Advertisement

Membranes and Membrane Technologies

, Volume 1, Issue 5, pp 331–339 | Cite as

Poly(trimethylsilylpropyne) Membranes for Removal of Alcohol Fermentation Products by Thermopervaporation with a Porous Condenser

  • G. S. GolubevEmail author
  • I. L. Borisov
  • A. V. Volkov
  • V. V. Volkov
Article
  • 45 Downloads

Abstract

Dense membranes made of poly[1-(trimethylsilyl)-1-propyne] (PTMSP) have been studied in thermopervaporation (TPV) separation of a butanol–water binary mixture and a real acetone–butanol–ethanol (ABE) broth. When comparing PTMSP with commercially used membranes based on polydimethylsiloxane (PDMS) in the TPV process, a significant advantage of PTMSP with respect to the butanol/water selectivity is shown. The value of the separation factor for the PTMSP membrane (114) with a thickness of 61 μm is more than ninefold higher than that of the most selective Pervap 4060 commercial membrane (Sulzer Chemtech, Switzerland). The permselectivity of PTMSP membranes with respect to the butanol/water components exceeds that of commercial membranes more than fivefold. The removal of acetone, butanol, and ethanol from real fermentation broths using PTMSP membranes in the TPV process with a porous condensation surface has been studied for the first time. It has been shown that the transport and separation characteristics of PTMSP membranes in the process of separation of real fermentation broths decline with time, while demonstrating high values of the butanol/water permselectivity (4.7).

Keywords:

thermopervaporation porous condenser fermentation broth butanol PTMSP 

INTRODUCTION

Butanol is a high-volume product of the chemical industry, which is used as a solvent in the manufacture of paints and varnishes, coatings, and adhesives [1, 2, 3]. Despite the fact that, currently, almost all butanol is produced from propylene according to the technology of oxosynthesis, the interest of researchers in biobutanol obtained via acetone–butanol–ethanol fermentation from renewable feedstock remains strong [4]. Although the ratio of the three main products can differ depending on the strain of Clostridium, their total concentration in the fermentation broth rarely exceeds ~2 wt % because of the inhibiting action of butanol on microorganisms [5]. Due to this reason, the recovery of the products by distillation is economically unreasonable, since a large amount of liquid wastes requiring disposal is formed [6, 7].

The problem of removal of fermentation products is considered as one of the main obstacles for the commercialization of bioalcohols. Distillation that is a traditionally used technology of recovery of butanol from aqueous media is an economically expensive process due to the high energy consumption [6, 8]. The energy expenditure for concentration can be decreased if some target products are removed from the fermentation broth and it is returned into the bioreactor for further fermentation [9]. One of the most promising technologies for recovery of bioalcohols from aqueous media is pervaporation because it does not require additional reagents and regeneration of the support phase and has no negative influence on the microorganisms [8].

However, when estimating the efficiency of traditional vacuum pervaporation for the recovery of butanol from fermentation broths, costs associated with the maintenance of vacuum and condensation of the permeate vapors are often neglected. In most research works, vacuum from 0.2 to several mbars is applied [10]. These values cannot be reached in industry, where vacuum of about 10 to 100 mbar is used [11]. In addition, the operational costs for vacuum equipment increase due to the formation of noncondensable gases, CO2 and H2, during the fermentation [12]. To condense the permeate vapors under laboratory conditions, liquid nitrogen (−196°C) or dry ice (−80°C) are most often used [13, 14]. Because of this, the use of the fluxes of components and separation factors experimentally obtained at such low pressures of vacuum and temperatures of condensation leads to the incorrect estimates of the efficiency of the industrial pervaporation units being designed.

To solve the problem of generation of vacuum and low temperatures of condensation of the permeate in the process of pervaporation, a method of thermopervaporation (TPV) has been proposed for the removal of alcohols from the fermentation broths [15, 16, 17, 18, 19, 20, 21, 22, 23, 24]. In this thermal gradient process, the moving force is generated by means of condensation of the permeate on a cold surface directly in the membrane module near the membrane. TPV makes it possible to perform separation under atmospheric pressure and at the temperatures of condensation of the permeate of 0–15°C [18].

We have proposed [25, 26] and patented [27] a new design of a TPV module which makes it possible to intensify the TPV process and implement it on an industrial scale. The key technical element of this design is the porous surface of condensation of the penetrated vapors (permeate), which makes it possible to scale up the process for industrial application, simplify the design, and decrease the dimensions of the membrane module [23]. An important problem that limits the application of both TPV and thermopervaporation with a porous condenser (TPV-PC) is the absence of commercial hydrophobic membranes providing acceptable selectivity of butanol/water separation [18, 26]. This limitation is associated with the fact that the coefficients of diffusion of organic substances in membrane materials are by orders of magnitude smaller when compared to water [28]. Due to this reason, the preferable transfer of organic molecules through a membrane only occurs if the selectivity of alcohol/water solubility substantially exceeds the selectivity of water/alcohol diffusion. Generally, this condition is fulfilled in materials with a low selectivity of diffusion [28]. The latter is characteristic for rubbers (e.g., polydimethylsiloxane (PDMS), polyoctylmethyl-siloxane (POMS), and polyester block polyamide (PEBA)) [13, 29, 30, 31], hydrophobic zeolites [32, 33] as well as highly permeable polymeric glass [34, 35].

