Sie sind auf Seite 1von 16

Journal of Membrane Science 169 (2000) 159174

A review on the separation of benzene/cyclohexane mixtures by pervaporation processes

J.P. Garcia Villaluenga, A. Tabe-Mohammadi
National Research Council Canada, Institute for Chemical Process and Environmental Technology, Montreal Rd., Ottawa, Ont., Canada K1A 0R6 Received 6 April 1999; received in revised form 12 July 1999; accepted 13 July 1999

Abstract Separation of benzene (Bz) and cyclohexane (Chx) is one of the most challenging processes in the chemical industry. On account of the only 0.6 C difference in volatilities of the two components, conventional distillation processes are not practical. Azeotropic distillation and extractive distillation, on the other hand, although feasible and in use in many industries, are accompanied by high capital and operating costs and also are complex processes in nature. Membrane pervaporation (PV) is a viable alternative to the conventional techniques both from economical and technical points of view. While extractive distillation/PV hybrid has attracted industrial attention, progresses in PV separation of Bz/Chx suggest that this technique can handle a greater share in the process. In this paper, an extensive review is presented on the properties, production, conventional and alternative separation processes of benzene and cyclohexane. Special emphasis is given to the performance of membrane PV processes. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Separation of benzene/cyclohexane; Pervaporation processes; Permselectivity

1. Introduction Separation of benzene (Bz) and cyclohexane (Chx) is among the most important and most difcult processes in the petrochemical industry. Cyclohexane is produced by catalytic hydrogenation of benzene. The unreacted benzene is present in the reactors efuent stream and must be removed for pure cyclohexane recovery. Separation of benzene and cyclohexane is difcult by a conventional distillation process because
Corresponding author. Tel.: +613-993-9574; fax: +613-9412529. E-mail address: (A. Tabe-Mohammadi).

these components form close boiling point mixtures at the entire range of their compositions. Presently, azeotropic distillation and extractive distillation are used for this separation. These two processes, however, suffer from complexity and high energy consumption. For all these reasons, the industry has always been eager to look for a viable alternative to the conventional Bz/Chx separation processes. In the recent years, pervaporation (PV) separation has emerged as an economical and simple alternative to many organic/organic separation applications. This process is especially attractive in azeotropic and close boiling point separation applications. That is so because, unlike distillation processes, the separation mechanism in PV is not based on the relative volatility

0376-7388/00/$ see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 6 - 7 3 8 8 ( 9 9 ) 0 0 3 3 7 - 3


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174 Table 1 Physical properties of benzene Constant fp bp ( C) Density at 25 C (g/mol) Refractive index, n25 D Viscosity (absolute) at 20 C (cP) Surface tension at 25 C (dyn/cm) Critical temperature ( C) ( C) Value 5.533 80.100 0.8737 1.49792 0.6468 28.18 289.45

Fig. 1. Chemical structures of benzene (top) and cyclohexane (bottom). Benzene has a planar structure with six equal CC bonds and bond angles. Cyclohexane posses a chair conformation.

of components, but on the difference in sorption and diffusion properties of the feed substances as well as permselectivity of the membrane.

2. Properties of benzene and cyclohexane Benzene is an aromatic hydrocarbon with the chemical formula C6 H6 . It is a non-polar, volatile, colorless, ammable liquid, with a characteristic odor and great thermal stability. The chemical structure of benzene is shown in Fig. 1 and its representative properties are given in Table 1 [1]. Benzene has a stable planar structure with a six-member Kekule ring. The carbon bonds have equal lengths of 1.39 and bond angles of 120 .

The principal use of benzene is as a chemical raw material in the synthesis of compounds such as styrene (polystyrene plastics and synthetic rubber), phenol (phenolic resins), cyclohexane (Nylon), aniline, maleic anhydride (polyester resins), alkylbenzenes (detergents), and chlorobenzenes. Benzene is also used in the production of drugs, dyes, insecticides, and plastics. Benzene, toluene, and xylenes (BTX) can be found in motor gasoline as light high-octane aromatic hydrocarbons. Benzene was also used as a solvent, but safer solvents have replaced it in the recent years. Table 2 shows selected applications of benzene [2]. An extensive application of benzene in different chemical products can be found in the same reference. Cyclohexane, with the chemical formula C6 H12 , is a volatile, colorless, ammable, water-insoluble, non-corrosive, non-polar liquid having a pungent odor. It is less toxic than benzene. The most stable chemical structure of cyclohexane, known as chair conformation, is shown in Fig. 1. Table 3 gives its physical properties. Cyclohexane is used in paints and varnishes, as a solvent in the plastic industry,

Table 2 Selected used of benzene compounds [2]

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174 Table 3 Physical properties of cyclohexane Constant fp bp ( C) Density at 25 C (g/mol) Refractive index, n25 D Viscosity (absolute) at 20 C (cP) Surface tension at 25 C (dyn/cm) Critical temperature ( C) ( C) Value 6.554 80.738 0.77855 1.42623 0.980 25.3 281.0


particularly for resins used in wire coating, and for the extraction of essential oils. However, the importance of cyclohexane lays mainly in its conversion to the intermediate cyclohexanone, a feedstock for Nylon precursors such as adipic acid, -caprolactam and hexamethylenediamine [2,3]. Cyclohexane forms binary, tertiary, and quaternary azeotropes with a variety of chemicals. Table 4 shows the three most well known binary azeotropes formed with cyclohexane. More details of the cyclohexane azeotropes can be found in [4]. 3. Production of benzene and cyclohexane 3.1. Benzene On account of the increasing demand for plastics, detergents, and organic chemicals, the production of benzene has increased rapidly since 1950. There are many roots to the production of benzene, including Extraction from crude oil or coal Dehydrogenation of cycloparafns Dealkylation of chain alkyl aromatics, e.g., toluene and xylenes Transalkylation of toluene Reforming-separation Extraction from coal Some expensive processes such as dealkylation of toluene, xylenes, or longer chain alkyl aromatics were
Table 4 Binary azeotropes of cyclohexane [4] Second component Water Benzene Methanol Cyclohexane (wt.%) 91.6 45 39 bp ( C) 69.0 77.5 54.2

