Evolutional Design of Ethyl Lactate Reactive Distillation Process with

Dec 18, 2018 - Watch How Flexible Electronic Skin Could Help Humans and Machines Interact. New research published in ACS Applied Materials & Interface...
3 downloads 0 Views 636KB Size
Subscriber access provided by University of South Dakota

Process Systems Engineering

Evolutional Design of Ethyl Lactate Reactive Distillation Process with various Separation Strategies Shi-Bao Dai, Hao-Yeh Lee, and Cheng-Liang Chen Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.8b03343 • Publication Date (Web): 18 Dec 2018 Downloaded from http://pubs.acs.org on December 19, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Submitted to Industrial and Engineering Chemistry Research (1st Revision)

Design and Economic Evaluation for Production of Ethyl Lactate via Reactive Distillation Combined with Various Separation Configurations

Shi-Bao Dai 1, Hao-Yeh Lee 2, and Cheng-Liang Chen 1*

1

Department of Chemical Engineering National Taiwan University Taipei 10617, Taiwan

2

Department of Chemical Engineering

National Taiwan University of Science and Technology Taipei 10607, Taiwan

24 June 2018 4 Dec 2018 (1st Revision)

*

Corresponding author: Cheng-Liang Chen, Tel: +886-2-33663039; Fax: +886-2-23623040; E-mail: [email protected]

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Two commercial scale ethyl lactate (L1E) production processes are studied in the work. The L1E processes can be divided into the reaction part and the separation section. For the reaction part, instead of the three-column design presented by Miller et al1, the proposed configuration only contains two reactive distillation (RD) columns, where the L1E product is taken from the first RD column as a sidedraw. This novel improvement can reduce 22.26% of energy consumption in the reaction part. Additionally, disparate separation approaches such as extractive distillation (ED) and the pervaporation (PV) are then implemented to deal with the ethanol/water azeotrope. Economics for alternative configurations are analyzed to find the most competitive and cost-effective process. As a result, the RD with PV design can save at least 31.46% of total annual cost compared to the RD with ED configuration.

Keywords: Ethyl Lactate; Process design; Reactive distillation; Extractive distillation; Pervaporation.

2

ACS Paragon Plus Environment

Page 2 of 39

Page 3 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

1. Introduction Ethyl lactate is a common solvent in chemical industry. The identity of this solvent is that it holds promise as a biodegradable and nontoxic replacement for petroleum-based solvents such as chlorofluorocarbons, methylene chloride, ethylene glycol ethers and chloroform that have long dominated world markets. Besides, the application of this green solvent ranges widely from coating, food, perfumery, polyurethane and pharmaceutical industry to some specific usages, for example, paint stripper and graffiti remover.2 Additionally, the potential global ethyl lactate market value will rise continually to one billion US dollars in 2019.3 However, the market price of ethyl lactate is almost two times higher than those traditional solvents.4 This can be categorized by two reasons. Firstly, no synthetic ethyl lactate is on the market currently, which means both reactants, ethanol (EtOH) and lactic acid (L1) are derived from an expensive natural-based feedstock. Secondly, the cost of separation and purification of the process has been estimated to account for half of the total cost. Therefore, improvements are needed in the process of producing ethyl lactate.2,

5

Ethyl lactate is usually produced through the main esterification reaction of ethanol and lactic acid. However, a key issue is that when the reactant concentration of lactic acid is higher than 20 wt%, several types of oligomer will appear, namely dilactic acid (L2), trilactic acid (L3), and their esters (L2E, L3E). All possible reaction routes are listed as follow.6 Equation (1) is the main reaction; Equations (2) and (3) are the side reactions to produce oligomers. Equations (4) to (8) are the esterification reactions and transesterification reactions of oligomers. Consequently, process design definitely becomes more challenging.



Main reaction: 3

ACS Paragon Plus Environment

1

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 39

Oligomeric side reaction (a):



2

Oligomeric side reaction (b):



3

Esterification of L2:



4

Esterification of L3:



5

⇌2

Transesterification of L2E: Transesterification of L3E (a):

2

⇌3 ⇌

Transesterification of L3E (b):

6 7 8

Reactive Distillation (RD) is a technique which combines reaction and separation units into a single one thus can highly reduce capital cost and reuse reaction heat directly. Meanwhile, high conversion and selectivity can be achieved by shifting the chemical equilibrium boundaries. Gao et al. reported a paper on the L1E process by using single reactive distillation column.7 Whereas the process is on an experimental scale and the reaction kinetics only considered the main esterification of L1 with EtOH. Daengpradab and Rattanaphanee provided a process regarding commercial scale production of L1E, which consists one RD column and three separation columns8. Nevertheless, the kinetics described in the process is also too simplified that neglected the oligomeric reactions. Asthana et al. proposed a process concept on producing ethyl lactate6. The process mainly includes two RD columns, one for undergoing esterification of L1E, another for transesterification reactions. However, transesterification is nearly impossible to achieve thus no further literature is unveiled regarding transesterification. Miller et al.1 disclosed a commercial scale process concept with completed reaction kinetics to produce L1E. However, it required two RD columns and one product separation column in the reaction part, which costs a lot. Furthermore, they did not specify which configuration would be implemented in the downstream separation part. 4

ACS Paragon Plus Environment

Page 5 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Extractive distillation (ED) is a common method for breaking azeotrope. To successfully extract the component from the azeotrope, adding an additional amount of non-volatile agent, which called entrainer to the system is necessary. By altering the relative volatility, one can separate the desired product from the azeotropic mixture easily.9 There are several advantages through ED. Neither heterogeneous liquid-liquid equilibrium nor distillation boundaries are formed while introducing the entrainer. Due to the complexity of different types of azeotrope, there are many kinds of entrainer being studied to tackle with corresponding process. Weiss and Herfurth reported a paper on using ethylene glycol as a solvent for EtOH/H2O system.10 Pinto et al.11 disclosed an idea about saline extractive distillation. In their study, compared to traditional process (extractive and azeotropic distillation), four saline agents performed well in the case of obtaining anhydrous EtOH from fermentation broth. Lynn and Hanson provided an uncommon process that combining extractive distillation and multi-effect evaporation.12 In their study, the steam consumption only took 0.94-1.47 kg per kg of EtOH product. Arifin and Chien proposed using dimethyl sulfoxide as an entrainer to isopropyl alcohol dehydration process. While comparing with the heterogeneous azeotropic distillation, ED saves about 32.7% of TAC and 30.3% of steam cost.13 In the current world, the idea of using energy wisely gains attention progressively. Meaning that further intensification on current processes is required. Hence, there are more and more researchers focusing on either the modification of existing processes or the development of new designs. Among them, membrane technology has been considered to be the most potential candidate. To illustrate, membrane holds various advantages. For example, it can be treated as an effective technique for separation because it is not limited by the volatility of components. Moreover, membrane unit normally requires lower energy consumption than conventional methods. Other 5

