Preparation and characterization of polyacrylonitrile ultrafiltration membranes

Preparation and characterization of polyacrylonitrile ultrafiltration membranes

Journal of Membrane Science 222 (2003) 87–98 Preparation and characterization of polyacrylonitrile ultrafiltration membranes Sen Yang, Zhongzhou Liu∗...

284KB Sizes 0 Downloads 38 Views

Journal of Membrane Science 222 (2003) 87–98

Preparation and characterization of polyacrylonitrile ultrafiltration membranes Sen Yang, Zhongzhou Liu∗ Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, P.O. Box 2871, Beijing 100085, PR China Received 12 December 2002; accepted 5 March 2003

Abstract Asymmetric polyacrylonitrile (PAN) ultrafiltration (UF) membranes are prepared from three kinds of coagulant: water, aqueous solution of sodium chloride (NaCl) and aqueous solution of sodium carbonate (Na2 CO3 ), using dimethylacetamide (DMAC) as solvent and calcium chloride (CaCl2 ) as additive by phase inversion method. The membranes are characterized in terms of the pure water flux, molecular weight cut-off (MWCO) profile and direct field emission scanning electron microscopy (FESEM) observations. The addition of CaCl2 to the casting solution, up to 3 wt.%, increases the resulting membrane permeability while maintaining their retentive properties. A decrease of the permeability and retention properties is observed when the coagulation bath is aqueous solution of NaCl. Na2 CO3 in the coagulation bath reacts with CaCl2 in the casting solution and produces precipitate of CaCO3 . The effect of Na2 CO3 concentration on the PAN membrane permeability and retention properties is examined. It is found that the pure water fluxes of the PAN membranes increase drastically when the concentration of Na2 CO3 is high enough. © 2003 Elsevier B.V. All rights reserved. Keywords: Chemical reaction; Coagulant; Additive; Ultrafiltration; Polyacrylonitrile

1. Introduction The majority of polymeric membranes with asymmetric structure have been prepared by nonsolvent induced phase inversion (NIPI) process [1]. In this process, a homogeneous polymer solution is spread directly onto a suitable support by using a casting knife, and then immersed into a nonsolvent coagulation bath. The casting solution phase separation, responsible for a membrane formation occurs by ∗ Corresponding author. Tel.: +86-10-62849195; fax: +86-10-62923563. E-mail addresses: [email protected] (S. Yang), [email protected] (Z. Liu).

the diffusional exchange of solvent and nonsolvent across the interface between casting solution and nonsolvent. It is well known that the formulation of the casting solution and the coagulation bath can affect the final membrane structure and properties. Introduction of suitable additive to casting solution is one convenient and efficient method to obtain membranes with special properties. The additive may be water [2,3], inorganic salts [4,5], low molecular weight organics [6–8], surfactants [9], polymer [10,11], mineral fillers [12–14], or mixtures of them [3,11]. There are several potential mechanisms through which such additives can affect the resulting membrane via

0376-7388/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0376-7388(03)00220-5

88

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

changing solvent capacity [15], precipitation kinetics and thermodynamic properties [10,16–18]. The addition of organic or inorganic components to a nonsolvent coagulation bath is another important method used in membrane modification. By the choice of different nonsolvent (coagulant), the polymeric membrane can be changed from asymmetric to symmetric. Usually, water as the nonsolvent coagulant is the most common choice in preparing asymmetric membranes. Adding solvent into the water is a common method to suppression the formation of macrovoids (finger-like pores), particularly in preparing hollow fiber membranes [19–21]. Chun et al. [22] reported that increasing the amount of dimethylacetamide (DMAC) in the coagulation bath could delay the demixing phase inversion, increase the stability of the polymer ternary system, and reduce the pore specific volume of sub-layer and the pore size in the skin layers. Some researchers used the interaction between additive in the casting solution and the coagulation medium to explain the effect of additive on the properties and morphology of membranes. Chuang et al. [23] proposed a mechanism describing the affinity between additive and coagulation medium to investigate the effect of dextran and poly(vinyl pyrrolidone) (PVP) additives on the formation of poly(vinyl alcohol) (PVA) membranes. However, most of these interactions are physical interactions in nature. To our knowledge there is no report on the preparation of asymmetric polymeric membrane by a chemical reaction between additive and coagulant. Polyacrylonitrile (PAN) is one of the versatile polymers that are widely used for making membranes due to its good solvent resistance. PAN has been used as a substrate for ultrafiltration (UF), microfiltration (MF) and reverse osmosis (RO) [24–26]. Schamagl and Buschatz [25] reported pure water flux of PAN membranes was 688 and 6150 l/m2 h bar with 20,000 and 15.7×106 u (20,000 and 15.7×106 Da) of molecular weight cut-off (100%). In the present work, the polyacrylonitrile membranes were prepared from three kinds of coagulant: water, aqueous solution of sodium chloride (NaCl) and aqueous solution of sodium carbonate (Na2 CO3 ), utilizing calcium chloride (CaCl2 ) as additive and dimethylacetamide as solvent. When the casting film was immersed in the aqueous solution of sodium