considered only in the past couple of decades when the high price of benzene justied the process. The above processes are described briey in this paper. More details could be found in [1,2,514]. Benzene can be practically found in all crude oils in small amounts. However, its separation becomes attractive only when its concentration in the petroleum fraction is increased. Another source of benzene is the light gas produced from coal. Benzene is one of the principal compounds present in the light oil recovered from coal carbonization gases. The composition of light oil depends mainly on the temperature of the evolved gases. A higher proportion of benzene is found at high temperatures. Dehydrogenation of cycloparafns was developed prior to 1950. In this process, a naphtha-range petroleum fraction is catalytically reformed to produce a high-octane product for use in gasoline blending. Then, solvent extraction and fractional distillation of the extract removes benzene and other aromatic hydrocarbons. In the dealkylation process, methyl radical in a toluene molecule is replaced by hydrogen atom to produce benzene [1]. More recently, benzene has been produced from toluene by transalkylation, also called toluene disproportionation. In this process, the methyl group from one toluene molecule is transferred to another toluene molecule to produce benzene and mixed xylenes [13,14]. The high value of p-xylene has made, at times, this process more protable than dealkylation. Mobil is presently using this method in its TransPlusTM as well as MTDP-3 technology [15,16]. The production of benzene by reforming-separation processes is associated with the production of toluene and xylene (BTX plants) ([2] and references therein; [5]). In reforming-separation processes, the aromatic fraction of crude oil is reformed catalytically to produce BTX compounds. Benzene, toluene, and xylenes are then recovered by extractive distillation. A novel example of this technology is Chevrons Aromax reforming process, which uses a zeolite catalyst to convert light naphtha to benzene and toluene [1719]. 3.2. Cyclohexane The production of cyclohexane is of special interest to us because it is directly related to the sub-


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174

Fig. 2. Top: multistage reactor, vapor phase process, (a) reactors; (b) separator drum; (c) compressor; (d) stabilizer. Bottom: liquid and vapor phase reactor process. (a) Liquid phase reactor; (b) vapor phase reactor; (c) separator drum; (d) compressor; (e) stabilizer.

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174


ject of this paper. Although cyclohexane was rst obtained by direct fractionation of crude gasoline cuts, two major factors pushed for the introduction of new processes. One was the increasing demand for cyclohexane, which could not be met by the amount of natural occurrence of this substance, and the other was the low purity of product by direct distillation (8598%). The demand for cyclohexane mainly comes from the plastic industry where cyclohexane is used as a feedstock for Nylon 6 and Nylon 6,6. Presently, virtually all cyclohexane is produced by the hydrogenation of benzene (ca. 8085%) [3,4]. Benzene is hydrogenated catalytically to cyclohexane in liquid or vapor phase in the presence of hydrogen. Several cyclohexane processes, which use nickel, platinum, or palladium as the catalyst, have been developed [4]. Hydrogenation proceeds readily and is highly exothermic ( H500K = 214 kJ/mol). Temperature and pressure control is essential in shifting the equilibrium in favor of benzene or cyclohexane [20,21]. Typical set-points are 300 C and 300450 psi. High purity cyclohexane can only be obtained under a complex process control. Therefore, economical and efcient heat removal has been the subject of many reactor designs [2224]. Fig. 2 shows the schematics of typical liquid phase and vapor phase benzene hyrogenation reactors. Cyclohexane can also be distilled directly from naphtha. The starting concentration of cyclohexane in the naphtha fraction of crude oil is in the range of 0.55.0 vol%. The presence of several hydrocarbons boiling in the same range, such as n-hexane, isohexanes, methylcyclopentane, benzene, and dimethylpentanes, makes the recovery of cyclohexane from petroleum by fractional distillation or recrystallization difcult and uneconomical. However, cyclohexane at 85 wt.% purity is manufactured in commercial quantities by fractionating the naphthas from natural gasoline. The C5 naphtha produced, then, contains 1014% cyclohexane. This stream is nally treated in high efciency distillation towers, which further concentrates the cyclohexane to 85% purity. The difculty of separating cyclohexane from naphtha is underscored by the fact that only one company, Phillips Petroleum, produces cyclohexane by distillation.

4. Conventional separation processes for Bz/Chx Separation of Bz/Chx mixtures is among the most complicated processes in the petrochemical industry. As the difference in the boiling points of benzene and cyclohexane is only 0.6 C, and the interaction parameter between them is small, the liquidvapor equilibrium curve of their binary liquid mixture, as shown in Fig. 3, is nearly a close-boiling system. Also, benzene and cyclohexane form azeotrope at 45 vol% cyclohexane, as shown in Table 4. Consequently, fractional distillation becomes expensive and cumbersome for Bz/Chx separation. The most common separation techniques developed for aromatic/aliphatic mixtures are listed in Table 5 [3]. The separation problems and the requirements for the technological or economical operation of the individual processes are also summarized in Table 5. Distillation processes such as azeotropic distillation and extractive distillation are the main technologies presently available to separate Bz/Chx mixtures. The selectivity of these processes is limited by the vaporliquid equilibrium. Also, these processes are partly carried out by adding a third component to the Bz/Chx mixtures. Removal of the third component from distillate or residue adds to the complexity and cost of the process.

Fig. 3. Bz/Chx vaporliquid equilibrium curve.


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174

Table 5 Processes for aromatic recovery [31] Process Azeotropic distillation Extractive distillation Liquidliquid extraction Crystallization Adsorption on solids Separation problem BTX separation from pyrolysis gasoline BTX separation from pyrolysis gasoline BTX separation from reformate gasoline Isolation of p-xylene from m/p-mixtures Isolation of p-xylene from C8 aromatic fractions Requirements for basic or economical operation High aromatic content (>90%) Medium aromatic content (6590%) Lower aromatic content (2065%) Distillative pre-separation of o-xylene and ethylbenzene from C8 aromatic fractions Continuous, reversible and selective adsorption