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

physical properties such as high selectivity, compact and modular design are also the merits that attracting more and more value in recent years. There are numerous types of membrane. Pervaporation (PV) which firstly named by Kober in 191714 is one of the most promising alternatives among them. It refers to a process that one or more components in fluid mixture permeate through a dense membrane selectively. Despite the fact that membrane holds plenty of advantages, it still has some limitations when it comes to practical application. This is because the high capital cost and low capacity usually hinder the industry to use single membrane module directly. As a result, the most general way is to combine PV with other conventional configurations, which called PV-based hybrid process. The hybrid process provides fascinating benefits over the conventional ones. One of the pioneers studying PV-based hybrid processes was Lipnizki et al.15 They give an overview of the applications, designs, and economics of the process. Since then, more and more literature that discuss the hybrid process were published. Furthermore, the integration of pervaporation with traditional esterification process is the most attractive issue among them. With constantly removing water from the system, the reaction is no longer limited by chemical equilibrium. Therefore, higher conversion can be accomplished without intensive energy consumption. Waldburger et al.16 reported a paper on discussing the continuous tube membrane reactor. Compared to the traditional distillation processes, the pervaporation-assisted process reduces the amount of energy requirement about 75% and 50% of total cost. Jyoti et al.17 disclosed a review paper on pervaporation. In the article, various factors such as temperature, catalyst concentration, etc. are being examined fully. The authors also suggest combining PV with other reaction units to achieve high product purity. Lee et al.18 proposed a process describing the esterification of ethyl acetate. This process is the first case by using sidedraw from the RD column and successively feeding to the pervaporation module. The optimal design 6

ACS Paragon Plus Environment

Page 6 of 39

Page 7 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

of the hybrid process saves 13% of energy compared to the two-column process. The objective of the study is to construct a commercial scale L1E production process, which taking oligomers into account. Furthermore, for downstream treatment, several separation strategies, namely, extractive distillation and pervaporation are being implemented to evaluate the most economical configuration.

2. Model Building 2.1. Thermodynamic Property Nonrandom two liquids (NRTL) activity coefficient model is used to account for the non-ideal vapor-liquid equilibrium (VLE) and possible vapor-liquid-liquid (VLLE) phase behaviors in this system. Also, to account for the dimerization and trimerization of lactic acid in the vapor phase, the second virial coefficients of Hayden -O’Connell19 is used. Table 1 gives the source and the binary interaction parameters in the ethyl lactate system. While the parameters could not found in the built-in Aspen database, the UNIFAC functional group contributions and the Dortmund method are used to estimate the parameters. Since the VLE experimental data for the system is limited, performing thermodynamic verification is confined to the following pairs. Figure 1 (A) and (B) shows the thermodynamic verification of L1E/EtOH20 and H2O/EtOH21. The above experimental pairs show great consistency by using the UNIFAC model. Therefore, we believe the binary interaction parameter prediction from UNIFAC model is good enough to represent the system. Another reason is that only UNIFAC model could provide all other needed parameters, especially bij and bji. In the supporting information, the binary interaction parameters regressed by the experimental data are also provided in Table S1. Results show that neither VLLE nor LLE are found in the system. Temperature and composition of azeotropes are given in 7

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 2.

2.2. Reaction Kinetics Although the reaction routes of the L1E system are complicated, one can simplify the system by applying some methodologies. According to Asthana et al.,22 only trace amount of L2E and L3E are found since the equilibrium constants of Equations (4) and (5) are relatively smaller than others. Besides, no literature is reported to support transesterification routes, Equations (6) to (8). Equations (1) to (3) are thus enough to represent the reaction system. Asthana et al.22 firstly reported the kinetic of the system. Su et al.23 then utilized the expertise of the parameters from them. In the study, Table 3 lists the parameters of the L1E system that we apply from Su et al.23 For this system, Amberlyst 15 is used as the catalyst. It should be noticed that this reaction kinetic model is catalyst-weight-based (mcat). Thus the conversion of tray volume between catalyst weight is necessary. One can solve by assuming that the solid catalyst occupies 50% of the liquid holdup in the RD column trays and the density of the catalyst is 770 kg/m3. Figure 2 shows the verified results of the kinetic model. Solid lines and different symbols represent the simulation results and experimental data, respectively.

2.3. Pervaporation Model In this work, a simple lumped system model (Luyben, 2009)24 is used. It would be done by dividing the membrane unit into several equal size units which called lumps. Moreover, material and energy balances, Equations (9) to (11), for each lump are applied. Total mass balance at the steady state is given in Equation (9): dM dt

0

F

,

F

,

F

,

9

where MR is the molar holdup in each cell, which is assumed to be constant25, Fn is 8

ACS Paragon Plus Environment

Page 8 of 39

Page 9 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

molar flow rate from cell n, and subscript R and P represent the retentate and the permeate sides, respectively. Component mass balance for species i is shown below: M

dZ , dt

,

F

,

Z

,

,

F

Z

,

F

, ,

,

Z

, ,

10

where zi is the molar fraction for species i. Energy difference of the inlet flow and outlet flow: dh , dt

M

F

,

h

,

F

,

h

,

F

,

H

,

11

where h is the molar enthalpy of liquid in retentate side, H is the molar enthalpy of vapor in permeate side. The membrane called PERVAP® is a commercial membrane provided by Sulzer ChemTech. Table 4 lists several operating conditions of the membrane. The permeation rate of water (Jw) and other species i (Ji) are depicted as follows: J

k exp J

k exp

E RT E RT

exp k w

,

exp k w

,

1 12 13

The pervaporation parameters shown in Table 5 are taken from Delgado et al.26 It is worth to notice that the membrane is really unfavorable for organic components. Also, the feed stream to the pervaporation module composes very small amount of L2 and L3 in the system. Therefore, the flux of L2 and L3 can be reasonably neglected.

3. Steady State Design The process designed here is to construct the commercial L1E production process. The commercial simulator, Aspen Plus v9.0 is used for simulation. The membrane module for PV is developed and implemented into Aspen Plus interface via Aspen Custom Modeler. The annual L1E production rate is set as 25 million pounds (roughly as 13.88 mole/min). The specification of L1E is 0.990 (molar basis) to meet the 9

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

product requirement. Besides, the purity of three feed streams, EtOH, L1, and H2O are 0.900, 0.152 and 0.995 (molar basis), respectively.

3.1. Reaction Section Figure 3 shows the base case simulation results of the reaction part from the process concept of Miller et al. (2010)1. It consists of two RD columns as denoted as RDC1 and RDC2 and one conventional distillation column (C1) for separating L1E from the top. By analyzing the composition profile in RDC1 (Figure 4), we found that the highest purity of the desired product, L1E does not appear in the bottom. Instead, it arises on the 57th stage with a purity of 0.964 that is higher than 0.920 in the bottom stream. Therefore, it can be fairly considered to draw a side stream from the middle of RDC1 to replace C1. Figure 5 shows the modified design of the reaction part. The product (L1E) is taken from the 39th stage of RDC1 as a sidedraw. Additionally, by changing the operating condition, such as the reboiler duty, L1E in the sidedraw can meet desired specification of 0.990 mole fraction. By comparing this novel configuration with the original design, the energy is reduced 22.3 %. For the further design of the process, the Mix stream from reaction section should be fairly treated. The Mix stream composes mainly of EtOH and H2O thus separating them by single distillation column is almost impossible because of typical EtOH/H2O azeotrope. In addition, in order to lower down the raw material cost of EtOH and H2O in the reaction section, effective separation technique for treating Mix stream is necessary. As a consequence, different separation configurations will be discussed in the following.

