- Email: [email protected]
poly(acrylonitrile) (PAN) by epichlorohydrin cross-linking
Journal of Membrane Science 297 (2007) 51–58
A novel composite nanofiltration (NF) membrane prepared from glycolchitin/poly(acrylonitrile) (PAN) by epichlorohydrin cross-linking Hongwei Sun, Guohua Chen ∗ , Ruihua Huang, Congjie Gao College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, China Received 10 October 2006; received in revised form 9 February 2007; accepted 18 February 2007 Available online 21 February 2007
Abstract A novel composite nanofiltration (NF) membrane was prepared by over-coating the PAN ultrafiltration (UF) membrane with a glycolchitin (GC) thin layer. The effects of the membrane preparation techniques and operating conditions on the rejection performance of the composite membranes were studied. The structure of the composite NF membrane was characterized by scanning electron microscopy (SEM) and infrared spectroscopy (IR). And the surface roughness of the composite membrane was also characterized by atomic force microscope (AFM). The pressure osmobic coefficient (β) was about −2 mV/MPa and the molecular weight cut-off (MWCO) was obtained at 540 Da or so. Rejections of the composite membrane of Na2 SO4 , K2 SO4 , MgSO4 , NaCl, KCl, MgCl2 solutions (1.0 g/L) were 95.2%, 91.5%, 41%, 31.1%, 31% and 20.1% at 25 ◦ C under 1.0 MPa, and permeation fluxes are 10.0, 10.1, 19.2, 19.7, 20.8 and 13.0 L/m2 h, respectively. The results suggest that rejection performance is governed by solute–membrane and solute–solute electrostatic interactions. This charge distribution mainly determined the performances of the membranes. © 2007 Elsevier B.V. All rights reserved. Keywords: Polyacrylontrile UF membranes; Glycolchitin; Epichlorohydrin; Composite NF membranes; Molecular weight cut-off; Rejection performances
1. Introduction Nanofiltration membranes based technological process is between ultrafiltration membranes and reverse osmosis (RO) membranes. NF membrane’s molecular weight cut-off ranges from 200 to 1000 Da and it’s pore size is nominally 1 nm in dimension. The separation mechanism of NF membrane is a combination of Donnan effect and sieving effect. More specifically, NF membrane can be used to remove small neutral organic molecules while surface electrostatic properties allowed monovalent ions to be reasonably well transmitted with multivalent ions mostly retained [1]. NF membrane’s operating pressure ranged from 0.5 to 1.5 MPa, which was much lower than RO membranes. NF membrane have been applied widely in many industry fields such as purification for drinking water [2] and industry wastewater [3]. Research on composite membrane and its application are highly attention-getting in recent years. Especially composite technique for making NF membranes were mostly commercial and has the most yield [4]. ∗
Corresponding author. Tel.: +86 532 82032223; fax: + 86 595 2686969. E-mail addresses: [email protected], [email protected] (H. Sun), [email protected] (G. Chen). 0376-7388/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2007.02.027
Chitin is not a hydrophilic, but some of its derivatives are soluble in water, it shows excellent membrane forming characteristics and biocompatibility. So chitin can be grafted to different molecular groups for preparing composite membrane of different capabilities [5]. Chitosan in acetic acid has been used to prepare reverse osmosis [6], nanofiltration [7,8], microfiltration [9], ultrafiltration [10] and pervaporation membranes [11]. To our knowledge that there is no reported literature on using water-soluble derivatives of chitin to prepare composite NF membrane except us [12]. In this study a composite NF membrane was prepared using glycolchitin (GC) as filming-forming material and epichlorohydrin as cross-linking reagent. 2. Experiments 2.1. Materials and apparatus 2.1.1. Materials Chitin (supplied by Haihui Bioengineering Limited, Qingdao, China), PAN UF Membrane with MWCO of 1 × 105 Da was supplied by Development Center of Water Treatment Technology, State Oceanic Administration, Hangzhou, China. All other reagents and chemicals were of analytical grade and used without further purification. De-ionized water with a
52
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
conductivity of 2 × 10−4 S/m was used for membrane preparation and permeation experiments. Glycolchitin prepared following a method described in the literature [13]. 2.1.2. Apparatus Scanning electron microscopy (SEM) apparatus (JEOL JMS-840, Japan), salt concentrations were determined with a Model DDS-11A conductivity meter (Shanghai Lida Instrument, China), IR spectra were measured with an Aratar360 IR spectrometer purchased from Nicolet, Atomic force microscope Easyscan Nanosurf), Membrane potential apparatus and PMI membrane evaluation apparatus were provided by the Development Center of Water Treatment technology, State Oceanic Administration (Hangzhou, China). 2.2. Preparation of GC /PAN composite NF membranes The casting solution was prepared by dissolving a certain amount (0.6–1.8 wt%) of GC in de-ionized water. After filtered with a G3 sand filter the GC solution was over-coated on the surface-dried PAN UF membrane with a glass rod. The 0.08 mm diameter brass wires attached to the two ends of the glass rod, which determined the thickness of the active functional layer. Then the composite membrane was vaporized at 50 ◦ C for a certain time, and cross-linked with epichlorohydrin (ECH) in acetone aqueous solution in an airtight container. The solvent of ECH is acetone aqueous solution (pH 10). The crosslinked membrane was heat-treated for 20 min at 50 ◦ C again, then washed thoroughly with de-ionized water and immersed in de-ionized water for 24 h. 2.3. Permeation experiments The permeation flux of the membrane was determined by the weight of the permeated fluid through the membrane during a certain period of time and calculated as the following equation: W ×t (1) A where F is the permeate flux, A the effective area of the membrane, t the time for permeation and W is the weight of the permeated fluid passing through the membrane. Rejection was calculated with the following equation: F=
R=1−
CP CM
of individual solutes (Na2 SO4 , K2 SO4 , MgSO4 , NaCl, KCl, MgCl2 ) from their aqueous solution. The performance tests for the prepared membrane were carried out under 0.8 MPa at 25 ◦ C in 1.0 g/L Na2 SO4 solution without specification. The recirculation rate of feed was kept 40 L/h. Both the retentate and permeate were re-circulated to the feed tank to maintain a constant feed concentration during the permeation experiments. Salt concentration were determined with a Model DDS-11A conductivity meter. Before testing the membrane was pre-pressurized for 1.0 h at 0.8 MPa, and the effective membrane area was 19.6 cm2 . 3.1.1. Effect of acetone solvent concentration on rejection and flux of GC/PAN composite NF membranes Acetone is a kind of non-polar solvent and can dissolve in water, it has a low boiling point. So acetone water solution was choose as solvent in this experiment. To investigate the effect of acetone solvent, a series of GC/PAN membranes were prepared by changing solvent acetone water solution concentration from 50 to 83 wt%. As shown in Fig. 1, rejection increased with the increase of acetone concentration until it was 75 wt%, after that it decreased. It could be explained that the polarity of solvent could effect cross-linking reaction [14], Polarities between reactants match well when acetone concentration was 75 wt%. So the optimal acetone concentration was 75 wt%. 3.1.2. Effect of GC casting solution concentration on rejection and flux of GC/PAN composite NF membranes There is necessary to choose a proper GC concentration because viscosity of GC solution is relevant to concentration, and it is not good for preparation of NF membrane if the solution is too dilute or too concentrated. To determine the casting solution concentration, a series of membranes were prepared by changing GC concentration from 0.6 to 1.8 wt%. As shown in Fig. 2, when GC concentration changed from 0.6 to 1.5 wt%, rejection increased linearly from 38.1% to 81%, whereas permeation flux decreased linearly from 10.6 to 3.93 L/m2 h. This is due to the thickness of the selective layer increased with the increase of GC concentration, which
