Journal of Membrane Science 313 (2008) 68–74
Study on the influence of the free volume of hybrid membrane on
pervaporation performance by positron annihilation spectroscopy
Chi-Lan Li
b, Shu-Hsien Huang
a, Wei-Song Hung
a, Se-Tsung Kao
a,c,
Da-Ming Wang
a,d, Y.C. Jean
a,e, Kueir-Rarn Lee
a,∗, Juin-Yih Lai
aaR&D Center for Membrane Technology, Department of Chemical Engineering, Chung Yuan University, Chung Li 32023, Taiwan bDepartment of Chemical and Material Engineering, Nanya Institute of Technology, Chung Li 32034, Taiwan
cInstitute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan dDepartment of Chemical Engineering, National Taiwan University, Taipei 10617, Taiwan eDepartment of Chemistry, University of Missouri Kansas City, Kansas City, MO 64110, USA
Received 14 November 2007; received in revised form 17 December 2007; accepted 21 December 2007 Available online 5 January 2008
Abstract
To improve the pervaporation separation performance, an organic–inorganic hybrid membranes were prepared by adding zeolite 13X into the polyimide (BAPP–BODA), which was synthesized using one-step polycondensation polymerization of 2,2-bis(4-(4-aminophenoxy)phenyl)propane (BAPP) with bicyclo[2.2.2]oct-7-ene-2,3,5,6-tetracarboxylic dianhydride (BODA). The influence of the zeolite 13X content on the free volume and pervaporation performance of hybrid membranes were systematically analyzed by positron annihilation spectroscopy (PAS). The trend of ortho-positronium (o-Ps) lifetime and free volume holes changes consistently well with the permeation rate variation of aqueous isopropanol mixtures dehydration. The BAPP–BODA/13X hybrid membrane morphology was characterized by AFM and SEM. Compared with the BAPP–BODA membrane, the BAPP–BODA/13X hybrid membrane had good pervaporation performance for aqueous alcohol solution dehydration. The sorption selectivity dominates the behavior of pervaporation while adding zeolite 13X into the polyimide matrix.
© 2008 Elsevier B.V. All rights reserved.
Keywords: Polyimide; Zeolite; Positron annihilation spectroscopy; Pervaporation; Alcohol–water mixture
1. Introduction
Membrane pervaporation has become one of the most promis-ing candidates for energy-savpromis-ing, low-cost separation processes, especially for mixtures with close volatility, and mixtures with thermal or chemical sensitivity [1]. Applications can be found in (a) organic/water (azeotropic) mixture dehydration, (b) removal/recovery of organic compounds from water, and (c) separation of organic/organic mixtures. Pervaporation pro-cess is fundamentally based on differences in permeation speed in various substances or classes of substances through a given membrane. The transport mechanisms of various classes of sub-stances through a pervaporation membrane can be described using a sorption–diffusion model[2]. According to this model, the transport of a given substance through a membrane takes
∗Corresponding author. Tel.: +886 3 2654190; fax: +886 3 2654198.
E-mail address:krlee@cycu.edu.tw(K.-R. Lee).
place within three consecutive processes: (1) sorption of the given substance onto the membrane, (2) transport through the membrane, and (3) desorption/evaporation from the membrane. The driving force for pervaporation is the pressure gradient that exists on opposite side of the membrane.
Reminiscent of the pervaporation performance of some glassy polymers, most of these materials exhibit a fairly high separation factor, albeit low permeation flux when separating aqueous alcohol mixtures. In an effort to increase the flux and/or improve the selectivity, many studies have incorporated metal complexes and zeolites in the membranes [3–6]. The zeolite-incorporated polymer membranes have received much attention recently in gas and pervaporation separation studies. The incorporation of zeolite or porous fillers in dense mem-branes can improve the separation performance [4–8] due to combined molecular sieving action, selective adsorption and difference in diffusion rates. In addition, zeolites have high mechanical strength, good thermal and chemical stability, and thus, these membranes can be used over the wide range of 0376-7388/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
operating conditions when incorporated with these fillers. The zeolite 13X, Faujasite type, used in this study has a three-dimensional pore structure. The pore entrances are composed of six-membered rings. The zoelite 13X particle and pore sizes are 2.67m and 8.16 ˚A, respectively. The silicon/aluminum (Si/Al) ratio of zeolite 13X is 1.43. The Si/Al ratio indicates the hydrophilicity of zeolite [9,10]. In general, the hydrophilicity increases when the Si/Al ratio decreases. Thus, 13X can be considered hydrophilic, resulting from the lower Si/Al ratio.