3.2. Separation Section 10

ACS Paragon Plus Environment

Page 10 of 39

Page 11 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

In this section, two approaches of separating EtOH/H2O azeotrope are investigated. Namely, the conventional technique called extractive distillation (ED) and the novel pervaporation (PV) method. Figures 6 (A) and (B) show the optimal results of RD + ED and RD + PV systems, respectively.

3.2.1. Extractive Distillation For the RD + ED system, the detailed optimizing procedure will be discussed in section 3.3. The entrainer used in the extractive section is glycerol. Figure 7 shows the residual curve map of adding glycerol into EtOH and H2O at 1 bar. The reason of choosing glycerol is that it is non-toxic and effective in extracting EtOH from H2O. According to Lee and Pahl27, the effect of introducing glycerol can be observed from Figure 8, which is the pseudo-binary vapor–liquid equilibrium diagram for ethanol– water–glycerol system. Glycerol obviously eliminates the EtOH/H2O azeotrope and changes the VLE curve. Besides, one should note that glycerol is relatively sensitive to temperature. The component will crack down at 293˚C. Hence, the pressure of the entrainer recovery column, C2 in Figure 6 (A), is set as 0.3 bar to ensure the design is practical. For the separation part, C1 represents the column to recycle the reactant, EtOH from the top (D3). The rest of water and entrainer will come out from B3 then fed to C2. The goal of C2 is purifying glycerol coming out at B4. Furthermore, since the boiling point of water is less than glycerol, water will come out from the top (D4). Last but not the least, because the amount of required H2O (1.200 kmol/hr) in RDC2 is less than the amount from D4 (5.905 kmol/hr), purging excess H2O is necessary.

3.2.2. Pervaporation The detailed optimizing procedure for RD + PV will be elucidated in section 3.3. Since the H2O content in Mix excesses the upper operating limitation of the 11

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

membrane, introducing a pre-concentrator, C1 in Figure 6 (B), before the pervaporation module is inevitable. The aim of the pre-concentrator is to enrich the content of EtOH in a stream. In other words, most of H2O in Mix will come out from the bottom of C1. Again, the required H2O (1.200 kmol/hr) in RDC2 is in relatively small amount, so splitting excess H2O from the bottom is essential. On the other hand, most of EtOH will come out as a distillate (D3) then sent to the membrane module. Four identical membranes are in a parallel arrangement. Each of them has the area of 21.65 m2. As reported by Lee et al.18, factor such as the composition in the feed will drastically influence the needed membrane area. Accordingly, arranging in a parallel way of this special type of membrane needs less area than series arrangement. Because the membrane is preferably water-permeable, the retentate side comprises mostly of EtOH, which can be recycled back to RDC1 as a reactant. Diversely, the rest of EtOH and a majority amount of H2O will constitute the permeate side, which will be sent back to C1 for further purification.

3.3. Process Optimization The objective of optimization for these two processes, RD + ED and RD + PV, is to identify which one is much more economical-competitive. The total annual cost (TAC) analysis is used in this study. The function of calculating TAC is shown in Equation (9), which is a combination of the annual operating cost (AOC) and the annualized total capital cost (TCC). TAC

AOC

TCC 14 payback period

The payback period is set as 8 years in this study. The reason for choosing 8 instead of 3 years is that in real-world industry, 8 years are more practical. The calculation for column and heat exchanger is based on Douglas28. The piping and pumping costs are 12

ACS Paragon Plus Environment

Page 12 of 39

Page 13 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

ignored for the analysis. The membrane price is taken from Van Hoof et al.29, and the lifetime of the membrane is assumed as 3 years. Furthermore, the cost of the membrane is composed of two parts: the membrane material with cost per area and the membrane modules with cost per unit.30 The capital and operating costs of the vacuum system are based on Woods31 and Oliveira et al.32, respectively. As for the vacuum cost of the ED system, the calculation is rooted on Seider et al.33 Eventually, all variables are divided into two parts for subsequent discussion.

3.4. Variables in Reaction Section In optimizing RDC1 in the reaction section, there are several variables need to be classified, including column pressure, sidedraw flow rate, sidedraw stage, reactants feed ratio, reactants feed stages, and the number of trays in rectifying, reactive, and stripping sections. As for RDC2, most of the variables are the same as RDC1 except for the sidedraw stage. Next, some assumptions are made to simplify our procedure: (1) The column is operated under normal pressure while cooling water can be used. (2) The flow rate of sidedraw is fixed as 0.833 kmol/h to meet the requirement of product volume and the product specification. Finally, the remaining variables in the reaction section are listed in the supporting information.

3.5. Variables in Separation Section To discuss variables in the separation section, we will discuss the ED system first, then investigation of the PV system follows. There are two columns in the ED system, namely C1 and C2. The feed ratio of GL/Mix will directly affect the cost of C1 because the more glycerol introduced to the column the more energy is needed for separation. Also, the amount of glycerol will influence the duty applied in C2. Hence, 13

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

for optimizing the ED system, one must lump C1 and C2 together for full consideration. As for the PV system, the EtOH purity at D3 will not only have the impact on the cost of C1, it likewise affects the membrane area. To illustrate, while choosing a higher purity of EtOH at distillate, smaller membrane area is required and vice versa. Besides, the inlet temperature to the membrane is fixed as 95˚C to prevent exceeding the upper operating limitation34. The permeate side pressure is kept at 0.16 bar.30 According to Tusel and Brüschke 35, the variation of inlet pressure does not have a significant effect. Therefore, the inlet pressure is set as 5 bar to avoid vaporization of D3 at 95 oC. Again, the overall variables needed to be optimized for both ED and PV system are provided in the supporting information.

3.6. Optimization Strategy The design procedure follows the direct search method proposed by Hooke and Jeeves.36 During the optimization, sensitivity test for the design and control variables can be recorded and make the researchers easily understand the effect of these variables on the process. The direct search method can be applied to most processes which are simulated by some commercial simulators such as Aspen Plus. The optimization algorithm for both RD + ED and RD + PV are provided in the supporting information.

