Methanol dehydration with pervaporation: Experiments and modelling

Accepted Manuscript Methanol dehydration with pervaporation: experiments and modelling Eniko Haaz, Andras Jozsef Toth PII: DOI: Reference:

S1383-5866(18)30484-2 https://doi.org/10.1016/j.seppur.2018.04.088 SEPPUR 14584

To appear in:

Separation and Purification Technology

Received Date: Revised Date: Accepted Date:

7 February 2018 11 April 2018 11 April 2018

Please cite this article as: E. Haaz, A. Jozsef Toth, Methanol dehydration with pervaporation: experiments and modelling, Separation and Purification Technology (2018), doi: https://doi.org/10.1016/j.seppur.2018.04.088

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Methanol dehydration with pervaporation: experiments and modelling Eniko Haaz, Andras Jozsef Toth* Environmental and Process Engineering Research Group, Department of Chemical and Environmental Process Engineering, Budapest University of Technology and Economics, H-1111, Hungary, Budapest, Műegyetem rkp. 3. * Corresponding author. E-mail address: [email protected] Tel: +36 1 463 1490; Fax: +36 1 463 3197

Abstract The research is motivated by a pharmaceutical industry problem where water should be removed from the aqueous mixture of methanol. To complete this target, pervaporation system is designed using hydrophilic Sulzer PERVAP™ 1510 membranes and examined to obtain information about the separation of methanol–water mixture. The aim of this work is to rigorously model and optimize the dewatering process. Separation factors, total and partial permeation fluxes, permeances and selectivities are experimentally determined. Pervaporation separation index (PSI) data are compared to those of other published membranes in the recent literature and it is found that PERVAP 1510 has the highest PSI value. The measured data are evaluated with improved pervaporation model by Valentinyi et al. [1] and it is found that the model can be applied also for this hydrophilic separation case. The separation system is rigorously modelled with ChemCAD and optimized with the dynamic programming optimization method. Such methanol–water separation has not been published in this professional flowsheet environment yet. The objective function of the process is the effective membrane area. The methanol dehydration is also investigated with distillation in flowsheet environment. It can be determined that pervaporation system is capable for the dehydration of methanol and it can become the alternative of distillation based separation, because it has lower heat duties in the case of same product compositions.

Keywords hydrophilic membrane; pervaporation, methanol dehydration; mathematical modelling; parameter estimation; professional flowsheeting environment

1. Introduction Pervaporation (PV) is relatively new membrane process with huge industrial potential for separating liquid mixtures forming azeotrope and close boiling mixtures [2]. The method is mainly applied for dehydration of organic mixtures, organic-organic separation and removal of organics from aqueous mixtures [3]. The first commercial application of the pervaporation process was the dehydration of ethanol [4]. Pervaporation can be used as hybrid system with a distillation column. Liquid mixtures can be separated on the diffusion and selective sorption of permeating component(s) through the dense membrane material. Pervaporation has the specialties such as no-pollution, simply actualization, energy-saving and high separation efficiency in the case of azeotropic mixtures, which are difficult to obtain by alternative conventional separation methods, e. g. distillation. The separated mixture is vaporized at low pressure on the downstream side of the PV membranes [5, 6]. Depending on the permeating component two main areas of pervaporation can be identified: PV dehydration and organophilic PV. For dehydration purposes, hydrophilic membranes are used. Pervaporation can be characterized by certain factors and quantities. The flux is determined using the following equation [3, 7]: (1) where is the amount of component in the permeate, is the time of duration of experiment and is the membrane area. Separation factor is calculated by Eq. (2) [3, 7]: (2) where is separation factor (dimensionless), is weight fraction of water in feed and fraction of water of permeate. The pervaporation separation index (PSI) is defined [7]:

is weight

(3) The performance of pervaporation membranes can be described by the permeance as component flux normalized for driving force the pressure difference-normalized flux [3, 8-10]: (4) The ideal membrane selectivity

is calculated as the ratio of permeances [8-10]: (5)

Despite of the fact that the most common application is the dewatering of organic solvents and more specifically of low molecular weight alcohols, is the most common application, there are many publications on the separation of ethanol/water mixtures [11-13]. Although some references reported that pervaporation technology which is used for separating ethanol/water mixture is on an industrial scale, there is nearly no study and application on separating methanol/water mixture. Compared with ethanol, methanol (MeOH) is more similar with water in polarity and molecular weight which makes methanol to compete with water on adsorbing in the membrane. So the

2

pervaporation membrane which is available for separating ethanol/water mixture is not ideal for separating methanol/water mixture [13]. The focus of this article is to introduce the emerging process of pervaporation dehydration and provide sufficient understanding of the process for successful interpretation of experimental data. The methanol dehydration is the actual task. As for the literature survey some papers were published in the separation of water from methanol by pervaporation. Until now mostly polymeric membranes have been used on an industrial scale. In recent years, research focuses on the development of ceramic membranes. Successful hydrophilic ceramic membranes have been made from silica [14, 15] from zeolites [16]. Full-scale plants using zeolite NaA membranes are already described [17]. Table 1 and Table 2 summarize a comparison of experimental data for the pervaporation dehydration of the methanol-water mixture.

