Full Paper Formation of Plasma-Polymerized Top Layers on Composite Membranes: Influence on Separation Efficiency Dung Thi Tran,* Shinsuke Mori, Daisuke Tsuboi, Masaaki Suzuki Plasma-polymerized reverse osmosis membranes were prepared by deposition of an allylamine plasma-polymerized top layer onto a cellulose ester surface Their separation performance was highly dependent on the formation of the top layer, which was itself determined by the plasma polymerization conditions The density and the thickness of the plasma-deposited polymer top layer influence the membrane flux, while the ability of the membrane to reject salt relies on the degree of polymer cross-linking, which is in turns related to the transfer of plasma energy into monomer molecules during plasma polymerization The separation efficiency of the multilayer reverse osmosis composite membrane can be optimized by choosing appopriate plasma conditions Introduction Plasma polymerization can create ultra-thin and pinholefree polymer films with a very high degree of cross-linking and a uniform structure.[1–6] The characteristics of plasma deposited polymers are highly dependent on the plasma polymerization conditions.[7–10] In the field of separation membranes, plasma polymerization can be used to prepare composite membranes for nanofiltration, reverse osmosis (RO), pervaporation and gas separation processes.[11–23] It is well known that the top layer plays the most important role in terms of separation perforD T Tran Department of Chemical Technology, Hanoi University of Science, VNU, 334 Nguyen Trai, Thanh Xuan, HaNoi, Vietnam E-mail: ttdung@vnu.edu.vn D T Tran, S Mori, D Tsuboi, M Suzuki Department of Chemical Engineering, Tokyo Institute of Technology, 2-12-1 O-okayama, Meguro-ku, Tokyo 152-8552, Japan 110 Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim mance and durability of composite membranes, and thus many efforts have been dedicated to the enhancement of the active barrier layer of composite membranes Kai et al developed composite reverse osmosis membranes for organic liquid mixtures by plasma-graft polymerization using vinyl acrylate and N, N0 -methylenebis(acrylamide) as the crosslinkers to prepare cross-linked pore filling-type membranes.[25] The separation performance and pressure durability can be improved when the chemical structure of the crosslinker and the composition of the monomer/ cross-linker mixture are optimized Chen et al prepared nanofiltration membranes from polyacrylonitrile ultrafiltration membranes by Ar low-temperature plasma treatment and subsequent grafting of acrylic acid and styrene monomers in the vapor phase.[26–28] These polyacrylonitrile membranes can be potentially used to recover the dewaxing solvent from dewaxed lube oils Lin, Chen and Xu improved the oxygen/nitrogen permselectivity of poly(1-trimethylsilyl-1-propyne) membranes by plasma polymerization using fluorine-containing monomers.[29] DOI: 10.1002/ppap.200800093 Formation of Plasma-Polymerized Top Layers on Composite Membranes: C.-J Chuong et al prepared gas separation membranes by deposition of plasma polymer thin films formed from vinyl acetic acid or perfluorohexane monomers onto nanoporous polycarbonate track-etched (PCTE) membranes to form sufficient composite membranes for O2 and CO2 separation.[30] H I Kim and S S Kim indicated that plasma treatment using several hydrophilic monomers and polysulfone substrates enhanced the rejection, as well as the flux of conventional polyamide composite membranes prepared by interfacial polymerization, and plasma polymerization using these monomers also led to the formation of composite RO membranes which have a good desalination property.[31] It is well known that for desalination purposes, the top layer of a composite RO membrane must be thin with a highly cross-linked structure so that it can reject salt effectively from feed solution and also have adequate water flux.[32] The aim of this work is to investigate the influence of the formation of plasma-polymerized top layers on the separation efficiency of cellulose ester (CE) based composite RO membranes The possibilities for the improvement of the separation efficiency of membranes were considered Overlap deposition of plasma polymer layers was carried out to prepare multilayer composite membranes, whose separation efficiencies were demonstrated to be an improvement upon that of the dual layer composite membranes Experimental Part Materials Microporous CE membrane filters (Advantec) purchased from Toyo Roshi Kaisha, Ltd Co (Japan), with a thickness of 110 mm, a skin pore size of 0.1mm, and a porosity of 65%, were used as substrates for the deposition of a plasma-polymerized top layer for composite membranes Allylamine solution (99.5 wt.