In azeotropic distillation, the separation process is facilitated by addition of strongly polar auxiliary agents such as amines, alcohols, ketones, and water. The auxiliary agent forms azeotrope with one of the two components and alters the vaporliquid equilibrium curve. This mixture is collected at the overhead of the distillation column. The second component is then recovered at high purity from the bottom. The new azeotrope is separated in a secondary stage, where the auxiliary agent is recycled to the rst distillation column and the second component is recovered at high purity. Azeotropic distillation is economical when the aromatic content is greater than 90%, i.e., only a small amount of non aromatics is to be separated, such as with the pyrolysis of gasoline or crude benzene from coal coking [2,4]. In the case of Bz/Chx separation, azeotropic distillation is practical only at high benzene concentrations, e.g., >90%. The separation is facilitated by the addition of acetone [25], as shown in Fig. 4 [26]. This is a hybrid process consisting of azeotropic distillation and extraction. In this process, two entrainers are used. The rst is acetone to mix with cyclohexane and replace Bz/Chx azeotrope. Acetone forms a minimum azeotrope with cyclohexane at approximately 77% cyclohexane, with a boiling point of 53.1 C [26]. The second entrainer is water, which recovers the acetone in the extraction column. As shown in Fig. 4, the border distillation line runs from Bz/Chx azeotrope to acetone/cyclohexane azeotrope. In the rst distillation tower, acetone/cyclohexane azeotrope is collected from the overhead, and benzene from the bottom. The azeotrope is fed into the extractor where acetone is washed with water. As cyclohexane is insoluble in water, nearly pure cyclohexane is produced. Finally, the water/acetone mixture is distilled in the second distillation tower. Acetone is

recycled to the rst distillation column as entrainer, and water is recycled back to the extractor for acetone wash. Extractive distillation is similar to azeotropic distillation in the sense that they both use an auxiliary agent to facilitate the separation. In extractive distillation, however, the additive is used to decrease the volatility of one component. The other component is distilled overhead in pure form, and the bottom mixture is fed into a stripping column for separation. Extraction distillation is particularly economical for the isolation of aromatics from pyrolysis gasoline (approximately 6590% aromatics). Extractive distillation is widely used by the industry to separate benzene and cyclohexane. An example of this process is shown in Fig. 5 [27]. In this process, furfural, which is highly benzene selective, is added to the mixture to shift the vaporliquid equilibrium. Highly pure cyclohexane is distilled overhead. The benzene/furfural mixture is fed into a second column where benzene is collected from the overhead and furfural is recycled back to the rst distillation column. A major problem with this system is that the production of high purity cyclohexane, i.e., >98.5%, is accompanied by rapidly increasing energy consumption [27]. Like azeotropic distillation, extractive distillation suffers from disadvantages such as high capital and operating cost, complexity of the process, need for a secondary distillation of the auxiliary agent, and limited range of feed concentration. The sum of these disadvantages became the driving force for research into alternative separation processes including the emerging membrane PV. Proper membranes must show afnity towards aromatic substances reected by a distinctly higher permeability for aromatics than for aliphatic organics.

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174


Fig. 4. General owsheet for azeotropic distillation. The process consists of a rst distillation stage to remove lower boiling fraction with the aid of an auxiliary agent. The agent is then water washed from the distillate and recovered in a second distillation process. Top: owsheet, bottom: triangular phase diagram for the distillation process.


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174

Fig. 5. Separation of Bz/Chx by extractive distillation. An auxiliary agent is added to shift the vaporliquid equilibrium curve. Cyclohexane is distilled overhead. Benzene/agent is treated in a secondary distillation process.

5. PV separation of Bz/Chx 5.1. PV processes PV is derived from combining permeation and evaporation, the two mechanisms involved in the process. In a PV process, components of a liquid feed permeate through a membrane and evaporate into the downstream at different rates. The fact that feed components undergo phase change makes PV unique among membrane processes. The driving force behind the process is the difference in chemical activity of components in the feed and the permeate. The separation occurs because of the different rates of sorption and diffusion of the feed components through the membrane. The main advantage of PV over distillation is that PV is independent of relative volatilities of components, and therefore, it is not limited by the vaporliquid equilibrium. This characteristic makes PV an attractive alternative to separate azeotropic and close-boiling point mixtures. PV processes cover a wide range of applications including organic dehydration, organic removal from aqueous solutions, and organic/organic separation. Separation performance of a membrane is characterized by two parameters: ux and selectivity. Flux is dened as the rate of permeation of feed components through unit area of membrane at unit time. Two commonly used ux units in PV processes are kg/m2 h, and g/cm2 s.

Two types of selectivities are dened for a membrane: ideal selectivity and actual selectivity. The ideal selectivity is simply the ratio of uxes of pure substances through the membrane. The actual selectivity of a membrane in a binary system is dened as the ratio of concentrations of components in the permeate to that in the feed: = ya /yb xa /xb

where is selectivity (separation factor), x and y are concentrations of components in feed and permeate, respectively, and a and b subscripts denote the two components to be separated. Selectivity of a membrane is strongly inuenced by two factors: afnity of the membrane towards one (or more) component(s) of the feed, and ease of diffusion of the permeating molecules through the membrane matrix. The overall selectivity of a membrane is dened as the product of sorption selectivity, s , and diffusion selectivity, D : = S D Either or both of these selectivities contribute to the preferential permeation of feed component(s) through a given membrane. However, in most cases, selectivity of a membrane is mainly governed by the sorption component of the selectivity. Studies by Aptel et al. [28], Nel [29] and Binning et al. [30] suggested that the unswollen fraction of the skin layer in a PV membrane controls the permselectivity of the membrane.

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174


Therefore, the choice of proper membrane material is a crucial factor for a specic separation. On one hand, a polymer with great afnity towards one component of the feed is preferred because it leads to higher selectivity. On the other hand, if this afnity exceeds a certain level, the membrane gets swollen by the compound and loses its integrity and therefore its selectivity. Consequently, in the preparation of a proper membrane, it is important to suppress or control its degree of swelling. In addition to the specic characteristics of the feed components and the membrane, the operating parameters inuence the overall performance of a PV process. These parameters include feed temperature, feed concentration, and downstream pressure. It is generally believed that feed pressure has insignicant effect on the permeability and selectivity of PV membranes. 5.2. Parameters affecting the separation of Bz/Chx by PV The effects of both sorption selectivity and diffusion selectivity of benzene and cyclohexane have been studied extensively. Table 6 [31,32] compares the molar volumes, collision diameters, and solubility parameters of benzene and cyclohexane. Benzenes smaller size, smaller collision diameter, and planar shape enhance its diffusivity. Many researchers have realized the domination of diffusion selectivity or sorption selectivity in different membranes. Huang and Lin [33] reported that the permeation and separation of Bz/Chx are signicantly inuenced by their molecular sizes. Also, Inui et al. [34] showed that the membrane selectivity decreased with increasing molecular size of aromatic hydrocarbons in an aromatic/cyclohexane separation study. The later group, in a series of separate studies [35,36], emphasized that the diffusion se-