(2)
where CP is the permeate concentration and CM is the feed concentration. Three membrane disks were cut from sheets, and the data presented are the averages of these measurements. 3. Results and discussion 3.1. Effect of preparation parameters upon rejection and flux of GC/PAN composite NF membranes The GC/PAN NF membrane prepared under optimum conditions was characterized by determining the rejection and flux
Fig. 1. Effect of acetone solvent concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.0 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 2. Effect of GC casting solution concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
results in higher rejection and lower permeation flux. But the rejections began to decrease when GC concentration exceeded 1.5 wt%, which might result from the decrease of cross-linking degree of the membrane surface [12]. Considering rejection and flux of Na2 SO4 together, the optimal concentration of GC is 1.5 wt%. 3.1.3. Effect of cross-linking reagent (ECH) concentration on rejection and flux of GC/PAN composite NF membranes ECH is a type of familiar reagent, it can enhance compact degree of the membrane surface. To investigate this preparation parameter, a series of GC/PAN composite NF membranes were prepared by changing ECH concentration from 0.2 to 1.8 wt%. The results are shown in Fig. 3, as ECH concentration changes from 0.2 to 1.0 wt%, rejection increased from 87.4% to 93%, and permeation flux decreased from 15.76 to 11.54 L/m2 h. Which because the reaction was intensified between ECH solution and surface of membrane with the increase of ECH concentration. Resulted in a compact surface and leading to decrease in flux.
Fig. 3. Effect of cross-linking reagent (ECH) concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC: vaporising for 2 h at 50 ◦ C; cross-linking with ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
53
Fig. 4. Effect of cross-linking temperature on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
However, when it exceeds 1.0 wt% rejection decreased and flux increased. It may be explained that reversible reaction or side reaction occurred. Considering rejection and flux of Na2 SO4 together, 1.0 wt% is proper. 3.1.4. Effect of cross-linking temperature on rejection and flux of GC/PAN composite NF membranes For this test, cross-linking temperature varying from 20 to 60 ◦ C were investigated because the boiling point of acetone solution is low. As is shown in Fig. 4, rejection increased from 52.76% to 91.3% and permeation flux decreased from 39.4 to 13.64 L/m2 h when the cross-linking temperature changes from 20 to 50 ◦ C. It could be explained that carbonyls take part into the cross-linking reaction [14] and this reaction will be sufficiency when the temperature is above 40 ◦ C. Besides, the membrane surface formed compact reticular structure as a result of cross-linking, leading to decrease of pore size [15], so rejection increased and permeation flux decreased. However, when it is above 50 ◦ C, the rejection decreased and flux increased as a result of the decrease in the degree of cross-linking due to the decomposition of cross-linking bonds. Considering rejection and flux of Na2 SO4 together, the optimal temperature for cross-linking is 50 ◦ C. 3.1.5. Effect of vaporizing time on rejection and flux of GC/PAN composite NF membranes To investigate the effect, a series of GC/PAN composite NF membranes were prepared with different vaporizing times in the range of 1.5–3.5 h. The result was shown in Fig. 5, permeation flux decreased linearly with the increase of vaporizing time, and rejection get to the highest when it was 2 h. It could be explained that long time of vaporizing can result in a compact surface of composite NF membrane, so flux decreased. However, the surface is so compact that it may be not good for cross-linking, which might decreased the rejection. So considering rejection and flux of Na2 SO4 together, the optimal vaporizing time is 2 h.
54
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 5. Effect of vaporizing time on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
3.1.6. Effect of cross-linking time on rejection and flux of GC/PAN composite NF membranes To investigate this effect, a series of GC/PAN composite NF membranes were prepared when cross-linking time changed from 16 to 32 h. As shown in Fig. 6, rejection increased and permeation flux decreased with the increase of cross-linking time until it was 24 h, which may result from the increase in cross-linking degree. However, when cross-linking time exceeds 24 h, rejection decreased and permeation flux increased. Which because the decomposition of cross-linking bonds occurs, leading to a decreasing rejection and increasing flux. Thus, a cross-linking time of 24 h is proper (Fig. 7). It draws a conclusion from the experiments above that the GC/PAN composite NF membrane prepared from 1.5 wt% GC solution, vaporizing for 2 h at 50 ◦ C, cross-linked at 50 ◦ C with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) and heat-treated at 50 ◦ C for 20 min show excellent properties. The resultant membrane was used to carry out experiments hereinafter.
Fig. 7. IR Spectra of PAN UF membrane (a), GC/PAN composite NF membrane cross-linked with ECH (b) and IR Spectrum of glycolchitin (c).
3.2. Characteristics of the GC/PAN composite NF membrane 3.2.1. IR spectra of GC/PAN membranes and glycolchitin The IR spectra of PAN UF membrane (a), GC/PAN membrane (b) cross-linked with ECH and glycolchitin were measured in the range of 500–3000 cm−1 ,the absorption band at 2242.74 cm−1 was ascribed to the –C N– stretching vibration of PAN, the absorption band at 1729.82 cm−1 in (b) correspond to –COOH bond suggesting PAN UF membrane hydrolyzed some nitrile groups to carboxylic acid groups in the PAN molecular chain, the absorption band of 1454.35 cm−1 in (b) was characteristic of –CH2 scissors vibration [10], which is very small in a. It can be explained that more –CH2 bonds were brought in because of cross-linking. There was a new absorption band at 1166.23 cm−1 in (b) correspond to–C–O–C– bond, suggesting ether linkage was formed as the result of cross-linking with ECH. 3.2.2. Structure characteristic of the GC/PAN composite NF membrane The cross-section and surface of the composite NF membrane were characterized with a scanning electron microscope and the results are shown in Figs. 8 and 9. The two pictures show that there was a thin active functional layer with a compact surface on a finger-like supporting layer of PAN UF base membrane. Suggesting the thin functional layer probably determined rejection performance and permeation flux of GC/PAN composite NF membrane.