Polyimides have excellent thermal stability, good chem-ical resistance and mechanchem-ical strength. They have a low permeation rate when they are used as a membrane mate-rial for pervaporation separation. Hydrophilic zeolite 13X was introduced into the polyimide BAPP–BODA, which was synthesized using one-step polycondensation polymer-ization of 2,2-bis(4-(4-aminophenoxy)phenyl)propane (BAPP) with bicyclo[2.2.2]oct-7-ene-2,3,5,6-tetracarboxylic dianhy-dride (BODA). This was expected to improve the membrane pervaporation performance through the molecular sieving effect and hydrophilicity of the zeolite. Moreover, the influ-ence of the zeolite 13X content on the free volume and pervaporation performance of hybrid membranes were sys-tematically analyzed by positron annihilation spectroscopy (PAS).
2. Experimental
2.1. Synthesis of polyimide
The polyimide BAPP–BODA was prepared by one-step poly-merization of the diamine, BAPP and the dianhydride, BODA. The synthetic route of the polyimide is illustrated in Fig. 1. A mixture of 30 mmol of diamine BAPP, 30 mmol of dianhy-dride BODA and 100 ml of m-cresol was heated with stirring at 100◦C for 1 h, 140◦C for 1 h, and 180◦C for 1 h. Then, the solvent was distilled at 230◦C, keeping the volume of the reac-tion mixture nearly constant by adding an appropriate volume of fresh m-cresol. The azeotropic removal of the water formed dur-ing the imidization was continued for 4 h. The polymer solution was poured into a large amount of methanol. The precipitate, polyimide BAPP–BODA, was collected by filtration, washed thoroughly with methanol, and dried at 70◦C under vacuum.
lite (13X/BAPP–BODA = 0.25–0.75) into the 10 wt% polymer solution (BAPP–BODA/dichloromethane). The following steps were the same with the preparation of the dense polyimide membranes. The resulting polyimide and zeolite 13X-filled polyimide membranes had a thickness of 48–60m.
2.3. Characterization of BAPP–BODA/13X hybrid membrane
The cross-sectional morphology of membranes was observed with scanning electron microscope (SEM). The surface mor-phology of the membrane was observed with an AFM (Digital Instruments, DI-NS3a USA). In order to investigate the vari-ation of the free volume in the BAPP–BODA/13X hybrid membrane, the positron annihilation spectroscopy (PAS) exper-iments were performed. The positron annihilation spectroscopy of BAPP–BODA/13X hybrid membranes were determined by detecting the prompt ␥-rays (1.28 MeV) from the nuclear decay that accompanies the emission of a positron from the 22Na radioisotope and the subsequent annihilation ␥-rays (0.511 MeV). All of the PAL spectra were analyzed by a finite-term lifetime analysis method using the PATFIT program and a continuous lifetime distributions using the MELT program. It had been reported in our previous paper[11,12]. We employ the results of o-Ps lifetime to obtain the volume of free volume holes.
2.4. Pervaporation study
A traditional pervaporation process was used[13]. The effec-tive membrane area was 11.6 cm2 and the feed temperature studied was 25◦C. The permeation rate was determined by mea-suring the weight of the permeate. The compositions of the feed solutions, the permeates, and the solutions adsorbed in the membrane were measured by gas chromatography (GC China Chromatography 8700T). The separation factor (αwater/alcohol) was calculated as follows:
αwater/alcohol= Y
water/Yalcohol
Xwater/Xalcohol
where Xwater, Xalcoholand Ywater, Yalcoholare the weight fraction of water and alcohol in the feed and permeate, respectively.
Fig. 2. o-Ps lifetime distributions of the BAPP–BODA/13X hybrid membranes.