lectivity dominates the overall selectivity of the membranes under study. However, in most cases, the experimental evidences indicate that the separation of benzene and cyclohexane is mainly governed by sorption selectivity due to the chemical interaction between benzene molecule and membrane [32,3741]. The domination of sorption selectivity is due to the afnity between the double bonds of benzene molecule and the polar groups of the polymeric membrane. That is, benzene has electrons, which show stronger afnity to polar molecules. Therefore, a polymer possessing polar groups facilitates the permeation of benzene through the membrane. The solubility parameters, reported in Table 6, indicate that the hydrogen bonding component ( H ) of benzene solubility is stronger than that of cyclohexane. It is expected that the high hydrogen bonding interacts with free polar groups in a membrane. An interesting study to support the above explanation was done by Yamasaki et al. [39]. They compared the retention time of benzene, n-hexane, and cyclohexane through a column lled with PVA polymer, and observed that the retention time of benzene was considerably higher than that for the other two components. They concluded that the difference in retention time was due to the stronger interaction between benzene and PVA. Tanihara et al. [32] compared the overall selectivity of a membrane with its diffusivity selectivity component to show that the latter has an insignicant effect on the separation process. Sun and Ruckenstein [41] also showed that the swelling of different membranes was mainly due to the sorption of benzene than cyclohexane. Furthermore, the swelling effect of benzene was noticed in a number of studies [38,42,43]. Suzuki and Onozato [42] compared the ideal selectivity and actual selectivity of benzene and cyclohexane through poly( -methyl l-glutamate) (PMLG) membrane and

Table 6 Molar volume, collision diameter, and solubility parameter of benzene and cyclohexane [26,27]a Solvent Molar volume (cm3 /mol) Collision diameter (nm) Solubility parameter ((MPa)1/2 ) D Benzene Cyclohexane 89.4 108.7 0.526 0.606 18.4 16.8 P 0 0 H 2.0 0.2 18.6 16.8

a : Hansen solubility parameter, : dispersive forces contribution component, : polar component, : hydrogen bonding component. D P H 2 2 2 2 = D + P + H .


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174

realized that the selectivity reduced from an ideal value of 370 to between 5 and 12 when a mixture of benzene and cyclohexane was used as the feed. They attributed this observation to the swelling of the membrane by benzene, which resulted in enhanced permeation of cyclohexane through the membrane. In another study, Ray et al. [44] measured ux and selectivity of 5% benzene in cyclohexane mixture through three copolymers of polyacrylonitrile with styrene, methyl methacrylate (MMA), and vinyl acetate. They observed that the membranes selectivity decreased with increasing content of active monomer in the copolymer. The selectivity loss was attributed to the plasticization effect of benzene in the monomer moieties of the membrane. The above evidences indicate that, because of small difference between diffusion of benzene and cyclohexane, the key to the choice of a proper membrane is the benzene/membrane chemical afnity. A proper membrane for Bz/Chx separation must possess both polar groups to facilitate the sorption of benzene and a rigid molecular structure resistive to swelling to retain the membranes integrity. For example, it is evidenced that the solubility of benzene in water is one order of magnitude larger than that of cyclohexane [45]. Therefore, benzene selectivity may improve by using membranes having hydrophilic groups in their polymer structure. Such separation studies using hydrophilic membranes were reported by Yamasaki et al. [39].

5.3. Types of membranes used for Bz/Chx separation The rst attempts to separate benzene and cyclohexane by membranes were not successful [33,46] because of low selectivity of membranes. Kucharski and Stelmaszek [46] slightly enhanced the selectivity of polypropylene towards benzene by modifying the orientation of polymers. Huang and Lin [33] proposed that the polymers functional groups are major contributors to the benzene selectivity of membranes. A successful attempt was performed by McCandless et al. [47] who obtained a selectivity of 20 using polyvinylidine uoride membrane. However, they were not successful in increasing the permeability by adding solvents to the feed mixture. These preliminary studies lighted the road to many membrane preparation techniques to obtain reasonable

selectivity and ux. In the recent years, polymer alloy membranes became popular for PV separation of Bz/ Chx. The concept of alloy membranes refers to the preparation of membranes by mixing a benzene-soluble polymer with a benzene-non-soluble one. The insoluble polymer forms the backbone of the polymeric membrane, while the soluble one provides afnity for benzene. Selectivity of these membranes was enhanced by techniques such as cross-linking, graftpolymerization, microphase separation, concentrated emulsion polymerization, and copolymerization. Cabasso et al. [43,48] tried out alloy membranes of polyphosphonates (PPN) and acetyl cellulose (AC). They succeeded in preparing membranes with selectivities between 12 and 40, and uxes between 800 and 1000 ml/m2 h. More recently, methacrylate family of polymers has become popular because of its favorable characteristics towards Bz/Chx separation. Examples of this type of polymers include MMA and ethyl methacrylate (EMA). This family has been used as common starting materials for many membrane alloys. Terada et al. [49] alloyed methyl acrylate (MA) with 2-hydroxyethyl methacrylate (HEMA). Yamaguchi et al. [50] tried graft copolymerization of MA/high-density polyethylene (HDPE). Inui and his coworkers have focused on polymethyl methacrylate (PMMA) in a number of studies. They prepared PMMA alloyed membranes with polyethyl methacrylate (PEMA) [34], dimethyl acrylamidemethyl methacrylate (DMMA) [35], and polymethyl methacrylatemethacrylic acid (MMAMAA) [36]. Finally, Ray and coworkers prepared polyacrylonitril (PAN)PMMA alloy membrane for Bz/Chx separation [44]. All these membranes showed moderate to high selectivity and ux. Also, a non-alloy use of polymethacrylic acid (PMA) is given by Sakohara et al. [51] who used PMA in porous silica membrane. A technique to prepare composite membrane was suggested by the Ruckenstein group by late 80s and is known as concentrated emulsion polymerization [41,55,56]. In this method, the membrane is prepared using two incompatible polymers. One polymer is soluble in one feed component, e.g., benzene. The other is insoluble and preserves the integrity of the membrane. The soluble polymer is emulsied and dispersed in the insoluble one. Benzene preferentially swells this polymer and permeates through the membrane, while