Fig. 6. Effect of cross-linking time on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
3.2.3. Roughness measurement with AFM Roughness is one of the most important surface properties because it has a strong influence on adhesion and local mass transfer. Figs. 10 and 11 show typical surface structure and 3D images of the membrane. The colorific intensity shows the vertical profile of the membrane surface. The light regions was for
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
55
Fig. 11. Three-dimensional AFM image of composite NF membrane. Fig. 8. SEM photograph of cross-section of the composite NF membrane.
Fig. 12. The effect of pressure on streaming potential of the composite membrane. Fig. 9. SEM photograph of surface texture of the composite NF membrane.
the highest points and the dark regions was for the depressions. The total scanning area was 1.0796 m2 and total roughness was 0.0191 m. The average roughness of the composite membrane surface was about 0.0108 m [16]. 3.2.4. The streaming potential of GC/PAN composite NF membranes The membrane streaming potential was measured in 0.1 mol KCl aqueous solution filled between the sides of the membrane under the pressure range of 0.1–0.3 MPa. The curve for streaming potential against pressure is shown in Fig. 12,
streaming potential decreased linearly with the increase of operating pressure and the pressure osmobic coefficient (β) was about—2 mV/MPa, suggesting that GC/PAN composite NF membrane s active layer was negatively charged [17]. It could be explained that the membrane surface layer absorbed anions from the electrolyte solution. Besides, PAN UF membrane hydrolyzed some nitrile groups to carboxylic acid groups in the PAN molecular chain [10]. 3.2.5. MWCO of GC/PAN composite NF membrane Molecular weight cut-off of composite membrane was measured with 1.0 g/L aqueous solutions of glucose, sucrose and polyethylene glycols (MW 600–1000) at 25 ◦ C under 1.0 MPa. The concentrations of these neutral organic matters in feed and permeated samples were determined by absorptionmetry, from which rejection can be obtained. The MWCO of membrane was the molecular weight of organic substances with retention of 90% [18]. The curve about rejection against molecular-weight is shown in Fig. 13. Obviously the MWCO of this membrane is approximately 540 Da and the pore size is about 0.73 nm [15]. 3.3. Effects of operating conditions on rejection and flux of GC/PAN composite NF membranes
Fig. 10. Two-dimensional AFM image of composite NF membrane.
For these tests, the resultant membrane was used. Before testing the membrane was pre-pressurized for 1.0 h, and the recirculation rate of feed was kept 40 L/h.
56
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 13. Rejection of the composite NF membrane to neutral organic matter of different molecular weight.
3.3.1. Effect of operating pressure on rejection and flux of GC/PAN composite NF membrane The effect of operating pressure on the permeation flux and rejection of the membrane of 1.0 g/L Na2 SO4 is shown in Fig. 14. Permeation flux increased all of the time with the increase of operating pressure, rejection increased with the increase of operating pressure until it was 1.2 MPa, after that rejection decreased. It can be explained by dissolution–diffusion model [19]: FW = A(P − π)
(3)
where FW is water flux, A the water permeation coefficient, P the operating pressure difference, β the polarization factor of the concentration difference, π is the osmosis pressure. FS = B(βC1 − C2 )
(4)
where FS is salt flux, B the salt permeation coefficient, C1 and C2 are the salt concentrations on the upstream and downstream sides of the membrane, respectively. β=
Cb Cm
(5)
where Cb is the salt concentration on the membrane surface and Cm is the salt concentration of the feed. It can be seen from formula (3) that FW increases linearly with the increase of P. And formula (4) shows that FS is a
Fig. 14. Effect of operating pressure on rejection and flux of the composite membrane.
Fig. 15. Effect of NaCl concentration on rejection and permeation flux of the composite membrane.
function of salt concentration on both sides of the membrane, it has no direct relation to P. Therefore, as P increases, so does the water flux, but salt flux remains constant thereby resulting in an increase in salt rejection. On the other hand, C2 is reduced by virtue of the increase of FW and the concentration difference on the two sides of the membrane increases, that is β(C1 − C2 ) increased, leading to an increase of FS , and the rejection decreased, So these two sides cooperated to make the rejection increased in the first instance and decreased afterward [20]. 3.3.2. Effect of feed solution concentration on rejection and flux of GC/PAN composite NF membranes The effect of feed concentration on flux and rejection of the membrane of NaCl and Na2 SO4 is shown in Figs. 15 and 16. It can be seen that both rejections decreased and permeation fluxes decreased with the increase of NaCl and Na2 SO4 concentration, which can be explained by Donnan effect. The concentration of Na+ on the membrane surface increased with the increase of feed concentration, which result in increasing of permeation rate of Na+ , so the electro-neutrality between the two sides of the membrane was broken, and in order to maintain the electroneutrality, more SO4 2− and Cl− will permeate from upstream side to downstream side, thereby those rejections decreased. But
Fig. 16. Effect of Na2 SO4 concentration on rejection and permeation flux of the membrane.