2.5. Sorption measurement
The membranes were immersed in several feed mixtures for 24 h at 25◦C. Then, they were erased between pieces of filter paper to remove excess solution and quickly placed in the left half of a twin-tube setup cooled with liquid nitrogen. The system was evacuated while the left tube was heated with hot water and the right tube was cooled in liquid nitrogen. GC determined the composition of the condensed liquid in the right tube. The separation factor of sorption (αsorp) was calculated as follows:
αsorp= Ywater/Yalcohol
Xwater/Xalcohol
where Xwater, Xalcoholand Ywater, Yalcoholare the weight fraction of water and alcohol in the feed and membrane, respectively. 2.6. Degree of swelling
The membranes were immersed in different feed mixtures for 24 h at 25◦C. The degree of swelling of the membrane was defined by the following equation:
degree of swelling= Ww− Wd
Wd × 100%
where Wdand Wwdenote the weight of dry and swollen mem-branes, respectively.
3. Results and discussion
3.1. Characterization of the BAPP–BODA/13X hybrid membrane
To understand the free volume size distribution of BAPP–BODA/13X hybrid membranes, the position lifetime dis-tributions and theτ3results are analyzed through the positron annihilation spectroscopy, as shown inFig. 2. We employ the results of o-Ps annihilation lifetime (τ3) to analyze the vol-ume of free volvol-ume hole. Fig. 2 shows the effect of zeolite 13X content of the BAPP–BODA/13X hybrid membranes on
Fig. 3. Effect of zeolite 13X content on free volume hole radius (R).
the o-Ps annihilation lifetime. It shows that the o-Ps annihila-tion lifetime increases with increasing the zeolite 13X content of the BAPP–BODA/13X hybrid membranes. This result indi-cates that the size of free volume hole is strongly composition dependent. On the other hand, the lifetime distribution curves shift to higher lifetime direction, as the zeolite 13X content increases. These phenomena might be due to the fact that the free volume radius of pure BAPP–BODA and zeolite 13X parti-cle is 2.85 and 4.08 ˚A, respectively. Thus, the free volume radius of the BAPP–BODA/13X hybrid membranes lie in between the pure BAPP–BODA and zeolite 13X particle, as shown in Fig. 3. Another interesting phenomena revealed in this positron annihilation spectroscopy is that the relative intensity of o-Ps annihilation lifetime (I3) decreases with increasing the zeolite 13X content (Fig. 4). The decrease in the relative intensity of o-Ps annihilation lifetime (I3) suggests that the adding zeolite 13X particle into the BAPP–BODA polyimide matrix reduces the number of free volume holes. However, the volume of free volume holes (with corresponding o-Ps annihilation life-time,τ3) increases by increasing the zeolite 13X content of the BAPP–BODA/13X hybrid membranes, also shown in Fig. 5.
Fig. 4. Effect of zeolite 13X content on ()() intensity (I3) and (䊉)()
Fig. 5. Effect of zeolite 13X content on volume of free volume hole (fv).
These phenomena imply that the interfacial voids formed, result-ing from the incompatibility between polyimide matrix and zeolite particles as well as the hole volume of zeolite particles. It also causes the distribution of free volume hole size shifts from smaller size to larger size in the BAPP–BODA/13X hybrid membranes (corresponding withFig. 2).
3.2. Effect of zeolite content on the pervaporation performance
The zeolite content effect on the BAPP–BODA/13X hybrid membranes in the pervaporation of a 90 wt% aqueous iso-propanol mixture is shown in Fig. 6. It shows that the permeation rate increases with increasing zeolite content in the BAPP–BODA/13X hybrid membrane. This phenomenon might be due to the molecular sieving effect and hydrophilicity of the zeolite 13X introduced into the polyimide membrane. From the analysis of positron annihilation spectroscopy, the volume of free volume holes increases by increasing the zeolite 13X con-tent of the BAPP–BODA/13X hybrid membranes (Fig. 5). These results correspond well with the results mentioned above, as
Fig. 6. Effect of zeolite 13X content on pervaporation separation of 90 wt% aqueous isopropanol mixture for the BAPP–BODA/13X hybrid membranes at 25◦C.