-%) purchased from Wako Pure Chemicals Industries, Ltd (Japan) was used without further purification to form the vaporized monomer for plasma polymerization Plasma Polymerization The plasma reactor system described in our previous work[12] was also employed in this study (Figure 1) It consists of a tubular-type reactor (with a diameter of 30 mm and a length of 400 mm) with two external electrodes (10 mm wide and 100 mm apart) connected with a cold trap, a monomer reservoir with a metering valve and mass flow meter, a radio-frequency (RF) power generator of 13.56 MHz with a matching network, a pressure gauge and a vacuum pump The substrate is placed in the plasma reactor by carefully pushing the sample from the end of reactor to a position located between two electrodes Next, the reactor is closed and evacuated to a base pressure below Pa; vaporized monomer is then introduced into the reactor at a determined flow Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Figure Experimental setup used in this work rate Plasma polymerization was carried out under different plasma conditions and the optical emission spectra of the plasma discharge were recorded using a high-resolution spectrometer (HR 4000 CG-UV-NIR) with an optical window located in the end of plasma reactor Plasma polymerization was carried out under the different conditions The influence of the plasma parameters on the characteristics of the plasma deposited polymer was assessed through analysis of the plasma energy input using composite parameter W/FM, where W is plasma power input given in W, F is volume flow rate and M is monomer molecular weight.[33] Characterization of Plasma-Deposited Polymer and Composite Membranes The functionality of allylamine plasma deposited polymer was determined using Fourier-transform infrared (FTIR) spectroscopy using a Jeol SPX-200 spectrometer For this work, the plasma polymer films were deposited onto CaF2 transparent slides and the spectra recorded with one hundred scans taken at a resolution of cmÀ1 The mass density of the plasma-deposited polymer (mg Á cmÀ2) was calculated by determining the weight of the plasma polymer deposited onto a substrate surface area unit The deposition rate (mg Á cmÀ2 Á minÀ1) of plasma polymer was calculated based on the weight of polymer deposited on a unit of substrate surface area per minute The morphology of membranes was analyzed by scanning electron microscopy (SEM, Keyence, VE-8800) To prevent the surface charging, a thin film (5 nm) of Pt was sputtered onto the sample surface by means of an ion sputter unit (Hitachi, E-1030) prior to imaging The separation characteristics of membranes were investigated through desalination experiments using a membrane cell (Osmonic, USA) with sodium chloride feed solution of 500 ppm at an applied pressure of 3.5 MPa The desalination experiments were carried out with a cross-flow of salt solution on the membrane surface The membrane separation factors, i.e., the selectivity (salt rejection R) and the flow (water flux J), are determined using the equations: R=% ẳ fẵC0 Cị=C0 100g (1) J=l m2 h1 ị ẳ ½V=ðS Â tÞ (2) www.plasma-polymers.org 111 D T Tran, S Mori, D Tsuboi, M Suzuki where C0 and C are the salt concentrations in the feed solution and filtrate, and V, S and t are the filtrate volume, membrane area and separation time, respectively The cross-links are chemical bonds that link one polymer chain with another In conventional polymers, the cross-linking degree can be determined by the percentage of cross-linker agent present in the polymer In plasma polymers, the cross-linking degree can be performed through the tightness of three-dimensional networks or the mass density of polymer In this work, the crosslinking degree of the plasma deposited polymer top layer prepared under various plasma conditions was estimated qualitatively through the relationship between the mass density of the deposited polymer top layer (mg Á cmÀ2) and the composite membrane separation efficiency With the same amount of plasma deposited polymer, the plasma polymer top layer with more cross-linking showed a higher selectivity as compared to those with less cross-linking The comparison in separation performance of membranes prepared under different plasma polymerization conditions was investigated by examining the relationship between the salt rejection, R, and the water flux, J, of the membranes Results and Discussion chains, which then recombine to form more highly crosslinked plasma deposited polymer FTIR Results The functionality of allylamine plasma polymer was studied by FTIR spectroscopy The results (Figure 3) show the NH stretching vibration peaks centered at 300– 400 cmÀ1 due to the presence of primary amine and secondary amine or imine groups in the plasma deposited polymer The stretching vibration of CH