carbonate, the following chemical reaction occurred: Na2 CO3 + CaCl2 = CaCO3 ↓ +2NaCl The effects of additive, coagulant and chemical reaction on the membrane properties and morphology were investigated. 2. Experimental 2.1. Materials Polyacrylonitrile was purchased from Shanghai Jinshan Chemical Industry Factory and dimethylacetamide from Shanghai Organic Chemical Industry Factory. Sodium carbonate, sodium chloride and calcium chloride were obtained from Beijing Chemical Industry Factory. Lysozyme (Shanghai Lizhu), pepsin (TBO, Tokyo), albumin egg (Sigma) and bovine serum albumin (BSA) (Beijing Shuangxuan) were used in the retention test. 2.2. Viscosity studies The viscosity of salt-free and salt-containing samples of solvent and dilute polymer solutions (≤0.3 wt.% PAN) was measured using a Ubbelohde capillary viscometer in a thermostatted water bath at 20 ± 0.1 ◦ C. For each polymer solution, two relative viscosity (η) values were obtained. One relative viscosity value was determined by normalizing the efflux time measured for the polymer solution with respect to the efflux time of pure DMAC. The other value was obtained by normalizing the solution efflux time with respect to that of CaCl2 -containing DMAC. The viscosity of 12 wt.% PAN solutions with and without CaCl2 was obtained by using a Falling Ball viscometer (Thermo Haake) at 20 ± 0.1 ◦ C. 2.3. Preparation of membranes The casting solutions consisted of 12 wt.% solution of PAN in DMAC to which quantity of CaCl2 was added. Solutions were allowed to stand overnight before casting and then cast at room temperature on a glass plate by spreading them between thin wires (0.195 mm in diameter) with a glass knife in order to control the thickness of the films without a preceding

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

89

dry phase inversion in atmosphere; and then the films were immediately immersed in a coagulation bath at 20 ◦ C. The coagulation mediums were demineralized water or aqueous solutions of sodium salt (NaCl or Na2 CO3 ) at different concentrations. After immersion, the membranes were kept in the coagulation bath for at least one night to complete the formation process. In order to remove the remaining solvent, additive or the CaCO3 settlings out of the membrane structure, the membranes were rinsed with demineralized water and wet stored until tested. The actual thickness of the membrane was measured using micrometer.

(iii) albumin egg (0.1 g/dl) and (iv) bovine serum albumin (0.1 g/dl) solutions prepared in distilled water. The protein concentrations in the feed and permeate samples were determined using a spectrophotometer (Shimadzu, UV-120-02) at 280 nm. And the molecular weight cut-off (MWCO) profiles were constructed.