Fig. 7. AFM images of the polyimide membranes: (a) BAPP–BODA and (b) BAPP–BODA/13X (13X/BAPP–BODA = 0.5/1 (g/g)).
indicated inFig. 6. In general, a rougher membrane surface can improve the affinity to water, resulting in an increased perme-ation rate. To investigate the membrane surface morphology, AFM studies are conducted, as shown in Fig. 7. The surface roughness for the BAPP–BODA membrane is lower than that of the BAPP–BODA/13X hybrid membrane. As a result, the effective pervaporation membrane area increases with increas-ing membrane roughness. These results also correspond well with the results indicated in Fig. 6. Moreover, a cross-section image (Fig. 8) of the zeolite 13X-filled polyimide membrane with a zeolite/polyimide ratio of 0.5/1 by weight shows that many interfacial voids appear around the zeolite particles. The hard segment of the polyimide chains might cause the interfa-cial voids, that is, the polyimide chains might be too stiff to wrap around the zeolite well. Therefore, the permeation rate increases with increasing zeolite content in the BAPP–BODA/13X hybrid membrane. Fig. 6also shows that the water concentration in the permeate almost unchanges while the zeolite content in the BAPP–BODA/13X hybrid membrane increases. These phe-nomena might be because the degree of swelling decreases, as
Fig. 8. Cross-section morphology of the BAPP–BODA/13X hybrid membrane: 13X/BAPP–BODA = 0.5/1 (g/g).
shown in Fig. 9, by incorporating zeolite 13X into the poly-imide membrane to inhibit polymer chain mobility and increase resistance to the larger permeates, isopropanol, through the BAPP–BODA/13X hybrid membrane. In the past, the selectivity (corresponding with water concentration in the permeate) of per-vaporation separation process has been traditionally understood in terms of the interaction between the permeates and the poly-meric membranes. Recently, positron annihilation spectroscopy studies could provide a new approach in terms of size selectivity of free volume. Fig. 3also shows the results of the free vol-ume hole radii of hybrid membranes. It is very interesting that the free volume radius distribution of the BAPP–BODA/13X hybrid membranes lies between the kinetic radius of water (1.2 ˚A) and isopropanol (3.86 ˚A). Thus, the water concentra-tion in the permeate almost unchanges while the zeolite content in the BAPP–BODA/13X hybrid membrane increases.
3.3. Pervaporation of alcohol–water mixtures through BAPP–BODA/13X hybrid membrane
Table 1 shows the pervaporation performance of 90 wt% alcohol–water mixtures through the BAPP–BODA/13X hybrid membrane. The permeation rate of both polyimide membranes decreases, but the water concentration in the permeate increases with increasing alcohol molar volume. The alcohol molar vol-umes can explain these phenomena. The molar volvol-umes of methanol, ethanol, isopropanol, and n-butanol are 40.7, 58.68,
Fig. 9. Effect of zeolite 13X content on degree of swelling of 90 wt% aqueous isopropanol mixture for the BAPP–BODA/13X hybrid membranes at 25◦C.
75.14, and 94.88 ml/mol, respectively. The water concentration in the permeate is found to depend on the molecular length of linear alcohols. In addition, the permeation rate of n-butanol is higher than that of isopropanol. It is possibly because the steric hindrance of the former is lower than that of the latter. A higher water concentration in the permeate and a lower perme-ation rate are achieved for a higher molecular weight alcohol. Compared with the BAPP–BODA polyimide membrane, the BAPP–BODA/13X hybrid membrane has a higher permeation rate with similar water concentrations in the permeate. This is possibly due to the molecular sieving effect and hydrophilicity of the zeolites. These results correspond well with the position lifetime distributions and theτ3results indicated inFigs. 2 and 4. The positron annihilation lifetime (τ3) of the BAPP–BODA polyimide is lower than that of the BAPP–BODA/13X hybrid membrane. Sorption experiments are conducted to investigate the solubility and diffusivity effects on the permselectivity of the BAPP–BODA/13X hybrid membranes.Fig. 10 shows the alcohol concentration effect on the water concentration in the membrane. It is found that the water concentration in the BAPP–BODA/13X hybrid membrane is higher than that for the BAPP–BODA polyimide membrane. These results might be due to the hydrophilicity of the zeolite. As the carbon atomic number of the alcohol increases, the water concentration in the BAPP–BODA polyimide membrane decreases rapidly and that in the BAPP–BODA/13X hybrid membrane increases slightly. Table 1
Results on the pervaporation separation performance of aqueous alcohol mixtures through the pristine BAPP–BODA and BAPP–BODA/13X hybrid membranesa
Aqueous alcohol mixtures (90 wt%) BAPP–BODA BAPP–BODA/13X
A (g/m2h) B (wt%) C (%) A (g/m2h) B (wt%) C (%)
Methanol 238 24.1 5.5 320 22.2 4.5
Ethanol 152 57.2 12.3 210 48.3 6.1
Isopropanol 90 94.6 16.2 150 96.8 10.2
n-Butanol 145 95.8 21.4 190 96.7 13.5
A: permeation rate (g/m2h); B: water concentration in permeate (wt%); C: degree of swelling (%); operation temperature: 25◦C. aBAPP–BODA/13X hybrid membrane composition: 13X/BAPP–BODA = 0.5/1 (g/g).