groups can be seen at 800–2 950 cmÀ1 The contributions at 650 cmÀ1 include the N –H band, C – – N and C – – C stretch vibrations The doublet at 350 cmÀ1 comes from residual air in the spectrometer (CO2) The band at 230 cmÀ1 can be associated with the presence of nitrile groups (C N stretch) or alkyne groups (C C stretch) in plasma polymer film The possible plasma reactions leading to the formation of imine and/or nitrile functional groups could be as it follows: CH2 ¼ CHCH2NH2 ỵ e ! CH2CHCH2NH2 (3) Optical Emission Spectroscopy (OES) The optical emission spectra of allylamine plasma discharge created under different plasma conditions were recorded Figure shows the spectrum of an allylamine glow discharge generated at the monomer flow rate of sccm and discharge power input of 30 W The spectrum shows the signals of NH (337 nm), CN (359, 388, 421 nm), Ha (656.5 nm), Hb (486 nm) and also of H2 molecules (band around 600 nm) in the gas phase The results reveal the dissociation of the monomer molecules and the formation of the excited species within the plasma Besides the dissociation of the original monomer molecules and the molecules excited by the molecular detachment, hydrogen detachment also plays a very important role in creating free radicals and/or the active points on pre-polymer Figure Optical spectrum of allylamine plasma discharge emission (2 sccm, 30 W) 112 Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim R ỵ CH2 ẳ CHCH2NH2 ! RCH2CHCH2NH2 (4) CH2NH2 ! CH2NH ỵ H ! CHẳNH ỵ H2 (5) CH ẳ NH ! CH ẳ N ỵ H ! CN ỵ H2 (6) Plasma Polymer Deposition Rates Allylamine plasma polymers were prepared under different plasma conditions The dependence of the polymer deposition rate (mg Á cmÀ2 Á minÀ1) on the effective plasma Figure FTIR spectrum of allylamine plasma polymer (2 sccm, 30 W) DOI: 10.1002/ppap.200800093 Formation of Plasma-Polymerized Top Layers on Composite Membranes: energy (W/FM) is shown in Figure The results indicate that the deposition rate of plasma polymers is a function of the composite parameter W/FM The deposition rate increases nearly linearly in the low-W/FM region and gradually approaches a ceiling deposition rate, which depends on the monomer flow rate used; i.e., the lower the flow rate, the lower the ceiling deposition rate Figure shows the SEM images of the substrate (microporous CE membrane filter) and the CE-based plasma-polymerized composite membranes prepared under different plasma conditions The pictures indicate that the deposition of plasma-polymerized films leads to the formation of a top layer in the case of composite membranes The size of the membrane skin pores gradually decreases and thus, the plasma deposited layer thickness increases with increasing the polymerization time The deposition of the plasma polymer occurred not only on the surface but also on the walls of the substrate pores The pictures indicate that the deposition rate of plasma polymer could increase when the monomer flow rate and discharge power input increases Dual-Layer Plasma-Polymerized Composite Membranes The influence of the plasma polymerization conditions on the degree of cross-linking of the plasma deposited Figure Dependence of the polymer deposition rate on the plasma energy input polymer was investigated through examining the relationship between the mass density (mg Á cmÀ2) of the plasma deposited polymer top layer formed under different plasma conditions and the composite membrane separation efficiency The obtained results (Figure 6) indicate that the membrane selectivity increases, while its flux decreases as the mass density of the plasma deposited polymer increases Furthermore, for the same plasma polymer mass density, the composite membranes whose top layer are Figure SEM images of CE-based plasma-polymerized composite membranes with allylamine plasma-polymerized top layer Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim www.plasma-polymers.org 113 D T Tran, S Mori, D Tsuboi, M Suzuki Figure Relationship between the mass density of polymer top layer prepared using different plasma energy inputs and the separation efficiencies of composite membranes formed at a higher effective plasma energy show higher salt rejection and lower water flux This shows that a more highly cross-linked top layer could be formed at higher effective plasma energies Figure shows the separation characteristics of the dual layer plasma-polymerized composite membranes, whose top layer were formed under various plasma conditions The experimental results show that the salt rejection increases and water flux decreases with increasing the polymerization time In addition, the polymerization time required to form an effective barrier layer (R % 90%) increases when the effective plasma energy decreases In the case of monomer flow