2.4. Membrane characteristics

3.1.1. Viscosity of the casting solution Viscosity of the casting solution can hinder severely the exchange rate of solvent and nonsolvent during phase inversion process, and therefore, it can be used as an important parameter to influence the precipitation kinetics and thus, the formation of resulting membrane morphology [18]. The viscosities of the salt-free and salt-containing PAN solutions were normalized with respect to the pure DMAC solution viscosity and the relative viscosity (ηr ) trends presented in Fig. 1a. The curves plotted in Fig. 1a show that the relative viscosity of the PAN solutions increases with increasing polymer concentration. At given PAN concentration, the relative viscosity also increases with increasing CaCl2 concentration. However, when the PAN solution viscosities were normalized with respect to the CaCl2 -containing DMAC viscosity instead of the salt-free DMAC viscosity (η r ), the curves plotted in Fig. 1b show that the relative viscosity data for all PAN solutions containing CaCl2 now lie practically on the same curve. This result suggests that at a given polymer concentration, the increase in the viscosity of PAN solutions with increasing salt content is determined predominantly by the viscosity characteristics of the salt–solvent medium. That is, the CaCl2 interacts more strongly with DMAC solvent than with the PAN polymer chains, and it suggests that the effective solvating power of DMAC for PAN is progressively reduced with increasing salt concentration. This is also consistent with the observation [27] that at a given polymer concentration, the increase in the viscosity of the PAA–NMP solutions with increasing LiCl content is mainly determined by the viscosity characteristics of the salt–solvent mixture medium. Lee et al. [27]

2.4.1. Membrane structure The morphology of the prepared membranes was inspected with field emission scanning electron microscopy (FESEM, AMARY, 1910FE). For this purpose, all samples were soaked in 40 vol.% glycerol aqueous solution for 24 h, dried in vacuum, frozen in liquid nitrogen and fractured. After plated with gold, they were transferred into the microscope. 2.4.2. Flux and separation experiments A common permeation test apparatus was used to measure pure water flux and protein separation of PAN membranes. The obtained membrane sheets were cut into circle membrane species of 3.35 cm diameter before use. The pure water flux and protein retention were measured at 100 kPa, room temperature and 471.24 × 10−1 rad/s (450 rpm). The filtration was repeated three times on different membrane sheets so that the total analyzed surface was 0.0079 m2 . Three sets of membrane samples were made for each casting condition specified in this paper and the average flux and solute separation data were reported. After each run the whole test apparatus was rinsed thoroughly with demineralized water and membrane was washed to remove any deposition. Each membrane sample was initially compacted at 200 kPa with deionized water, until three successive flux measurements were constant. Flux was determined by mass collected over a measured time (1–10 min). Ultrafiltration experiments at 100 kPa and room temperature were carried out in the same assembly with (i) lysozyme (0.05 g/dl), (ii) pepsin (0.1 g/dl),

3. Results and discussion 3.1. Effect of different coagulation mediums on membrane performance

90

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

Fig. 1. Relative viscosity of PAN–DMAC solutions as a function of PAN and CaCl2 concentrations. Polymer solution efflux times were normalized to (a) the efflux time of pure DMAC solvent (ηr ), and (b) the efflux times of CaCl2 -containing DMAC (η r ). CaCl2 concentration is 0, 1, 2 and 3 wt.% from bottom to top.

thought the reason is that LiCl interact more strongly with NMP than with PAA, leading to the formation of LiCl–NMP complexes and, hence, a decrease in the solvation power of NMP for PAA. The viscosities (η) of the casting solutions are shown in Table 1. The addition of CaCl2 to 12 wt.% PAN/DMAC solution causes a decrease in viscosity, and the viscosity is the lowest (651 mPa s) when CaCl2 concentration is 2 wt.%. This is consistent with the decrease of solvent power of DMAC/CaCl2 for PAN. 3.1.2. Water as the coagulant Four polyacrylonitrile membranes, designated as PAN1a to PAN4a were prepared using different concentrations of CaCl2 . The compositions of casting

solution and total membrane thickness are given in Table 1. The cast thickness of all the four membranes, before gelling, is 195 ␮m based on the thickness of cast knife and the final membrane thickness is found to be thinner than the cast thickness indicating densification of polymer network during gelation [28]. Table 1 shows that the PAN membranes prepared from CaCl2 is thinner than that of the PAN membranes prepared from neat DMAC, and the thickness of PAN membrane decreases with increasing of CaCl2 concentration. Fig. 2 shows the relative pure water fluxes in the various membranes. PAN membranes prepared from CaCl2 have considerably higher flux than membranes prepared from neat DMAC. PAN2a, PAN3a, and

Table 1 PAN membranes coagulated in water Membrane

PAN1a PAN2a PAN3a PAN4a

Composition of casting solution (wt.%) P

S

A

12 12 12 12

88 87 86 85

– 1 2 3

Solution viscosity η (mPa s)

Total membrane thickness (␮m)

1399 963 651 708

135 125 120 117

P, polyacrylonitrile; S, DMAC; A, CaCl2 . Coagulation bath: demineralized water at 20 ◦ C.