Membranes based on siloxane polymers are quite easy to obtain and relatively cheap [26, 31]. Because of this, all the commercial membranes for hydrophobic pervaporation are produced based on polysiloxanes. However, they demonstrate low butanol/water selectivity in both vacuum pervaporation (the separation factor of 9–36) [29, 30] and thermopervaporation (the separation factor of 7.5–11.9) [23, 26].

Zeolite membranes possess a high alcohol selectivity (the butanol/water separation factor of 160–400) [32, 33]. Their main disadvantage in the pervaporation separation of fermentation broths is the contamination and degradation of the porous structure of the zeolite in the presence of fermentation byproducts (low-volatile organic compounds, surface-active substances (SASs), acids, etc.) which leads to a decrease in the transport and separation characteristics of the membranes. It should be noted that the cost of such membranes is high due to the difficulty of forming of the defect-free thin layer of the zeolite [36].

The third type of materials which is being actively studied for the task of recovery of the ABE fermentation products is highly permeable glass-like polymers such as polyacetylenes (polytrimethylsilylpropyne (PTMSP) andpolymethylpentyne(PMP)) [35, 37, 38, 39], polymers with internal microporosity (PIM-1), and polynorbornenes [40, 41]. Materials of such a type are capable of forming strong thin polymer films and possess a microporous structure formed by the elements of unrelaxed free volume [42]. Because of this, the butanol/water permeability and selectivity for highly permeable glass is closer to hydrophobic zeolites rather than to glass-like polymers that are water selective in most cases [43].

PTMSP is the most studied in the process of pervaporation recovery of butanol from fermentation broths because it possesses the greatest butanol/water permeability and selectivity out of all the microporous polymers [18]. Thus, it was shown in [36, 39] that PTMSP membranes demonstrate the separation factor of 80–135 in a range of temperatures of 25–60°C in the separation of an aqueous solution of butanol.

Since PTMSP membranes have great potential for the task of TPV recovery of butanol from fermentation broths, the aim of this work was to study PTMSP membranes via TPV-PC in the recovery and concentration of biobutanol, acetone, and ethanol from model mixtures and real ABE fermentation broths.

EXPERIMENTAL

Membranes

Dense membranes made of poly[1-(trimethylsilyl)-1-propyne] (PTMSP) were studied in this work. PTMSP was purchased from Gelest, Inc. (United States).

The membranes for the studies on a laboratory thermopervaporation (TPV) unit with a classic configuration of the module were obtained via casting of a 1.5 wt %solution of the polymer in chloroform onto cellophane followed by drying at room temperature for 200 h. The diameter of the initial films was 7.5 cm and the thicknesses were 33 and 61 μm.

The PTMSP membrane for the studies on a laboratory thermopervaporation with a porous condenser (TPV-PC) bench was obtained via casting of a 1.5 wt % solution of the polymer in chloroform onto a rectangular mold with a Teflon coating. Then the solvent was evaporated for 50–100 h. The size of the initial film was 8 × 28 cm and the thickness was 34 μm.

The thickness of the obtained membranes was measured using a Mitutoyo 293 Digimatic QuickMike electronic micrometer with an accuracy of ±1 μm.

Feed Mixtures

All the solutions were prepared from reagentgrade organic solvents via a gravimetric method. A binary solution of butanol in water with the concentration of butanol of 1.0 wt % and a real acetone–butanol–ethanol (ABE) fermentation broth were used as the mixtures to be separated.

The fermentation broths were obtained at the State Research Institute of Genetics and Selection of Industrial Microorganisms of the National Research Center “Kurchatov Institute.” A B-10939 strain of bacteria Clostridium in a flour medium with the concentration of rye flour of 100 g/L and concentration of CaCO3 of 2 g/L was used as the producer of biobutanol. The final concentration of the target components in the fermentation broth was as follows: 0.77 wt % acetone, 1.61 wt % butanol, and 0.32 wt % ethanol. The broth was subjected to centrifugation prior to the experiments.

The concentration of the initial mixture and permeate was determined via gas chromatography on a Crystallux-4000M chromatograph equipped with a thermal conductivity detector. The chromatography parameters were as follows: the evaporator temperature of 230°C, column temperature of 180°C, detector temperature of 230°C, and length of the packed column with Porapak Q of 1 m. For the analysis of the water–butanol permeate that formed a two-phase system, water was added for the homogenization of the sample.