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174


cyclohexane permeation is retarded. Sun and Ruckenstein report a selectivity of around 25 at a ux of 2 kg m/m2 h with a feed concentration of 10 wt.% benzene in cyclohexane at 20 C [41]. Other examples of polymer alloy membranes include polyethyleneglycol (PEG)/Nylon 6 [52], and PEMA/Nylon 6 [53], and grafted polyglycidyl methacrylate (GMA) [38]. In addition, some researchers examined polymer mixing and hybrid polymers. An example is the study done by Kusakabe et al. [54] who prepared polyurethane (PU)/silica and PU/tetraethylorthosilicate (TEOS) to separate Bz/Chx. They observed a ux of 0.1 kg/m2 h and a selectivity of 19 for a 50% mixture of benzene in cyclohexane. Besides alloy membranes, a number of studies examined other types of membranes for this separation. For example, Inui et al. used benzoylchitosan [37] at different degrees of benzoylation. The same group also tried nematic and smectic liquid crystalline polymers (n-LCP and s-LCP) [40,57]. Also, Bryant et al. explored the possibility of facilitated transport with Ag(I) ions in a polyvinyl alcohol (PVA) holder [58]. Finally, Darkow et al. [59] synthesized polyacrylonitril-co-butadien-co-styrene-co-diaryltetrazolyl (ABSV) membrane, a new type of polymer containing a photosensitive moiety. The functionalization and cross-linking were initiated by photoirradiation. The PV results showed that although the photochemical treatment enhanced the permselectivity of the membranes, the separation factor was still low ( = 2.46). 5.4. Flux and selectivity of membranes Table 7 compares the ux and selectivity of a number of membranes for Bz/Chx separation. All ux values in Table 7 were converted to kg m/m2 h. These values are also presented graphically in Fig. 6. It can be observed in Table 7 that both ux and selectivity vary in a wide range. Flux deviates between as low as 0.03 kg m/m2 h and as high as 64.6 kg m/m2 h. Similarly, selectivity starts from almost no separation, e.g., 1.3, to perfect separation, e.g., . However, a majority of the experimental uxes lay in the vicinity of 110 kg m/m2 h, at selectivities of 520. These ux and selectivity values are comparable to those obtained in most industrial applications of PV processes.

Most membranes used in the laboratory experiments were of homogeneous type. Only few studies were performed using asymmetric membranes. Three examples are shown in Table 7 for comparison [34,39]. In one case, Yamasaki et al. [39] compared the performances of homogeneous and asymmetric PVA membranes for Bz/Chx separation and observed a selectivity of 10 for an asymmetric membrane compared to 3 for a homogeneous one. The authors explained this unusual phenomenon by the density difference in the skin layer of an asymmetric membrane and a homogeneous one: the asymmetric membrane probably had denser polymer packing, resulting in smaller free volume for the diffusion of permeating molecules. Therefore, sorption selectivity dominated the overall selectivity, and because benzene had greater afnity towards PVA, the overall selectivity increased with increasing polymer density. 5.5. Effects of operating parameters Few researchers studied the effects of temperature on the permselectivity of membranes in Bz/Chx separation [36,41,4648,65]. All the experimental evidences conrm an increase in permeability and decrease in selectivity of membranes with increasing temperature. Fig. 7 represents the general trend in the effects of temperature on the ux and selectivity of PV membranes. Acharya et al. [65] and Inui et al. [36] show that the permeability of PV membranes decrease linearly with reciprocal of temperature following an Arrhenius type relation of the form Qi = Q0 exp i Ep RT

where Q0 is a constant, Ep is activation energy for peri meation, R is universal gas constant, and T is absolute temperature. The effect of temperature on the selectivity of membrane was explained by Sun and Ruckenstein [41] by considering two parallel effects of temperature on the membrane. One is increasing polymer chain mobility and facilitating the diffusion of both components. The other is weakening the interaction between benzene molecule and the membrane lowering the sorption of benzene. In addition, the Cabasso group observed irregularities in the sorption behavior of AC/PPN membrane


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174

Table 7 Results on the separation of Bz/Chx mixtures by PV Membranea Modied CE PP (oriented) PP (double oriented) LDPE PVDF PVDF PE MA-g-HEMA PMG CA CA/PPN PEU (asymmetric) PSP/CA CA/PPOBrP HDPE-g-MA PVA/PAAm BP3MPDIPD BP3MPD/PD BPPEO PAS Modied Nylon 6 Modied Nylon 6 DSDADDBT DSDADDBT PUTEOS PGMA grafted PS/PAA PEMAEGDM PMMAEGDM PVA (asymmetric) Benzene in feed (wt.%) 53 55 55 50 53 53 10 50 50 50 50 50 55 51 50 50 50 50 60 50 50 50 60 60 50 60 50 10 10 25 Temperature ( C) 80 25 25 25 60 56 35 25 25 77.8 77.8 30 78 53.5 25 25 70 70 2570 3050 50 50 78 78 50 4070 20 40 40 50 Q1 (kg m/m2 h) 50.3 54.4 64.6 10.8 1.6 1.5 30.0 7.4 3 0.34 13.6 1.5b 0.8 4.0 10 2.3 2.7 0.4 2.1 2.6 0.057 0.16 0.93 2.4 0.65 8.7 48.4 8.7 55.1 0.002b 5.2 1.3 2.4 1.6 6.2 5.4 2.5 7 19 40 3.8 13.3 9.7 7 11.9 14 27 9.1 22.5 4.5 >104 32 20 19 22 9.6 6.7 3.9 10 Reference Martin 1964 Kucharski 1967 Kucharski 1967 Huang 1968 McCandless 1973 McCandless 1974 Rautenbach 1980 Terada 1982 Suzuki 1982 Cabasso 1983 Cabasso 1983 Enneking 1993 N el 1985 e Acharya 1988 Yamaguchi 1991 Park 1994 Tanihara 1994 Tanihara 1994 Tanihara 1995 Ray 1997 Yoshikawa 1997 Yoshikawa 1997 Hao 1997 Hao 1997 Kusakabe 1998 Wang 1998 Sun 1995 Inui 1997 Inui 1997 Yamasaki 1997 Reference # [60] [46] [46] [33] [61] [47] [62] [49] [42] [43] [43] [63] [64] [65] [50] [66] [32] [32] [67] [44] [52] [52] [68] [68] [54] [38] [41] [34] [34] [39]