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

91

Fig. 2. Pure water flux of PAN membranes prepared from different concentrations of CaCl2 (wt.%).

PAN4a have 191, 229 and 289% of the PAN1a flux, respectively. The results indicate that the pure water flux of PAN membranes can be increased by addition of CaCl2 as additive, and increasing the content of CaCl2 in the casting solution from 1 to 3 wt.%, the pure water permeability rate increases almost linearly. 3.1.3. Aqueous solution of NaCl as the coagulant The effect of aqueous solution of NaCl as the coagulant on the membrane performance is studied. NaCl concentration varies from 1 to 20 wt.%, and the composition of the casting solution keeps unchanged. The composition of the coagulation bath, the gelation time measured by stopwatch (the gelation time is defined as from the casting solution being put into the coagulation bath to the appearance of the totally opaque

film) and the thickness of the membranes are shown in Table 2. The gelation time increases from 10 to 25 s as NaCl concentration increases from 1 to 20 wt.%, and the membrane prepared from aqueous solution of NaCl is thicker than the membrane prepared from water, especially when the concentration of NaCl is more than 5 wt.%. A method for decreasing activity of the aqueous coagulation bath is addition of an inorganic salt to the coagulation bath [29]. Addition of inorganic salt to the coagulation bath reduces the chemical potential (µ) of water due to the salt effect, and then reduces the driving force for the film precipitation. Inflow of H2 O and NaCl (coagulant) and the outflow of DMAC (solvent) therefore become slow. This is consistent with the result that NaCl as additive in the coagulation bath

Table 2 PAN membranes coagulated in NaCl solution Membrane

PAN1b PAN2b PAN3b PAN4b

Composition of the coagulation bath (wt.%) NaCl

Water

1 5 10 20

99 95 90 80

Gelation time (s)

Total membrane thickness (␮m)

10 12 17 25

125 128 210 291

The composition of the casting solution: polyacrylonitrile, 12 wt.%; DMAC, 85 wt.%; CaCl2 , 3 wt.%. Temperature of coagulation bath: 20 ◦ C.

92

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

Fig. 3. Pure water flux of PAN membranes prepared from different concentration of aqueous solution of NaCl (

decreases the gelation rate. When the concentration of NaCl is above 5 wt.% the thickness of the PAN membrane become thicker, resulting in a loose morphology of membrane (Fig. 7b). The pure water fluxes of PAN membranes prepared from different concentration of NaCl are shown in Fig. 3. It can be seen that the pure water fluxes are approximately equal to that of the membrane prepared from water when NaCl concentration is lower than 5 wt.%. The pure water flux decreases from 688 to 470 l/m2 h while the concentration of NaCl increases from 5 to 20 wt.%.

) and Na2 CO3 (

) (wt.%).

3.1.4. Aqueous solution of Na2 CO3 as the coagulant CaCl2 in the casting solution can react with Na2 CO3 in the coagulation bath to produce CaCO3 precipitate. The effects of this chemical process on the PAN membrane performance are investigated. Composition of the coagulation bath, thickness of the membrane, gelation time and amount of precipitate are shown in Table 3. In this system, Na2 CO3 as additive in the coagulation bath should reduce the driving force for the membrane formation and the diffusion rate between solvent and nonsolvent just like NaCl in the coagulation bath. However, the gelation

Table 3 PAN membranes coagulated in Na2 CO3 solution Membrane

PAN1c PAN2c PAN3c PAN4c PAN5c PAN6c

Composition of the coagulation bath (wt.%) Na2 CO3

Water

0.1 1 2 5 10 20

99.9 99 98 95 90 80

Gelation time (s)

The amount of precipitate

Total membrane thickness (␮m)

11 10 10 6 5 3

Little Little Much More More Morea

134 130 104 88 96 162

The composition of the casting solution: polyacrylonitrile, 12 wt.%; DMAC, 85 wt.%; CaCl2 , 3 wt.%. Temperature of coagulation bath: 20 ◦ C. a The precipitate is locked in the membrane.

