Acetic Acid Esterification with Ethanol by Reactive


Mixed Succinic Acid/Acetic Acid Esterification with Ethanol by Reactive...

4 downloads 110 Views 2MB Size

ARTICLE pubs.acs.org/IECR

Mixed Succinic Acid/Acetic Acid Esterification with Ethanol by Reactive Distillation. Alvaro Orjuela,†,‡ Aspi Kolah,† Carl T. Lira,† and Dennis J. Miller*,† †

Department of Chemical Engineering and Materials Science, Michigan State University, 2527 Engineering Building, East Lansing, Michigan 48824, United States ‡ Departamento de Ingenieria Quimica, Universidad Nacional de Colombia, Ciudad Universitaria, Bogota, Columbia ABSTRACT: Esterification of mixtures of succinic acid and acetic acid, commonly produced via fermentation of biomass carbohydrates, with ethanol in a continuous reactive distillation unit has been studied. Experiments were carried out in a 6 m tall, 51 mm diameter pilot-scale stainless steel column with the reactive zone consisting of Katapak-SP11 structured packing containing Amberlyst 70 cation exchange resin as catalyst. Noncatalytic BX structured gauze packing was used in stripping and enrichment sections. Steady-state experiments were performed under different conditions by varying the composition and number of inlet streams, the column pressure, the reboiler power, and the reflux ratio. Conversions approaching 100% for both succinic acid and acetic acid were verified experimentally; succinate esters were obtained as bottom products with purities of diethyl succinate as high as 98%, and ethyl acetate was recovered in the distillate. Computer simulations based upon experimental phase equilibrium data and chemical kinetics were performed in Aspen Plus using RadFrac to reproduce steady-state results. Good agreement between experiment and simulation was observed under diverse operating conditions. The model developed here can be used to design commercial-scale systems for succinic acid production when acetic acid is also formed during the fermentation.

1. INTRODUCTION Fermentation of renewable substrates is an alternative approach to produce many chemical commodities currently obtained via petrochemical routes. The synthesis of carboxylic acids has gained major attention because it occurs naturally in metabolism of many microorganisms and because organic acids can be used as building blocks for many chemicals of commercial interest.15 Among these, succinic acid (SA) has been recognized as a potential substitute for maleic anhydride to be used in the synthesis of 1,4 butanediol, γ-butyrolactone, and tetrahydrofuran. SA can also be used to produce biodegradable deicing agents, biopolymers, surfactants, plasticizers, green solvents, and a large variety of value-added products.69 Despite the large variety of organisms known to produce SA, the low titer obtained in the broth (50100 kg/m3) and contamination with byproducts represent major challenges in the product separation and purification stages. Reports indicate that acetic acid (540 kg/m3), pyruvic acid (020 kg/m3), lactic acid (014 kg/m3), formic acid (05 kg/m3), traces of other acids, and even ethanol are also found in final products depending on the microorganisms used during fermentation.2,1012 Among many separation alternatives, reaction of fermentation products with alcohols such as with ethanol (EtOH) has been successfully applied to isolate SA in the form of esters.13 In this process, separation occurs by esterification of the acid mixture while inorganic salts are removed from the organic solution. Using this approach, different esters can be obtained simultaneously and separation of the complex mixture can be performed by distillation. However, the extent of esterification is limited by chemical equilibrium, and thus removal of reaction products is required to drive reaction to completion. r 2011 American Chemical Society

To circumvent this thermodynamic limitation, reactive distillation (RD) can be used as a means of process intensification1417 to overcome equilibrium constraints by removing water and simultaneously separating the different esters recovered from fermentation. Reactive distillation also reduces capital and energy costs because it combines reaction and separation in a single process unit. Although RD has been successfully applied in esterification of single acids with EtOH, major challenges appear when processing SA mixed with other carboxylic acids. Because SA has low volatility and low solubility in most solvents, fast esterification kinetics are required to avoid accumulation or precipitation at the bottom of the column, where succinate species are withdrawn as products. Rapid kinetics are also required for other heavy acids present in the mixture (e.g., lactic acid, pyruvic acid, and malic acid) to ensure they do not end up as free acids in the bottoms and catalyze hydrolysis of the esters formed. On the other hand, if SA is mixed with carboxylic acids (e.g., formic acid, acetic acid) that form volatile esters, the specification of feed location, boilup rate, and reflux ratio can strongly affect column performance because light acids and their esters are removed in the distillate stream. To achieve complete esterification of volatile acids, they must have sufficient residence time in the reactive zone of the column. Reflux may be required to avoid free acids removal, but refluxing also returns water to the column which adversely affects performance. Received: January 19, 2011 Accepted: May 26, 2011 Revised: May 19, 2011 Published: July 06, 2011 9209

dx.doi.org/10.1021/ie200133w | Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Figure 1. Reactions in mixed acid esterification: esterification of succinic acid (SA) to monoethyl succinate (MES) and diethyl succinate (DES); esterification of acetic acid (AcAc) to ethyl acetate (EtAc); dehydration of EtOH to diethyl ether (DEE).

