Polyamide Ultrafiltration Membranes Modified with ... - Springer Link

3 downloads 0 Views 158KB Size Report
Polyamide Ultrafiltration Membranes Modified with Nanocarbon Additives. G. A. Polotskaya, A. V. Pen'kova, N. N. Sudareva, A. E. Polotskii, and A. M. Toikka.
ISSN 1070-4272, Russian Journal of Applied Chemistry, 2008, Vol. 81, No. 2, pp. 236!240. + Pleiades Publishing, Ltd., 2008. Original Russian Text + G.A. Polotskaya, A.V. Pen’kova, N.N. Sudareva, A.E. Polotskii, A.M. Toikka, 2008, published in Zhurnal Prikladnoi Khimii, 2008, Vol. 81, No. 2, pp. 246 !250.

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

PHYSICOCHEMICAL STUDIES OF SYSTEMS AND PROCESSES

ÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍÍ

Polyamide Ultrafiltration Membranes Modified with Nanocarbon Additives G. A. Polotskaya, A. V. Pen’kova, N. N. Sudareva, A. E. Polotskii, and A. M. Toikka Institute of Macromolecular Compounds, Russian Academy of Sciences, St. Petersburg, Russia St. Petersburg State University, St. Petersburg, Russia State Research Institute of Ultrapure Biopreparations, St. Petersburg, Russia Received July 24, 2007

-

Abstract Possibility of modification of polyphenyleneisophthalamide with 5 wt % nanocarbon additive (fullerene, astralene, or graphite soot) was studied. The resulting composites were used to prepare asymmetric membranes, and the effect of carbon additives on the performance of these membranes was examined. Finally, the membranes were tested using a mixture of proteins of various molecular weights. DOI: 10.1134/S1070427208020146

Ultrafiltration membrane processes are widely used in microbiology, food industry, pharmaceutics, and other branches of industry, and also for solving environmental problems [1]. The diversity of systems to be separated requires extension of the spectrum of membrane materials. Modification of polymers commonly used as membrane materials shows promise as a way to develop novel membranes.

of fullerene soot. This material contains 20 to 90 wt % fulleroid nanoparticles (primarily with cage structure): multi- and single-wall nanotubes, nanobarrels, nanobulbs, and nanocones [10]. There is scarce information on gas-separation and pervaporation membranes based on polymer nanocomposites with fullerene [11313]. Preparation and characterization of ultrafilters with membranes of fullerene-containing polyphenylene oxide were reported in [14]. Despite that fullerenes are produced on a commercial scale in a number of countries, only protonconducting Nafion 117 fuel-cell membranes with addition of C60 fullerene are ready for commercial introduction [15].

Among the common commercial polymers used for production of fibers and membranes, a firm place is held by aromatic polyamides and, in particular, products of condensation of isoterephthaloyl chloride with p- and m-phenylenediamine. The structure and properties of polyphenyleneisophthalamide (PA), as well as some features of supramolecular structures in its concentrated solutions, were reported in [2, 3]. The formation of polyamide-based ultrafiltration membranes was studied by Bil’dyukevich et al. [4 3 6].

The goals of this study were to prepare polyamide composites with nanocarbon additives, to form asymmetric membranes on their basis, and to examine the effect of nanocarbon additives (graphite soot, C60, and astralene) on the ultrafiltration performance of the membranes.

In this study, we modified PA with nanocarbon additives: fullerene C60, astralene, and graphite soot, which are essentially different in their structure and properties. Graphite soot is a chemically inert amorphous form of carbon. Fullerene C60, prepared by laser ablation of graphite, has a closed-shell spheroid configuration consisting of 20 hexagons and 12 pentagons [7]. It is this configuration that is responsible for the high chemical activity and noticeable solubility of C60 in organic solvents [8, 9]. Astralene is a carbonaceous nanomaterial formed in synthesis

EXPERIMENTAL Polyphenyleneisophthalamide (Fenilon4) (PA) was used as a test subject. Composites of PA with 5 wt % nanoadditive (fullerene C60, astralene, or graphite soot) were prepared by mixing PA and nanocarbon in a porcelain mortar. The resulting mixture was dissolved in N,N-dimethylacetamide. A 0.12% polymer 236