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

rate accelerates with the increase of Na2 CO3 concentration in the coagulation bath. That can be explained by the reaction between CaCl2 and Na2 CO3 . Increasing Na2 CO3 concentration, the chemical reaction become faster, the gelation time become shorter and the membrane become thinner except that the membrane became thicker dramatically when the concentration is 20 wt.%. An interesting phenomenon is noticed that the gelation rate of PAN membrane is so quick that the precipitate cannot diffuse out of the membrane structure and is locked in it when Na2 CO3 concentration is 20 wt.%. This can explain that the membrane become thicker and rough slightly compared with other membranes. Fig. 3 shows the relative water fluxes of various membranes. It can be seen that the pure water flux also increases as Na2 CO3 concentration increase in the coagulant. Compared to PAN membrane prepared from water with the same composition of casting solution, the pure water fluxes of PAN1c, PAN2c and PAN3c are lower and those of PAN4c, PAN5c, and PAN6c are higher. These results may be due to the competition between the salt effect and the effect of the chemical reaction. An interesting phenomenon is noticed that the low Na2 CO3 concentration shows a same effect on the permeation of PAN membranes just like that of high NaCl concentration. Further studies are in progress to explain the interesting results.

93

3.2. Membrane characterization 3.2.1. Molecular weight cut-off profile The molecular weight cut-off profiles are constructed for PAN membranes by measuring solute separation for four proteins of molecular weight ranging from 14,400 to 67,000 u (14,400 to 67,000 Da). The results are given in Figs. 4–6. For the various membranes prepared from CaCl2 as additive and water as the coagulant, the molecular weight cut-off profiles (Fig. 4) are superposed at molecular weight of protein higher than 35,000 u (35,000 Da). The results indicate that CaCl2 as additive can increase the pure water flux of PAN membrane and does not change the average pore size significantly compared with the PAN membrane prepared from water as the coagulant and without any additive in the casting solution. Shinde et al. [30] examined the effect of various inorganic halides (LiCl, ZnCl2 and AlCl3 ) added to a casting solution of PAN in N,N-dimethyl formamide (DMF). They reached the same conclusion that addition of di- and trivalent salts resulted in membranes with a pore size similar to the membrane prepared without any additive. Comparing Fig. 4 with Fig. 5, an interesting result can be drawn that the aqueous solution of NaCl as the coagulant reduces the protein retention slightly. Deshmukh and Li [31] studied the effect of coagulation

Fig. 4. MWCO profile of PAN membranes prepared from different concentrations of CaCl2 . CaCl2 concentration: (a) 0 wt.%; (b) 1 wt.%; (c) 2 wt.% and (d) 3 wt.%.

94

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

Fig. 5. MWCO profile of PAN membranes prepared from aqueous solution of NaCl. NaCl concentration: (a) 1 wt.%; (b) 5 wt.%; (c) 10 wt.% and (d) 20 wt.%.

medium, ethanol (10–50%) and water (90–50%), on the PVDF hollow fiber membranes. They found that the presence of ethanol in the coagulation bath reduced the polymer precipitation rate in phase inversion process, and the effective porosity of the resulting membranes decreased as ethanol concentration in the coagulation bath increased. This can explain the slightly decreased retention capability and the

drastically reduced permeation property when NaCl concentration is high enough. The retention properties of PAN membranes prepared from aqueous solution of Na2 CO3 are showed in Fig. 6. PAN membranes prepared from the low Na2 CO3 concentration (1 and 5 wt.%) have the similar MWCO profiles compared with PAN membranes prepared from water as coagulant (Fig. 4). That means the

Fig. 6. MWCO profile of PAN membranes prepared from aqueous solution of Na2 CO3 . Na2 CO3 concentration: (a) 1 wt.%; (b) 5 wt.%; (c) 10 wt.% and (d) 20 wt.%.