Esterification of mixed acetic, propionic, and butyric acids with methanol using sulfuric acid as catalyst was reported as one of the first continuous RD processes.18 Since then, few applications have been reported, but recently the esterification of maleic acid in aqueous solution with other carboxylic acids has been examined.19 Previous studies from our group indicate that esterification of SA with EtOH can be carried out in a RD unit to achieve high SA conversion and selectivity to diethyl succinate (DES) under continuous operation.20,21 Considering that the most common impurity in SA fermentation broths is acetic acid (AcAc), this work explores esterification of mixtures of SA and AcAc with EtOH using a continuous RD system. Figure 1 presents a chemical description of the mixed acid esterification including dehydration of EtOH to diethyl ether (DEE). Experiments to evaluate the feasibility of mixed acid esterification were conducted in a pilot-scale RD column under different conditions. Simulations to verify experimental observations and to explore improvements in the system were performed using AspenPlus process design software. The effect of major processing variables on SA and AcAc conversions, separation efficiency, and DES purity were studied. On the basis of these experimental and modeling studies, a preferred configuration of a reactive distillation process to obtain pure succinate esters is presented.

2. MATERIALS AND METHODS 2.1. Materials. Succinic acid (>99.5%, Sigma-Aldrich), acetic acid (99.9%, Aristar), diethyl succinate (99.92%, Sigma-Aldrich), monoethyl succinate (89.3%, Sigma-Aldrich), ethyl acetate (HPLC grade, J. T. Baker), ethanol (200 proof, Decon Laboratories), water (HPLC grade, J.T. Baker), diethyl ether (EMD Chemicals, 99.9%) n-butanol (99.9%, Mallinckrodt), and acetonitrile (HPLC grade, EMD Chemicals) were used for experiments and analysis. Hydranalcoulomat E solution (Riedel-de Ha€en) was used in Karl Fischer analysis. Amberlyst 70 and Amberlyst 15 ion exchange resins used as catalysts were purchased from Dow Chemical Co.

2.2. Analysis. Analysis was performed using gas chromatography (GC) for volatile components, liquid chromatography (HPLC) for succinate species, and Karl Fisher titration for quantification of water. Samples were diluted 50-fold in water for HPLC analysis. When samples contained high concentrations of DES, EtOH was also added (∼2 wt %) to ensure a homogeneous solution. A HPLC system with a Waters 717 autosampler, Waters 410 refractive index, and Perkin-Elmer LC90 UV detectors was used for characterization of succinate species. A 100  7.6 mm fast acid analysis column (Bio-Rad) was used for the separation. A 5 mM aqueous solution of H2SO4 flowing at 1 mL/min was used as the mobile phase. For GC analysis of ethanol, acetate species, diethyl ether, and diethyl succinate, samples were diluted 20-fold in acetonitrile using n-butanol as an internal standard (5 wt %). A Shimadzu 2010 gas chromatograph equipped with a flame ionization detector and AOC-5000 autoinjector was used with a 15 m Alltech EC-WAX column (0.53 mm i.d., 1.20 μm film thickness) heated under the following temperature program: column initial temperature of 313 K (1.37 min), ramp at 30 K/min to 353 K, ramp at 40 K/min to 523 K, and hold at 523 K for 4 min. The injector port was maintained at 553 K with a split ratio of 5:1, and detector temperature was 573 K. Hydrogen was used as carrier gas (50 cm3(STP)/min), with volume injections of 1 μL. Calibration curves for both GC and HPLC were developed by analysis of samples of known composition in the range of interest. Repeatability within 0.5% by mass was obtained. Karl Fischer analysis was performed using an Aquacount Coulometric titrator AQ-2100. 2.3. Reactive Distillation Column Description. Continuous RD experiments were carried out in a 51 mm i.d. stainless steel pilot-scale column. Table 1 describes major characteristics of the experimental equipment, and Figure 2 shows a scheme of the RD system. High separation efficiency structured packings (Sulzer Chemtech Ltd.) were used in both catalytic and noncatalytic zones. As shown in Figure 2, the column was built with six removable flanged sections, four reactive and two nonreactive. Reactive 9210

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Table 1. Characteristics of Reactive Distillate Column parameter

parameter

material of construction

SS-316

rated pressure (kPa)

2000

condenser type

rated temperature (K)

573

geometry

coiled concentric tubes

diameter (cm)

5.1

inner tube diameter (cm)

0.63

outer tube diameter (cm)

1.27

height (cm)

74

cooling fluid

waterglycol

packing

Sulzer BX

cooling fluid chiller

Julabo FL 2506

360

operating temperature range (K) cooling power at 263 K (kW)

258 313 0.3

stripping zone

reactive zone height (cm)

total

packing

Katapak-SP-11

catalyst

Amberlyst 70

prereactor construction material

Pyrex glass

catalyst particle size (mm)

>0.5

volume (m3)

0.075

catalyst loading (kg/m3)

76

maximum power (kW)

2

enrichment zone height (cm) packing reboiler type

stirrer power (kW)

0.25

42

catalyst

Amberlyst 15

Sulzer BX

catalyst loading (wt %)

2

kettle

heating element

stainless steel coated electrical resistance

maximum power (kW)

2.4

hold-up (m3)