3

Table 1 Comparison of experimental data with composite PVA membranes for pervaporation dehydration of methanol-water mixture PVA membranes PVA with 0.05% nano SiO2 PVA with 0.075% nano SiO2 PVA with 0.1% nano SiO2 PVA with 0.125% nano SiO2 PVA PVA with citric acid Sodium alginate/PVA Sulzer PERVAP-2201

T Fwater Jtotal [°C] [m/m%] [kg/m2h] 60 2 0.10 60 2 0.10 60 2 0.40 60 2 0.10 60 70 0.90 30 10 0.20 60 10 0.03 60 10 0.50

β PSI [-] [kg/m2h] 180 18 950 95 1458 583 980 98 2.5 1 8.6 2 135 3 3 1

Reference Liu et al. [6] Liu et al. [6] Liu et al. [6] Liu et al. [6] Shah et al. [16] Burshe et al. [18] Bano et al. [19] Van Baelen et al. [20]

Table 2 Comparison of experimental data with composite other membranes for pervaporation dehydration of methanol-water mixture Other membranes Polyamide-6 5% sPPSU PAI-PEI Hollow fiber Amorphous silica (ECN) Crosslinked chitosan Tubular membr. Pervatech+silica NaA-Modified zeolite

T Fwater Jtotal [°C] [m/m%] [kg/m2h] 30 10 0.02 60 14.74 0.03 60 15 1.03 90 10 2.20 30 16.1 0.49 50 15 0.70 60 70 1.80

β PSI Reference [-] [kg/m2h] 891 15 El-Gendia et al. [21] 11.1 0.33 Tang et al. [22] 4.71 4 Wang et al. [23] 55 119 Sommer and Melin [24] 5.3 2 Won et al. [25] 7 4 ten Elshof et al. [26] 140 250 Shah et al. [16]

Polyvinyl alcohol (PVA) is a benchmark hydrophilic PV membrane material. It can be determined, PVA membranes have the highest separation factor and PSI published in the literature. It can be concluded, using PVA membranes can be achieved dehydration with high separation factor and PSI. There is an actual problem in the pharmaceutical industry, which is the treatment of ethyl-acetate– methanol–water mixture. This complex, highly non-ideal mixture is generated in large quantities and it means serious environmental problem for industry sector. The first step, ethyl-acetate–water separation is already described in papers of [27, 28]. The binary heterogeneous azeotrope can be enriched in overhead product without methanol with extractive heterogeneous-azeotropic distillation technique [27] and the actual task of this paper is to separate the bottom fraction, which is methanol dehydration. Although methanol and water do not form an azeotrope and can be separated by conventional distillation, but the distillation separation is cost demanding because of the low relative volatilities of methanol and water. On the other hand, dehydration of alcohols, particularly ethanol and isopropanol, is one of the most developed applications of pervaporation, but there is little work on methanol dehydration. Compared with other alcohols, the difference in the molecular size of water and methanol is almost double as well as the difference in their solubility parameters (see in Table 3) [29]. Table 3 Solvent properties [29] Solvent

Calculated molecular Solubility diameter [nm] parameter [MPa]1/2 Water 0.26 47.9 Methanol 0.41 29.7 Ethanol 0.52 26.2 Isopropanol 0.58 23.4

The solubility parameters are cohesive properties characteristic of the liquids. Consequently, most pervaporation membranes are less selective to methanol/water separation than dehydration of ethanol and isopropanol. The more polar and smaller a molecule is, the easier it can be absorbed and passed through the membrane. As water is the smallest and most polar molecule, the separation of water from organic compounds functions the best. The driving force for the separation is the difference in the partial pressure of the water vapour between the feed and the rear side of the membrane. The separation principal of pervaporation is based on the difference in the polarity of the compounds which need separation, their molecular size, and the affinity of the most polar substances for the interface of the membrane. As far as the polarity is concerned, methanol is the water’s closet of neighbour. Based on the experience gained, Sulzer Chemtech has developed a new generation membranes which can dehydrate methanol from organic substances [10]. PV is considered as a competitive separation alternative of distillation [10, 29]. The aim of this research is to study the separation of methanol–water mixture with pervaporation dehydration with

rigorous modelling in professional flowsheeting environment and to compare the heat duty with distillation process.