POLYAMIDE ULTRAFILTRATION MEMBRANES

237

solution was cast onto a glass plate using a draw die with a 0.30-mm passageway. Finally, the glass plate was immersed in a precipitation bath to form an asymmetric porous membrane through phase inversion. The rigidity of the structure of the membranes was monitored when passing solvent across them in the order ethanol, water, isopropanol, hexane, i.e., in accordance with their mutual solubility. The experiments on determination of the solvent flux across the membranes and on ultrafiltration of protein solutions were carried out in stirred cells under a pressure of 0.1 MPa. The rejection curves of the ultrafiltration membranes were obtained according to the procedure described in [16], using a 0.4 wt % aqueous solution of a model mixture of proteins of various molecular weights: vitamin B12 (MW 1340), cytochrome C (MW 12 500), chemotrypsinogen (MW 24 000), egg albumin (MW 44 500), bovine serum albumin (MW 67 000), and g-globulin (MW 160 000). The selectivity of the membranes was characterized by the rejection factor j, which was estimated as j = 1 3 cp /cf, where cp and cf are the concentrations of the permeate and feed solution, respectively. The experiments on static sorption of proteins on the membranes were carried out in hermetically sealed weighing bottles by immersing the membranes into 4 ml of the solution of protein mixture, with intermittent stirring. The protein concentration in the solution was monitored by the optical density at 280 and 415 nm at fixed intervals. The equilibrium concentration was established in 4 h, and the total experiment time was 20 h. The contact angles of the membrane surface with water were determined by the Wilhelmy method on a KRUSS contact-angle-measuring system with computer-controlled sample immersion rate and data processing. To gain insight into the effect of a modifier, we performed a comparative study of the PA-based membranes containing 5 wt % of various nanocarbon additives. The membrane morphology was examined by scanning electron microscopy (SEM). The SEM results revealed that nanocarbon additives had virtually no effect on the pore structure. Figure 1 shows a cross-sectional SEM micrograph of a PA membrane modified with 5 wt % fullerene. It should be pointed out that the performance characteristics of membranes are not always manifested on the morphological level as judged from SEM data. For example, polysulfone gas-separation membranes modified with various carbonaceous additives have similar morphology, but RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Vol. 81

Fig. 1. Cross-sectional SEM image of a polyimide membrane modified with 5 wt % fullerene C60.

Fig. 2. Flux Q vs. the solvent viscosity h for asymmetric membranes: (1) PA, (2) PA + 5 wt % graphite soot, (3) PA + 5 wt % fullerene, and (4) PA + 5 wt % astralene; the same for Figs. 3 and 4.

demonstrate considerable differences in their transport characteristics [17]. The modifier nature affects the structure of the skin layer. The skin porosity can be characterized by the solvent flux across the membrane. In this study, the durability of the membranes and their resistance with respect to various media were characterized by the flux of water and several organic solvents (hexane, ethanol, and isopropanol) across the ultrafiltration membranes modified with 5 wt % nanocarbon additives. The solvents used were different in their viscosity (hhexane = 0.30 0 10!3, hw = 1.0 0 10!3, hEtOH = 1.13 0 10!3, and hi-PrOH = 2.43 0 10!3 Pa s) and, therefore, in their flux across the membranes. Figure 2 shows that, in accordance with the Hagen3 Poiseuille law [18], the flux decreases with increasing viscosity of the fluid for all the membranes studied. It is noteworthy that, for all the solvents studied, No. 2

2008

238

POLOTSKAYA et al.

tion of donor3acceptor and maybe covalent bonds, thus influencing the polymer structure [19, 20]. With the astralene-modified membranes, the flux of all the solvents decreased even stronger than with the fullerene additive, which is probably due to the high reactivity of the fulleroid particles with respect to the PA matrix. Fig. 3. Normalized flux Qn (cm Pa atm!1) vs. the fluid viscosity h (Pa s) for the asymmetric membranes.