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
57
Table 1 Rejections and permeation fluxes of the composite membrane to different types of feed solutions Type of feeds
Na2 SO4
K2 SO4
MgSO4
NaCl
KCl
MgCl2
Rejection (%) Permeation flux (L/m2 h)
95.3 10.0
91.5 10.1
41.0 19.2
31.1 19.7
31.0 20.8
20.1 13.0
the permeation velocity of Cl− is quicker than that of SO4 2− , So the rejection of NaCl solution decreased more than that of Na2 SO4 solution. 3.3.3. Effect of different type of feeds on rejection and flux of GC/PAN composite NF membrane To test the rejection performance of GC/PAN composite membrane, NaCl, KC1, MgCl2 , Na2 SO4 , MgSO4 and K2 SO4 solutions were used as feed solutions. The test was carried out at 25 ◦ C under 1.0 MPa and the concentration of feed solution is 1.0 g/L. The result was shown in Table 1, the rejections to different inorganic electrolytes follow the order of Na2 SO4 > K2 SO4 > MgSO4 > NaCl > KCl > MgCl2 , which shows the static electrification characteristic of negatively charged NF membrane [21]. The active functional layer of the composite membrane has stronger repulsion to SO4 2− than Cl− . Therefore, SO4 2− was rejected, so the rejection order is NaCl ≈ KCl < MgSO4 . On the other hand, Mg2+ is easy to get combination with anions on the membrane surface which maybe decrease the effective surface charge of the membrane, and so MgSO4 < Na2 SO4 ≈ K2 SO4 and MgCl2 < NaCl ≈ KCl. To compare with other NF membranes [22,23], the GC/PAN NF membrane has higher selective rejections. For example at the operating pressure of 1.0 MPa and 25 ◦ C, the rejections to Na2 SO4 and K2 SO4 approach 95.3% and 91.5%, whereas the ones to the other model solutions are lower than 13.0%. Based on its excellent selective rejections, it can be used to separate mono/divalent salts. 4. Conclusions A novel composite NF membrane with excellent performance have been prepared from 1.5 wt% casting solution by vaporizing for 2 h and cross-linking with ECH at 50 for 24 h in 75 wt% acetone solution (pH 10). The structure of the resultant membrane was characterized by SEM and AFM. In addition, its molecular weight cut-off was obtained at 540 Da or so, with a pore size of 0.73 nm. In addition, some performance tests have been performed for the composite NF membrane. It was found its properties were influenced to a great extent by feed concentration and the type of model solutions. The order of their rejections is Na2 SO4 > K2 SO4 > MgSO4 > NaCl > KCl > MgCl2 , which is typical characteristics of a negatively charged membrane. Additionally, the curve about the streaming potential illustrates the negatively charged characteristics of this membrane, with a pressure osmobic coefficient of −2 mV/MPa. The membrane has a high rejection to divalent salts and can be used to wastewater treatment in the future.
Acknowledgements Financial support from the major state Basic Research Development Program of China (973) program, No. 2003 CB615706) was gratefully acknowledged. References [1] W.R. Bowen, J.S. Welfoot, Modelling the performance of membrane nanofiltration-critical assessment and model development, Chem. Eng. Sci. 57 (2002) 1121–1137. [2] W.R. Bowen, A.W. Mohammad, A theoretical basis for specifying nanofiltration membranes—dye/salt/water streams, Desalination 117 (1998) 257–264. [3] B. Nicolaisen, Developments in membrane technology for water treatment, Desalination 153 (1–3) (2003) 355–360. [4] G.-L. Huan, Y.-F. Zhang, Q.-Y. Du, Y. Wu, Optimization of composite conditions for composite nanofiltration membrane by orthogonal design, Membr. Sci. Technol. 24 (5) (2004) 77–79 (in Chinese). [5] T.-D. Jiang, Chitin, Environment Science Book Press, Beijing, 1995, in Chinese. [6] T. Yang, R.R. Eall, Chiton memberans for reverse osmosis application, J. Food Sci. 49 (1984) 91–93. [7] D.A. Musale, A. Kumar, Effects of surface crosslinking on sieving characteristics of chitosan/poly(acrylonitrile) composite nanofiltration membranes, Sep. Purif. Technol. 21 (2000) 27–38. [8] S.-Z. Tan, Z.-H. Chen, Z.-H. Chen, Preparation of polyacrynitrile nanofiltration membrane and its separation property in the treatment of papermaking effluent, Trans. China Pulp Paper 17 (2) (2002) 27–38 (in Chinese). [9] Z. Zhiping, W. Zhi, N. ye, et al., A novel N,O-carboxymothyl amphoteric chitosan/poly(ethersufone) composite MF membrane and its charged characteristics, Desalination 144 (2002) 35–39. [10] D.A. Musal, A. Kumar, G. Pleizier, Formation and characterization of poly(acrynitrile)/chitosan composite ultrafiltration membranes, J. Membr. Sci. 154 (1999) 163–173. [11] J. Zhou, B.-H. Sun, C.-X. Chen, Optimizing the conditions of extracting berberine hydrochloride by liquid membrane with the method of orthoyonal design, Membr. Sci. Technol. 21 (1) (2001) 21–24 (in Chinese). [12] J. Miao, G.-H. Chen, C.-J. Gao, A novel kind of amphoteric composite nanofiltration membrane prepared from sulfated chitosan (SCS), Desalination 181 (2005) 173–183. [13] Y. Zhao, G.-H. Chen, X.-L. Lu, X.-H. Lv, C.-J. Gao, Study on preparation and properties of temperature and pH-sensitive hydroxythyl chitin hydrogels, Period. Ocean Univ. China 35 (2) (2005) 309–312 (in Chinese). [14] H.-Y. Zhu, L. Xiao, Z.-X. Liu, Influence of solvent treatment on the properties and structure of chitosan films, Polym. Mater. Sci. Eng. 20 (2) (2004) 110–113 (in Chinese). [15] B.O. Krajewska, A. Olech, Pore structure of gel chitosan membranes. I. Solute diffusion measurements, Poly. Gels Networks 4 (1996) 33–43. [16] S.A.L.-J. shi, A. Neville, An investigation into the relationship between flux and roughness on RO membranes using scanning probe microscopy, Desalination 189 (2006) 221–228. [17] J.-X. Mo, S.-M. Liu, Theory experimental method and result of the streaming potential, Technol. Water Treat. 17 (3) (1991) 153–161 (in Chinese).
58
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
[18] S.-H. Wu, H. He, S.-Q. Yuan, Barium chloride method and its application in determination of polysulphone membrane performance, Technol. Water Treat. 21 (4) (1995) 193–197 (in Chinese). [19] Y.-J. Song, B.-H. Sun, Study of factors affecting separation performance of nanofiltration membrane, Technol. Water Treat. 23 (2) (1997) 78–82 (in Chinese). [20] X.-M. Liang, X.-F. Lu, L.-Q. Shi, Z.-Y. Liu, The preparation of polyarylate composite nanofiltration membrane by interfacial polymerization, J. East China Univ. Sci. Technol. 25 (5) (1999) 489–493 (in Chinese).
[21] R.J. Peterson, Composite reverse osmosis and nanofiltration membranes, J. Membr. Sci. 83 (1993) 81–150. [22] W. Pronk, H. Palmquist, M. Biebow, M. Boller, Nanofiltration for the separation of pharmaceuticals from nutrients in source-separated urine, Water Res. 40 (2006) 1405–1412. [23] L.-C. Li, C.-L. Ma, H.-M. Tan, T.-L. Chen, Preparation of sulfonated phenolphthalein poly(etherketone) composite nanofiltration membrane, Technol. Water Treat. 31 (10) (2005) 13–16 (in Chinese).