rate of sccm and discharge power input of 30 W, although the plasma energy input is Figure Relationship between R and J of dual layer composite membranes with the top layer prepared under different plasma conditions higher, the polymerization time must be extended because of a rather low plasma polymer deposition rate The relationship between R and J values of the dual layer plasma-polymerized composite membranes is given in Figure It can be seen that an increase of the salt rejection is always associated with a decrease of the water flux In this experiment, the membranes whose top layers were formed at power input of 30 W and monomer flow rates of and sccm show higher water flux as compared to others with the same salt rejection However, the flux of these membranes is worse when selectivity is higher than 90% From the experimental results, it can be surmised that, in order to improve the separation performance of plasmapolymerized composite reverse osmosis membranes, plasma polymerization should be performed in such a way that the resultant deposited top layer is as thin as possible but presenting a highly cross-linked structure, so that the final membranes can reject salt effectively but still have adequate water flux In principle, the thickness of plasma deposited polymer top layer also relies on the skin pore size of the substrate The plasma deposited polymer layer first must be thick enough to cover up the skin pores to form a top layer for the composite membranes The smaller the skin pore size of the substrate, the thinner the effective top layer needed With the substrate whose skin pore size is rather large, to reduce the thickness and increase the degree of cross-linking in the barrier layer of composite membranes, overlap deposition of plasma polymer layers was carried out to form multilayer plasma-polymerized composite membranes Multilayer Plasma-Polymerized Composite Membranes Figure Separation properties of the dual layer composite membranes prepared under different plasma conditions 114 Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim From the obtained experimental results, it can be seen that dual layer composite membranes with less cross-linking in the top layer could have high water flux but rather poor DOI: 10.1002/ppap.200800093 Formation of Plasma-Polymerized Top Layers on Composite Membranes: salt rejection, whereas those with a more highly crosslinked top layer would have better selectivity but lower flux Hence, in order to obtain a sufficient selectivity, the top layer with less cross-linking has to be rather thick; however, this has a negative influence on the flux of the membrane For multilayer composite membranes, therefore, the top layer should be prepared by overlap deposition of plasma deposited polymer layers The first layer (sublayer) with a less cross-linked structure was deposited onto the substrate surface before deposition of the second (barrier) layer with a more highly cross-linked structure In this experiment, for multilayer composite membranes, the sublayer was formed at a low plasma energy with a monomer flow rate of sccm and power input of 10 W within Next, the overlap deposition of the barrier layer was performed at higher plasma energy inputs The comparison in separation efficiency of the dual layer and the multilayer plasma-polymerized composite membranes is given in Figure The results indicate that the multilayer composite membranes demonstrate better separation efficiency in comparison with that of the dual layer membranes, whose top layer was also prepared at the same effective plasma energy In addition, it can be seen that the plasma polymerization time required to create an active top layer for the multilayer composite membranes is shorter than that for the dual layer membranes Figure 10 shows the separation efficiency of some multilayer composite membranes, namely, M0, M1, M2 and M3 The sublayer of these membranes was formed at a discharge power input of 10 W and monomer flow rate of sccm within The barrier layers were prepared at different plasma polymerization parameters In this work, the plasma energy input for barrier layer formation increases gradually from M0 to M3 The experimental results indicate that the membrane M3 shows better separation efficiency because of higher water flux and its salt rejection is almost similar to that of M0, M1 and M2 This could be due to the barrier layer of M3 having a more highly, yet thinner, cross-linked structure than the others It can be seen that the polymerization time required to form an active barrier layer for membranes M0 and M3 is longer than for membranes of M1 and M2 This could be due to the lower degree of cross-linking in the barrier layer of M0 and the lower deposition rate for the plasma polymer top layer in the case