Thermopervaporation

Thermopervaporation experiments were performed on the unit shown in Fig. 1. Two fluid circuits with different temperatures were piped into thermopervaporation module 3. The first circuit, in which a coolant circulated, was closed to cryostat 7. The second circuit consisted of a vessel with the mixture under separation (a 2-L flask) placed into thermostat 1 which heated the feed mixture to the set temperature and peristaltic pump 2, with the use of which the circulation of the liquid in the circuit was executed. The error of the temperature maintained was ±0.2°C. The temperature of the feed mixture and coolant at the inlet and outlet of the TPV module was measured using thermometers.

Fig. 1.

Block diagram of the thermopervaporation unit: (1) thermostat with feed mixture, (2) peristaltic pump, (3) membrane module, (4) permeate collector, (5) membrane, (6) metallic condensation surface, and (7) cryostat.

The module consisted of two parts which were mirror-symmetric to each other and were separated by membrane 5 and condensation surface 6, between which there was an air gap. The permeate condensed on a metal plate and was removed from the module by the force of gravity into permeate collector 4. The effective membrane area was 25.5 cm2.

Thermopervaporation was performed at a feed mixture temperature of Ts = 60 ± 0.2°C, with the downstream pressure after the membrane being atmospheric. The permeate condensed at Tc = 10°C. The width of the air gap was 2.5 mm.

Thermopervaporation with Porous Condenser

Experiments on thermopervaporation with a porous condenser (TPV-PC) were performed on a laboratory bench shown in Fig. 2. The TPV-PC separation proceeded as follows: by means of pump 3, the mixture to be separated continuously circulated between the hot part of membrane module 4 and feed vessel 1. The membrane was in direct and continuous contact with the feed mixture. The permeate passed through membrane 5, evaporated in air gap 6, and then condensed on porous metal partition 7. The condensate wetted the condensation surface, penetrated into the pores, and passed into the chamber with the coolant. The coolants (distilled water at the beginning of the experiment followed by water + permeate) continuously circulated between the chamber with a porous condenser of TPV-PC module 4 and separator 13. With the increase in the concentration of butanol in the coolant/permeate, the separation to the organic phase (>70 wt % butanol) and the aqueous phase (<8 wt % butanol) occurred in separator 13. It should be noted that the pressure in the chamber with the coolant was lower than in the condensation chamber due to the fact that pump 10 which provided the circulation of the coolant in the system was located behind the module.

Fig. 2.

Block diagram of the thermopervaporation unit with a porous condenser: (1) vessel with feed mixture, (2) thermostat, (3) peristaltic pump, (4) membrane module, (5) membrane, (6) air gap, (7) porous metal partition, (8) cryostat, (9) temperature sensors, (10) coolant/permeate pump, and (11) separator with the coolant/permeate.

Membrane module 4 with the active membrane area of 260 cm2 was equipped with porous metal plate 7 made of stainless steel (VMZ-Tekhno, Russia, a thickness of 200 μm and porosity of 30%). The width of air gap 6 was 1 mm. The heated butanol–water binary mixture with a volume of 2 L or a real ABE fermentation broth with the initial volume of 10 L circulated at a constant rate of 0.6 L/min through the membrane part of the module at 60°C. Distilled water in an amount of 0.5 kg was used as a suitable coolant and circulated at a constant flow rate of 0.3 L/min at 10°C.

The total permeate flux J, kg m−2 h−1, was determined gravimetrically according to the formula
$$J = \frac{m}{{St}},$$
(1)
where m is the total weight of the permeate, kg, which penetrated through the membrane with the area S, m2, over time t, h.

The separation factor was determined according to the formula

$$\beta = \frac{{[{{{{C}_{{\text{o}}}}} \mathord{\left/ {\vphantom {{{{C}_{{\text{o}}}}} {{{C}_{{\text{w}}}}{{]}^{{\text{p}}}}}}} \right. \kern-0em} {{{C}_{{\text{w}}}}{{]}^{{\text{p}}}}}}}}{{[{{{{C}_{{\text{o}}}}} \mathord{\left/ {\vphantom {{{{C}_{{\text{o}}}}} {{{C}_{{\text{w}}}}{{]}^{{\text{f}}}}}}} \right. \kern-0em} {{{C}_{{\text{w}}}}{{]}^{{\text{f}}}}}}}} = \frac{{[{{{{x}_{{\text{o}}}}} \mathord{\left/ {\vphantom {{{{x}_{{\text{o}}}}} {{{x}_{{\text{w}}}}{{]}^{{\text{p}}}}}}} \right. \kern-0em} {{{x}_{{\text{w}}}}{{]}^{{\text{p}}}}}}}}{{[{{{{x}_{{\text{o}}}}} \mathord{\left/ {\vphantom {{{{x}_{{\text{o}}}}} {{{x}_{{\text{w}}}}{{]}^{{\text{f}}}}}}} \right. \kern-0em} {{{x}_{{\text{w}}}}{{]}^{{\text{f}}}}}}}},$$
(2)

where Co and Cw are the weight fractions of the organic component and water in the feed mixture (f) and the permeate (p), respectively, and xo and xw are the molar fractions of the organic component and water in the feed mixture (f) and the permeate (p).