0.0015

sections were separated by liquid re-distributors that allowed sample collection. The column and the reboiler were wrapped with electric heating tapes and glass wood bands. Internal chromelalumel probes sheathed in stainless steel and external surface thermocouples were used to register temperature profiles inside and outside the column. The tapes along the column height were individually controlled to match column temperature and minimize heat loss to the surroundings. Reboiler power was supplied by an electrical heater submerged in the liquid and controlled by an Omega controller. The liquid level was maintained with an overflow outlet controlled with a level indicator and a pneumatic valve. Column pressure was maintained with a solenoid valve that controlled the outlet vapor flow at the top of the column. Downstream from the condenser, a reservoir was used to collect liquid condensate product; reflux was dispensed from this reservoir via a controlled flow pump, and the flow rate was registered with a coriolis flowmeter. Liquid condensate that was not recycled was sent to the distillate product collection tank. Because succinate species and acetic acid have lower volatilities than EtOH, acid solution was fed from the upper part of the column. When the column operated without reflux, the acid feed stream was fed closer to the top of the column. Using reflux, acid feed was introduced just above the catalytic packing. In some experiments, an additional EtOH feed stream was introduced below the catalytic zone. Inlet streams were taken from tanks positioned on electronic balances via heat-traced and insulated pipelines and diaphragm pumps. Inlet temperatures were maintained as close as possible to the corresponding internal temperature of the column at the feed location. 2.4. Reactive Distillation Column Operation. Succinic acid was fed with ethanol in the feed stream at different concentrations. In most experiments, the SA concentration in the feed stream to the column was close to the solubility limit in EtOH at

room temperature (∼8 wt %). The concentration of AcAc in the same stream was 0.10.3 that of SA, typical of the concentration ratio observed from fermentation. When the acid feed stream contained a high concentration of SA, the feed was prereacted in a 70 L stirred glass reactor under reflux at atmospheric pressure in the presence of Amberlyst 15 catalyst for 24 h before the experiment to obtain a nearly equilibrated mixture of acids and ethyl esters in EtOH. In most experiments, EtOH was intentionally routed to the reboiler to maintain a reboiler temperature low enough to avoid unsafe conditions or thermal decomposition of succinate species. To reduce the time required to achieve steady state, liquid holdup drained from a previous run was kept in the reboiler. In typical operation, the column was started by slowly turning on the reboiler heater and the external heating tapes as the vapor reached the different zones of the column. After vapor reached the condenser, total reflux was maintained until a stable temperature profile and pressure were attained. Then feed pumps were started at the specified rates for the experiment, and collection of products from the distillate and bottoms streams was initiated. Stream flow rates, the acid value of the bottom product, the column pressure, and temperature profiles were recorded at time intervals of 1530 min throughout the run. Outlet flows from distillate and bottoms streams were obtained by measuring the volume and density of the liquid collected in volumetric flasks. The acid value (AV, mg of KOH/g) of the bottoms product was obtained by titration with a 0.1 M solution of NaOH in EtOHH2O until the end point was observed with phenolphthalein. After reaching steady conditions, liquid samples from column redistributors, reboiler, and distillate reservoirs were collected in sealed vials every hour and refrigerated until they could be analyzed. Experimental steady state was assumed after reaching bottoms AVs within a maximum 5% variation in consecutive samples, temperature profile changes over time of no greater than (0.5 K, 9211

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Figure 2. Pilot plant reactive distillation column with internal catalytic packing Katapack-SP11 (middle zone) and BX structured packing (top and bottom).

and after constant pressure is reached in the column ((7 kPa). Additional steady-state criteria were inlet and outlet flow rates within 5% of their set points and mass balance closures of 100 ( 5%. This last criterion was evaluated after the run ended because of the time required for sample preparation and chromatographic analysis. At the end of an experiment, final samples were collected, feed pumps were shut down, inlet and outlet valves were closed, and the reboiler and column heaters were turned off. Evaporation was stopped by pressurizing the column at ∼700 kPa with nitrogen. Finally, refrigeration at the condenser was stopped, and the general power supply was shut down.

3. RESULTS AND DISCUSSION 3.1. Reactive Distillation Experiments. Operating conditions for the RD experiments are listed in Table 2. In most experiments,

the flow rate and the temperature of the acid feed stream were maintained as constant as possible to facilitate comparison among different operating conditions. Typical dynamic behavior of the RD system is presented in Figures 3 and 4. In general, steady conditions were obtained after ∼20 h of operation. As observed in Figure 4, changing a process variable (EtOH feed rate, Table 2) from run 14A to run 14B drove the column to a new steady state in considerably less time (∼6 h). In most experiments, this intentional altering of process variables reduced the total column operating time and also minimized waste generation. Scatter observed in flow rates is a result of using overflow level control in the reboiler and the condensate reservoir, as this type of level control is affected by excessive foaming and also by uneven evaporation. During experiments, brief flow discontinuities in top and bottoms outlet streams were observed even at steady 9212

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Table 2. Experimental Conditions and Results for Continuous Pilot-Scale Reactive Distillation Studies top feed (balance EtOH) reboiler power column pressure reflux

a

SA

H2O

flow

AcAc

ratio (wt %) (wt %) (wt %)

bottoms bottoms feed (EtOH) distillate flow flow (g/min)

flow

T AVc

run prereacted

(kW)

(kPa gauge)