a 3MPD: 2,4,6-trimethyl-1,3-phenylenediamine; BP: 3,3 ,4,4 -biphenyltetracarboxylic dianhydride; CA: cellulose acetate; CE: cellulose ester; DDBT: dimethyl-3,7-diaminibenzothiophene-5,5 -dioxide; DSDA: 3,3 ,4,4 -diphenilsulpone tetracarboxylic dianhydride; EGDM: ethlyne glycol dimethacrylate; HDPE-g-MA: high density polyethylene graft methylacrylate; LDPE: low density polyethylene; MA-g-HEMA: polymethylacrylate graft 2-hydroxylethyl methacrylate; PAA: polyacrylic acid; PAAm: polyallyl amine; PAS: polyacrylonitril-co-styrene; PD: 1,3-phenylenediamine; PE: polyethylene; PEMA: polyethyl methacrylate; PEO: polyethylene oxide; PEU: polyesterurethane; PGMA: polyglycidyl methacrylate; PMG: poly -methyl l-glutamate; PMMA: polymethyl methacrylate; PP: polypropylene; PPN: polystyrene diethylphosphonate; PPOBrP: polybromophenylene oxide dimethylphosphonate ester; PS: polystyrene; PSP: copolymer vinylidin chloride and styrene diethylphosphonate; PU: polyurethane; PVA: polyvinyl alcohol; PVDF: polyvinylidine uoride; TEOS: tetraethylorthosilicate. b Unit for asymmetric membrane ux is kg/m2 h.

[48], which was later conrmed by Acharya et al. [65] through PV experiments. Cabasso realized that rate of sorption of benzene depended on the thermal history of the membrane. That is, the sorption increased with decreasing temperature. Reversing the direction also reversed the sorption behavior. However, when the temperature started from low values and then increased, the sorption increased again, until the starting temperature of the rst experiment was reached. At this point, reversing the temperature direction did

not reverse the sorption behavior, but caused it to increase along the same curve as in the rst experiment. Also, Acharya et al. [65] realized that ux was higher during cooling than when heating the feed within similar temperature limits. Their observations can be explained by the temperature history effect proposed by Cabasso. Permeability and selectivity of membranes are also functions of feed concentration [32,34,35,37,43,44,47, 63,65,67,69]. Fig. 8 shows that permeability increases

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174


and selectivity decreases sharply, with increasing benzene content of the feed. The decline in the selectivity is explained by the plasticization effect of benzene on the membrane. With increasing benzene content, the membrane is swollen and the relaxed polymer chains allows for enhanced permeation of cyclohexane. As evidenced in Fig. 8, the permeability of cyclohexane was enhanced more profoundly compared to benzene when the benzene concentration of the feed increased.

6. Economy Membrane processes have not been used in the industry for Bz/Chx separation yet. Therefore, not much economical evaluation and comparison with other separation technologies are available in the open literature. Acharya et al. [65] compared two membrane processes, namely, PV and perstraction to suggest that the earlier is advantageous because of smaller membrane area requirements. Rautenbach and Albrecht [27,70] did a thorough economical comparison between a conventional extractive distillation process alone, a PV system alone, and a hybrid system consisting of the two processes. The hybrid system was designed to separate a 1 : 1 Bz/Chx mixture to 99.5 mol% benzene and 99.2 mol% cyclohexane. The extractive distillation used furfural as carrier in a two-stage distillation process. In the rst stage, furfural was added to break the Bz/Chx azeotrope. The product from this stage was a cyclohexane rich stream and a mixture of benzene/furfural. The benzene impurity in the cyclohexane rich stream was removed using a PV system. Finally, The benzene/furfural mixture was distilled in the second stage to recover furfural and produce highly pure benzene. The authors analyses indicated that a hybrid system is more economical than either of the two processes alone. That is especially true when high purity products are required. The cost of production using an extractive distillation process is mainly determined by the cost of heating steam (58%), followed by the investment cost on the sieve tray column (19%), and evaporator/heat exchanger (14%). On the other hand, similar cost using a PV system is governed by the membrane replacement cost (51%), followed

Fig. 6. Comparison of performances of PV membranes for the separation of Bz/Chx at 1 : 1 feed composition and 50 C.

Fig. 7. Effects of temperature on the permeability and selectivity of membrane in Bz/Chx separation.

Fig. 8. Effects of benzene concentration in the feed on the permeability and selectivity of membrane in Bz/Chx separation.