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

95

Fig. 7. FSEM-pictures of the cross-sections of different membranes prepared from the systems 12/3/85 (w/w/w) PAN/CaCl2 /DMAC. The compositions of the coagulant: (a) demineralized water; (b) 10 wt.% NaCl; (c) 1 wt.% Na2 CO3 ; (d) 5 wt.% Na2 CO3 and (e) 20 wt.% Na2 CO3 .

96

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

low Na2 CO3 concentration (1 and 5 wt.%) has little effect on the pore size distribution. PAN membranes prepared from the high Na2 CO3 concentration (10 and 20 wt.%) has the similar MWCO profiles, it becomes flat compared with PAN membranes prepared from water as coagulant (Fig. 4) and the higher the concentration, the flatter the MWCO profile. That means there are some big defects on the membrane surface. These defects are caused by some solid particles (CaCO3 ) formed in the membrane after the solidification takes place. This explains the fact that the membrane becomes thicker and rough slightly as compared with other membranes. It can be seen from Figs. 4–6 that the separation rate increases with increase in molecular weight of protein solute and the profiles are not sharp but are gradual and diffusive for PAN membranes prepared from different coagulation mediums. The cut-off of most PAN membranes is found to be around 60,000 u (60,000 Da). 3.2.2. Membrane morphology In order to compare the morphology of PAN membranes prepared from different coagulation mediums, the structural changes of the membrane are observed by using FESEM. Fig. 7 shows some images of PAN membrane cross-sections. PAN membranes (Fig. 7a) prepared from water exhibits a typical asymmetric structure [32], composed of a thin and dense skin layer and a porous bulk that contains independent finger-like cavities enclosed in a porous solid matrix. The skin layer is responsible for the permeation or retention of solutes whereas the porous bulk acts as a mechanical support. Young and Chen [33] reported that the skin layer became less dense and the finger-like macrovoids became less evident, as the DMSO (solvent) content in the coagulation bath increased. The macrovoids even can be eliminated when the bath contains a significant amount of solvent. NaCl in the coagulation bath shows a similar effect on the structure of PAN membranes. When the content of NaCl is low in the coagulation bath, the structure of PAN membranes does not show any obvious change (the pictures are not shown); when the content is more than 10 wt.% the structure exhibits a kind of transition from a typical asymmetric structure to a structure between the asymmetric and the symmetric (untypical asymmetric structure). FESEM

(Fig. 7b) shows that the macrovoids are lessened drastically and the surface layers become somewhat porous in the membranes prepared from 10 wt.% NaCl. The change of PAN membrane structure shows a reverse tendency while the coagulation bath is aqueous solution of Na2 CO3 . The FESEM pictures (Fig. 7c–e) show the transition from an untypical asymmetric structure to a typical asymmetric structure with increasing content of Na2 CO3 . This is consistent with the acceleration of the reaction rate. These pictures (Fig. 7c and d) illustrate that the membrane thickness decreases when more Na2 CO3 is added to the coagulation bath; this is consistent with the thickness result of various membranes. This indicates a faster transportation of solvent/nonsolvent during the membrane formation process when Na2 CO3 is added to the coagulation bath, and the transportation is even faster while more Na2 CO3 is added to the coagulation bath. Combining with the cut-off, the similar structure of PAN membranes explains the almost same permeation prepared from water, 1 and 2 wt.% of NaCl and 5 wt.% of Na2 CO3 . The finger-like pores decrease in size and number, the sponge-like structure of the less dense skin layer explains the low permeation capability of PAN membranes prepared from high concentration of NaCl and low concentration of Na2 CO3 .