of M3 In other experiments, higher plasma energy inputs were used by reducing the monomer flow rate to less than sccm while the discharge power input was maintained at 30 W The obtained experimental results reveal that, in this case, because the deposition rate of plasma polymer was reduced, the polymerization time required to form an effective barrier layer for composite membranes was Figure Comparison in the separation efficiencies of the dual and multilayer composite membranes Figure 10 Separation efficiencies of multilayer composite membranes with barrier layer formed under different plasma conditions Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim www.plasma-polymers.org 115 D T Tran, S Mori, D Tsuboi, M Suzuki extended (it takes about 90 for the top layer to form at a monomer flow rate of 0.5 sccm and discharge power of 30 W, for example); however, improvement of the separation efficiency of the formed composite membranes is negligible as compared to the one formed at sccm and 30 W A similar trend also occurred by reducing both discharge power input and monomer flow rate (for example, it takes about 60 for the top layer to form at a monomer flow rate of 0.5 sccm and discharge power of 15 W) In order to increase the polymer deposition rate, plasma polymerization can be carried out at higher monomer flow rates and higher discharge power inputs However, to prevent damage to the polymer substrate, the discharge power input should be not too high In addition, if the plasma power is too high, ablation etching effects could result in negative influence on the formation of the deposited polymer top layer Figure 11 shows the relationship between R and J of the multilayer composite membranes M0, M1, M2 and M3 The results indicate that, similar to the dual layer composite membranes, the enhancement of salt rejection is also associated with the decline of water flux For multilayer composite membranes, however, the decline tendency is slower and thus, with the same selectivity, the flux of these membranes is always higher than that of the dual layer composite membranes Figure 12 shows the optical emission spectra of allylamine plasma discharge generated at the same plasma energy applied for preparation of barrier layer for M0, M1, M2 and M3 composite membranes The results reveal that the emission intensity of H and H2 increases with increasing the plasma energy input and thus, a larger number of the excited species could be formed due to a higher level of dissociation of monomer molecules and by a formation of active points in pre-polymer chains Figure 13 shows the FTIR spectra of allylamine plasma Figure 11 Relationship between R and J of the multilayer composite membranes with barrier layer formed under different plasma parameters 116 Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim Figure 12 Optical emission spectra of allylamine plasma discharge generated under different plasma energy input polymers prepared under different plasma energy inputs The obtained results indicate that the intensity of the absorption peaks decreased with increasing the plasma energy input This could be due to the decrease of the deposition rate of polymer formed by the recombination of the smaller fragments generated at higher plasma energy input and thus, the higher cross-linked polymer top layer for composite membranes would be formed This assumption is also in good agreement with the obtained results of OES and separation data From the obtained results above, it can be seen that the plasma energy input is an important parameter, which strongly influences the characteristics of the plasma deposited polymer and the separation performance of plasma-polymerized composite membranes By applying overlap deposition, a continuous graded plasma polymer top layer was prepared by changing the plasma energy input to form multilayer composite membranes, whose separation efficiency appear to be better than that of the dual layer composite membranes DOI: 10.1002/ppap.200800093 Formation of Plasma-Polymerized Top Layers on Composite Membranes: ciency of plasma-polymerized composite reverse osmosis membranes could be improved in comparison with that of dual layer composite membranes Acknowledgements: The authors would like to thank the Japanese Society for the Promotion of Science (JSPS) for the support it has given to this project Received: June 12, 2008; Revised: September 21, 2008; Accepted: October 21, 2008; DOI: 10.1002/ppap.200800093 Keywords: composite membranes; cross-linking degree; deposition; multilayers; plasma polymerization; separation efficiency Figure 13 FTIR spectra of allylamine plasma polymers prepared under different plasma energy input (polymerization time of 20 min) Conclusion The separation efficiency of plasma-polymerized composite