BAPP–BODA/13X hybrid membrane results from the increased hydrophilicity from zeolite 13X incorporation into the polyimide membrane.
Pervaporation membranes generally follow the solut-ion–diffusion model. According to the solutsolut-ion–diffusion model, the permeability (P) is determined by the solubility (S) and the diffusivity (D). Therefore, the overall selectivity is determined by both the solution and diffusion selectivities; i.e. P = SD, and αwater/alcohol= Pwater/Palcohol= (Swater/Salcohol) (Dwater/Dalcohol) =αSαD. The solubility S is obtained using
sorp-tion measurements. The diffusion D is calculated from the P to S ratio. Table 2shows the molar volume effect of aqueous alcohol mixtures on overall selectivity (α), solution selectivity (αS) and diffusion selectivity (αD) for the pristine BAPP–BODA
and BAPP–BODA/13X hybrid membranes. It shows thatα and αD increase but αS decreases with increasing alcohol carbon
atomic number of feed mixtures for the BAPP–BODA poly-imide membrane. Obviously, the diffusion selectivity dominates the pervaporation behavior. However, by incorporating zeolite 13X into the polyimide membrane, the variation of α corre-sponds withαSwith increasing alcohol carbon atomic number.
This means that the sorption selectivity dominates the behavior of pervaporation because the separation factor of pervaporation
Fig. 10. Sorption experiments of 90 wt% alcohol–water mixtures at 25◦C: () BAPP–BODA; (䊉) BAPP–BODA/13X (13X/BAPP–BODA = 0.5/1 (g/g)). C1: methanol; C2: ethanol; C3: isopropanol; C4: n-butanol.
Fig. 11. Long-term experiment results for pervaporation separation of 70 wt% aqueous isopropanol mixture through the BAPP–BODA/13X hybrid membrane (13X/BAPP–BODA = 0.5/1 (g/g)) at 25◦C. () Pi/P1(the permeation rate ratio
of the ith day to the 1st day); () water concentration in permeate.
through the BAPP–BODA/13X hybrid membranes follows the same trend as the sorption selectivity but not the trend of the diffusion selectivity. The increasingαSis due to incorporating
the hydrophilic, larger pore size (8.16 ˚A) zeolite 13X. Zeolite 13X dominates the overall selectivity for the BAPP–BODA/13X hybrid membranes.
3.4. Durability of BAPP–BODA/13X hybrid membrane Long-term membrane stability experiments are performed in this study.Fig. 11shows the durability of the BAPP–BODA/13X hybrid membrane for a 70 wt% aqueous isopropanol mixture at 25◦C. It shows that the water concentration in the permeate and the Pi/P1ratio (the permeation rate ratio of the ith day to the 1st day) show no evident change during the long-term test procedure. It exhibits that BAPP–BODA/13X has good perva-poration performance although incompatibility exists between the polyimide matrix and zeolite particles. The pervaporation performance of a 70 wt% isopropanol–water mixture at 25◦C is maintained for at least 180 days for the BAPP–BODA/13X hybrid membrane.