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174 [14] M. Sittig, Aromatic Hydrocarbons, Manufacture and Technology, Noyes Data Coro., NJ, 1976. [15] [16] [17] D.V. Law, P.W. Tamm, C.M. Detz, Selective catalytic process for conversion of light naphtha to aromatics, paper presented at the AIChE Spring National Meeting, Houston, TX, 1987. [18] M. Roberts, Saudi decree encourages MTBE, Chevron aromatics plant, in: Chem. Week, Vol. 152, 7, 14 and 24 February, 1993. [19] R.V. Jasra, S.G.T. Bhat, Adsorptive bulk separations by zeolite molecular sieves, Sep. Sci. Technol. 23 (10/11) (1988) 945 989. [20] S. Matar, L.F. Hatch, Chemistry of Petrochemical Processes, Gulf Publishing Company, Houston, 1994, pp. 274276. [21] R.F. Goldstein, A.L. Waddams, The Petroleum Chemicals Industry, 3rd Edition, E. & F.N. Spon, London, 1967. [22] L.S. Kasel, Hydrogenation of benzene to cyclohexane, US Patent 2,755,317 (1956). [23] S. Field, M.H. Dolson, Hydrocarbon Process. 46 (1967) 169. [24] O.D. Vu, R. Odello, Manufacture of naphtenic hydrocarbons by hydrogenation of the corresponding aromatic hydrocarbons, Institut Franais de Ptrole, US Patent 3,597,489 (1971). [25] J.D. Seader, Distillation, in: R.H. Perry, D.W. Green, J.O. Maloney (Eds.), Perrys Chemical Engineering Handbook, McGraw-Hill, New York, 1984. [26] J.G. Stichlmair, J.R. Fair, Distillation, Principles and Practice, Wiley/VCH, NY, 1998, pp. 246249. [27] R. Rautenbach, R. Albrecht, Membrane Processes, Wiley, New York, 1989. [28] P. Aptel, J. Cuny, J. Jozefonvicz, G. Morel, J. Neel, Liquid transport through membranes prepared by grafting of polar monomers onto poly(tetrauoroethylene) lms III. Steady-state distribution in membrane during pervaporation, J. Appl. Polym. Sci. 18 (1974) 365. [29] J. Nel, in: R.Y.M. Huang (Ed.), Pervaporation Membrane Separation Process, 1991, Chap. 1, p. 1. [30] R.C. Binning, R.J. Lee, J.F. Jennings, E.C. Martin, Separation of liquid mixtures by permeation. Liquid phase permeation through thin plastic lms has commercial potential for separating azeotropes and other organic mixtures, Ind. Eng. Chem. 53 (1961) 47. [31] J. Brandrup, E.H. Immergut, Polymer Handbook, 3rd Edition, Wiley/Interscience, New York, 1989, Chap. 7, p. 540. [32] N. Tanihara, K. Tanaka, H. Kita, K. Okamoto, Pervaporation of organic liquid mixtures through membranes of polyimides containing methyl-substituted phenylenediamine moieties, J. Membr. Sci. 95 (1994) 161169. [33] R.Y.M. Huang, V.J.C. Lin, Separation of liquid mixtures by using polymer membranes I. Permeation of binary organic liquid mixtures through polyethylene, J. Appl. Polym. Sci. 12 (1968) 2615. [34] K. Inui, H. Okumura, T. Miyata, T. Uragami, Permeation and separation of benzene/cyclohexane mixtures through cross-linked poly(alkyl methacrylate) membranes, J. Membr. Sci. 132 (1997) 193202.

by steam (17%) and cooling water (15%). Also, the overall cost of production using a PV system is more than twice as much as that using an extractive distillation. A comparison of the numbers given above show that the two processes complement each other in a hybrid system. The production cost using a hybrid system is 74.51 DM/t (10.55 DM/t for PV + 63.96 DM/t for distillation). Which is 20% less than that using an extractive distillation alone (93.65 DM/t). This saving comes from a reduced cost of membrane replacement on one hand, and reduced energy consumption compared to the conventional process on the other.

7. Final remarks The attractiveness of PV systems in Bz/Chx separation is evidenced by the increasing number of papers appearing in the scientic journals on the subject. Many studies have proved the technical feasibility of PV processes in Bz/Chx separation. However, lack of a more recent economical evaluation can be felt in this eld. Such study is an essential factor to encourage the industry to invest in researches focusing on the pilot plant and large scale studies of the process. References
[1] E.G. Handcock (Ed.), Benzene and Its Industrial Derivatives, Wiley, New York, 1975. [2] Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Edition, Wiley, New York, 1978. [3] K. Weissermel, H.-J. Arpe, Industrial Organic Chemistry, 2nd Edition, VCH Publishers, Weinheim, New York, 1993. [4] M.L. Campbell, in: W. Gerhartz (Ed.), Ullmanns Encyclopedia of Industrial Chemistry, 5th Edition, Vol. A8, VCH Publishers, Weinheim, New York, 1987. [5] Hydrocarbon Process. 55 (1976) 171. [6] Oil Gas J. 66 (1968) 140. [7] Oil Gas J. 69 (1971) 45. [8] Oil Gas J. 69 (1971) 48. [9] Oil Gas J. 69 (1971) 58. [10] Oil Gas J. 70 (1972) 50. [11] Oil Gas J. 74 (1976) 48. [12] C.S. McCoy, P. Munk, The Rheniformer process, Chem. Eng. Prog. 67 (10) (1971) 7882. [13] J.J. McKetta (Ed.), Advances in Petroleum Chemistry and Rening, Vols. 110, Wiley, New York, 1965.

J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174 [35] K. Inui, H. Okumura, T. Miyata, T. Uragami, Characteristics of permeation and separation of dimethyl acrylamidemethyl methacrylate random and graft copolymer membranes for a benzene/cyclohexane mixture, Polym. Bull. 39 (1997) 733 740. [36] K. Inui, T. Noguchi, T. Miyata, T. Uragami, Pervaporation characteristics of methyl methacrylatemethacrylic acid copolymer membranes ionically cross-linked with methal ions for a benzene/cyclohexane mixture, J. Appl. Polym. Sci. 71 (1999) 233241. [37] K. Inui, K. Tsukamoto, T. Miyata, T. Uragami, Permeation and separation of a benzene/cyclohexane mixture through benzoylchitosan membranes, J. Membr. Sci. 138 (1998) 67 75. [38] H. Wang, X. Lin, K. Tanaka, H. Kita, K. Okamoto, Preparation of plasma-grafted polymer membranes and their morphology and pervaporation properties toward benzene/cyclohexane mixtures, J. Polym. Sci. A: Polym. Chem. 36 (1998) 22472259. [39] A. Yamasaki, T. Shinbo, K. Mizoguchi, Pervaporation of benzene/cyclohexane and benzene/n-hexane mixtures through PVA membrane, J. Appl. Polym. Sci. 64 (1997) 10611065. [40] K. Inui, T. Miyata, T. Uragami, Permeation and separation of benzene/cyclohexane mixtures through liquid crystalline polymer membranes, J. Polym. Sci., Polym. Phys. 35 (1997) 699. [41] F. Sun, E. Ruckenstein, Sorption and pervaporation of benzene/cyclohexane mixtures through composite membranes prepared via concentrated emulsion polymerization, J. Membr. Sci. 99 (1995) 273284. [42] F. Suzuki, K. Onozato, Pervaporation of benzenecyclohexane mixture by poy( -methyl l-glutamate) membrane and synergetic effect of their mixture on diffusion rate, J. Appl. Polym. Sci. 27 (1982) 42294238. [43] I. Cabasso, Organic liquid mixtures separation by permselective polymer membrane I. Selection and characteristics of dense isotropic membrane employed in the pervaporation process, Ind. Eng. Chem. Prod. Res. Dev. 33 (1983) 313319. [44] S.K. Ray, S.B. Sawant, J.B. Joshi, V.G. Pangarkar, Development of new synthetic membranes for separation of benzenecyclohexane mixtures by pervaporation: a solubility parameter approach, Ind. Eng. Chem. Res. 36 (1997) 5265 5276. [45] Y. Marcus, Introduction to Liquid State Chemistry, Wiley, NY, 1977. [46] M. Kucharski, J. Stelmaszek, Separation of liquid mixture by permeation, Int. Chem. Eng. 7 (1967) 618622. [47] F.P. McCandless, D.P. Alzeheimer, R.B. Hartman, Solvent membrane separation of benzene and cyclohexane, Ind. Eng. Chem. Process Des. Dev. 13 (1974) 310312. [48] I. Cabasso, J.J. Grodzinski, D. Vofsi, Polymeric alloys of polyphosphonates and acetyl cellulose I. Sorption and diffusion of benzene and cyclohexane, J. Appl. Polym. Sci. 18 (1974) 21172136. [49] J. Terada, T. Hohjoh, S. Yoshimasu, M. Ikemi, I. Shinohara, Separation of benzenecyclohexane azeotropic mixture