12A

no

0.96

310.3

0

6.7

2.2

30.3 ( 0.3

12B

no

1.15

310.3

0.59

6.7

2.2

30.2 ( 0.2

12C

no

1.41

310.3

0.36

6.7

2.2

30.3 ( 0.2

13A

no

1.03

310.3

0

6.7

2.2

30.7 ( 0.5

13B

no

1.44

310.3

0.97

6.7

2.2

30.3 ( 0.2

13C

no

1.68

310.3

0.65

6.7

2.2

30.4 ( 0.1

12.5 ( 0.2

14A 14B

yes yes

1.63 1.85

310.3 310.3

0 0

25.0 a 25.0 a

2.5 a 2.5 a

31.2 ( 0.4 30.9 ( 0.1

16.6 ( 0.6 28.7 ( 0.6

20A

yes

1.92

34.5

0.35

25.0 a

2.5 a

29.6 ( 0.3

30.5 ( 0.9

49.7 ( 1.0

9.6 ( 1.2

7.8 498.2

21A

no

1.49

34.5

0

8.0

0.8

41.9 ( 2.6

23.2 ( 0.4b

57.4 ( 3.6

7.8 ( 2.4

75.8 364.6

22A

yes

1.50

34.5

0

8.0 a

5.5 a

0.8 a

45.2 ( 0.7

23.5 ( 0.4b

58.8 ( 1.2

9.4 ( 1.6

42.9 365.6

a

a

a

45.0 ( 0.4

23.8 ( 0.4b

59.5 ( 0.5

8.5 ( 0.7

47.7 363.3

29.8 ( 0.3

44.0 ( 4.1

5.5 5.5

0.8

(g/min)

(g/min)

(g/min)

0.0

20.6 ( 0.4

10.1 ( 0.2

12.5 394.2

0.0

20.1 ( 0.5

10.6 ( 0.4

11.8 394.2

32.2 ( 0.7

9.4 ( 0.7

5.6 394.2

0.0

25.2 ( 0.7

5.4 ( 0.7

22.3 394.2

0.0

24.9 ( 0.4

5.8 ( 0.5

19.4 394.2

36.1 ( 0.4

6.3 ( 0.4

5.8 394.2

35.9 ( 0.7 46.5 ( 0.7

11.7 ( 0.8 11.7 ( 0.5

5.9 505.3 3.7 515.8

11.2 ( 0.5

(K)

22B

yes

1.94

34.5

0.29

8.0

23A

no

1.92

34.5

0

8.0

91.2

0.8

30.6 ( 0.6

24A

no

2.40

34.5

0

6.2

93.0

0.8

30.9 ( 0.4

55.4 ( 0.9

67.9 ( 1.7

17.3 ( 2.0

26A

no

2.35

34.5

0.58

6.7

92.7

0.6

29.4 ( 1.1

28.9 ( 1.2

40.9 ( 0.7

14.7 ( 1.0 119.2 381.3

17.2 ( 6.0 131.2 379.1 50.6 362.9

Top feed concentrations prior to prereaction. b 6 wt % H2O in EtOH. c Acid value, mg of KOH/g.

Figure 3. Outlet flow rates and bottoms acid value versus time during approach to steady state in runs 14A (01100 min) and 14B (11001600 min): (black open boxes) distillate flow, (blue filled circles) bottoms flow, and (red open triangles) bottoms acid value.

state, but fairly constant concentrations with time in bottoms and distillates were obtained in most experiments. A typical concentration profile with time at steady state is shown in Figure 4 (run 14A). This suggests that on-site acid value measurement is a rapid and good indicator of steady-state conditions. A summary of results obtained from RD experiments is given in Table 3. Reported concentrations and overall mass balances were obtained by reconciliation of different analytical methods. Succinate species are reported as the average of HPLC and GC analysis, while concentrations of other components are obtained by GC. In general, high conversions of SA and AcAc together with high selectivity to DES were achieved in most experiments. To evaluate the feasibility of processing acid streams diluted in water (resembling acidified fermentation broth), aqueous acid solutions were used as feed in runs 23A, 24A, and 26A. These were the only experiments in which low conversions were observed— operating under these low conversion conditions was challenging because unreacted succinic acid precipitates and obstructs outlet

Figure 4. Composition of bottoms and distillate products during approach to steady state (run 14A): (open circles) DES, (open boxes) MES, (open triangles) SA, (gray filled triangles) AcAc, (gray filled circles) EtAc, (black filled circles) H2O, (black filled triangles) EtOH, and (þ) DEE.

lines in the cooled bottom product. This affects liquid withdrawal from the reboiler, causing large flow fluctuations during operation. It is noted that analysis of runs 23A, 24A, and 26A showed that they did not reach steady state as defined earlier in this paper. 9213

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Table 3. Summary of Outlet Stream Compositions Obtained in Reactive Distillation Experiments distillate composition (wt %)

bottoms composition (wt %)

conversion SA

run

a

H2O DEE EtAc EtOH AcAc MES DES SA

H2O DEE EtAc EtOH AcAc MES DES

SA

AcAc

SDES a (%)