A novel composite nanofiltration (NF) membrane prepared from glycolchitin/poly(acrylonitrile) (PAN) by epichlorohydrin cross-linking Hongwei Sun, Guohua Chen ∗ , Ruihua Huang, Congjie Gao College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, China Received 10 October 2006; received in revised form 9 February 2007; accepted 18 February 2007 Available online 21 February 2007
Abstract A novel composite nanofiltration (NF) membrane was prepared by over-coating the PAN ultrafiltration (UF) membrane with a glycolchitin (GC) thin layer. The effects of the membrane preparation techniques and operating conditions on the rejection performance of the composite membranes were studied. The structure of the composite NF membrane was characterized by scanning electron microscopy (SEM) and infrared spectroscopy (IR). And the surface roughness of the composite membrane was also characterized by atomic force microscope (AFM). The pressure osmobic coefficient (β) was about −2 mV/MPa and the molecular weight cut-off (MWCO) was obtained at 540 Da or so. Rejections of the composite membrane of Na2 SO4 , K2 SO4 , MgSO4 , NaCl, KCl, MgCl2 solutions (1.0 g/L) were 95.2%, 91.5%, 41%, 31.1%, 31% and 20.1% at 25 ◦ C under 1.0 MPa, and permeation fluxes are 10.0, 10.1, 19.2, 19.7, 20.8 and 13.0 L/m2 h, respectively. The results suggest that rejection performance is governed by solute–membrane and solute–solute electrostatic interactions. This charge distribution mainly determined the performances of the membranes. © 2007 Elsevier B.V. All rights reserved. Keywords: Polyacrylontrile UF membranes; Glycolchitin; Epichlorohydrin; Composite NF membranes; Molecular weight cut-off; Rejection performances
1. Introduction Nanofiltration membranes based technological process is between ultrafiltration membranes and reverse osmosis (RO) membranes. NF membrane’s molecular weight cut-off ranges from 200 to 1000 Da and it’s pore size is nominally 1 nm in dimension. The separation mechanism of NF membrane is a combination of Donnan effect and sieving effect. More specifically, NF membrane can be used to remove small neutral organic molecules while surface electrostatic properties allowed monovalent ions to be reasonably well transmitted with multivalent ions mostly retained [1]. NF membrane’s operating pressure ranged from 0.5 to 1.5 MPa, which was much lower than RO membranes. NF membrane have been applied widely in many industry fields such as purification for drinking water [2] and industry wastewater [3]. Research on composite membrane and its application are highly attention-getting in recent years. Especially composite technique for making NF membranes were mostly commercial and has the most yield [4]. ∗
Corresponding author. Tel.: +86 532 82032223; fax: + 86 595 2686969. E-mail addresses: [email protected], [email protected] (H. Sun), [email protected] (G. Chen). 0376-7388/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2007.02.027
Chitin is not a hydrophilic, but some of its derivatives are soluble in water, it shows excellent membrane forming characteristics and biocompatibility. So chitin can be grafted to different molecular groups for preparing composite membrane of different capabilities [5]. Chitosan in acetic acid has been used to prepare reverse osmosis [6], nanofiltration [7,8], microfiltration [9], ultrafiltration [10] and pervaporation membranes [11]. To our knowledge that there is no reported literature on using water-soluble derivatives of chitin to prepare composite NF membrane except us [12]. In this study a composite NF membrane was prepared using glycolchitin (GC) as filming-forming material and epichlorohydrin as cross-linking reagent. 2. Experiments 2.1. Materials and apparatus 2.1.1. Materials Chitin (supplied by Haihui Bioengineering Limited, Qingdao, China), PAN UF Membrane with MWCO of 1 × 105 Da was supplied by Development Center of Water Treatment Technology, State Oceanic Administration, Hangzhou, China. All other reagents and chemicals were of analytical grade and used without further purification. De-ionized water with a
52
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
conductivity of 2 × 10−4 S/m was used for membrane preparation and permeation experiments. Glycolchitin prepared following a method described in the literature [13]. 2.1.2. Apparatus Scanning electron microscopy (SEM) apparatus (JEOL JMS-840, Japan), salt concentrations were determined with a Model DDS-11A conductivity meter (Shanghai Lida Instrument, China), IR spectra were measured with an Aratar360 IR spectrometer purchased from Nicolet, Atomic force microscope Easyscan Nanosurf), Membrane potential apparatus and PMI membrane evaluation apparatus were provided by the Development Center of Water Treatment technology, State Oceanic Administration (Hangzhou, China). 2.2. Preparation of GC /PAN composite NF membranes The casting solution was prepared by dissolving a certain amount (0.6–1.8 wt%) of GC in de-ionized water. After filtered with a G3 sand filter the GC solution was over-coated on the surface-dried PAN UF membrane with a glass rod. The 0.08 mm diameter brass wires attached to the two ends of the glass rod, which determined the thickness of the active functional layer. Then the composite membrane was vaporized at 50 ◦ C for a certain time, and cross-linked with epichlorohydrin (ECH) in acetone aqueous solution in an airtight container. The solvent of ECH is acetone aqueous solution (pH 10). The crosslinked membrane was heat-treated for 20 min at 50 ◦ C again, then washed thoroughly with de-ionized water and immersed in de-ionized water for 24 h. 2.3. Permeation experiments The permeation flux of the membrane was determined by the weight of the permeated fluid through the membrane during a certain period of time and calculated as the following equation: W ×t (1) A where F is the permeate flux, A the effective area of the membrane, t the time for permeation and W is the weight of the permeated fluid passing through the membrane. Rejection was calculated with the following equation: F=
R=1−
CP CM
of individual solutes (Na2 SO4 , K2 SO4 , MgSO4 , NaCl, KCl, MgCl2 ) from their aqueous solution. The performance tests for the prepared membrane were carried out under 0.8 MPa at 25 ◦ C in 1.0 g/L Na2 SO4 solution without specification. The recirculation rate of feed was kept 40 L/h. Both the retentate and permeate were re-circulated to the feed tank to maintain a constant feed concentration during the permeation experiments. Salt concentration were determined with a Model DDS-11A conductivity meter. Before testing the membrane was pre-pressurized for 1.0 h at 0.8 MPa, and the effective membrane area was 19.6 cm2 . 3.1.1. Effect of acetone solvent concentration on rejection and flux of GC/PAN composite NF membranes Acetone is a kind of non-polar solvent and can dissolve in water, it has a low boiling point. So acetone water solution was choose as solvent in this experiment. To investigate the effect of acetone solvent, a series of GC/PAN membranes were prepared by changing solvent acetone water solution concentration from 50 to 83 wt%. As shown in Fig. 1, rejection increased with the increase of acetone concentration until it was 75 wt%, after that it decreased. It could be explained that the polarity of solvent could effect cross-linking reaction [14], Polarities between reactants match well when acetone concentration was 75 wt%. So the optimal acetone concentration was 75 wt%. 3.1.2. Effect of GC casting solution concentration on rejection and flux of GC/PAN composite NF membranes There is necessary to choose a proper GC concentration because viscosity of GC solution is relevant to concentration, and it is not good for preparation of NF membrane if the solution is too dilute or too concentrated. To determine the casting solution concentration, a series of membranes were prepared by changing GC concentration from 0.6 to 1.8 wt%. As shown in Fig. 2, when GC concentration changed from 0.6 to 1.5 wt%, rejection increased linearly from 38.1% to 81%, whereas permeation flux decreased linearly from 10.6 to 3.93 L/m2 h. This is due to the thickness of the selective layer increased with the increase of GC concentration, which