4. Results and Discussion Before diving into the detailed optimization results between RD + ED and RD + PV systems. It is vital to clarify the optimal recycled EtOH purity from the separation part. This is because when the recycled EtOH purity decreases, the cost of separation would lower down. Conversely, the cost of the reaction part would increase. Moreover, 14

ACS Paragon Plus Environment

Page 14 of 39

Page 15 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

the purity of EtOH recycled back to RDC1 could not be too small since it should be assured enough to carry out the reaction in the reactive distillation column. Hence, one should scrutinize this important trade-off first. Three different recycled EtOH purities (0.91, 0.95, and 0.99) are studied in the following. Figure 9 shows the effect of recycled EtOH purity on TAC for both RD + ED and RD + PV systems. For the RD + ED system, the EtOH purity imposes slight influence on the TAC. It only declines from 500.94 to 494.36 (1000 USD) when the EtOH purity increases. The inverse correlation arises from the inherent characteristic of ED. To be more specific, EtOH is the lightest component in C1 thus it will come out as a distillate easily as glycerol is presented. Therefore, one can obtain a high purity of EtOH from D3. Nevertheless, when reducing the EtOH purity from D3, more energy needs to be applied to C1 to drives more H2O to the top. Last but not least, no matter which purity of EtOH is chosen, C2 must operate under atmospheric pressure to avoid cracking of glycerol. Therefore, the high cost of the vacuum system in C2 is irresistible as shown in Figure 10. On the other hand, for the RD + PV system, a positive correlation could be observed. Explicitly speaking, when the purity rises, one requires far more membrane area to attain high purity EtOH. Especially, the phenomenon shows an exceedingly jump on TAC (355.04 to 420.87) from 0.95 to 0.99. The reason can be explained from Figure 11, as the purity increases, the cost of the membrane increases accordingly. Based on the investigation above, 0.99 and 0.91 are selected as the recycled EtOH purity for RD + ED and RD + PV, respectively. Next we begin to study the variables in reactive part. For RD + ED and RD + PV configurations, both need two reactive distillation columns. Since the reactive part is almost the same in these two configurations, we would take the RE + PV as an example. At first glance, it seems that the variables are 15

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

quite much as listed in Table S2 in the supporting information. While conducting the optimization process, some variables could be clearly distinguished. Firstly, since both systems are highly complicated, some variables, such as the sidedraw stage, the feed ratios, and the number stages of the enriching section are sensitive to the constraint of the system. Secondly, some variables, such as the feed stages have a relatively slight effect on TAC (around 10-3 ~ 10-4 in difference). Therefore, the “really influential variables” for reactive distillation columns would be the number stages of reactive section and the number stages of stripping section. To evaluate the TAC, one could reasonably find that the TAC is mainly influenced by the total number of trays since they are positively correlated. Figures 12 (A) and 12 (B) are the results showing the influence of these two variables on TAC. When looking further into the system, some important points could be illustrated from Figure 13. Figure 13 shows the effect of the number stages of reactive section (NR1) on reboiler duty (Qr), conversion of L1, and the product (L1E) purity for RDC1. To begin with, increasing NR1 would increase the reboiler duty since there are much more L1E being produced thus uplifting the system loading. Besides, the positive correlation could also be observed between the conversion and the NR1. Moreover, when decreasing NR1 to be 13, the conversion drops suddenly to around 68.09%. Therefore, the product purity is only 0.71 instead of the specification of 0.99. This phenomenon shows that how the variable interacts with the constraint of the system. Following we will discuss the optimization process in the separation part. For variables in the RD + ED configuration, the total stages of C1 should be studied carefully. Figure 14 indicates the effect of total stages of C1. While increasing total stages of C1, the total implemental cost (TIC) of C1 increases in considerably larger scale than the total operating cost (TOC) of C2. Owing to the built-in identity of EtOH/H2O/Glycerol system, the entrainer performs well by separating EtOH from 16

ACS Paragon Plus Environment

Page 16 of 39

Page 17 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

H2O under any stage of C1. This means that nearly trace amount of EtOH in B3 will go to C2, thus the TOC of C2 remains almost the same. In this case, the total stages of C1 are selected as 27. As for the configuration of RD + PV, the EtOH purity at D3 is influential in designing the parameters of C1 and pervaporation unit. As increasing the EtOH purity at D3, a lot more energy (reboiler duty) is needed for obtaining high EtOH specification. On the contrary, less membrane area is required for further purification. Therefore, the inverse correlation between the required membrane area and the reboiler duty of C1 under different EtOH purity at D3 can be reasonable understood. Following the discussion above, Figure 15 demonstrates the membrane cost, TAC of C1, and TAC under various EtOH purity. The optimal EtOH purity that fed to the pervaporation unit is 0.83 since it achieves the lowest TAC. Table 6 lists the detailed cost distribution of the RD + ED and RD + PV configurations. Compared to the separation part, the column has more stages to carry out both esterification and transesterification in reaction part. Consequently, the cost for the reaction part is larger than the separation section in both configurations. While investing in the operating cost in RD + ED, it is interesting to find that the cost of vacuum system almost made up of 74% of the operating cost. In other words, in order to prevent glycerol from losing its function, C2 must be operated under atmospheric pressure, which pays a lot in using ED. As for the configuration of RD + PV, the operating cost is just 2.70 (1000 USD). The considerable reduction can be contributed by the following factors. Firstly, by implementing the pervaporation unit instead of ED, only one column presents in the separation part. Hence, the capital cost is saved. Secondly, considering that the amount of molar flow fed to PV is not that large, the required membrane area is relatively small even if the cost for the membrane per square meter is high. Additionally, the steam/cooling water cost in both configuration 17

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

includes the steam/cooling water utilized in distillation columns and heaters/coolers. To be more specific, the steam cost in the reaction section in RD + ED and RD + PV is quite high (18.34 and 24.98, separately.). This is because in order to lower down the loading of the reboiler in RDC1, it is wise to preheat the EtOH feed stream. Furthermore, the cooling water cost in the separation section in RD + ED is about 17.03. The cost is mainly generated from cooling down the entrainer stream to C1. In summary, by utilizing RD + PV configuration, all merits that possess will cause it to save effectively around 31.46% of TAC compared to RD + ED. Table S3 in the supporting information listed the result based on regressed experimental data. The final configurations of RD + ED and RD + PV are compared. The overall trend of results is similar with the case by applying the UNIFAC model. The reason might result from the parameters regarding L1. Firstly, from the study by Delgado et al.37, the VLE data is 4 components T-xy form. After regression, it is not easy to verify the binary interaction parameters. Secondly, they also mentioned that they did not consider higher concentration of L1 to avoid polymerization. Therefore, further thermodynamic studies should be completed, including higher order polymers of lactic acid, in order to achieve a global understanding of the real system. Owing to the reasons above, UNIFAC is chosen to predict most of the parameters for the conceptual design in the very beginning stage. Notice that, Su et al.23 also used UNIFAC to predict the binary interaction parameters in the existing literature regarding L1E process

18

ACS Paragon Plus Environment

Page 18 of 39

Page 19 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

5. Conclusion This paper proposes a novel L1E production process consisting of two different separation configurations, namely extractive distillation and pervaporation. The process is targeting to synthesize 99 mol% L1E with EtOH and L1 as reactants. For the reaction part, a stream has being side-drawing from RDC1 to acquire L1E product. Compared with the process concept by Miller et al.1, the advantage of adopting a sidedraw reduces the cost of another column for product purification. For the separation section, pervaporation intensifies the system appreciably more than extractive distillation since it decreases a tremendous amount of operating cost in the system. To illustrate, about 97.80% of operating cost in separation part and 31.46% of TAC are saved. In conclusion, the RD + PV hybrid configuration is endowed with inherent economical-effective potential in terms of process intensification.