12A

3.1

2.1

4.2

90.4

0.1

0.1

0.0

0.0

2.9

0.0

0.1

61.6

0.1

2.9

32.3

0.1

99.2

94.9 89.5

12B

3.2

2.0

4.3

90.3

0.1

0.1

0.0

0.0

3.4

0.0

0.0

66.5

0.1

2.7

27.1

0.1

99.2

95.0 88.6

12C

3.3

1.3

2.6

92.5

0.1

0.1

0.0

0.0

0.8

0.0

0.0

66.5

0.1

1.4

31.0

0.1

99.4

94.7 93.2

13A 13B

3.0 3.9

1.4 1.2

3.6 3.4

91.8 82.4

0.1 0.1

0.1 0.1

0.0 0.0

0.0 0.1

2.1 1.9

0.0 0.0

0.0 0.0

37.3 55.8

0.1 0.1

5.5 4.3

54.5 37.7

0.3 0.1

98.7 98.2

94.4 88.7 95.1 86.9

13C

2.9

1.6

2.3

93.0

0.1

0.1

0.0

0.1

0.3

0.0

0.0

48.6

0.1

1.2

49.6

0.2

98.6

94.1 96.3

14A

8.2

1.8

2.7

87.2

0.0

0.1

0.0

0.0

0.1

0.0

0.0

0.5

0.1

1.2

98.0

0.2 95.3 b 99.7 82.1 b 97.9 98.4

14B

6.1

1.8

2.0

90.0

0.0

0.1

0.0

0.0

0.1

0.0

0.0

0.4

0.1

0.9

98.3

0.2 95.3 b 99.7 82.1 b 97.4 98.7

20A

5.1

0.1

1.7

92.9

0.1

0.0

0.0

0.2

0.0

0.0

0.0

6.5

0.0

2.3

90.4

0.6 96.8 b 97.6 86.1 b 95.5 97.0

21A

7.6

0.0

0.7

91.6

0.1

0.0

0.0

0.0

6.0

0.0

0.0

35.4

0.1

13.8

42.9

22A

7.4

0.0

0.5

92.1

0.0

0.0

7.4

0.0

3.7

0.0

0.0

34.4

0.3

9.7

3.7

0.0 96.8 b 98.2 78.5 b 87.0 81.5

22B 23A

6.9 40.3

0.0 0.0

0.7 0.3

92.3 59.2

0.1 0.1

0.0 0.0

6.9 0.0

0.0 0.1

6.1 80.5

0.0 0.0

0.0 0.8

35.4 5.4

0.4 0.0

10.8 4.3

6.1 0.1

0.0 96.8 b 98.0 78.5 b 78.9 78.2 8.8 28.4 77.7 2.2

24A

34.0

0.0

0.2

65.5

0.1

0.0

0.0

0.1

34.8

0.0

0.0

53.2

0.1

5.3

4.4

2.3

72.6

60.9 41.0

26A

36.1

0.0

0.4

63.3

0.1

0.0

0.0

0.0

85.9

0.0

0.0

0.0

0.1

1.9

0.1

12.0

11.7

70.1 5.3

1.8

95.7

85.4 72.2

Selectivity in (mol of DES formed)/(mol of SA converted). b Conversion obtained in prereactor.

In runs 21A, 22A, and 22B, acid feed was prepared in 190 proof EtOH (∼6 wt % H2O) and prereacted before being fed. These experiments were performed to evaluate the feasibility of operation using recycled ethanolwater azeotrope for esterification. Because of the large excess of EtOH in the system, the column operated within a few degrees of the EtOH boiling point at the column pressure. However, the reboiler temperature was higher because of the high concentration of succinate species and was strongly affected by EtOH concentration. This is clearly observed in runs 14B and 20A, carried out under the same pressure with similar feeds. An increase of ∼20 °C in reboiler temperature was observed with only a small change in EtOH concentration. Even though the reboiler operated at high temperature in some experiments, no degradation products were observed in chromatographic analysis. 3.1.1. Effect of Reflux Ratio. In experiments where the feed streams contained no water, conversions of SA and AcAc were not significantly affected by increasing reflux ratio (runs 12A and 12B) because of the large excess of EtOH used. However, a slight decrease was observed in DES selectivity with reflux ratio as distillate flow rate was increased (runs 13A and 13B). This effect was magnified when water was introduced into the feed streams (runs 22A and 22B). In this case, conversion of AcAc dropped significantly as well as selectivity to DES because water is recycled into the system. 3.1.2. Effect of Column Pressure. As column pressure increased and column temperature followed, esterification rates and overall conversions were not particularly enhanced. However, enhanced dehydration of EtOH to diethyl ether (DEE) was observed. This behavior is expected given the large excess of EtOH present and the strong acidity of Amberlyst 70 resin. 3.1.3. Effect of EtOH Feed Location. Higher conversion of acids and high selectivity to DES were obtained by feeding EtOH below the catalytic zone of the column. This can be observed by comparing runs 12B and 12C, and also runs 13B and 13C. This behavior is further demonstrated with prereacted acid feeds in