(2)
where CP is the permeate concentration and CM is the feed concentration. Three membrane disks were cut from sheets, and the data presented are the averages of these measurements. 3. Results and discussion 3.1. Effect of preparation parameters upon rejection and flux of GC/PAN composite NF membranes The GC/PAN NF membrane prepared under optimum conditions was characterized by determining the rejection and flux
Fig. 1. Effect of acetone solvent concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.0 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 2. Effect of GC casting solution concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
results in higher rejection and lower permeation flux. But the rejections began to decrease when GC concentration exceeded 1.5 wt%, which might result from the decrease of cross-linking degree of the membrane surface [12]. Considering rejection and flux of Na2 SO4 together, the optimal concentration of GC is 1.5 wt%. 3.1.3. Effect of cross-linking reagent (ECH) concentration on rejection and flux of GC/PAN composite NF membranes ECH is a type of familiar reagent, it can enhance compact degree of the membrane surface. To investigate this preparation parameter, a series of GC/PAN composite NF membranes were prepared by changing ECH concentration from 0.2 to 1.8 wt%. The results are shown in Fig. 3, as ECH concentration changes from 0.2 to 1.0 wt%, rejection increased from 87.4% to 93%, and permeation flux decreased from 15.76 to 11.54 L/m2 h. Which because the reaction was intensified between ECH solution and surface of membrane with the increase of ECH concentration. Resulted in a compact surface and leading to decrease in flux.
Fig. 3. Effect of cross-linking reagent (ECH) concentration on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC: vaporising for 2 h at 50 ◦ C; cross-linking with ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
53
Fig. 4. Effect of cross-linking temperature on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
However, when it exceeds 1.0 wt% rejection decreased and flux increased. It may be explained that reversible reaction or side reaction occurred. Considering rejection and flux of Na2 SO4 together, 1.0 wt% is proper. 3.1.4. Effect of cross-linking temperature on rejection and flux of GC/PAN composite NF membranes For this test, cross-linking temperature varying from 20 to 60 ◦ C were investigated because the boiling point of acetone solution is low. As is shown in Fig. 4, rejection increased from 52.76% to 91.3% and permeation flux decreased from 39.4 to 13.64 L/m2 h when the cross-linking temperature changes from 20 to 50 ◦ C. It could be explained that carbonyls take part into the cross-linking reaction [14] and this reaction will be sufficiency when the temperature is above 40 ◦ C. Besides, the membrane surface formed compact reticular structure as a result of cross-linking, leading to decrease of pore size [15], so rejection increased and permeation flux decreased. However, when it is above 50 ◦ C, the rejection decreased and flux increased as a result of the decrease in the degree of cross-linking due to the decomposition of cross-linking bonds. Considering rejection and flux of Na2 SO4 together, the optimal temperature for cross-linking is 50 ◦ C. 3.1.5. Effect of vaporizing time on rejection and flux of GC/PAN composite NF membranes To investigate the effect, a series of GC/PAN composite NF membranes were prepared with different vaporizing times in the range of 1.5–3.5 h. The result was shown in Fig. 5, permeation flux decreased linearly with the increase of vaporizing time, and rejection get to the highest when it was 2 h. It could be explained that long time of vaporizing can result in a compact surface of composite NF membrane, so flux decreased. However, the surface is so compact that it may be not good for cross-linking, which might decreased the rejection. So considering rejection and flux of Na2 SO4 together, the optimal vaporizing time is 2 h.
54
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 5. Effect of vaporizing time on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C for 22 h; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
3.1.6. Effect of cross-linking time on rejection and flux of GC/PAN composite NF membranes To investigate this effect, a series of GC/PAN composite NF membranes were prepared when cross-linking time changed from 16 to 32 h. As shown in Fig. 6, rejection increased and permeation flux decreased with the increase of cross-linking time until it was 24 h, which may result from the increase in cross-linking degree. However, when cross-linking time exceeds 24 h, rejection decreased and permeation flux increased. Which because the decomposition of cross-linking bonds occurs, leading to a decreasing rejection and increasing flux. Thus, a cross-linking time of 24 h is proper (Fig. 7). It draws a conclusion from the experiments above that the GC/PAN composite NF membrane prepared from 1.5 wt% GC solution, vaporizing for 2 h at 50 ◦ C, cross-linked at 50 ◦ C with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) and heat-treated at 50 ◦ C for 20 min show excellent properties. The resultant membrane was used to carry out experiments hereinafter.
Fig. 7. IR Spectra of PAN UF membrane (a), GC/PAN composite NF membrane cross-linked with ECH (b) and IR Spectrum of glycolchitin (c).
3.2. Characteristics of the GC/PAN composite NF membrane 3.2.1. IR spectra of GC/PAN membranes and glycolchitin The IR spectra of PAN UF membrane (a), GC/PAN membrane (b) cross-linked with ECH and glycolchitin were measured in the range of 500–3000 cm−1 ,the absorption band at 2242.74 cm−1 was ascribed to the –C N– stretching vibration of PAN, the absorption band at 1729.82 cm−1 in (b) correspond to –COOH bond suggesting PAN UF membrane hydrolyzed some nitrile groups to carboxylic acid groups in the PAN molecular chain, the absorption band of 1454.35 cm−1 in (b) was characteristic of –CH2 scissors vibration [10], which is very small in a. It can be explained that more –CH2 bonds were brought in because of cross-linking. There was a new absorption band at 1166.23 cm−1 in (b) correspond to–C–O–C– bond, suggesting ether linkage was formed as the result of cross-linking with ECH. 3.2.2. Structure characteristic of the GC/PAN composite NF membrane The cross-section and surface of the composite NF membrane were characterized with a scanning electron microscope and the results are shown in Figs. 8 and 9. The two pictures show that there was a thin active functional layer with a compact surface on a finger-like supporting layer of PAN UF base membrane. Suggesting the thin functional layer probably determined rejection performance and permeation flux of GC/PAN composite NF membrane.
