is

Y = 98.03 + 8.31X1 + 9.57X 2 − 4.01X3 − 0.71X1X 2 + 1.30X1X3 + 6.48X 2X3 − 6.12X12 − 7.54X 22 − 2.08X32 (10)

Y is the MA conversion. X1, X2, and X3 represent coded values of variables Rm, He, and Sm, respectively. The determination coefficient R2 of eq 8 is 0.9710, indicating that eq 8 can predict within the range of experimental variables. The analysis of 9182


Article


equilibrium behavior in the reactive distillation column for MA hydrolysis.20,24 Pöpken et al.24 showed that the assumption of chemical equilibrium on each theoretical stage is unreasonable for the MA hydrolysis catalyzed by the heterogeneous ionexchange resin catalyst, so the reaction kinetic model should be considered in the simulation. The kinetic models given in Section 4.2 were applied in the simulation. The comparison between the simulation results and the experimental data is shown in Figures 3a−d for MA conversion. An error of less than 2% was found in the comparison between the simulation results and experimental data, which verified the creditability of the models used. 5.2.2. Superiority of RDWC to TRDC for MA Hydrolysis. MA hydrolysis in TRDC was simulated under conditions similar to those for MA hydrolysis in RDWC. In the simulation of TRDC, some parameters should be carefully specified to make the simulation results comparable to those of RDWC. Detailed specifications of these parameters are given in the Supporting Information. Table 4 lists the performance comparison between RDWC and TRDC for MA conversion of 99 and 99.5%. It should be

Table 2. BBD Design and Results for the Optimization of Operation Conditions variable

conversion (%)

run

Rm (mol/mol)

He (kw)

Sm (g/g)

by Aspen Plus

by eq 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

5 5 5 5 3 5 5 7 7 7 3 7 5 5 5 3 3

0.06 0.09 0.06 0.12 0.06 0.09 0.09 0.06 0.12 0.09 0.09 0.09 0.09 0.09 0.12 0.12 0.09

1.5 2.0 2.5 2.5 2.0 2.0 2.0 2.0 2.0 2.5 2.5 1.5 2.0 2.0 1.5 2.0 1.5

90.87 98.03 65.10 98.89 67.69 98.03 98.03 83.60 99.63 98.89 77.55 99.51 98.03 98.03 98.77 86.56 83.35

89.31 98.02 66.35 98.44 65.77 98.02 98.02 83.81 99.54 95.42 76.21 99.84 98.02 98.02 96.51 86.34 85.82

Table 4. Performance Comparison between the RDWC and TRDC

variance (ANOVA) for the response surface quadratic equation is presented in Table 3. The p-value is used to evaluate the

MA conversion XMA

Table 3. ANOVA for the Fitted Polynomial Quadratic Model source

sum of squares

df

mean square

F-value

p-value

model X1 X2 X3 X1X2 X1X3 X2X3 X12 X22 X32

2043.45 552.41 733.27 128.54 2.01 6.71 167.71 157.73 239.30 18.24

9 1 1 1 1 1 1 1 1 1

227.05 552.41 733.27 128.54 2.01 6.71 167.71 157.73 239.30 18.24

26.02 63.31 84.04 14.73 0.23 0.77 19.22 18.08 27.43 2.09

0.0001