runs 14A, 14B, and 20A. Ethanol enhances the entrainment of water into the distillate stream while promoting esterification of esters in the reactive zone. 3.1.4. Operation with Prereactor. In general, a higher ester yield was obtained when operating with prereacted acid feed. Concentrations obtained after processing the acid feed stream in the prereactor were close to those at equilibrium for both SA and AcAc esterification. Despite this high conversion of SA in prereactor, selectivity to DES was only around 5060%. In runs 14A, 14B, and 20A, the column operated with high concentrations of SA in the acid feed. In these experiments, high concentrations of succinate species were observed in the reboiler with a resulting high reboiler temperature. Run 14B shows that the column can operate at nearly complete acid conversion and produce a high-purity (>98%) DES bottoms stream, while simultaneously recovering EtAc in the distillate. 3.1.5. Effect of Water in the Feed. As observed in runs 21A, 22A, and 22B, operating with 190 proof EtOH reduces conversion of both SA and AcAc as well as the selectivity to DES. In these experiments, high concentrations of MES and H2O were present in the reboiler; this condition was exacerbated when the column was run with reflux (run 22B). In this case, even the separation efficiency between succinate and acetate species was affected, because unconverted AcAc was driven to the bottom of the column by the water refluxed back into the column. When the acid feed was an aqueous solution (runs 23A, 24A, and 26A), low conversions (∼70% SA and ∼60% AcAc) were obtained, and MES was the predominant succinate species in the bottoms product. In these experiments, high boilup rates were used to remove most of the water from the bottoms, but because of reboiler power limitations, it was not possible to drive all water to the distillate stream. This result indicates that direct esterification of aqueous solutions of organic acids via reactive distillation will require sufficient reboiler duty to transform aqueous succinic acid to its pure diethyl ester in a single processing unit. 9214

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

Table 4. Parameters for Activity-Based Kinetics Used in Simulation of RD Experiments reaction SA þ EtOH = MES þ H2O

parameter

Amberlyst 70

Amberlyst 15

self-catalyzed

k0 ((kmol/kgCAT)/s)

1.04  104

5.17  103

1.57  105 a

Ea (kJ/kmol)

46 200

46 900

57 300

KEQ MES þ EtOH = DES þ H2O

48.9

k0 ((kmol/kgCAT)/s)

2.11  103

7.0  104

17.8 a

Ea (kJ/kmol)

46 600

61 400

34 400

KEQ AcAc þ EtOH = EtAc þ H2O

2 EtOH f DEE þ H2O

10.14

k0 ((kmol/kgCAT)/s)

1.03  104

1.28  103

Ea (kJ/kmol) KEQ

47 600 12.11

44 600 2.12

k0 ((kmol/kgCAT)/s)

4.99  104

8.06  104 b

85 400

86 900 b

Ea (kJ/kmol) a

3

b

Preexponential factor for self-catalyzed reaction has units of (kmol/m )/s. Mole fraction model assuming activity coefficients of unity.

Table 5. Parameters Used for Simulation of the RD Pilot-Scale Column parameter total number of stages (N) rectifying stages

parameter 12 2

Murphree stage efficiency (stages 211) liquid holdup (cm3) (stages 2 to N  1) (10 vol %)

0.5 98.2

reboiler holdup (cm3)

1500

catalytic

39

product removal stage

noncatalytic

1012

reactive stages

feed stages

top

1

bottom

12

acid feed

above 3

catalyst loading (kgCAT/m3column)

78

EtOH feed

on 10

catalyst loading in prereactor (wt %)

2

column pressure drop per stage (kPa)

0.07

3.2. Simulation of RD Experiments. Simulations of succinate ester formation via reactive distillation has been carried out to expand the range of conditions considered and to provide a commercial-scale column design. Steady-state simulations were performed using the rigorous equilibrium stage model RadFrac in Aspen Plus (Version 7.1, Aspen Tech) with reaction kinetically controlled. To model reaction, each stage in the column is modeled as a stirred-tank reactor of liquid volume determined by column hold-up. Phase equilibria were modeled using the NRTL equation for liquid-phase activities22 and the HaydenO’Connell (HOC) equation for vapor phase fugacities.23 Binary parameters were obtained from experiments or literature and were reported in previous work from our group.2429 Reaction rates in the reactive stages of the column are described by a second-order, pseudohomogeneous activity-based kinetic model for Amberlyst 70 resin.29 The prereactor was simulated as an ideal stirred tank reactor with 24 h residence time using a similar activity-based kinetic model for Amberlyst 15 resin.30 A third activity-based kinetic model30 for self-catalyzed SA esterification was used to describe reaction in the noncatalytic stages of the stripping zone and in the reboiler. Kinetic parameters for the three models are summarized in Table 4, and parameters used for the AspenPlus simulation of pilot-scale RD experiments are listed in Table 5. The number of stages in the pilot-scale column simulation was determined from the height equivalent to a theoretical stage (HETP ∼ 0.5 m) reported for the structured packing.31,32 Column hydrodynamic parameters were estimated from correlations for structured packings in Aspen Plus. Because neither