Fig. 6. Effect of cross-linking time on rejection and flux of GC/PAN membrane. Preparation conditions of membrane: 1.5 wt% GC; vaporizing for 2 h at 50 ◦ C; cross-linking with 1.0 wt% ECH in 75 wt% acetone solution (pH 10) at 50 ◦ C; heat-treated for 20 min at 50 ◦ C. Operation: 1.0 g/L Na2 SO4 solution; at the operation pressure of 1.0 MPa and cycling flow of 40 L/h.
3.2.3. Roughness measurement with AFM Roughness is one of the most important surface properties because it has a strong influence on adhesion and local mass transfer. Figs. 10 and 11 show typical surface structure and 3D images of the membrane. The colorific intensity shows the vertical profile of the membrane surface. The light regions was for
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
55
Fig. 11. Three-dimensional AFM image of composite NF membrane. Fig. 8. SEM photograph of cross-section of the composite NF membrane.
Fig. 12. The effect of pressure on streaming potential of the composite membrane. Fig. 9. SEM photograph of surface texture of the composite NF membrane.
the highest points and the dark regions was for the depressions. The total scanning area was 1.0796 m2 and total roughness was 0.0191 m. The average roughness of the composite membrane surface was about 0.0108 m [16]. 3.2.4. The streaming potential of GC/PAN composite NF membranes The membrane streaming potential was measured in 0.1 mol KCl aqueous solution filled between the sides of the membrane under the pressure range of 0.1–0.3 MPa. The curve for streaming potential against pressure is shown in Fig. 12,
streaming potential decreased linearly with the increase of operating pressure and the pressure osmobic coefficient (β) was about—2 mV/MPa, suggesting that GC/PAN composite NF membrane s active layer was negatively charged [17]. It could be explained that the membrane surface layer absorbed anions from the electrolyte solution. Besides, PAN UF membrane hydrolyzed some nitrile groups to carboxylic acid groups in the PAN molecular chain [10]. 3.2.5. MWCO of GC/PAN composite NF membrane Molecular weight cut-off of composite membrane was measured with 1.0 g/L aqueous solutions of glucose, sucrose and polyethylene glycols (MW 600–1000) at 25 ◦ C under 1.0 MPa. The concentrations of these neutral organic matters in feed and permeated samples were determined by absorptionmetry, from which rejection can be obtained. The MWCO of membrane was the molecular weight of organic substances with retention of 90% [18]. The curve about rejection against molecular-weight is shown in Fig. 13. Obviously the MWCO of this membrane is approximately 540 Da and the pore size is about 0.73 nm [15]. 3.3. Effects of operating conditions on rejection and flux of GC/PAN composite NF membranes
Fig. 10. Two-dimensional AFM image of composite NF membrane.
For these tests, the resultant membrane was used. Before testing the membrane was pre-pressurized for 1.0 h, and the recirculation rate of feed was kept 40 L/h.
56
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
Fig. 13. Rejection of the composite NF membrane to neutral organic matter of different molecular weight.
3.3.1. Effect of operating pressure on rejection and flux of GC/PAN composite NF membrane The effect of operating pressure on the permeation flux and rejection of the membrane of 1.0 g/L Na2 SO4 is shown in Fig. 14. Permeation flux increased all of the time with the increase of operating pressure, rejection increased with the increase of operating pressure until it was 1.2 MPa, after that rejection decreased. It can be explained by dissolution–diffusion model [19]: FW = A(P − π)
(3)
where FW is water flux, A the water permeation coefficient, P the operating pressure difference, β the polarization factor of the concentration difference, π is the osmosis pressure. FS = B(βC1 − C2 )
(4)
where FS is salt flux, B the salt permeation coefficient, C1 and C2 are the salt concentrations on the upstream and downstream sides of the membrane, respectively. β=
Cb Cm
(5)
where Cb is the salt concentration on the membrane surface and Cm is the salt concentration of the feed. It can be seen from formula (3) that FW increases linearly with the increase of P. And formula (4) shows that FS is a
Fig. 14. Effect of operating pressure on rejection and flux of the composite membrane.
Fig. 15. Effect of NaCl concentration on rejection and permeation flux of the composite membrane.
function of salt concentration on both sides of the membrane, it has no direct relation to P. Therefore, as P increases, so does the water flux, but salt flux remains constant thereby resulting in an increase in salt rejection. On the other hand, C2 is reduced by virtue of the increase of FW and the concentration difference on the two sides of the membrane increases, that is β(C1 − C2 ) increased, leading to an increase of FS , and the rejection decreased, So these two sides cooperated to make the rejection increased in the first instance and decreased afterward [20]. 3.3.2. Effect of feed solution concentration on rejection and flux of GC/PAN composite NF membranes The effect of feed concentration on flux and rejection of the membrane of NaCl and Na2 SO4 is shown in Figs. 15 and 16. It can be seen that both rejections decreased and permeation fluxes decreased with the increase of NaCl and Na2 SO4 concentration, which can be explained by Donnan effect. The concentration of Na+ on the membrane surface increased with the increase of feed concentration, which result in increasing of permeation rate of Na+ , so the electro-neutrality between the two sides of the membrane was broken, and in order to maintain the electroneutrality, more SO4 2− and Cl− will permeate from upstream side to downstream side, thereby those rejections decreased. But
Fig. 16. Effect of Na2 SO4 concentration on rejection and permeation flux of the membrane.