2. Materials and methods Figure 1 depicts the flowsheet of the modelling and simulation of methanol dehydration with pervaporation. This algorithm illustrates also the background of model improvement as usual and it was described in detail in paper of Toth et al. [10]. First of all, the aims and the problems must be defined, which is 1000 kg/h methanol–water feed flow with 90 m/m% methanol–water mixture which, should be separated. 99.7 m/m% product purity should be achieved for water. The retentate, which is the methanol product is recommended for recycle at the beginning of the process. It can be seen the modelling of pervaporation part stands up 3 main steps, as follows: identification, parameter estimation and verification. The pervaporation model of Valentinyi et al. [1] is selected. There is a semi-empirical model, where parameter estimation from laboratory experiments are required to determine the parameters of the pervaporation model. Thereafter the modelled and measured data have to be compared, so the determined parameters are verified. If the model parameters are accurate and corresponding, they can be used for rigorous modelling in flowsheeting environment. ChemCAD program is applied for modelling of pervaporation. In the modelling part, at first model validation is necessary. The simulator has to run with measured data and if the results are appropriate, then the optimization can be carried out. In our case the membrane transfer area (A) is determined [10].

Figure 1 Algorithm of modelling and simulation of pervaporation dehydration in the case of methanol dehydration

2.1 Pervaporation experiments The composite PVA (Sulzer PERVAP™ 1510) flat sheet membrane is applied in dehydration experiments. The laboratory apparatus is P-28 membrane unit from CM-Celfa Membrantechnik AG

6

(see Figure 2) with 28 cm2 effective area (A). The size of the feed tank is 0.5 l. Cross-flow circulation is achieved at constant value of ∼182 l/h [10].

Figure 2 Schematic figure of CM-Celfa P-28 Membrantechnik AG in pervaporation mode [10, 29]

The vacuum on the permeate side is kept up with VACUUMBRAND PC2003 VARIO vacuum pump and maintained at 2 Torr (3 mbar). The isotherm conditions are settled with an ultrathermostat. The permeate is collected in two traps connected in series and cooled with liquid nitrogen to prevent loss of the permeate [10]. The methanol concentration of the feed (F), retentate (R) and permeate (P) are measured with Shimadzu GC2010Plus+AOC-20 autosampler gas chromatograph with a CP-SIL-5CB column connected to a flame ionization detector, EGB HS 600 headspace apparatus is used for sample preparation. The water content is measured with Hanna HI 904 coulometric Karl Fischer titrator [10, 30]. Swelling tests of membrane are carried out. The PERVAP™ 1510 membrane, which is completely dried at room temperature and weighed and then immersed in methanol solutions of 5, 10 and 15 m/m% water in a sealed vessel at 50°C. After 0.5, 1, 2, 3, 4, 5, 6 and 8 hours, the membranes are quickly taken out of the vessel, wiped rapidly to remove the solution residue, and weighed again [29]. The pervaporation measurements are carried out at six different feed concentrations and three temperatures, as follows: 1, 3, 5, 10, 15 and 20 m/m% water in feed and 50, 60 and 70°C.

2.2 Modelling of pervaporation The methodology of Valentinyi et al. [1] is selected for modelling of pervaporation. Eq. 6 shows the basic formula of this model: (6) This PV model is a development of Rautenbach model [31]. The improvements take into account the temperature dependencies of the PV and concentration dependencies of the transport coefficient

7

[10, 32]. The basic Rautenbach model (Model I) and the improved one (Model II) are used for modelling of our experiments. Partial pressures (

) are calculated according to the Antoine equation: (7)