Supporting Information A Supporting Information file will be provided which includes Source and binary interaction parameters of L1E system mainly by regression, main variables for optimization, and detailed cost comparison with different thermodynamic models. The optimization algorithm for reactive section and separation section are also included. This information is available free of charge via the Internet at http://pubs.acs.org/.

Acknowledgement The authors thank the Ministry of Science and Technology of ROC for supporting this research under Grant MOST 106-2221-E-002-177-.

19

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Reference 1.

Miller, D. J.; Asthana, N.; Kolah, A.; Lira, C. T. Process for Production of

2.

Organic Acid Esters. US 7,652,167, Jan. 26, 2010. Pereira, C. S. M.; Silva, V. M. T. M.; Rodrigues, A. E. Ethyl lactate as a solvent: Properties, applications and production processes - a review. Green Chem. 2011, 13, 2658-2671.

3. 4.

5.

Global Ethyl Lactate Market Research Report 2016; Gos International Inc.: 2016. de Jong, E.; Hisgson, A.; Walsh, P.; Wellisch, M. Bio-Based Chemicals: Value Added Products From Biorefineries; IEA Bioenergy, Task 42, Biorefineries: Wageningen, The Netherlands, 2013. (Available online at http://www.ieabioenergy.com/wp-content/uploads/2013/10/Task-42-Biobased-Ch emicals-value-added-products-from-biorefineries.pdf) Biddy, M. J.; Scarlata, C.; Kinchin, C. Chemicals from Biomass: A Market Assessment of Bioproducts with Near-Term Potential; National Renewable

6.

Energy Laboratory: Golden, CO, United States, 2016 . (Available online at http://www.nrel.gov/docs/fy16osti/65509.pdf). Asthana, N.; Kolah, A.; Vu, D. T.; Lira, C. T.; Miller, D. J. A Continuous Reactive

7.

Separation Process for Ethyl Lactate Formation. Org. Process Res. Dev. 2005, 9, 599-607. Gao, J.; Zhao, X. M.; Zhou, L. Y.; Huang, Z. H. Investigation of Ethyl Lactate

8.

Reactive Distillation Process. Chem. Eng. Res. Des. 2007, 85, 525-529. Daengpradab, B.; Rattanaphanee, P. Process Intensification for Production of Ethyl Lactate from Fermentation-Derived Magnesium Lactate: A Preliminary

Design. Int. J. Chem. React. Eng. 2015, 13, 407-412. 9. Perry, R., Chemical engineer’s handbook. McGraw Hill: New York, 1992. 10. Meirelles, A.; Weiss, S.; Herfurth, H. Ethanol dehydration by extractive distillation. J. Chem. Technol. Biotechnol. 1992, 53, 181-188. 11. Pinto, R. T. P.; Wolf-Maciel, M. R.; Lintomen, L. Saline extractive distillation process for ethanol purification. Comput. Chem. Eng. 2000, 24, 1689-1694. 12. Lynn, S.; Hanson, D. N. Multieffect extractive distillation for separating aqueous azeotropes. Ind. Eng. Chem. Proc. Des. Dev. 1986, 25, 936-941. 13. Arifin, S.; Chien, I. L. Design and Control of an Isopropyl Alcohol Dehydration Process via Extractive Distillation Using Dimethyl Sulfoxide as an Entrainer. Ind. Eng. Chem. Res. 2008, 47, 790-803. 14. Kober, P. A. Pervaporation, perstillation and percrystallization. J. Membr. Sci. 1995, 100, 61-64. 15. Lipnizki, F.; Field, R. W.; Ten, P.-K. Pervaporation-based hybrid process: a 20

ACS Paragon Plus Environment

Page 20 of 39

Page 21 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

review of process design, applications and economics. J. Membr. Sci. 1999, 153, 183-210. 16. Waldburger Raoul, M.; Widmer, F. Membrane reactors in chemical production processes and the application to the pervaporation assisted esterification. Chem. Eng. Technol. 1996, 19, 117-126. 17. Jyoti, G.; Keshav, A.; Anandkumar, J. Review on Pervaporation: Theory, Membrane Performance, and Application to Intensification of Esterification Reaction. J. Eng. 2015, 2015, 24. 18. Lee, H.-Y.; Li, S.-Y.; Chen, C.-L. Evolutional Design and Control of the Equilibrium-Limited Ethyl Acetate Process via Reactive Distillation−Pervaporation Hybrid Configuration. Ind. Eng. Chem. Res. 2016, 55, 8802-8817. 19. Hayden, J. G.; O'Connell, J. P. A Generalized Method for Predicting Second Virial Coefficients. Ind. Eng. Chem. Proc. Des. Dev. 1975, 14, 209-216. 20. Peña-Tejedor, S.; Murga, R.; Sanz, M. T.; Beltrán, S. Vapor–liquid equilibria and excess volumes of the binary systems ethanol+ethyl lactate, isopropanol+isopropyl lactate and n-butanol+n-butyl lactate at 101.325kPa. Fluid Phase Equilib. 2005, 230, 197-203. 21. Lai, H.-S.; Lin, Y.-F.; Tu, C.-H. Isobaric (vapor+liquid) equilibria for the ternary system of (ethanol+water+1,3-propanediol) and three constituent binary systems at P=101.3kPa. J. Chem. Thermodyn. 2014, 68, 13-19. 22. Asthana, N. S.; Kolah, A. K.; Vu, D. T.; Lira, C. T.; Miller, D. J. A Kinetic Model for the Esterification of Lactic Acid and Its Oligomers. Ind. Eng. Chem. Res. 2006, 45, 5251-5257. 23. Su, C.-Y.; Yu, C.-C.; Chien, I. L.; Ward, J. D. Plant-Wide Economic Comparison of Lactic Acid Recovery Processes by Reactive Distillation with Different Alcohols. Ind. Eng. Chem. Res. 2013, 52, 11070-11083. 24. Luyben, W. L. Control of a Column/Pervaporation Process for Separating the Ethanol/Water Azeotrope. Ind. Eng. Chem. Res. 2009, 48, 3484-3495. 25. Geankoplis, C., Transport processes and separation process principles. Prentice Hall Professional Technical Reference: New York, 1993. 26. Delgado, P.; Sanz, M. T.; Beltrán, S. Pervaporation of the quaternary mixture present during the esterification of lactic acid with ethanol. J. Membr. Sci. 2009, 332, 113-120. 27. Lee, F. M.; Pahl, R. H. Solvent screening study and conceptual extractive distillation process to produce anhydrous ethanol from fermentation broth. Ind. Eng. Chem. Proc. Des. Dev. 1985, 24, 168-172. 28. Douglas, J. M., Conceptual design of chemical processes. McGraw-Hill: New 21