In the case of an acetone–butanol–ethanol–water quaternary system, the separation factors of the alcohols were calculated like for acetone–water, butanol–water, and ethanol–water binary systems.

The mass fluxes of the components in the permeate were determined as Ji = Jyo.

The efficiency of the thermopervaporation process was characterized by the pervaporation separation index (PSI) [13] which takes account of two quantities, the permeate flux and the separation factor:
$${\text{PSI}} = J(\alpha - 1).$$
(3)
To compare the results obtained under different conditions, the data were presented in the terms of permeability and permselectivity which were calculated by the following expressions [24]:
$$\left( {\frac{{{{P}_{i}}}}{l}} \right) = \frac{{{{J}_{i}}}}{{{{M}_{{wi}}}\left( {p_{i}^{{\text{f}}} - p_{i}^{{\text{p}}}} \right)}},$$
(4)
$$\alpha _{i}^{{\text{p}}} = {{\left( {\frac{{{{P}_{i}}}}{l}} \right)} \mathord{\left/ {\vphantom {{\left( {\frac{{{{P}_{i}}}}{l}} \right)} {\left( {\frac{{{{P}_{w}}}}{l}} \right)}}} \right. \kern-0em} {\left( {\frac{{{{P}_{w}}}}{l}} \right)}},$$
(5)

where \(\left( {\frac{{{{P}_{i}}}}{l}} \right)\) and \(\left( {\frac{{{{P}_{{\text{w}}}}}}{l}} \right)\) are the permeabilities of the ith component and water, Mwi is the molecular weight of the ith component, \(p_{i}^{{\text{f}}}\) and \(p_{i}^{{\text{p}}}\) are the vapor pressures of the ith component in the feed mixture and the permeate, and \(\alpha _{i}^{{\text{p}}}\) is the permselectivity of the ith component relative to water.

The partial pressures of the components are related to the pressures of their saturated vapors and activities in the solution:
$${{p}_{i}}(T) = {{y}_{i}}{{x}_{i}}p_{i}^{0}(T),$$
(6)
where yi is the activity coefficient of the ith component, xi is the molar fraction of the ith component, and \(p_{i}^{0}\) is the saturated vapor pressure of the ith component at the temperature of the system.

In the case of the acetone–butanol–ethanol–water quaternary system, the permeabilities to the components were calculated like those for binary systems. To determine the vapor pressures of the permeate and feed mixture, the activity coefficients for acetone–water and ethanol–water solutions were calculated with the use of the non-random two liquid (NRTL) model and the Aspen Plus V8.6 software. This calculation method is widely used in published works for aqueous alcohol systems [18, 25]. In the case of the butanol–water binary system, the approximated four-parameter Margules equation, which allows taking into account the complex interactions between of butanol and water molecules, was used for calculating the activity coefficients [14].

RESULTS AND DISCUSSION

Thermopervaporation Comparison of PTMSP Membranes with Commercial PDMS Membranes

In this work, thermopervaporative properties of dense PTMSP membranes were studied, and the obtained results were compared with the published data on the TPV recovery of butanol from aqueous media with the use of commercial organophilic membranes based on PDMS [22, 26] (Table 1). Experiments on the characterization of the TPV properties of the membranes were performed using a module with the working membrane area of 25.5 cm2. A 1 wt % butanol solution in water was separated. The PTMSP membranes demonstrated the maximum values of both the butanol flux and the separation factor, of which the latter for a PTMSP membrane with a thickness of 61 μm exceeded by more than ninefold the separation factor of the most selective commercial membrane Pervap 4060 (Sulzer Chemtech, Switzerland). To compare the efficiency of the membranes, the pervaporation separation index (PSI) was used, which takes into account both the permeate flux and the separation factor. For PTMSP membranes, PSI is greater than 31 kg m−2 h−1 and exceeds the PSI of commercial membranes two- to threefold.

Table 1.  

TPV characteristics of hydrophobic membranes in the separation of 1 wt % butanol solution in water. Ts = 60°C and Tc = 10°C

Membrane

Flux,

kg m−2 h−1

BuOH flux, kg m−2 h−1

Separation factor, β

PSI,

kg m−2 h−1

Ref.