membranes is strongly influenced by the plasma deposited polymer top layer formation, which is itself determined by the plasma polymerization parameters From the obtained results, the following can be concluded: The thickness and mass density of the plasma deposited polymer layer is determined by the plasma polymer deposition rate and plasma polymerization time, while the degree of cross-linking of polymer depends on the application of an effective plasma energy, which is related to monomer flow rate and plasma power input used for plasma polymerization A higher effective plasma energy could lead to the formation of a more highly cross-linked polymer top layer, whereas a lower effective plasma energy results in less cross-linking in the top layers For plasmapolymerized composite reverse osmosis membranes, with the same mass density of plasma deposited polymer, the top layer formed at lower plasma energy input has a higher water flux but rather poor salt rejection; whereas a membranes whose top layer is formed at a higher plasma energy input have better selectivity but lower flux Plasma polymerization should be performed in such a way that the deposited polymer top layer is as thin as possible with a highly cross-linked structure By performing multilayer deposition of a plasma-polymerized top layer onto a porous substrate surface, in which the sublayer is prepared at a lower effective plasma energy and the barrier layer is formed at a higher effective plasma energy, the separation effiPlasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim [1] H Yasuda, ‘‘Plasma Polymerization’’, Academic Press, Orlando 1985 [2] H Biederman, Y Osada, ‘‘Plasma Polymerization Processes’’, Elsevier, Amsterdam 1992 [3] N Inagaki, ‘‘Plasma Surface Modification and Plasma Polymerization’’, Technomic, Lancaster 1996 [4] H V Boenig, ‘‘Fundamentals of Plasma Chemistry and Technology’’, Technomic, Lancaster 1988 [5] ‘‘Plasma Processes and Polymers’’, R d’Agostino, P Favia, C Oehr, M R Wertheimer, Eds., Wiley-VCH, Weinheim 2005 [6] H Yasuda, J Membr Sci 1984, 18, 273 [7] V Sciarratta, U Vohrer, D Hegemann, M Muller, C Oehr, Surf Coat Technol 2003, 174, 805 [8] E R Fisher, Plasma Process Polym 2004, 1, 13 [9] J Friedrich, G Kuhn, R Mix, W Unger, Plasma Process Polym 2004, 1, 28 [10] A Choukourov, H Biederman, I Kholodkov, D Slavinska, M Trchova, A Hollander, J Appl Polym Sci 2004, 92, 979 [11] M Bryjak, G Pozniak, I Gancarz, W Tylus, Desalination 2004, 163, 231 [12] T D Tran, S Mori, M Suzuki, Thin Solid Films 2007, 515, 4148 [13] P W Kramer, Y.-S Yeh, H Yasuda, J Membr Sci 1989, 46, [14] S.-G Miao, B.-G Wang, Language 2005, 19, 297 [15] Y Kong, X Lin, Y Wu, J Chen, J Xu, J Appl Polym Sci 1992, 46, 191 [16] M Hirai, I Tajima, M Yamamoto, Proc ACS Div Polym Mater 1987, 56, 802 [17] C.-L Li, C.-Y Tu, N Inagaki, K.-R Lee, J.-Y Lai, J Appl Polym Sci 2006, 102, 909 [18] J Weichart, Surf Coat Technol 1993, 59, 342 [19] J Sakata, M Yamamoto, M Hirai, Language 1988, 45, 499 [20] T Masuoka, T Iwatsubo, K Mizoguchi, J Membr Sci 1992, 69, 109 [21] J Sakata, M Yamamoto, I Tajima, J Polym Sci., Part A: Polym Chem 1988, 26, 1721 [22] W.-I Sohn, D.-H Ryu, S.-J Oh, J.-K Koo, J Membr Sci 2000, 175, 163 [23] D T Tran, S Mori, M Suzuki, Thin Solid Films 2008, 516, 4384 [24] R G Gutman, ‘‘Membrane Filtration’’, Adam Hilger, Bristol 1987 [25] T Kai, H Goto, Y Shimizu, T Yamaguchi, S.-I Nakao, S Kimura, J Membr Sci 2005, 265, 101 www.plasma-polymers.org 117 D T Tran, S Mori, D Tsuboi, M Suzuki [26] Z.-P Zhao, J Li, D.-X Zhang, C.-X Chen, J Membr Sci 2004, 232, [27] Z.-P Zhao, J Li, D.-X Zhang, C.-X Chen, J Membr Sci 2005, 251, 239 [28] Z.-P Zhao, J Li, D.-X Zhang, C.-X Chen, Surf Coat Technol 2007, 201, 6789 [29] X Lin, J Chen, J Xu, J Membr Sci 1994, 90, 81 118 Plasma Process Polym 2009, 6, 110–118 ß 2009 WILEY-VCH Verlag GmbH & Co KGaA, Weinheim [30] C L Chapman, P Bhattacharyya, R C Eberhart, R B Timmons, C.-J Chuong, J Membr Sci 1994, 318, 137 [31] H I Kim, S S Kim, J Membr Sci 2006, 286, 193 [32] R W Baker, ‘‘Membrane Technology and Applications’’, John Wiley & Sons Ltd., Chichester 2004 [33] H K Yasuda, Plasma Process Polym 2005, 2, 293 DOI: 10.1002/ppap.200800093 ... the influence of the formation of plasma-polymerized top layers on the separation efficiency of cellulose ester (CE) based composite RO membranes The possibilities for the improvement of the separation. .. multilayer composite membranes, whose separation efficiency appear to be better than that of the dual layer composite membranes DOI: 10.1002/ppap.200800093 Formation of Plasma-Polymerized Top Layers on. .. (polymerization time of 20 min) Conclusion The separation efficiency of plasma-polymerized composite membranes is strongly influenced by the plasma deposited polymer top layer formation, which