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

York, 1988. 29. Van Hoof, V.; Van den Abeele, L.; Buekenhoudt, A.; Dotremont, C.; Leysen, R. Economic comparison between azeotropic distillation and different hybrid systems combining distillation with pervaporation for the dehydration of isopropanol. Sep. Purif. Technol. 2004, 37, 33-49. 30. Santoso, A. Design and Control of Hybrid Distillation-Membrane Systems for Separating Azeotropic Mixtures. Master thesis, National Taiwan University, Taiwan, 2010. 31. Woods, D. R., Cost Estimation for the Process Industries. McMaster University: Hamilton, Ontario, Canada, 1983. 32. Oliveira, T. A. C.; Cocchini, U.; Scarpello, J. T.; Livingston, A. G. Pervaporation mass transfer with liquid flow in the transition regime. J. Membr. Sci. 2001, 183, 119-133. 33. Seider, W. D.; Seader, J. D.; Lewin, D. R.; Widagdo, S., Product and process design principles: synthesis, analysis, and evaluation. WILEY: New York, 2009. 34. Sert, E.; Atalay, F. S. n-Butyl acrylate production by esterification of acrylic acid with n-butanol combined with pervaporation. Chem. Eng. Process. Process Intensif. 2014, 81, 41-47. 35. Tusel, G. F.; Brüschke, H. E. A. Use of pervaporation systems in the chemical industry. Desalination 1985, 53, 327-338. 36. Hooke, R.; Jeeves, T. A. `` Direct Search'' Solution of Numerical and Statistical Problems. J. ACM 1961, 8, 212-229. 37. Delgado, P.; Sanz, M. T.; Beltrán, S. Isobaric vapor–liquid equilibria for the quaternary reactive system: Ethanol+water+ethyl lactate+lactic acid at 101.33kPa. Fluid Phase Equilib. 2007, 255, 17-23.

22

ACS Paragon Plus Environment

Page 22 of 39

Page 23 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Industrial & Engineering Chemistry Research

Table 1. Source and binary interaction parameters of L1E system Component 1

Component 2

Source

aij

aji

bij (K)

bji (K)

cij

L1

EtOH

UNIFAC

0

0

13.3045

30.4187

0.3

L1E

UNIFAC

0

0

382.505

-287.146

0.3

H2O

UNIFAC

0

0

-363.348

823.798

0.3

L2

UNIFAC

0

0

199.205

-130.146

0.3

L3

UNIFAC

0

0

-433.467

618.048

0.3

L 1E

UNIFAC

0

0

343.39

-233.071

0.3

H2O

Aspen built-in

-0.8009

3.4578

246.18

-586.081

0.3

L2

UNIFAC

0

0

342.207

-248.824

0.3

L3

UNIFAC

0

0

753.472

-412.698

0.3

H2O

UNIFAC

0

0

-260.95

1179.05

0.3

L2

UNIFAC

0

0

-632.572

1091.56

0.3

L3

UNIFAC

0

0

-361.803

494.394

0.3

L2

UNIFAC

0

0

1326.35

-404.623

0.3

EtOH

L1E

H2O

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 39

Table 2. Composition and temperature of the azeotropes for L1E system (at 1 atm) Experimental

Computed

Composition

Temp. (oC)

Composition

Temp. (oC)

H2O/ EtOH

(0.106, 0.894)

78.12

(0.112, 0.888)

78.12

H2O/ L1E

-

-

(0.969, 0.031)

99.85

L1/ L2

-

-

(0.404, 0.596)

215.38

 Unit in mole fraction

Table 3. Kinetic model for L1E system r

m

k , r k , r k ,

k ,

6.52 m

k

x

x k

k

,

10

2.72

,

5.54

k, k

,

10 exp

x x

10 exp

k , x x

4.56

,

10 exp

k , x x

1.10 m

x x

10 exp

x x

exp

k

,

2.28

10 exp

ri (kmol/s), mcat (kgcat), ki (kmol/(kgcats)), R = 8.314 (kJ/(kmol/K)), T (K), xi (mole fraction)

Table 4. The operating limitations of PERVAP® 220134 The properties of PERVAP® 2201 (as reported by the manufacturer) Manufacturer: Sulzer ChemTech. Maximum temperature (oC)

105

Maximum water content in the feed

< 50

Organic acids

< 50

pH

2-7

ACS Paragon Plus Environment

Page 25 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Table 5. Pervaporation parameters for L1E system Component

ka (kgh-1 m-2)

ED (kJmol-1)

kb (kgh-1 m-2)

H 2O

1.19

10

49.96

2.17

EtOH

4.25

10

51.41

8.10

L1 E

1.93

10

40.93

9.58

L1

9.72

10

76.89

6.34

Table 6. Detailed cost distribution of RD + ED configuration,* RD + ED Separation section

Reaction section Capital

Operating

Capital

RD + PV Separation section

Reaction section

Operating

Capital

Operating

Capital

Operating

286.25

43.37

257.94

16.68

18.05

1.82

16.26

0.76

Reboiler

0.05

0.47

0.04

0.18

Condenser

0.06

0.15

0.05

0.16

Heater/ Cooler Feed Pump Vacuum System Catalyst

1.39

0.02

1.58

0.43

0.02

0.01

Column Trays

Steam

0.03 0.24

0.24

105.47

0.58

0.93 18.34

0.18

24.98

0.94

0.47

17.03

0.48

1.09

Cooling Water

15.71

Membrane Summary TAC

0.67

325.57

168.75

301.93

494.32

36.87 338.80

 All figures are in 1000 USD/yr as a unit. * Catalyst cost is 7.7162 $/kg

25

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Caption of Figures Figure 1. Thermodynamic verification for (A) L1E/EtOH; (B) H2O/EtOH Figure 2. Verification of kinetic model between experimental data Figure 3. The simulation results from the process concept of Miller et al. (2010) Figure 4. The composition profile in RDC1 Figure 5. The simulation results of modified process for reactive part Figure 6. The optimal results for (A) RD + ED configuration; (B) RD + PV configuration Figure 7. The residual curve map of EtOH/H2O/Glycerol system Figure 8. Pseudo-binary vapor–liquid equilibrium for EtOH/H2O/Glycerol system (Lee & Pahl, 1985) Figure 9. Effect of various recycled EtOH purity on TAC for RD + ED and RD + PV Figure 10. Effect of various recycled EtOH purity on vacuum system cost and TAC for RD + ED Figure 11. Effect of various recycled EtOH purity on membrane cost and TAC for RD + PV Figure 12. “Influential” variables for optimizing RD + PV (A) RDC1; (B) RDC2 Figure 13. Effect of the number of stages of reactive section on reboiler duty, conversion, and L1E purity Figure 14. Effect of total stages of C1 on the cost of ED system Figure 15. Effect of EtOH purity at D3 stream on the cost of TAC, TIC of C1, and membrane cost

26

ACS Paragon Plus Environment

Page 26 of 39

Page 27 of 39

Figure 1. (A)

TEMPERATURE ( )

155

135

115

95

75 0

0.2

0.4

0.6

0.8

1

MOLE FRACTION, EtOH

Figure 1. (B)

105 100 TEMPERATURE ( )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

95 90 85 80 75 0

0.2

0.4

0.6

MOLE FRACTION, EtOH

27

ACS Paragon Plus Environment

0.8

1

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 39

Figure 2.