where are material depending constants. Transport coefficient ( ) depends on the temperature in an Arrhenius type exponential way. (8) In Eq. (8) is the reference temperature, equal to 293 and is the activation energy for component and is associated with the transport coefficient. The liquid activity coefficients can be calculated with different vapor-liquid equilibrium models or with the Wilson equation. Detailed description of the semi-empirical PV model can be found in [1, 33]. Activation energies, transport coefficients and in the case of Model II for both compounds the B parameters show the concentration dependencies of the transport coefficients, which are estimated based on our measured data [10]. Nonlinear estimation process is used by defining a user specified regression custom loss function (Eq. (7)) in STATISTICA® program environment. The model verification can be obtained with objective function (OF), which is minimized the deviation of the modelled and the measured values. (9) The improved model was tested by Toth et al. [29] for methanol removal from aqueous mixtures with organophilic membranes (Sulzer PERVAP™ 2211 and 4060 from PDMS) in the range of 0.05–20 m/m% feed methanol concentration. Furthermore, it was investigated in the case of isobutanolwater mixture separation, organophilic PV (Sulzer PERVAP™ 4060 from PDMS) between 0.5 and 7.0 m/m% isobutanol feed concentrations and PV dehydration (Sulzer PERVAP™ 1510 from PVA) in the range of 85–99 m/m% feed isobutanol concentrations by Toth et al. [10]. Ashraf et al. [34] applied a PV model, which is based on this improved model with two dehydration systems: 83–98 m/m% 1butanol and 85–97 m/m% isobutanol over Sulzer™ PERVAP 2510 (PVA). 2.3 Simulation of pervaporation and distillation User added PV subroutine is written and applied in ChemCAD professional flowsheeting software package both for the application of Model I and Model II [2]. It can be seen in Figure 1, after the successful model validation, pervaporation dehydration system can be rigorously modelled and optimized with dynamic programming optimization method [10, 35]. Figure 3 shows the flowsheet of the pervaporation dehydration system.

8

Figure 3 Simulated pervaporation system for methanol dehydration

Further devices are also necessary for PV dehydration system. The preheating of feed flow and increasing of its pressure are needed, because it has atmospheric conditions (1 bar, 20°C). Adiabatic pervaporation mode is applied, therefore retentate reheating must be designed after each PV unit, except for the last [10, 36, 37]. Heat exchangers regulate the temperature and pumps increases the pressure. It can be detected in Figure 3, the retentate flow is recycled into the beginning of the process and it is mixed with feed. Permeate flows are collected, mixed, condensed with cooler and their pressure is increased again from vacuum with pump [10]. Post coolers and valves decreases the temperature and pressure of the permeate and retentate products [10]. Distillation computer simulations are carried out with ChemCAD for methanol dehydration. The aim of the simulations is to find the configuration which can satisfy the purity requirement with minimal heat duty. The reflux ratio, number of theoretical plates (N), feed plate and heat duty are optimized. SCDS column is applied with NRTL model. The flowsheet can be seen in Figure 4.

9

Figure 4 Simulated distillation system for the methanol dehydration

10

4. Results and discussion The results of the swelling experiments of Sulzer PERVAP™ 1510 PVA membrane in methanol/water binary mixtures are shown in Figure 5. These results are the average of three parallel measurements. As it can be seen, with increasing water concentration, the degree of swelling increases. After 0.5-1 hours the degree of swelling is practically constant.

Figure 5 Degree of swelling of Sulzer PERVAP™ 1510 Figure 6 shows the effect of feed concentration on the pervaporation performance of the PERVAP™ 1510 hydrophilic membranes at different operating temperatures. It can be seen that increasing feed water concentration increases the total, water and methanol permeate fluxes. A possible reason is that increasing water content in feed results major swelling of flat sheet membrane (see Figure 5), thus allowing that higher flow can permeate through the membrane. The tendencies of partial fluxes are in accordance with each other. As it can be seen, increasing the temperature increases the fluxes of water and methanol too. However, increasing the water concentration decreases the separation factor values (Figure 6/A). A maximum separation factor of 1320 can be observed at 70°C and at the feed water concentration of 1.0 m/m%. It can be seen PERVAP™ 1510 has the second highest separation factor and the highest PSI in the group of PVA membranes, compared to other literature results (see Table 1). It can be determined at higher water concentration the separation efficiency of the PV membranes is increasingly decadent. A possible reason is that increasing the water content increases water sorption through the flat sheet PV membrane [38] and as a reaction, the membrane becomes more swollen owing to its hydrophilic character (see Figure 5). The swollen sheet because 11

of its increased free volume redound the diffusion of methanol through the membrane eventuating in decreasing separation factor [29, 39]. The selectivity and PSI follow the tendency of the separation factor. Similar trend has been already published by Won et al. [25], Van Baelen et al. [20], Tang et al. [22], Pang et al. [40] and Liu et al. [6].

Figure 6 Pervaporation performance as a function of feed water concentration at different operating temperatures for PERVAP™ 1510 membrane (50°C: ; 60°C: ; 70°C: )

The methanol and water permeate weight fractions for the pervaporation dehydration are plotted against feed methanol concentrations in Figure 7. The equilibrium vapour-liquid curve is also shown at atmospheric pressure (1 bar, full line) so that pervaporation and flash distillation could be compared [10, 29]. It can be seen that in the cases of PERVAP™ 1510 PVA-type membranes, there is no significant difference between permeate concentrations in the case of temperatures. It can be 12

determined that PV can be competitive alternative for flash distillation, based on the analysis of available compositions.