PTMSP 33 μm

0.4

0.18

81

32.0

This work

PTMSP 61 μm

0.28

0.15

114

31.6

This work

Pervap 4060, Sulzer Chemtech (Switzerland)

0.9

0.10

11.9

9.8

[26]

Pervatech PDMS, Pervatech (The Netherlands)

1.5

0.11

7.8

10.2

[26]

PolyAn, POL_OR_M2, PolyAn GmbH (Germany)

2.2

0.15

7.5

14.3

[26]

MDK-3, ZAO NTTsVladipor (Russia)

1.5

0.15

11.5

15.7

[26]

To compare the transport properties of the membranes in the TPV process, the permeabilities for butanol and water and the butanol/water selectivity were calculated (Table 2). The PTMSP membranes, having similar butanol permeability in comparison with the commercial membranes based on PDMS, demonstrate low water permeability, which eventually leads to a high selectivity of separation. It should be noted that the selectivities of PTMSP membranes with thicknesses of 33 and 61 μm are 4.2 and 5.9, respectively, whereas the selectivity of the best commercial membranes is smaller or equal to unity. Thus, the membrane contribution to separation selectivity is essential in the case of PTMSP, and commercial membranes only act as the evaporation boundary between the liquid and vapor–air phases. In the latter case, the separation selectivity is determined by the liquid–vapor phase equilibrium in the butanol–water system.

Table 2.  

Transport properties of hydrophobic membranes upon TPV separation of 1 wt % butanol solution in water. Ts = 60°C and Tc = 10°C

Membrane

Permeability for butanol,

mol m−2 h−1 kPa−1

Permeability for water,

mol m−2 h−1 kPa−1

Selectivity

Ref.

PTMSP 33 μm

2.8

0.66

4.2

This work

PTMSP 61 μm

2.3

0.39

5.9

This work

Pervap 4060

1.5

2.4

0.63

[26]

Pervap 4060*

2.0

1.6

1.25

[22]

Pervatech PDMS

1.7

4.2

0.4

[26]

Pervatech PDMS*

1.3

1.8

0.73

[22]

PolyAn, POL_OR_M2

2.3

6.1

0.38

[26]

MDK-3

2.3

4.0

0.58

[26]

* TPV experiments were performed under different conditions: the temperature of the feed mixture was 63 ± 1°C and the condensation temperature was 6 ± 1°C. The concentration of butanol in the feed mixture was 4 wt %, and the width of the air gap was 5 mm.

Testing PTMSP Membranes in TPV with Porous Condenser

It was shown earlier that the TPV process can be effectively implemented in the design of a membrane module, where a porous plate is used as the condensation surface [26]. The main advantage of this approach is the possibility for scaling up the TPV on modules with a larger active membrane area, and, therefore, studying the TPV process under conditions close to the industrial ones. To estimate the potential of PTMSP membranes in the TPV recovery of butanol from a fermentation broth, it was important to find out how the scale and design of the membrane module affect the separation process. In connection with this, the transport characteristics of PTMSP membranes with an average thickness of 34 μm were studied in a TPV-PC module with the active area of the membranes of 260 cm2 using a butanol–water binary mixture as a feed.

Figure 3 presents the dependences of the permeate and butanol fluxes and butanol/water separation factor on time in the case of separation of a 1 wt % butanol solution in water. As it can be seen, the initial values of the total flux (0.43 kg m−2 h−1), butanol flux (0.19 kg m−2 h−1), and separation factor (78) are not inferior to the values obtained for a PTMSP membrane with an average thickness of 33 μm in a TPV unit with the classical design of the module and active membrane area of 25.5 cm2. Thus, using a porous partition as the condensation surface and increasing the active membrane area tenfold do not decrease the efficiency of the TPV process. During 100 h of the experiment, the membrane demonstrates stable values of the butanol/water separation factor (109), although the butanol flux decreases by 15%. This change in the flux is m most likely due to the physical aging of PTMSP, which was observed not only in gas separation [44], but also in vacuum pervaporation [9] and TPV [19] processes.

Fig. 3.

Testing of a PTMSP membrane in the separation of a 1 wt % binary solution of butanol in water. Ts = 60°C, Tc = 10°C, lPTMSP = 34 μm, and the initial coolant is distilled water.

In addition to acetone, butanol, and ethanol, real fermentation broths contain a great number of components (surfactants, fermentation byproducts, and unprocessed substrate) which can affect the transport characteristics of the membrane. Because of this, the TPV-PC study of PTMSP membranes in the process of separation of real fermentation broths is a task important from a practical standpoint.

In this work, a real ABE fermentation broth with a volume of 10 L obtained on a flour nutrition medium was separated. Prior to the TPV experiments, the broth was centrifuged as the model of the broth in a continuous fermentation process with immobilized cells. The concentrations of acetone, butanol, and ethanol in the initial broth were 0.77, 1.61, and 0.32 wt %, respectively. Distilled water was initially used as the coolant. The temperature of the broth under separation was 60°C, and the temperature of the coolant was 10°C. The separation process was performed until the final concentration of butanol in the mixture under separation was 0.4 wt % because butanol has a toxic effect on the culture during fermentation at high concentrations [45]. The obtained broth depleted in the target components can be sent to repeated fermentation, thus achieving deeper processing of the feedstock.