28

ACS Paragon Plus Environment

Page 29 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Industrial & Engineering Chemistry Research

Figure 3.

D1: 12.445 kmol/hr L1 Feed: 5.921 kmol/hr Qc = ‐176.80 kW EtOH = 0.469 L1 = 0.152 L1E = 0.001 L2 = 0.005 H2O = 0.530 H2O = 0.843

Mix: 12.474 kmol/hr EtOH = 0.469 L1E = 0.001 H2O = 0.530

D2: 0.841 kmol/hr D3: 0.028 kmol/hr Qc = ‐12.33 kW RDC1 L1 = 0.005 Qc = ‐10.67 kW EtOH = 0.536 EtOH = 0.002 36 L1E = 0.002 39 L1E = 0.990 H2O = 0.462 NT = 30 EtOH Feed: 7.44 kmol/hr 2 L 2 = 0.003 ID = 0.110 m H2O Feed: 1.2 kmol/hr EtOH = 0.900 RDC2 15 EtOH = 0.002 H2O = 0.100 B1: 0.916 kmol/hr C1 B2: 0.075 kmol/hr 39 L1E = 0.003 Qr = 106.58 kW L1 = 0.557 NT = 60 L1 = 0.050 H2O = 0.995 L1E = 0.130 ID = 1.000 m EtOH = 0.002 B3: 1.246 kmol/hr L2 = 0.313 L1E = 0.920 L1 = 0.080 Qr = 12.41 kW L2 = 0.028 Qr = 9.69 kW H2O = 0.919 L2 = 0.001 3

NT = 58 ID = 1.000 m

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 39

Figure 4.

ACS Paragon Plus Environment

Page 31 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Industrial & Engineering Chemistry Research

Figure 5.

Qc = ‐159.44 kW

L1 Feed: 5.921 kmol/hr L1 = 0.152 L2 = 0.005 NT = 58 H2O = 0.843

ID = 1.000 m

Mix: 12.476 kmol/hr EtOH = 0.470 L1E = 0.001 H2O = 0.529

D1: 12.451 kmol/hr L1 = 0.002 EtOH = 0.469 H2O = 0.529

3 RDC1

EtOH Feed: 7.44 kmol/hr EtOH = 0.900 H2O = 0.100 Qr = 87.62 kW

36 39

Side draw: 0.833 kmol/h EtOH = 0.006 L1E = 0.990 L2 = 0.004 B1: 0.078 kmol/hr L1 = 0.518 L1E = 0.161 L2 = 0.321

Qc = ‐12.47 kW

2 RDC2 NT = 60 ID = 1.000 m 39

Qr = 12.42 kW

ACS Paragon Plus Environment

D2: 0.025 kmol/hr EtOH = 0.712 L1E = 0.001 H2O = 0.287 H2O Feed: 1.2 kmol/hr EtOH = 0.002 L1E = 0.003 H2O = 0.995 B2: 1.253 kmol/hr L1 = 0.083 H2O = 0.916 L2 = 0.001

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Page 32 of 39

Figure 6. (A) Fresh EtOH: 0.834 kmol/h Glycerol Make‐up: trace EtOH = 1.000 Glycerol: 1.000 Qc = ‐633.44 kW 82.64 °C P = 1.2 bar

L1 Feed: 6.004 kmol/h L1 = 0.150 Qc = ‐150.58 kW EtOH = 0.004 84.18 °C L1E = 0.010 P = 1.2 bar L2 = 0.005 H2O = 0.831 NT = 51 ID = 1.000 m RR = 0.240 HETP = 0.6 m

2 3 RDC1 33 37 49

D1: 10.788 kmol/h EtOH = 0.449 L1E = 0.002 H2O = 0.549

Mix: 10.819 kmol/h EtOH = 0.450 L1E = 0.002 H2O = 0.548 NT = 27 ID = 0.447m RR = 10.872

Side draw: 0.833 kmol/h L1 = 0.004 Qc = ‐36.27 kW D2: 0.031 kmol/h L1E = 0.990 82.51 °C P = 1.2 bar EtOH = 0.867 L2 = 0.005 L1E = 0.005 H 2O = 0.128 2

EtOH Feed: 5.749 kmol/h B1: 0.132 kmol/h EtOH = 0.991 H2O = 0.009 Qr = 105.57 kW L1 = 0.127 172.87 °C L1E = 0.697 L2 = 0.170 GL = 0.006

3 RDC2 35 36

Qr = 35.92 kW 106.34 °C

2

B3: 10.909 kmol/h L1 = 0.001 L1E = 0.002 H2O = 0.539 Glycerol = 0.458

C1

26

Qr = 675.95 kW 133.11 °C

H2O Feed: 1.200 kmol/h L1E = 0.005 H2O = 0.995

NT = 42 HETP = 0.6 m ID = 1.000 m RR = 44.104

D3: 4.915 kmol/h EtOH = 0.990 H2O = 0.010

Qc = ‐71.14 kW 68.84 °C

D4: 5.905 kmol/h L1E = 0.005 H2O = 0.995

2

P = 0.3 bar C2

9

NT = 10 ID = 0.425 m RR = 0.027

Qr = 95.86 kW 296.75 °C

B4: 5.005 kmol/h L1 = 0.001 Glycerol = 0.999

Fresh L1: 4.696 kmol/h B2: 1.257 kmol/h L1 = 0.173 H O Purged: 4.705 kmol/h L1 = 0.069 L2 = 0.016 2 H 2O = 0.825 L1E = 0.025 Glycerol = 0.001 L1E = 0.005 L2 = 0.002 H2O = 0.889 H2O = 0.995

32

ACS Paragon Plus Environment

Page 33 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Industrial & Engineering Chemistry Research

Figure 6. (B) Fresh EtOH: 0.846 kmol/h EtOH = 1.000 Retentate: 6.333 kmol/h EtOH = 0.910 H2O = 0.090 L1 Feed: 5.933 kmol/h L1 = 0.152 L2 = 0.005 H2O = 0.841 L1E = 0.002