4. Conclusions 1. Membranes with high permeation can be prepared from polyacrylonitrile polymer with CaCl2 as additive, DMAC as solvent and water as the coagulation bath. The permeation of PAN membrane can be controlled by the concentration of CaCl2 . 2. Changing concentration of CaCl2 from 1 to 3 wt.% increases the membrane permeation and has little influence on the average pore radius. 3. The aqueous solution of NaCl as the coagulation bath decreases the permeability of PAN membrane, and the higher the NaCl concentration, the more the permeability decrement. 4. The chemical reaction between the additive and the coagulation medium play a significant role in PAN membrane performance. The reaction rate can control the membrane performance. The higher rate results in the higher permeation. PAN membranes

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

with big pore size can be fabricated by accelerating the chemical reaction. 5. FESEM pictures clearly show that the membranes all have an analogous typical asymmetric structure, whenever the membranes were prepared from low concentration of NaCl, high concentration of Na2 CO3 or water; and the membranes prepared from the high concentration of NaCl and low concentration of Na2 CO3 have the analogous structure—a structure between the asymmetric and the symmetric. References [1] K. Scott, Handbook of Industrial Membranes, second edition, Elsevier Advanced Technology, pp. 205–207. [2] D. Wang, K. Li, W.K. Teo, Highly permeable polyethersulfone hollow fiber gas separation membranes prepared using water as non-solvent additive, J. Membr. Sci. 176 (2000) 147. [3] D. Wang, K. Li, W.K. Teo, Porous PVDF asymmetric hollow fiber membranes prepared with the use of small molecular additives, J. Membr. Sci. 178 (2000) 13. [4] M.A. Kraus, M. Nemas, M.A. Frommer, The effect of low molecular weight additives on the properties of aromatic polyamide membranes, J. Appl. Polym. Sci. 23 (1979) 445. [5] S.R. Kim, K.H. Lee, M.S. Jhon, The effect of ZnCl2 on the formation of polysulfone membrane, J. Membr. Sci. 119 (1996) 59. [6] S.P. Petrov, Conditions for obtaining ultrafiltration membranes from a solution of polyacrylonitrile in dimethylformamide in the presence of formamide, J. Appl. Polym. Sci. 62 (1996) 267. [7] W.-Y. Chuang, T.-H. Young, W.-Y. Chiu, The effect of acetic acid on the structure and filtration properties of poly(vinyl alcohol) membranes, J. Membr. Sci. 172 (2000) 241. [8] M.-J. Han, Effect of propionic acid in the casting solution on the characteristic of phase inversion polysulfone membranes, Desalination 121 (1999) 31. [9] H.A. Tsai, L.D. Li, K.R. Lee, Y.C. Wang, C.L. Li, J. Huang, J.Y. Lai, Effect of surfactant addition on the morphology and pervaporation performance of asymmetric polysulfone membranes, J. Membr. Sci. 176 (2000) 97. [10] P.S.T. Machado, A.C. Habert, C.P. Borges, Membrane formation mechanism based on precipitation kinetics and membrane morphology: flat and hollow fiber polysulfone membranes, J. Membr. Sci. 155 (1999) 171. [11] B. Torrestiana-Sanchez, R.I. Ortiz-Basurto, E.B. La Fuente, Effect of nonsolvents on properties of spinning solutions and polyethersulfone hollow fiber ultrafiltration membranes, J. Membr. Sci. 152 (1999) 19. [12] P. Aerts, E. Van Hoof, R. Leysen, I.F.J. Vankelecomb, P.A. Jacobs, Polysulfone–aerosil composite membranes Part 1. The influence of the addition of aerosil on the formation process and membrane morphology, J. Membr. Sci. 176 (2000) 63.