Figure 7 Compositions of our measured permeates of pervaporation dehydration

Table 4 summarizes the estimated values of transport coefficients, activation energies and exponential parameters of the two models. Table 4 Estimated parameters for methanol–water mixture PERVAP™ 1510 [kmol/m2h] [kJ/kmol] B [-]

Model I Model II Water MeOH Water MeOH -2 -5 -1 4.82 x 10 1.84 x 10 1.67 x 10 1.80 x 10-4 20336 33833 23498 30795 -6.51 -2.40 13

Comparison of the measured and calculated partial fluxes are presented in Figure 8.

Figure 8 Measured partial fluxes ( ) of water and methanol compared to fluxes calculated with Model I ( ) and Model II ( ) in a function of feed water content in weight percent with PERVAP™ 1510 hydrophilic membrane The minimized objective functions are shown in Table 5. Table 5 Objective functions resulted by the two models PERVAP™ 1510 OF-Water OF-MeOH Model I 5.291 0.968 Model II 1.385 0.074

Figure 8 and Table 5 show that Model II is much more appropriate for description of pervaporation than Model I. The Rautenbach model assumes constant transport coefficient. It can be determined this statement is not suitable, so any concentration dependencies of have to be considered. Model II tales into account the suggestion of many authors, there is an exponential relationship between diffusion coefficient and feed concentration [1, 10, 41, 42].

14

It can be determined, only Model II is capable for accurate rigorous modelling of pervaporation. Table 6 shows validation results, the comparison of permeate fluxes in measurements and model obtained for our pervaporator set up by flowsheet simulator program. The feed temperature is 70°C and pressure is 3 bar and the permeate pressure is 3 mbar. It can be seen that the values have good accuracy. Table 6 Model validation with total fluxes Fwater [m/m%] 1 3 5 10 15 20

Jtotal - Experiment [kg/m2h] 0.58 0.59 0.61 0.62 0.68 0.74

Jtotal - Model [kg/m2h] 0.59 0.60 0.61 0.62 0.68 0.75

Deviation [%] 1.1 0.9 0.5 -0.4 -0.7 0.8

The results of the rigorously modelled and optimized pervaporation dehydration is presented in Table 7. It can be seen beside the purity of permeate water content, the water concentration of retentate can be reduced with increasing the membrane surface. It can be seen 15 m 2 effective membrane area is fulfil the permeate requirement, which is 99.7 m/m%. It can be also determined, there is not worth increasing further the membrane area, because significant improvement can not be reached in water clarity of permeate side. Table 7 Calculated flows and concentrations in the case of membrane areas

Feed

Permeate

Retentate

Flow [kg/h] Methanol [m/m%] Water [m/m%] Flow [kg/h] Methanol [m/m%] Water [m/m%] Flow [kg/h] Methanol [m/m%] Water [m/m%]

6 1000 90 10 5 0.8 99.2 995 90.4 9.6

Effective membrane surface [m2] 9 12 15 30 1000 1000 1000 1000 90 90 90 90 10 10 10 10 8 10 12 23 0.7 0.6 0.3 0.3 99.3 99.4 99.7 99.7 992 990 988 977 90.7 90.9 91.1 92.1 9.3 9.1 8.9 7.9

60 1000 90 10 42 0.3 99.7 958 93.9 6.1

However, if sharper separation has to be reached more membrane surface is required. In many cases, the investment cost of membrane material is significant, therefore the recycling of retentate flow can be worth solution. Table 8 shows a possible purification of methanol-water mixture, which can be reached 99.7 m/m% water concentration in cumulated permeate product and 99.97 m/m% methanol in the last retentate product (see Figure 3).

15

Table 8 Sharp separation of methanol-water mixture with pervaporation

Step I, A=30 m

2

Step II, A=30 m2

Step III, A=15 m2

Flow [kg/h] Methanol [m/m%] Water [m/m%] Flow [kg/h] Methanol [m/m%] Water [m/m%] Flow [kg/h] Methanol [m/m%] Water [m/m%]

Feed 1000 90 10 977.0 92.1 7.9 929.2 96.8 3.2

Permeate Retentate 23.0 977.0 0.3 92.1 99.7 7.9 47.8 929.2 0.2 96.8 99.8 3.2 29.1 900.1 0.1 99.97 99.9 0.03

Table 9 shows the optimized parameters of distillation, where two configurations have to be emphasized. The 16 theoretical plates column can satisfy the water emission limit, which is 99.7 m/m%. To compare separations with each other, same composition is examined in methanol products in the case of PV and distillation (see Table 8). Table 9 Optimized parameters of distillation separation N - Sum [-] N - Feed [-] Reflux ratio [-] Flow [kg/h] Feed Methanol [m/m%] Water [m/m%] Flow [kg/h] Distillate Methanol [m/m%] product Water [m/m%] Flow [kg/h] Bottom Methanol [m/m%] product Water [m/m%] Reboiler duty [MJ/h]