Katapak-SP11 or BX packings are included in the Aspen Plus database, parameters for MELLAPAK-250Y (Sulzer) structured packing with similar liquid holdup, pressure drop, and flooding capacity33 were used in simulations. In each case, the fractional approach to maximum capacity (fMC) was varied to match the calculated column diameter with that of the pilotscale column. Because reboiler temperature is most strongly influenced by EtOH concentration, distillate flows were adjusted in each simulation to within the range of variation observed in experiment (∼5%) to match both EtOH concentration and temperature in the reboiler. Results obtained from simulations are summarized in Tables 6 and 7; good agreement with experimental results (Table 3) is observed for most runs. The average difference between predicted and experimental concentrations of DES in the bottoms stream and of EtOH in the distillate stream is less than 5% for the experiments simulated. Higher average deviations were observed for components present at low concentrations (0.5 mm) used in the RD structured packing is larger than that used (98% yield as a nearly pure bottoms stream. Ethyl acetate was removed in nearly quantitative yields from the distillate stream along with excess EtOH and product water. This proves that simultaneous esterification and separation of mixed acid streams can be accomplished in a single reactive distillation unit. Reactive distillation experiments were modeled with RadFrac in Aspen Plus, using a simple equilibrium stage model that includes homogeneous reaction kinetics and activity-based phase equilibria (NRTL-HOC). In general, the simulations reasonably reproduce the experimental observations and were thus used to evaluate the effect of different processing variables on the performance of the RD system. A preferable set of operating conditions to achieve high acids conversion with high selectivity to DES was identified. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Phone: (517) 353-3928. 9218

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

’ ACKNOWLEDGMENT This work was supported by the Michigan Economic Development Corp., 21st Century Jobs Fund Project No. 95096. ’ LIST OF SYMBOLS AV acid value (mg of KOH/g) AcAc acetic acid DEE diethyl ether DES diethyl succinate energy of activation (kJ/kmol) E0 EtAc ethyl acetate EtOH ethanol fractional approach to maximum capacity fMC water H2 O HETP height equivalent to a theoretical plate equilibrium constant KEQ preexponential factor ((kmol/kgCAT)/s) k0 MES monoethyl succinate N number of stages RD reactive distillation P gauge pressure (kPa) S selectivity SA succinic acid T temperature (K) mass fraction of component i wi X conversion ’ REFERENCES (1) Wolf, O.; Crank, M.; Patel, M.; Marscheider-Weidemann, F.; Schleich, J.; H€using, B.; Angerer, G. Techno-economic Feasibility of Large-Scale Production of Bio-based Polymers in Europe. European Science and Technology Observatory; EUR 22103 EN, 2005. (2) Rogers, P.; Chen, J.; Zidwick, M. Organic Acid and Solvent Production. In The Prokaryotes: Ecophysiology and Biochemistry,3rd ed., Vol. 2; Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K., Stackebrandt, E., Eds.; Springer Science: New York, 2006; Chapter 3.1, pp 511755. (3) Hermann, B.; Patel, M. Today’s and Tomorrow’s Bio-Based Bulk Chemicals from White Biotechnology. Appl. Biochem. Biotechnol. 2007, 136, 361. (4) Kamm, B.; Kamm, M. Advances in Biochemical Engineering and Biotechnology. In White Biotechnology, Vol. 105; Ulber, R., Sell, D., Eds.; Springer: Heidelberg, Germany, 2007; Chapter 4, pp 175203. (5) Sauer, M.; Porro, D.; Mattanovich, D.; Branduardi, P. Microbial Production of Organic Acids: Expanding the Markets. Trends Biotechnol. 2008, 26, 100. (6) McKinlay, J. B.; Vieille, C.; Zeikus, J. G. Prospects for a Bio-based Succinate Industry. Appl. Microbiol. Biotechnol. 2007, 76, 727. (7) Bechthold, I.; Bretz, K.; Kabasci, S.; Kopitzky, R.; Springer, A. Succinic Acid: A New Platform Chemical for Biobased Polymers from Renewable Resources. Chem. Eng. Technol. 2008, 31, 647. (8) Cukalovic, A.; Stevens, C. V. Biofuel Feasibility of Production Methods for Succinic Acid Derivatives: A Marriage of Renewable Resources and Chemical Technology. Biofuels, Bioprod. Bioref. 2008, 2, 505. (9) Delhomme, C.; Weuster-Botz, D.; K€uhn, F. E. Succinic Acid from Renewable Resources as a C4 Building-Block Chemical—A Review of the Catalytic Possibilities in Aqueous Media. Green Chem. 2009, 11, 13. (10) Zeikus, J. G.; Jain, M. K.; Elankovan, P. Biotechnology of Succinic Acid Production and Markets for Derived Industrial Products. Appl. Microbiol. Biotechnol. 1999, 51, 545.