[51] [52]








[60] [61]

[62] [63]

[64] [65]


through polymeric membranes with microphase separated structures, Polym. J. 14 (5) (1982) 347353. T. Yamaguchi, S. Nakao, S. Kimura, Plasma-graft lling polymerization: preparation of a new type of pervaporation membrane for organic liquid mixtures, Macromolecules 24 (1991) 55225527. S. Sakohara, T. Koshi, M. Asaeda, Kobunshi Ronbunshu 52 (1995) 155. M. Yoshikawa, S. Takeuchi, T. Kitao, Pervaporation separation of benzene/cyclohexane mixtures with nylon 6-graft-poly(oxyethylene) membranes, Die Angewandte Makromolular Chemie 245 (1997) 193202. M. Yoshikawa, T. Kitao, Specialty polymer membranes VI. Pervaporation separation of benzene/cyclohexane mixtures through nylon 6-graft-poly(ethyl methacrylate) membranes, Eur. Polym. J. 33 (1997) 2531. K. Kusakabe, S. Yoneshige, S. Morooka, Separation of benzene/cyclohexane mixtures using polyurethanesilica hybrid membranes, J. Membr. Sci. 149 (1998) 2937. E. Ruckenstein, Emulsion pathways to composite polymeric membranes for separation processes, Colloid Polym. Sci. 267 (1989) 792797. J.S. Park, E. Ruckenstein, Selective permeation through hydrophobichydrophilic membranes, J. Appl. Polym. Sci. 38 (1989) 453461. K. Inui, K. Okazaki, T. Miyata, T. Uragami, Effect of mesogenic groups on characteristics of permeation and separation for benzene/cyclohexane mixtures of side-chain liquid-crystalline polymer membranes, J. Membr. Sci. 143 (1998) 93104. D.L. Bryant, R.D. Noble, C.A. Kovac, Facilitated transport of benzene and cyclohexane with poly(vinyl alcohol)AgNO3 membranes, J. Membr. Sci. 127 (1997) 161170. R. Darkow, M. Yoshikawa, T. Kitao, G. Tomaschewski, J. Schellenberg, Photomodication of a poly(acrylonitrileco-butadiene-co-styrene) containing diaryltetrazolyl groups, J. Polym. Sci.: Polym. Chem. 32 (9) (1994) 16571664. E.C. Martin, R.C. Binning, L.M. Adams, R.J. Lee, US Patent 3,140,256 (1964). F.P. McCandless, Separation of aromatics and naphtalenes by permeation through modied vinylidene uoride lms, Ind. Eng. Chem. Proc. Des. Dev. 12 (1973) 354359. R. Rautenbach, R. Albrecht, Separation of organic binary mixtures by pervaporation, J. Membr. Sci. 7 (1980) 203223. L. Enneking, W. Stephan, A. Heintz, Sorption and diffusivity measurements of cyclohexanebenzene and cyclohexanetoluene mixtures in polyurethane membranes. Model calculations of the pervaporation process, Ber. Bunsenges. Phys. Chem. 97 (1993) 912. J. Nel, P. Aptel, R. Clement, Basic aspects of pervaporation, Desalination 53 (1985) 297326. H.R. Acharya, S.A. Stern, Z.Z. Liu, I. Cabasso, Separation of liquid benzene/cyclohexane mixtures by perstraction and pervaporation, J. Membr. Sci. 37 (1988) 205232. C.K. Park, B.K. Oh, M.J. Choi, Y.M. Lee, Separation of benzene/cyclohexane by pervaporation through chelate poly(vinyl alcohol)/poly(allyl amine) blend membrane, Polym. Bull. 33 (1994) 591.


J.P. Garcia Villaluenga, A. Tabe-Mohammadi / Journal of Membrane Science 169 (2000) 159174 [69] L. Enneking, A. Heintz, R.N. Lichtenthaler, Sorption equilibria of the ternary mixture benzene/cyclohexene/ cyclohexane in polyurethane- and PEBA-membrane polymers, J. Membr. Sci. 115 (2) (1996) 161170. [70] R. Rautenbach, R. Albrecht, The separation potential of pervaporation Part 2. Process design and economics, J. Membr. Sci. 25 (1985) 2554.

[67] N. Tanihara, N. Umeo, T. Kawabata, K. Tanaka, H. Kita, K. Okamoto, Pervaporation of organic liquid mixtures through poly(ether imide) segmented copolymer membranes, J. Membr. Sci. 104 (1995) 181192. [68] J. Hao, K. Tanaka, H. Kita, K. Okamoto, The pervaporation properties of sulfonyl-containing polyimide membranes to aromatic/aliphatic hydrocarbon mixtures, J. Membr. Sci. 132 (1997) 97108.