16 8 1 1000 90 10 904.6 99.5 0.5 95.4 0.3 99.7 2144

16 8 0 1000 90 10 936.8 96.1 3.9 63.2 0.3 99.7 1212

20 10 0 1000 90 10 936.2 96.1 3.9 63.8 0.3 99.7 1213

16 4 0 1000 90 10 936.1 96.1 3.9 63.9 0.3 99.7 1214

16 12 0 1000 90 10 937.1 96.1 3.9 62.9 0.3 99.7 1213

16 8 1 1000 90 10 904.6 99.5 0.5 95.4 0.3 99.7 2144

It can be determined, reflux is required to achieve corresponding composition, therefore the heat duty of distillation increases considerably. Table 10 shows the comparison of heat duties of PV and distillation. It can be seen in our case the pervaporation is more favourable than distillation.

16

16 8 2 1000 90 10 900.0 99.97 0.03 100.0 0.3 99.7 3121

Table 10 Comparison of heat duties of two separation methods

Hydrophilic PV (A=75 m2) Distillation (N=8)

Heat duty [MJ/h] 2664 3121

17

Final Methanol [m/m%] 99.97 99.97

Final Water [m/m%] 99.7 99.7

5. Conclusions It can be determined that pervaporation is now regarded as the latest state of the art on the chemical industry, where separation is concerned. The most frequent application is the removal of water from aqueous organic mixtures with hydrophilic membranes. Typical applications are the splitting of azeotropes and the final dehydration of the product. The total flux of the investigated Sulzer PERVAP 1510 membrane is found to vary from 0.27 to 0.74 kg/m2h over the feed water concentration range of 1.0–20.0 m/m% at 50–70 . The highest PSI of 765 kg/m2h and second highest separation factor (1320) are measured with flat sheet PVA membrane. The figures depict that selectivity and flux are in inverse relationship, which is expected by the theories. The results of parameter estimation and modelling of the pervaporation dehydration show that the model of Valentinyi et al. [1] (Model II) is also capable for the modelling of PV and results in a better fit to the experimental data. The rigorous flowsheet modelling suggests that the pervaporation is able to dehydrate the methanol. It can be also determined that our verified, adequate and optimized model can be a competitive alternative for the mature modelling of distillation. It can be found pervaporation has 85% heat duty of distillation separation in the case of same product compositions.

Acknowledgements The authors would like to acknowledge the financial support of János Bolyai Research Scholarship of the Hungarian Academy of Sciences and OTKA 112699 project.

18

Nomenclature Membrane transfer area Constant in Model II Transport coefficient of component Relative transport coefficient of component Activation energy of component transport coefficient

in Eq. (8) for temperature dependence of the

Feed Component number Component number Total flux Partial flux N

Number of theoretical plates

[-]

Permeate Pure component vapour pressure Partial pressure of component on the liquid phase membrane side Partial pressure of component on the vapour phase membrane side Pressure on the permeate side Permeance of component Retentate Ʀ

Gas constant Time Temperature Reference temperature: 293 Feed methanol weight fraction in vapour-liquid equilibrium (VLE) diagram (Figure 3) 19

Concentration of component in the feed Permeate methanol and water weight fraction in vapour-liquid equilibrium (VLE) diagram (Figure 3)

Abbreviations OF

Objective function

PDMS

Polydimethylsiloxane

PSI

Pervaporation Separation Index

PVA

Polyvinyl alcohol

PV

Pervaporation

VLE

Vapour-Liquid Equilibrium

Greek letters Selectivity Separation factor Average activity coefficient of component Activity coefficient of component in the feed Membrane thickness