Figure 4a presents the dependences of the decrease in the concentration of the target components in the fermentation broth. It has been found that the concentration of butanol in the broth under separation decreases from 1.61 to 0.41, the concentration of acetone, from 0.77 to 0.38, and the concentration of ethanol, from 0.32 to 0.23 wt % over 68.5 h of the experiment on the separation of the ABE fermentation broth. Therefore, the concentration of butanol decreases almost fourfold, and the obtained mixture can be used for further fermentation. As it can be seen from Fig. 4c, the separation factor with respect to all the organic components also decreases with time.

Fig. 4.

Time dependences of (a) acetone, butanol, and ethanol concentrations and (b) acetone/water, butanol/water, and ethanol/water separation factors. Ts= 60°C, Tc= 10°C, lPTMSP = 34 μm, and the initial coolant is distilled water.

Figure 5 presents the time dependence of the recovery of the permeate and the recovery of the organic matter which can be calculated by the following formulae:

$$D = \frac{{{{m}_{{\text{p}}}}}}{{{{m}_{{\text{f}}}}}},$$
(7)
$${{D}^{{{\text{org}}}}} = \frac{{m_{{\text{p}}}^{{{\text{org}}}}}}{{m_{{\text{f}}}^{{{\text{org}}}}}},$$
(8)
Fig. 5.

Time dependences of the fraction of the total amount of the permeate on the initial weight of the fermentation broth and fraction of ABE in the permeate on the initial amount of ABE in the initial fermentation broth.

where mp and \(m_{{\text{p}}}^{{{\text{org}}}}\) are the total weight of the permeate and weight of ABE in the permeate, respectively, and mf and \(m_{{\text{f}}}^{{{\text{org}}}}\) are the initial weight of the fermentation broth and initial weight of ABE in the fermentation broth, respectively.

It is seen that, in percentage terms, ABE are faster removed from the fermentation broth. Thus, after 68.5 h, degree of extraction of the organic matter is about 65%, and degree of extraction of the permeate is 33%. However, after 40 h of the experiment, degree of extraction of the organic matter has a trend to reaching a plateau, which suggests a decrease in the efficiency of recovery of the desired components. Such a dependence is associated with the decrease in the organic matter flux with time (Fig. 4a).

The transport properties of the membrane are illustrated in Fig. 6. As is seen, both the total permeate flux and the butanol flux decrease during the course of the experiment, with the water flux decreasing insignificantly. Thus, the total flux decreases from 0.34 to 0.12 kg m−2 h−1, and the butanol flux, from 0.17 to 0.023 kg m−2 h−1. Since the concentration of the organic matter in the fermentation broth changes with time, the time dependences of the permeability for water and butanol have been constructed (Fig. 6b). The water permeability decreases from 0.3 to 0.2 mol m−2 h−1 kPa−1, and that of butanol, from 1.55 to 1.0 mol m−2 h−1 kPa−1, with the butanol/water permselectivity remaining at a high level to be 4.7 on average. This decline in the transport properties of PTMSP membranes is most likely due to membrane fouling by fermentation byproducts. Thus, the clogging of PTMSP membranes during the vacuum pervaporative separation of real ABE fermentation broths was studied in [37, 46]. It was found that the pervaporation characteristics of PTMSP membranes (the total and partial fluxes and separation factors) decrease on passing from binary solutions to multicomponent artificial mixtures and further to real fermentation broths. The authors associated this decrease in the pervaporation properties with the bulk fouling of PTMSP membranes by low-volatile fermentation byproducts (e.g., diols).

Fig. 6.

Dependences of the (a) total and partial butanol and water fluxes and (b) calculated permeabilities of butanol and water and butanol/water selectivity on the duration of the experiment. Ts= 60°C and Tc= 10°C.

In addition, a similar experiment on the separation of an ABE fermentation broth with the same composition and volume was earlier performed in [26] with the use of an MDK-3 membrane in a TPV-PC unit. It was shown that the decrease in the concentration of butanol in the fermentation broth from 1.6 to 0.4 wt % takes 52 h, which is by 16.5 h less than for a PTMSP membrane. In addition, the permeability of the MDK-3 membrane for butanol (1.6 mol m−2 h−1 kPa−1) and water (2.4 mol m−2 h−1 kPa−1) remained stable throughout the entire experiment but at a butanol/water selectivity of 0.7. Therefore, in the case of separation of real ABE fermentation broths, PTMSP membranes are more susceptible to fouling by fermentation byproducts in comparison with PDMS membranes, but they demonstrate high values of the butanol/water permselectivity (4.7). Due to the high values of the butanol/water selectivity, PTMSP continues to be an attractive material for the TPV separation of real ABE fermentation broths but effective and technically simple methods of cleaning of PTMSP membranes to remove the fermentation byproducts are required for them to compete with PDMS membranes.