Qc = ‐1001.24 kW 82.50 °C P = 1.2 bar Qc = ‐155.86 kW 84.31 °C P = 1.2 bar

NT = 53 ID = 1.000 m RR = 0.132 HETP = 0.6 m

2 3

D1: 12.193 kmol/h EtOH = 0.472 L1E = 0.002 H2O = 0.526

4 membranes in parallel Total area = 86.600 m2

Mix: 12.216 kmol/h EtOH = 0.472 2 L1E = 0.002 D3: 6.968 kmol/h C1 H2O = 0.526 EtOH = 0.830 8 H2O = 0.170 NT = 12

Side draw: 0.833 kmol/h 9 ID = 0.596 m L1 = 0.006 Qc = ‐10.44 kW Permeate: 0.635 kmol/h RR = 12.192 D2: 0.021 kmol/h 82.59 °C 16 E = 0.990 L 1 45 EtOH = 0.031 P = 1.2 bar EtOH = 0.806 Qr = 1004.39 kW L2 = 0.004 106.91 °C EtOH Feed: 7.178 kmol/h 49 H2O = 0.969 L1E = 0.002 H2O Feed: 1.200 kmol/h EtOH = 0.921 H2O = 0.192 2 EtOH = 0.001 H2O = 0.079 HETP = 0.6 m B1: 0.085 kmol/h 3 NT = 31 L1E = 0.003 ID = 1.000 m Qr = 89.17 kW RDC2 L1 = 0.424 RR = 44.098 H2O = 0.995 197.62 °C 29 L1E = 0.279 Fresh L1: 4.669 kmol/h H2O Purged: 4.682 kmol/h B2: 1.263 kmol/h L2 = 0.297 EtOH = 0.001 L1 = 0.177 L1 = 0.060 H2O = 0.919 L1E = 0.003 O = 0.820 H EtOH = 0.001 L2 = 0.011 2 Qr = 9.76 kW H  = 0.003 L 2O = 0.995 E = 0.008 L 2 106.44 °C 1 RDC1

33

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

Figure 7.

Figure 8. 1.0

Vapor mole fraction of EtOH (solvent‐ free)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 39

0.8

0.6

 Effect of adding glycerol ━ No Solvent

0.4 0.3

0.4

0.5

0.6

0.7

0.8

0.9

Mole frac. of EtOH in liquid (solvent‐free)

ACS Paragon Plus Environment

1.0

Page 35 of 39

Figure 9. 550

450

500

400 501

498

494

355 339

350

450

TAC of RD + ED

TAC of RD + PV

421

RD+PV RD+ED

300 0.90

400 1.00

0.93 0.95 0.98 EtOH Purity from Separation Part

Figure 10. 550

115 Vacuum Syst. Cost

500

110 501

498

494

105

106

450 105

400 0.90

0.93 0.95 0.98 EtOH Purity from Separation Part

35

ACS Paragon Plus Environment

105

100 1.00

Vacuum System Cost

RD+ED

TAC of RD + ED

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Industrial & Engineering Chemistry Research

Figure 11. 150

450 421

RD+PV

100

400 98 355 339

350

50

Memrane Cost

TAC of RD + PV

Membrane Cost

32 16 300 0.90

0.93 0.95 0.98 EtOH Purity from Separation Part

0 1.00

Figure 12. (A)

TAC (1000USD)

372.00 FR2 = 14.86 FB1 = 2 FH2O = 28 NEn2 = 1

360.00

365.61 362.27 358.92 362.26 358.92 355.58 345.55

348.00

NS1 = 36

342.19

NS1 = 37

338.83

NS1 = 38

336.00 13

14

15 NR1

16

17

Figure 12. (B) 354.00

TAC (1000USD)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 36 of 39

NL1E = 49 FR1 = 1.21 FL1 = 2 FEtOH = 45 NEn1 = 1

348.00

352.24 348.88

348.89

345.53

345.53

345.55

342.18

342.19

342.00

NS2 = 2 NS2 = 3 NS2 = 4

338.83

336.00 26

26

27

27 NR2 36

ACS Paragon Plus Environment

28

28

29

Page 37 of 39

Figure 13.

Figure 14. 115

260 TOC of C2

254

TAC of C1+C2

110

111

250

250

246 107

105

106

105

240

105

103

100

230

26

27

28 NT of C1

29

37

ACS Paragon Plus Environment

30

TAC of C1 + C2 (1000 USD)

TIC of C1

Cost (1000 USD)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Industrial & Engineering Chemistry Research

Figure 15. 100

400 Membrane Cost

TAC of C1

TAC

75

375 56

50

42

350

340

342

25 17

0 0.80

34

339

325 25

20

0.82 0.84 EtOH Purity @ D3

38

ACS Paragon Plus Environment

300 0.86

Cost (1000 USD)

Membrane Cost or TAC of C1  (1000 USD)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 38 of 39

Page 39 of 39 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Table of Contents (TOC) Graphic Fresh EtOH: 0.846 kmol/h EtOH = 1.000 Retentate: 6.333 kmol/h EtOH = 0.910 H2O = 0.090 L1 Feed: 5.933 kmol/h L1 = 0.152 L2 = 0.005 H2O = 0.841 L1E = 0.002

Qc = ‐1001.24 kW 82.50 °C P = 1.2 bar Qc = ‐155.86 kW 84.31 °C P = 1.2 bar

NT = 53 ID = 1.000 m RR = 0.132 HETP = 0.6 m

2 3

D1: 12.193 kmol/h EtOH = 0.472 L1E = 0.002 H2O = 0.526

4 membranes in parallel Total area = 86.600 m2

Mix: 12.216 kmol/h EtOH = 0.472 2 L1E = 0.002 D3: 6.968 kmol/h C1 H2O = 0.526 EtOH = 0.830 8 H2O = 0.170 NT = 12

Side draw: 0.833 kmol/h 9 ID = 0.596 m L1 = 0.006 Qc = ‐10.44 kW Permeate: 0.635 kmol/h RR = 12.192 D2: 0.021 kmol/h 82.59 °C L1E = 0.990 EtOH = 0.031 P = 1.2 bar EtOH = 0.806 Qr = 1004.39 kW L2 = 0.004 106.91 °C EtOH Feed: 7.178 kmol/h H2O = 0.969 L1E = 0.002 H O Feed: 1.200 kmol/h 2 EtOH = 0.921 H2O = 0.192 2 EtOH = 0.001 H2O = 0.079 HETP = 0.6 m B1: 0.085 kmol/h 3 NT = 31 L1E = 0.003 ID = 1.000 m Qr = 89.17 kW RDC2 L1 = 0.424 RR = 44.098 H2O = 0.995 197.62 °C 29 L1E = 0.279 Fresh L1: 4.669 kmol/h H2O Purged: 4.682 kmol/h B2: 1.263 kmol/h L2 = 0.297 EtOH = 0.001 L1 = 0.177 L1 = 0.060 H2O = 0.919 L1E = 0.003 H2O = 0.820 EtOH = 0.001 L2 = 0.011 Qr = 9.76 kW H2O = 0.995 L2 = 0.003 L1E = 0.008 106.44 °C RDC1 16 45 49

39

ACS Paragon Plus Environment