20

References

[1] N. Valentínyi, E. Cséfalvay, P. Mizsey, Modelling of pervaporation: Parameter estimation and model development, Chemical Engineering Research and Design, 91 (2013) 174-183. [2] N. Valentínyi, P. Mizsey, Comparison of pervaporation models with simulation of hybrid separation processes, Periodica Polytechnica: Chemical Engineering, 58 (2014) 7-14. [3] R.W. Baker, Membrane Technology and Applications, 3rd edition ed., Wiley, 2012. [4] R.W. Baker, A.L. Athayde, R. Daniels, M. Le, I. Pinnau, J.H. Ly, J.G. Wijmans, J.H. Kaschemekat, V.D. Helm, Development of Pervaporation to Recover and Reuse Volatile Organic Compounds from Industrial Waste Streams, Report to US Department of Energy, 1997. [5] J. Neel, Introduction to pervaporation, in: R.Y.M. Huang (Ed.) Pervaporation Membrane Separation Processes, Amsterdam, 1991, pp. 1-109. [6] X. Liu, Y. Sun, X. Deng, Studies on the pervaporation membrane of permeation water from methanol/water mixture, Journal of Membrane Science, 325 (2008) 192-198. [7] J.G. Crespo, C. Brazinha, 1 - Fundamentals of pervaporation, in: Pervaporation, Vapour Permeation and Membrane Distillation, Woodhead Publishing, Oxford, 2015, pp. 3-17. [8] R.W. Baker, J.G. Wijmans, Y. Huang, Permeability, permeance and selectivity: A preferred way of reporting pervaporation performance data, Journal of Membrane Science, 348 (2010) 346-352. [9] W.F. Guo, T.S. Chung, T. Matsuura, Pervaporation study on the dehydration of aqueous butanol solutions: A comparison of flux vs. permeance, separation factor vs. selectivity, Journal of Membrane Science, 245 (2004) 199-210. [10] A.J. Toth, A. Andre, E. Haaz, P. Mizsey, New horizon for the membrane separation: Combination of organophilic and hydrophilic pervaporations, Separation and Purification Technology, 156 (2015) 432–443. [11] J.H. Kim, K.H. Lee, S.Y. Kim, Pervaporation separation of water from ethanol through polyimide composite membranes, Journal of Membrane Science, 169 (2000) 81-93. [12] M. Nomura, T. Yamaguchi, S.I. Nakao, Ethanol/water transport through silicalite membranes, Journal of Membrane Science, 144 (1998) 161-171. [13] M. Wesslein, A. Heintz, R.N. Lichtenthaler, Pervaporation of liqued mixtures through poly (vinyl alcohol) (PVA) membranes. II. The binary systems methanol/1- propanol and methanol/dioxane and the ternary system water/ methanol/1-propanol, Journal of Membrane Science, 51 (1990) 181-188. [14] R.W. van Gemert, F. Petrus Cuperus, Newly developed ceramic membranes for dehydration and separation of organic mixtures by pervaporation, Journal of Membrane Science, 105 (1995) 287-291. [15] H.M. van Veen, Y.C. van Delft, C.W.R. Engelen, P.P.A.C. Pex, Dewatering of organics by pervaporation with silica membranes, Separation and Purification Technology, 22-23 (2001) 361-366. [16] D. Shah, K. Kissick, A. Ghorpade, R. Hannah, D. Bhattacharyya, Pervaporation of alcohol–water and dimethylformamide–water mixtures using hydrophilic zeolite NaA membranes: mechanisms and experimental results, Journal of Membrane Science, 179 (2000) 185-205. [17] Y. Morigami, M. Kondo, J. Abe, H. Kita, K. Okamoto, The first large-scale pervaporation plant using tubular-type module with zeolite NaA membrane, Separation and Purification Technology, 25 (2001) 251-260.