CONCLUSIONS

It has been found that homogeneous PTMSP membranes demonstrate better separation and transport characteristics in the recovery of butanol from aqueous media by thermopervaporation in comparison with available commercial membranes based on PDMS. The separation factor for a PTMSP membrane (114) with a thickness of 61 μm turned out to be more than ninefold higher than that of the most selective commercial membrane Pervap 4060 (Sulzer Chemtech, Switzerland). The pervaporation separation index for PTMSP membranes turned out to be more than 31 kg m−2 h−1. This value is two- to threefold higher in comparison with commercial membranes. In addition, the butanol/water permselectivity of PTMSP membranes exceeds the selectivity of commercial membranes by more than fivefold.

Time dependences the permeate and butanol fluxes and the separation factor for a PTMSP membrane in the separation of a 1 wt % butanol solution in water on a TPV-PC bench have been first obtained. During 100 h of the experiment, the PTMSP membrane demonstrates stable values of the butanol/water separation factor (109), with the permeate flux decreasing by 15%.

The recovery of acetone, butanol, and ethanol from real fermentation broths with the use of PTMSP membranes in the TPV process with a porous condensation surface has been studied for the first time. It has been shown that the transport and separation characteristics of PTMSP membranes in the separation of real fermentation broths decrease with time due to the bulk fouling of PTMSP membranes by low-volatile fermentation byproducts. However, membranes made of PTMSP demonstrate high values of the butanol/water permselectivity (4.7). In this connection, PTMSP is an attractive material for the separation of real ABE fermentation broths, although effective and technically simple methods of cleaning PTMSP membranes to remove fermentation byproducts are required.

Notes

FUNDING

This work was carried out within the State Programs of the Topchiev Institute of Petrochemical Synthesis, Russian Academy of Sciences.