4. Conclusions
In the past, the selectivity of pervaporation separation pro-cess has been traditionally understood in terms of the interaction
between the permeates and the polymeric membranes. Recently, positron annihilation spectroscopy studies could provide a new approach in terms of size selectivity of free volume. By incor-porating zeolite 13X into the polyimide membrane, the o-Ps annihilation lifetime increases with increasing the zeolite 13X content of the BAPP–BODA/13X hybrid membranes. This result indicates that the size of free volume hole is strongly composition dependent. Compared with the BAPP–BODA membrane, the BAPP–BODA/13X hybrid membrane has good pervaporation performance for dehydrating aqueous alcohol solutions. In the long-term stability examination, the pervapo-ration of a 70 wt% isopropanol–water mixture at 25◦C could be maintained for at least 180 days for the BAPP–BODA/13X hybrid membrane.
Acknowledgements
The authors wish to sincerely thank the Ministry of Economic Affairs, Ministry of Education Affair and the National Science Council of Taiwan, for financially supporting this work.
References
[1] P. Shao, R.Y.M. Huang, Polymeric membrane pervaporation, J. Membr. Sci. 287 (2007) 162–179.
[2] M.H.V. Mulder, A. Smolders, Pervaporation, solubility aspects of the solution–diffusion model, Sep. Purif. Methods 15 (1986) 1– 19.
[3] K.J. Kim, S.H. Park, W.W. So, S.J. Moon, Pervaporation separation of aqueous organic mixtures through sulfated zirconia-poly(vinyl alcohol) membrane, J. Appl. Polym. Sci. 79 (2001) 1450–1455.
[4] X. Chen, Z.H. Ping, Y.C. Long, Separation properties of alcohol–water mix-ture through silicalite-i-filled silicone rubber membranes by pervaporation, J. Appl. Polym. Sci. 67 (1998) 629–636.
[5] X.M. He, W.H. Chan, C.F. Ng, Water–alcohol separation by pervaporation through zeolite-modified poly(amidesulfonamide), J. Appl. Polym. Sci. 82 (2001) 1323–1329.
[6] A.A. Kittur, M.Y. Kariduraganavar, U.S. Toti, K. Ramesh, T.M. Aminab-havi, Pervaporation separation of water–isopropanol mixtures using ZSM-5 zeolite incorporated poly(vinyl alcohol) membranes, J. Appl. Polym. Sci. 90 (2003) 2441–2448.
[7] M.Y. Kariduraganavar, A.A. Kittur, S.S. Kulkarni, K. Ramesh, Devel-opment of novel pervaporation membranes for the separation of water-isopropanol mixtures using sodium alginate and NaY zeolite, J. Membr. Sci. 238 (2004) 165–175.
[8] S.S. Kulkarni, S.M. Tambe, A.A. Kittur, M.Y. Kariduraganavar, Prepara-tion of novel composite membranes for the pervaporaPrepara-tion separaPrepara-tion of water–acetic acid mixtures, J. Membr. Sci. 285 (2006) 420–431. [9] M.Y. Teng, K.R. Lee, S.C. Fan, D.J. Liaw, J. Huang, J.Y. Lai, Development
of aromatic polyamide membranes for pervaporation and vapor permeation, J. Membr. Sci. 164 (2000) 241–249.
[10] C.F. Mellot, A.K. Cheetham, S. Harms, S. Savitz, R.J. Gorte, A.L. Myers, Calorimetric and computational studies of chlorocarbon adsorption in zeo-lites, J. Am. Chem. Soc. 120 (1998) 5788–5792.
[11] C.C. Hu, Y.J. Fu, S.W. Hsiao, K.R. Lee, J.Y. Lai, Effect of physical aging on the gas transport properties of poly(methyl methacrylate) membranes, J. Membr. Sci. 303 (2007) 29–36.
[12] H.M. Chen, W.S. Hung, C.H. Lo, S.H. Huang, M.L. Cheng, G. Liu, K.R. Lee, J.Y. Lai, Y.M. Sun, C.C. Hu, R. Suzuki, T. Ohdaira, Y.C. Jean, Free-volume depth profile of polymeric membranes studied by positron annihilation spectroscopy: layer structure from interfacial polymerization, Macromolecules 40 (2007) 7542–7557.
[13] M. Sakuth, J. Meyer, J. Gmehling, Vapor phase adsorption equilibria of toluene + 1-propanol mixtures on y-zeolites with different silicon to alu-minum ratios, J. Chem. Eng. Data 40 (1995) 895–899.