21

[18] M.C. Burshe, S.B. Sawant, J.B. Joshi, V.G. Pangarkar, Sorption and permeation of binary water-alcohol systems through PVA membranes crosslinked with multifunctional crosslinking agents, Separation and Purification Technology, 12 (1997) 145-156. [19] S. Bano, A. Mahmood, K.-H. Lee, Vapor Permeation Separation of Methanol–Water Mixtures: Effect of Experimental Conditions, Industrial and Engineering Chemistry Research, 52 (2013) 10450–10459. [20] D. Van Baelen, B. Van der Bruggen, K. Van den Dungen, J. Degreve, C. Vandecasteele, Pervaporation of water–alcohol mixtures and acetic acid–water mixtures, Chemical Engineering Science, 60 (2005) 1583-1590. [21] A. El-Gendia, H. Abdallaha, Selectivity performance for polyamide-6 membranes using pervaporation of water/methanol mixtures, Desalination and Water Treatment, 51 (2013) 3263-3272. [22] Y. Tang, N. Widjojo, G.M. Shi, T.S. Chung, M. Weber, C. Maletzko, Development o f flat-sheet membranes for C1-C4 alcohols dehydration via pervaporation from sulfonated polyphenylsulfone (sPPSU), Journal of Membrane Science, (2012) 415-416. [23] Y. Wang, S.H. Goh, T.S. Chung, P. Na, Polyamide-imide/polyetherimide dual-layer hollow fiber membranes for pervaporation dehydration of C1–C4 alcohols, Journal of Membrane Science, 326 (2009) 222-233. [24] S. Sommer, T. Melin, Influence of operation parameters on the separation of mixtures by pervaporation and vapor permeation with inorganic membranes. Part 1: Dehydration of solvents, Chemical Engineering Science, 60 (2005) 4509-4523. [25] W. Won, X. Feng, D. Lawless, Separation of dimethyl carbonate/methanol/water mixtures by pervaporation using crosslinked chitosan membranes, Separation and Purification Technology, 31 (2003) 129-140. [26] J.E. ten Elshof, C.R. Abadal, J. Sekulić, S.R. Chowdhury, D.H.A. Blank, Transport mechanisms of water and organic solvents through microporous silica in the pervaporation of binary liquids, Microporous and Mesoporous Materials, 65 (2003) 197-208. [27] A.J. Toth, A. Szanyi, K. Angyal-Koczka, P. Mizsey, Enhanced Separation of Highly Non-ideal Mixtures with Extractive Heterogeneous-azeotropic Distillation, Separation Science and Technology, 51 (2016) 1238–1247. [28] A.J. Toth, A. Szanyi, E. Haaz, P. Mizsey, Separation of process wastewater with extractive heterogeneous-azeotropic distillation, Hungarian Journal of Industry and Chemistry, 44 (2016) 29-32. [29] A.J. Toth, P. Mizsey, Methanol removal from aqueous mixture with organophilic pervaporation: Experiments and modelling, Chemical Engineering Research and Design, 98 (2015) 123-135. [30] A.J. Toth, E. Haaz, T. Nagy, R. Tari, A.J. Tarjani, D. Fozer, A. Szanyi, K.-A. Koczka, L. Racz, G. Ugro, P. Mizsey, Evaluation of the accuracy of modelling the separation of highly non-ideal mixtures: extractive heterogeneous-azeotropic distillation, in: A. Espuña, M. Graells, L. Puigjaner (Eds.) Computer Aided Chemical Engineering, Elsevier, 2017, pp. 241246. [31] R. Rautenbach, C. Herion, U. Meyer-Blumentoth, Pervaporation membrane separation processes, Membrane Science and Technology Series, 1 (1990) 181-191. [32] P. Mizsey, K. Koczka, A. Deák, Z. Fonyó, Simulation of pervaporation using the "solution-diffusion" model, Journal of Hungarian Chemistry, 7 (2005) 239-242. [33] A.J. Toth, Liquid Waste Treatment with Physicochemical Tools for Environmental Protection, PhD Thesis, Budapest University of Technology and Economics, Budapest, 2015, pp. 197.

22

[34] M.T. Ashraf, J.E. Schmidt, J. Kujawa, W. Kujawski, H.A. Arafat, One-dimensional modeling of pervaporation systems using a semi-empirical flux model, Separation and Purification Technology, 174 (2017) 502-512. [35] T.F. Edgar, D.M. Himmelblau, L.S. Lasdon, Optimization of chemical processes, McGraw-Hill, 2001. [36] K. Koczka, P. Mizsey, Z. Fonyo, Rigorous modelling and optimization of hybrid separation processes based on pervaporation, Central European Journal of Chemistry, 5 (2007) 1124-1147. [37] G.F. Tusel, H.E.A. Brüschke, Use of pervaporation systems in the chemical industry, Desalination, 53 (1985) 327-338. [38] T. Mohammadi, A. Aroujalian, A. Bakhshi, Pervaporation of dilute alcoholic mixtures using PDMS membrane, Chemical Engineering Science, 60 (2005) 1875-1880. [39] A. Bakhshi, T. Mohammadi, O.G. Nik, A. Aroujalian, Effect of operating conditions on pervaporation of methanol-water mixtures: Part 2, Membrane Technology, 2006 (2006) 7-11. [40] X. Pang, X. Deng, Y. Sun, Studies on the PVA/P(AA-co-AN/SiO2)/PVA composite membranes for pervaporation dehydration of methanol, Modern Applied Science, 4 (2010) 28-35. [41] R.A. Shelden, E.V. Thompson, Dependence of diffusive permeation rates and selectivities on upstream and downstream pressures. IV. Computer simulation of nonideal systems, Journal of Membrane Science, 19 (1984) 39-49. [42] R.B. Bird, W.E. Stewart, E.N. Lightfoot, D.J. Klingenberg, Introductory Transport Phenomena, Wiley, 2015.

23

Highlights 

Polyvinyl-alcohol membrane is able to remove the water from methanol.



Two different pervaporation model are tested on our experimental data.



The improved novel model results in a better fit to the experimental data.



Improved model is able to describe pervaporation in flowsheet environment.



Heat duties of pervaporation dehydration is lower than distillation.

24