Fig. 4. Rejection factors of the membranes lecular weight of proteins.

j vs. the mo-

In addition to the solvent flux Q, we estimated the normalized flux Qn written as a product Qn = Qh, where h is the solvent viscosity. The normalized flux is intended to characterize the stability of the membranes in the media investigated. Our results show (Fig. 3) that the normalized fluxes of all the solvents are similar, suggesting an insignificant swellability of the membranes in solvents of varied chemical nature. The modified PA membranes were tested with a solution of a protein mixture. The data obtained in the ultrafiltration experiments were then used to construct the rejection curves, which represent the dependences of the rejection factor on the molecular weight of proteins (Fig. 4). An important characteristic of the rejection curves is the molecular weight corresponding to the rejection factor of 0.9, which is known as the molecular-weight cut-off (ML). All the membranes studied are characterized by similar values of ML [(40 350) 0 103g mol!1]. As can be seen from Fig. 5, the flux of the protein solution is considerably lower than that of water. The primary factors responsible for the increase in the hydraulic resistance and the corresponding decrease in the flux are considered to be concentration polarization, formation of a deposit layer on the membrane, and adsorption. The last factor depends primarily on the chemical nature of the molecules to be filtered and on the polymer material of the membrane. It is the adsorption that essentially causes the decrease in the flux in the case of filtration of hydrophilic compounds [21], among them proteins.

Fig. 5. Flux Q (cm s!1 atm!1) of (I) water and (II) protein mixture across the membranes.

the flux across the PA membrane modified with 5 wt % graphite soot was slightly higher as compared to the unmodified membrane, suggesting that this chemically inert additive serves as a filler for both the casting solution and the resulting membrane. For all the solvents, the flux across the fullerene-modified membranes considerably decreased as compared to the PA membrane, which can be attributed to forma-

The suggestion that the adsorption of proteins on the surface and in the pores of the membranes is responsible for the flux decrease was checked by comparison of the water flux restoration after the membranes were in contact with protein mixture under the ultrafiltration conditions and also under the static sorption conditions, when the membranes were held in a solution of protein mixture for 20 h. In both cases, the water flux restoration ratio (FRR) was estimated as FRR = Q1/Q0, where Q0 and Q1 are the water fluxes before and after contacting the membranes with the protein solution.

RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Vol. 81

No. 2

2008

POLYAMIDE ULTRAFILTRATION MEMBRANES

239

Figure 6 shows data on the water flux restoration in the ultrafiltration and static sorption modes. The membranes modified with fullerene and astralene show high FRR in the range 0.8 3 0.9. For PA membranes unmodified and modified with graphite soot, FRR was below 0.45. Such a low FRR suggests that these membranes strongly absorb proteins in both ultrafiltration and static sorption modes. It should be pointed out that such a behavior is typical of most of the existing polymeric membranes. For the membranes modified with fullerene C60 and astralene, the sorption of proteins is considerably lower, which can be attributed to the fact that these additives block the sorption sites of PA. This suggestion is supported by data on the contact angles of the membranes with water. The contact angle increases from 81o for unmodified PA membranes to 85o for the membranes modified with 5 wt % fullerene, suggesting an increase in the hydrophobicity. Finally, modification of PA membranes with fullerene C60 and astralene improves such an important performance characteristic as the water flux restoration ratio, which facilitates regeneration of the membranes and reduces the loss of the target products.

ACKNOWLEDGMENTS The study was financially supported by the Russian Foundation for Basic Research (project no. 06-0332 493). REFERENCES