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
57
Table 1 Rejections and permeation fluxes of the composite membrane to different types of feed solutions Type of feeds
Na2 SO4
K2 SO4
MgSO4
NaCl
KCl
MgCl2
Rejection (%) Permeation flux (L/m2 h)
95.3 10.0
91.5 10.1
41.0 19.2
31.1 19.7
31.0 20.8
20.1 13.0
the permeation velocity of Cl− is quicker than that of SO4 2− , So the rejection of NaCl solution decreased more than that of Na2 SO4 solution. 3.3.3. Effect of different type of feeds on rejection and flux of GC/PAN composite NF membrane To test the rejection performance of GC/PAN composite membrane, NaCl, KC1, MgCl2 , Na2 SO4 , MgSO4 and K2 SO4 solutions were used as feed solutions. The test was carried out at 25 ◦ C under 1.0 MPa and the concentration of feed solution is 1.0 g/L. The result was shown in Table 1, the rejections to different inorganic electrolytes follow the order of Na2 SO4 > K2 SO4 > MgSO4 > NaCl > KCl > MgCl2 , which shows the static electrification characteristic of negatively charged NF membrane [21]. The active functional layer of the composite membrane has stronger repulsion to SO4 2− than Cl− . Therefore, SO4 2− was rejected, so the rejection order is NaCl ≈ KCl < MgSO4 . On the other hand, Mg2+ is easy to get combination with anions on the membrane surface which maybe decrease the effective surface charge of the membrane, and so MgSO4 < Na2 SO4 ≈ K2 SO4 and MgCl2 < NaCl ≈ KCl. To compare with other NF membranes [22,23], the GC/PAN NF membrane has higher selective rejections. For example at the operating pressure of 1.0 MPa and 25 ◦ C, the rejections to Na2 SO4 and K2 SO4 approach 95.3% and 91.5%, whereas the ones to the other model solutions are lower than 13.0%. Based on its excellent selective rejections, it can be used to separate mono/divalent salts. 4. Conclusions A novel composite NF membrane with excellent performance have been prepared from 1.5 wt% casting solution by vaporizing for 2 h and cross-linking with ECH at 50 for 24 h in 75 wt% acetone solution (pH 10). The structure of the resultant membrane was characterized by SEM and AFM. In addition, its molecular weight cut-off was obtained at 540 Da or so, with a pore size of 0.73 nm. In addition, some performance tests have been performed for the composite NF membrane. It was found its properties were influenced to a great extent by feed concentration and the type of model solutions. The order of their rejections is Na2 SO4 > K2 SO4 > MgSO4 > NaCl > KCl > MgCl2 , which is typical characteristics of a negatively charged membrane. Additionally, the curve about the streaming potential illustrates the negatively charged characteristics of this membrane, with a pressure osmobic coefficient of −2 mV/MPa. The membrane has a high rejection to divalent salts and can be used to wastewater treatment in the future.
Acknowledgements Financial support from the major state Basic Research Development Program of China (973) program, No. 2003 CB615706) was gratefully acknowledged. References [1] W.R. Bowen, J.S. Welfoot, Modelling the performance of membrane nanofiltration-critical assessment and model development, Chem. Eng. Sci. 57 (2002) 1121–1137. [2] W.R. Bowen, A.W. Mohammad, A theoretical basis for specifying nanofiltration membranes—dye/salt/water streams, Desalination 117 (1998) 257–264. [3] B. Nicolaisen, Developments in membrane technology for water treatment, Desalination 153 (1–3) (2003) 355–360. [4] G.-L. Huan, Y.-F. Zhang, Q.-Y. Du, Y. Wu, Optimization of composite conditions for composite nanofiltration membrane by orthogonal design, Membr. Sci. Technol. 24 (5) (2004) 77–79 (in Chinese). [5] T.-D. Jiang, Chitin, Environment Science Book Press, Beijing, 1995, in Chinese. [6] T. Yang, R.R. Eall, Chiton memberans for reverse osmosis application, J. Food Sci. 49 (1984) 91–93. [7] D.A. Musale, A. Kumar, Effects of surface crosslinking on sieving characteristics of chitosan/poly(acrylonitrile) composite nanofiltration membranes, Sep. Purif. Technol. 21 (2000) 27–38. [8] S.-Z. Tan, Z.-H. Chen, Z.-H. Chen, Preparation of polyacrynitrile nanofiltration membrane and its separation property in the treatment of papermaking effluent, Trans. China Pulp Paper 17 (2) (2002) 27–38 (in Chinese). [9] Z. Zhiping, W. Zhi, N. ye, et al., A novel N,O-carboxymothyl amphoteric chitosan/poly(ethersufone) composite MF membrane and its charged characteristics, Desalination 144 (2002) 35–39. [10] D.A. Musal, A. Kumar, G. Pleizier, Formation and characterization of poly(acrynitrile)/chitosan composite ultrafiltration membranes, J. Membr. Sci. 154 (1999) 163–173. [11] J. Zhou, B.-H. Sun, C.-X. Chen, Optimizing the conditions of extracting berberine hydrochloride by liquid membrane with the method of orthoyonal design, Membr. Sci. Technol. 21 (1) (2001) 21–24 (in Chinese). [12] J. Miao, G.-H. Chen, C.-J. Gao, A novel kind of amphoteric composite nanofiltration membrane prepared from sulfated chitosan (SCS), Desalination 181 (2005) 173–183. [13] Y. Zhao, G.-H. Chen, X.-L. Lu, X.-H. Lv, C.-J. Gao, Study on preparation and properties of temperature and pH-sensitive hydroxythyl chitin hydrogels, Period. Ocean Univ. China 35 (2) (2005) 309–312 (in Chinese). [14] H.-Y. Zhu, L. Xiao, Z.-X. Liu, Influence of solvent treatment on the properties and structure of chitosan films, Polym. Mater. Sci. Eng. 20 (2) (2004) 110–113 (in Chinese). [15] B.O. Krajewska, A. Olech, Pore structure of gel chitosan membranes. I. Solute diffusion measurements, Poly. Gels Networks 4 (1996) 33–43. [16] S.A.L.-J. shi, A. Neville, An investigation into the relationship between flux and roughness on RO membranes using scanning probe microscopy, Desalination 189 (2006) 221–228. [17] J.-X. Mo, S.-M. Liu, Theory experimental method and result of the streaming potential, Technol. Water Treat. 17 (3) (1991) 153–161 (in Chinese).
58
H. Sun et al. / Journal of Membrane Science 297 (2007) 51–58
[18] S.-H. Wu, H. He, S.-Q. Yuan, Barium chloride method and its application in determination of polysulphone membrane performance, Technol. Water Treat. 21 (4) (1995) 193–197 (in Chinese). [19] Y.-J. Song, B.-H. Sun, Study of factors affecting separation performance of nanofiltration membrane, Technol. Water Treat. 23 (2) (1997) 78–82 (in Chinese). [20] X.-M. Liang, X.-F. Lu, L.-Q. Shi, Z.-Y. Liu, The preparation of polyarylate composite nanofiltration membrane by interfacial polymerization, J. East China Univ. Sci. Technol. 25 (5) (1999) 489–493 (in Chinese).
[21] R.J. Peterson, Composite reverse osmosis and nanofiltration membranes, J. Membr. Sci. 83 (1993) 81–150. [22] W. Pronk, H. Palmquist, M. Biebow, M. Boller, Nanofiltration for the separation of pharmaceuticals from nutrients in source-separated urine, Water Res. 40 (2006) 1405–1412. [23] L.-C. Li, C.-L. Ma, H.-M. Tan, T.-L. Chen, Preparation of sulfonated phenolphthalein poly(etherketone) composite nanofiltration membrane, Technol. Water Treat. 31 (10) (2005) 13–16 (in Chinese).