ARTICLE

(11) Agarwal, L.; Isar, J.; Saxena, R. Rapid Screening Procedures for Identification of Succinic Acid Producers. J. Biochem. Biophys. Methods 2005, 63, 24. (12) Song, H.; Lee, S. Y. Production of Succinic Acid by Bacterial Fermentation. Enzyme Microb. Technol. 2006, 39, 352. (13) Orjuela, A.; Yanez, A.; Lira, C. T.; Miller, D. J. Carboxylic Acid Recovery from Fermentation Solutions. U.S. provisional patent application filed December 2009. (14) Tuchlenski, A.; Beckmann, A.; Reusch, D.; D€ussel, R.; Weidlich, U.; Janowsky, R. Reactive Distillation—Industrial Applications, Process Design & Scale-up. Chem. Eng. Sci. 2001, 56, 387. (15) Schoenmakers, S.; Bessling, B. Reactive and Catalytic Distillation from an Industrial Perspective. Chem. Eng. Proc. 2003, 42, 145. (16) Hiwale, R.; Bhate, N.; Mahajan, Y.; Mahajani, S. Industrial Applications of Reactive Distillation: Recent Trends. Int. J. Chem. React. Eng. 2004, 2 (Rev. R1), 1. (17) Jam Harmsen, G. Reactive Distillation: The Front-Runner of Industrial Process Intensification: A Full Review of Commercial Applications, Research, Scale-up, Design and Operation. Chem. Eng. Process. 2007, 46, 774. (18) Backhaus, A. A. Method for the Production of Esters. U.S. Patent 1,400,852, 1921. (19) Sutton, D.; Reed, G.; Hiles, G. Process for the Production of Esters of Mono-, Di-, or Polycarboxylic Acids. WO Patent 051885 A1, 2005 (U.S. Patent Appl. 0129565 A1, 2007). (20) Kolah, A.; Orjuela, A.; Hanna, N.; Lira, C. T.; Miller, D. J. Reactive Distillation for the Biorefinary: Pilot Plant Synthesis of Succinic Acid Esters. Presented at AIChE Spring Meeting and 6th Global Congress on Process Safety, 13p, San Antonio, TX, March 2125, 2010. (21) Orjuela, A.; Kolah, A.; Hong, Xi; Lira, C. T.; Miller, D. J. Diethyl Succinate Synthesis by Reactive Distillation. Submitted for publication in Sep. Pur. Technol.. (22) Renon, H.; Prausnitz, J. M. AIChE J. 1968, 14, 135. (23) Hayden, J. G.; O’Connell, J. P. A Generalized Method for Predicting Second Virial Coefficients. Ind. Eng. Chem. Process Des. Dev. 1975, 14, 209. (24) Tang, Y.; Huang, H.; Chien, I. J. Chem. Eng. Jpn. 2003, 36, 1352. (25) Orjuela, A.; Yanez, A.; Vu, D.; Bernard-Brunel, D.; Miller, D. J.; Lira, C. T. Phase Equilibria for Reactive Distillation of Diethyl Succinate: Part I. System Diethyl Duccinate þ Ethanol þ Water. Fluid Phase Equilib. 2010, 290, 63. (26) Orjuela, A.; Yanez, A.; Rossman, P.; Vu, D.; Bernard-Brunel, D.; Miller, D. J.; Lira, C. T. Phase Equilibria for Reactive Distillation of Diethyl Succinate. Part II: Systems Diethyl Succinate þ Ethyl Acetate þ Water and Diethyl Succinate þ Acetic Acid þ Water. Fluid Phase Equilib. 2010, 290, 68. (27) Orjuela, A.; Yanez, A.; Evans, J.; Miller, D. J.; Lira, C. T. Phase Equilibria in Binary Mixtures with Monoethyl Succinate. Fluid Phase Equilib. 2011manuscript in preparation. (28) Orjuela, A.; Yanez, A.; Lee, A.; Miller, D. J.; Lira, C. T. Solubility and Phase Equilibria for Mixtures Containing Succinic Acid. Manuscript in preparation, 2011. (29) Orjuela, A.; Yanez, A.; Lira, C. T.; Miller, D. J. Kinetics of Mixed Succinic Acid/Acetic Acid Esterification with Amberlyst 70 Ion Exchange Resin as Catalyst. Submitted for publication in Chem. Eng. J. (30) Kolah, A.; Asthana, N.; Vu, D.; Lira, C. T.; Miller, D. J. Reaction Kinetics for the Heterogeneously Catalyzed Esterification of Succinic Acid with Ethanol. Ind. Eng. Chem. Res. 2008, 47, 5313. (31) G€otze, L.; Bailer, O.; Moritz, P.; von Scala, C. Reactive Dstillation with KATAPAK. Catal. Today 2001, 69, 201. (32) Sieres, J.; Fernandez-Seara, J. Mass Transfer Characteristics of a Structured Packing for Ammonia Rectification in AmmoniaWater Absorption Refrigeration Systems. Int. J. Refrig. 2007, 30, 58. (33) Sulzer Chemtech. Structured Packings for Distillation, Adsoption, and Reactive Distillation (Online), http://www.sulzerchemtech.com/ portaldata/11/Resources//brochures/mtt/Structured_Packings_April_ 2010.pdf, Nov. 1, 2010. 9219

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220

Industrial & Engineering Chemistry Research

ARTICLE

(34) Behrens, M.; Olujic, Z.; Jansens, P. J. Liquid Holdup in Catalyst-Containing Pockets of a Modular Catalytic Structured Packing. Chem. Eng. Technol. 2008, 31, 1630. (35) Ratheesh, S.; Kannan, A. Holdup and Pressure Drop Studies in Structured Packings with Catalysts. Chem. Eng. J. 2004, 104, 45. (36) Hoffmann, A.; Noeres, C.; Gorak, A. Scale-up of Reactive Distillation Columns with Catalytic Packings. Chem. Eng. Process. 2004, 43, 383. (37) Kolodziej, A.; Jaroszynski, M.; Bylica, I. Mass Transfer and Hydraulics for KATAPAK-S. Chem. Eng. Process. 2004, 43, 457. (38) Olujic, Z.; Behrens, M. Holdup and Pressure Drop of Packed Beds Containing a Modular Catalytic Structured Packing. Chem. Eng. Technol. 2006, 29, 979.

9220

dx.doi.org/10.1021/ie200133w |Ind. Eng. Chem. Res. 2011, 50, 9209–9220