3

1. Baker, R.W., Membrane Technology and Applications, New York: McGraw Hill, 2000. 2. Kazaryan, L.G., Ivankin, D.Ya., Vasil’ev, V.M., et al., Vysokomol. Soedin., Ser. A, 1975, vol. 17, no. 7, pp. 1560 1568. 3. Klenin, V.I., Prozorov, L.V., and Zhizdyuk, B.I., Vysokomol. Soedin., Ser. A, 1974, vol. 16, no. 7, pp. 1665 1672. 4. Bil’dyukevich, A.V., Soldatov, V.S., and Kaputskii, F.N., Khim. Tekhnol. Vody, 1990, vol. 12, no. 6, pp. 542 548. 5. Pratsenko, S.A. and Bil’dyukevich, A.V., Vysokomol. Soedin., Ser. A, 1994, vol. 36, no. 3, pp. 457 460. 6. Pratsenko, S.A., Bil’dyukevich, A.V., Varslovan, E.S., et al., Zh. Prikl. Khim., 1995, vol. 68, no. 3, pp. 453 456. 7. Kerl, R.F., Usp. Fiz. Nauk, 1998, vol. 168, no. 3, pp. 331 342. 8. Bezmel’nitsyn, V.N., Eletskii, A.V., and Okun’, M.V., Usp. Fiz. Nauk, 1998, vol. 168, no. 11, pp. 1195 1234. 9. Konarev, D.V. and Lyubovskaya, R.N., Usp. Khim., 1999, vol. 68, no. 1, pp. 23 42. 10. Harris, P.J., Carbon Nanotubes and Related Structures, Cambridge: Cambridge Univ., 2003. 11. Gladchenko, S.V., Polotskaya, G.A., Gribanov, A.V., and Zgonnik, V.N., Zh. Tekh. Fiz., 2002, vol. 72, no. 1, pp. 105 109. 12. Higuchi, A., Agatsuma, T., Uemiya, S., et al., J. Appl. Polym. Sci., 2000, vol. 77, pp. 529 537. 13. Polotskaya, G.A., Penkova, A.V., Toikka, A.M., et al.,

3

CONCLUSIONS (1) Modification of polyphenyleneisophthalamide with nanocarbon additives has no significant effect on the initial porous sponge-like structure of the membranes. The modifier nature is manifested in a change in the surface porosity and sorption activity of the membranes. (2) The membrane permeability depends on the modifier nature. For each of the membranes studied, the normalized fluxes of hexane, water, ethanol, and isopropanol are similar, suggesting virtually no swelling of the membranes in these solvents. (3) The membranes modified with fullerene and astralene demonstrate good water flux restoration after being in contact with a protein mixture (FRR ; 0.83 0.9), in contrast to straight polyphenyleneisophthalamide membranes and those modified with carbon black (FRR < 0.45). This is due to the fact that addition of fullerene and astralene increases the hydrophobicity of the membrane matrix and decreases its sorption capacity for proteins. Modification of the polyamide with fullerene and astralene improves the ultrafiltration performance of the membranes. RUSSIAN JOURNAL OF APPLIED CHEMISTRY

Fig. 6. Water flux restoration ratio FRR in (1) ultrafiltration and (2) static sorption modes.

Vol. 81

3

3

3

3 3

3

3

3

No. 2

2008

3

240

14.

15. 16. 17.

POLOTSKAYA et al.

3

Sep. Sci. Technol., 2007, vol. 42, no. 2, pp. 333 348. Ong, S.L., Hu, J.Y., Biryulin, Yu.F., and Polotskaya, G.A., Fullerenes, Nanotubes, Carbon Nanostruct., 2006, vol. 14, nos. 2 3, pp. 463 466. Tasaki, K., De Sousa, R., Wang, H., et al., J. Membrane Sci., 2006, vol. 281, nos. 1 2, pp. 570 580. Cherkasov, A.N., Sep. Sci. Technol., 2005, vol. 40, pp. 2775 2801. Bhardwaj, V., Macintosh, A., Sharpe, I.D., et al., Adv. Membr. Technol., 2003, vol. 984, no. 1, pp. 318 328.

3

3

3

3

3

3

18. Cherkasov, A.N. and Pasechnik, V.A., Membrany i sorbenty v biotekhnologii (Membranes and Sorbents in Biotechnology), Leningrad: Khimiya, 1991. 19. Janaki, J., Premila, M., Gopalan, P., and Sundarm C.S., Thermochim. Acta, 2000, vol. 356, nos. 1 2, pp. 109 116. 20. Isobe, H., Tomita, N., and Nakamura, E., Org. Lett., 2000, vol. 2, no. 23, pp. 3663 3665. 21. Zularisam, A.W., Ismail, A.F., Salim, M.R., et al., Desalination, 2007, vol. 212, pp. 191 208.

3

RUSSIAN JOURNAL OF APPLIED CHEMISTRY

3

3

Vol. 81

3

No. 2

2008