REFERENCES

  1. 1.
    V. García, J. Päkkilä, H. Ojamo, et al., Renew. Sust. Energy Rev. 15, 964 (2011).CrossRefGoogle Scholar
  2. 2.
    E. M. Green, Curr. Opin. Biotechnol. 22, 337 (2011).CrossRefGoogle Scholar
  3. 3.
    C. Xue, J. Zhao, L. Chen, et al., Biotechnol. Adv. 35, 310 (2017).CrossRefGoogle Scholar
  4. 4.
    P. Dürre, J. Biotechnol. 2, 1525 (2007).CrossRefGoogle Scholar
  5. 5.
    T. C. Ezeji, N. Qureshi, and H. P. Blaschek, Curr. Opin. Biotechnol. 18, 220 (2007).CrossRefGoogle Scholar
  6. 6.
    Y. Ni and Z. Sun, Appl. Microbiol. Biotechnol. 83, 415 (2009).CrossRefGoogle Scholar
  7. 7.
    L. M. Vane, J. Chem. Technol. Biotechnol. 80, 603 (2005).CrossRefGoogle Scholar
  8. 8.
    A. Oudshoorn, L. A. M. van der Wielen, and A. J. J. Straathof, Ind. Eng. Chem. Res. 48, 7325 (2009).CrossRefGoogle Scholar
  9. 9.
    M. F. Dubreuil, P. Vandezande, W. H. van Hecke, et al., J. Membr. Sci. 447, 134 (2013).CrossRefGoogle Scholar
  10. 10.
    W. van Hecke, E. Joossen-Meyvis, H. Beckers, and H. De Wever, Appl. Energy 228, 437 (2018).CrossRefGoogle Scholar
  11. 11.
    H. Bannwarth, Liquid Ring Vacuum Pumps, Compressors and Systems: Conventional and Hermetic Design (Wiley–VCH, Weinheim, 2006).Google Scholar
  12. 12.
    W. van Hecke and H. De Wever, J. Membr. Sci. 540, 321 (2017).CrossRefGoogle Scholar
  13. 13.
    N. Qureshi, M. M. Meagher, J. Huang, and R. W. Hutkins, J. Membr. Sci. 187, 93 (2001).CrossRefGoogle Scholar
  14. 14.
    N. Qureshi and H. P. Blaschek, Biomass Bioenergy 17, 175 (1999).CrossRefGoogle Scholar
  15. 15.
    P. Aptel, N. Challard, J. Cuny, and J. Neel, J. Membr. Sci. 1, 271 (1976).CrossRefGoogle Scholar
  16. 16.
    A. C. M. Franken, M. H. V. Mulder, and C. A. Smolders, J. Membr. Sci. 53, 127 (1990).CrossRefGoogle Scholar
  17. 17.
    I. L. Borisov, V. V. Volkov, V. A. Kirsh, and V. I. Roldugin, Pet. Chem. 51, 542 (2011).CrossRefGoogle Scholar
  18. 18.
    I. L. Borisov and V. V. Volkov, Sep. Purif. Technol. 146, 33 (2015).CrossRefGoogle Scholar
  19. 19.
    G. S. Golubev, I. L. Borisov, and V. V. Volkov, Pet. Chem. 58, 975 (2018).CrossRefGoogle Scholar
  20. 20.
    I. L. Borisov, G. S. Golubev, E. V. Patrusheva, and S. P. Sineokoy, Chem. Eng. Trans. 64, 43 (2018).Google Scholar
  21. 21.
    G. S. Golubev, I. L. Borisov, E. G. Litvinova, et al., Pet. Chem. 57, 498 (2017).CrossRefGoogle Scholar
  22. 22.
    A. Kujawska, J. Kujawski, M. Bryjak, and W. Kujawski, Chem. Eng. Process.: Process Intens. 94, 62 (2015).CrossRefGoogle Scholar
  23. 23.
    A. Rozicka, S. Koter, and W. Kujawski, Monogr. Environ. Eng. Commit. 118, 69 (2014).Google Scholar
  24. 24.
    S. Koter, A. Kujawska, and W. Kujawski, J. Membr. Sci. 480, 129 (2015).CrossRefGoogle Scholar
  25. 25.
    A. V. Volkov, E. G. Novitsky, I. L. Borisov, et al., Sep. Purif. Technol. 171, 191 (2016).CrossRefGoogle Scholar
  26. 26.
    I. L. Borisov, G. S. Golubev, V. P. Vasilevsky, et al., J. Membr. Sci. 523, 291 (2017).CrossRefGoogle Scholar
  27. 27.
    I. L. Borisov, V. P. Vasilevsky, V. V. Volkov, and A. V. Volkov, WO Patent No. 2 014 077 739 (2014).Google Scholar
  28. 28.
    V. V. Volkov, Russ. Chem. Bull. 43, 187 (1994).CrossRefGoogle Scholar
  29. 29.
    A. Rozicka, J. Niemistö, R. L. Keiski, and W. Kujawski, J. Membr. Sci. 453, 108 (2014).CrossRefGoogle Scholar
  30. 30.
    J. Niemistö, W. Kujawski, and R. L. Keiski, J. Membr. Sci. 434, 55 (2013).CrossRefGoogle Scholar
  31. 31.
    D. Liu, G. Liu, L. Meng, et al., Sep. Purif. Technol. 146, 24 (2015).CrossRefGoogle Scholar
  32. 32.
    Q. Liu, R. D. Noble, J. L. Falconer, and H. H. Funke, J. Membr. Sci. 117, 163 (1996).CrossRefGoogle Scholar
  33. 33.
    S. Li, V. A. Tuan, J. L. Falconer, and R. D. Noble, Microporous Mesoporous Mater. 58, 137 (2003).CrossRefGoogle Scholar
  34. 34.
    T. Masuda and T. Higashimura, Adv. Polym. Sci. 81, 121 (1987).Google Scholar
  35. 35.
    A. V. Yakovlev, M. G. Shalygin, S. M. Matson, et al., J. Membr. Sci. 434, 99 (2013).CrossRefGoogle Scholar
  36. 36.
    V. van Hoof, C. Dotremont, and A. Buekenhoudt, Sep. Purif. Technol. 48, 304 (2006).CrossRefGoogle Scholar
  37. 37.
    A. G. Fadeev, M. M. Meagher, S. S. Kelley, and V. V. Volkov, J. Membr. Sci. 173, 133 (2000).CrossRefGoogle Scholar
  38. 38.
    A. G. Fadeev, S. S. Kelley, J. D. McMillan, et al., J. Membr. Sci. 214, 229 (2003).CrossRefGoogle Scholar
  39. 39.
    A. G. Fadeev, Y. A. Selinskaya, S. S. Kelley, et al., J. Membr. Sci. 186, 205 (2001).CrossRefGoogle Scholar
  40. 40.
    M. Žák, M. Klepic, L. C. Štastná, et al., Sep. Purif. Technol. 151, 108 (2015).CrossRefGoogle Scholar
  41. 41.
    D. G. Kim, T. Takigawa, T. Kashino, et al., Chem. Mater. 27, 6791 (2015).CrossRefGoogle Scholar
  42. 42.
    D. Hofmann, M. Heuchel, Y. Yampolskii, et al., Macromolecules 35, 2129 (2002).CrossRefGoogle Scholar
  43. 43.
    M. Lewandowska and W. Kujawski, J. Food Eng. 79, 430 (2007).CrossRefGoogle Scholar
  44. 44.
    K. Nagai and T. Nakagawa, J. Membr. Sci. 105, 261 (1995).CrossRefGoogle Scholar
  45. 45.
    N. Qureshi and T. C. Ezeji, Biofuels Bioprod. Biorefin. 2, 319 (2008).CrossRefGoogle Scholar
  46. 46.
    A. V. Volkov, V. V. Volkov, and V. S. Khotimskii, Polym. Sci., Ser. A 51, 1367 (2009).CrossRefGoogle Scholar

Copyright information

© Pleiades Publishing, Ltd. 2019

Authors and Affiliations

  • G. S. Golubev
    • 1
    Email author
  • I. L. Borisov
    • 1
  • A. V. Volkov
    • 1
  • V. V. Volkov
    • 1
  1. 1.Topchiev Institute of Petrochemical Synthesis, Russian Academy of SciencesMoscowRussia

Personalised recommendations