97

[13] P. Aerts, I. Genne, S. Kuypers, R. Leysen, I.F.J. Vankelecom, P.A. Jacobs, Polysulfone–aerosil composite membranes Part 2. The influence of the addition of aerosil on the skin characteristics and membrane properties, J. Membr. Sci. 178 (2000) 1. [14] A. Bottino, G. Capannelli, V.D. Asti, P. Piaggio, Preparation and properties of novel organic–inorganic porous membranes, Sep. Purif. Technol. 22-23 (2001) 269. [15] S. Munori, A. Bottino, G. Capannelli, P. Moretti, P.P. Bon, Preparation and characterization of polysulfone-polyvinylpyrrolidone based membranes, Desalination 70 (1988) 265. [16] T.-P. Hou, S.-H. Dong, L.-Y. Zheng, The study of mechanism of organic additive action in the polysulfone membrane casting solution, Desalination 83 (1991) 343. [17] Y.S. Kang, H.J. Kim, U.Y. Kim, Asymmetric membrane formation via immersion precipitation method. Part I. Kinetic effect, J. Membr. Sci. 60 (1991) 219. [18] J.-H. Kim, K.-H. Lee, Effect of PEG additive on membrane formation by phase inversion, J. Membr. Sci. 138 (1998) 153. [19] W. Albrecht, Th. Weigel, M. Schossig-Tiedemann, K. Kneifel, K.-V. Peinemann, D. Paul, Formation of hollow fiber membranes from poly(ether imide) at wet phase inversion using binary mixtures of solvents for the preparation of the dope, J. Membr. Sci. 192 (2001) 217. [20] H. Strathmann, K. Kock, The formation mechanism of phase inversion membranes, Desalination 21 (1977) 241. [21] J.M. Cheng, D.M. Wang, F.C. Lin, J.Y. Lai, Formation and gas flux of asymmetric PMMA membranes, J. Membr. Sci. 109 (1996) 93. [22] K.-Y. Chun, S.-H. Jang, H.-S. Kim, Y.-W. Kim, H.-S. Han, Y.-I. Joe, Effects of solvent on the pore formation in asymmetric 6FDA-4,4 ODA polyimide membrane: terms of thermodynamics, precipitation kinetics, and physical factors, J. Membr. Sci. 169 (2000) 197. [23] W.-Y. Chuang, T.-H. Young, W.-Y. Chiu, C.-Y. Lin, The effect of polymeric additives on the structure and permeability of poly(vinyl alcohol) asymmetric membranes, Polymer 41 (2000) 5633. [24] I.-C. Kim, H.-G. Yun, K.-H. Lee, Preparation of asymmetric polyacrylonitrile membrane with small pore size by phase inversion and post-treatment process, J. Membr. Sci. 199 (2002) 75. [25] N. Schamagl, H. Buschatz, Polyacrylonitrile (PAN) membranes for ultra- and microfiltration, Desalination 139 (2001) 191. [26] M.V. Chandorilar, P.C. Bhavsar, Evaluation of polyacrylonitrile and poly (methyl methacrylate) as membrane materials for reverse osmosis, Indian J. Technol. 21 (1983) 124. [27] H.J. Lee, J.W.H. Lee, Y.S. Kang, Solution properties of poly(amic acid)–NMP containing LiCl and their effects on membrane morphologies, J. Membr. Sci. 196 (2002) 267. [28] B.K. Chaturvedi, A.K. Ghosh, V. Ramachandhran, M.K. Trivedi, M.S. HanTab, B.M. Misra, Preparation, characterization and performance of polyethersulfone ultrafiltration membranes, Desalination 133 (2001) 31. [29] J.F. Hester 1, A.M. Mayes, Design and performance of foul-resistant poly(vinylidene fluoride) membranes prepared

98

S. Yang, Z. Liu / Journal of Membrane Science 222 (2003) 87–98

in a single-step by surface segregation, J. Membr. Sci. 202 (2002) 119. [30] M.H. Shinde, S.S. Kulkarni, D.A. Musale, S.G. Joshi, Improvement of the water purification capability of poly(acrylonitrile), ultrafiltration membranes, J. Membr. Sci. 162 (1999) 9. [31] S.P. Deshmukh, K. Li, Effect of ethanol composition in water coagulation bath on morphology of PVDF hollow fibre membranes, J. Membr. Sci. 150 (1998) 75.

[32] D.-T. Lin, L.-P. Cheng, Y.-J. Kang, L.-W. Chen, T.-H. Young, Effects of precipitation conditions on the membrane morphology and permeation characteristics, J. Membr. Sci. 140 (1998) 185. [33] T.-H. Young, L.-W. Chen, A two step mechanism of diffusion-controlled ethylene vinyl alcohol membrane formation, J. Membr. Sci. 57 (1991) 69.