Effect of UV-Ozone Treatment on Poly(dimethylsiloxane) Membranes

Dec 8, 2009 - The conversion of the cross-linked polysiloxane to SiOx increased with UV-ozone exposure time and cross-linking agent content, and the ...
5 downloads 3 Views 1MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

)

Effect of UV-Ozone Treatment on Poly(dimethylsiloxane) Membranes: Surface Characterization and Gas Separation Performance Ywu-Jang Fu,† Hsuan-zhi Qui,‡ Kuo-Sung Liao,‡ Shingjiang Jessie Lue,§ Chien-Chieh Hu,*, Kueir-Rarn Lee,*,‡ and Juin-Yih Lai‡

Department of Biotechnology, Vanung University, Chung-Li 32061, Taiwan, ‡R&D Center for Membr. Technol. and Department of Chemical Engineering, Chung Yuan University, Chung-Li 32023, Taiwan, § Department of Chemical and Materials Engineering and Pollution Prevention Group in Green Technology Research Center, Chang Gung University, Kwei-shan, Taoyuan 333, Taiwan, and Department of Chemical and Materials Engineering, Nanya Institute of Technology, Chung-Li 32034, Taiwan )



Received September 14, 2009. Revised Manuscript Received November 17, 2009 A thin SiOx selective surface layer was formed on a series of cross-linked poly(dimethylsiloxane) (PDMS) membranes by exposure to ultraviolet light at room temperature in the presence of ozone. The conversion of the cross-linked polysiloxane to SiOx was monitored by attenuated total reflectance Fourier transform infrared (ATRFTIR) spectroscopy, X-ray photoelectron spectroscopy (XPS), energy-dispersive X-ray (EDX) microanalysis, contact angle analysis, and atomic force microscopy (AFM). The conversion of the cross-linked polysiloxane to SiOx increased with UV-ozone exposure time and cross-linking agent content, and the surface possesses highest conversion. The formation of a SiOx layer increased surface roughness, but it decreased water contact angle. Gas permeation measurements on the UV-ozone exposure PDMS membranes documented interesting gas separation properties: the O2 permeability of the cross-linked PDMS membrane before UV-ozone exposure was 777 barrer, and the O2/N2 selectivity was 1.9; after UV-ozone exposure, the permeability decreased to 127 barrer while the selectivity increased to 5.4. The free volume depth profile of the SiOx layer was investigated by novel slow positron beam. The results show that free volume size increased with the depth, yet the degree of siloxane conversion to SiOx does not affect the amount of free volume.

Introduction Membrane-based technology has become competitive in both cost and performance in the gas separation industry. Gas separation by polymeric membranes seems to have tremendous potential as an energy-efficient alternative for cryogenic separations. To develop the predominance of polymeric membrane gas separation, new generations of polymeric membranes need to possess both higher permeability and selectivity.1-3 However, there is a trade-off between permeability and selectivity in polymeric membranes. Almost all industrial membrane gas separation processes utilize glassy polymers because of high gas selectivity and good mechanical properties. In glassy polymers, selectivity is controlled by the diffusivity of the penetrants so the more permeable species are those with low molecular diameter.4,5 Although glassy polymers possess adequate selectivity, permeability still in need of improvement. Rubbery polymers present high permeabilities, and their selectivity is mainly influenced by differences in condensability of the penetrants. Among the rubbery polymers, poly(dimethylsiloxane) (PDMS) is considered as a potential material *To whom correspondence should be addressed. (C.-C.H.) Fax: 886 3 4652040. E-mail: [email protected]. (K.-R.L.) Fax: 886 3 2654198. E-mail: [email protected]. (1) Koros, W. J.; Mahajan, R. J. Membr. Sci. 2000, 175, 181–196. (2) Baker, R. W. Ind. Eng. Chem. Res. 2002, 41, 1393–1411. (3) Shantarovich, V. P.; Kevdina, I. B.; Yampolskii, Y. P.; Alentiev, A. Y. Macromolecules 2000, 33, 7453–7466. (4) Burns, R. L.; Koros, W. J. Macromolecules 2003, 36, 2374–2381. (5) Zimmerman, C. M.; Koros, W. J. Polymer 1999, 40, 5655–5664. (6) Jiang, X.; Kumar, A. J. Membr. Sci. 2005, 254, 179–188. (7) Prabhakar, R. S.; Merkel, T. C.; Freeman, B. D.; Imizu, T.; Higuchi, A. Macromolecules 2005, 38, 1899–1910.

4392 DOI: 10.1021/la903445x

for gas separation due to its higher permeable natures for most of the penetrants.6-10 Currently, PDMS is the most commonly used membrane material for separation of organic vapors from permanent gases.11-13 PDMS has high permeability, owing to its large free-volume, and high selectivities for organic vapor. However, it exhibits low selectivity for permanent gases. If we increased the O2/N2 selectivity of PDMS to 4-6, the resultant membrane would have a much higher permanent gas separation performance than almost all other synthetic polymers known at present. Nowadays, the development of high performance membrane materials has involved the synthesis of new polymers, chemical modification of existing polymers, or physical/chemical modification of formed membranes.14-17 Surface modification of formed membranes is a fast and easier way in all of those methods; therefore, a highly selective poly(dimethylsiloxane) membrane was developed by using UV-ozone treatment in this study. Over the past decades, thin SiOx barrier coatings deposited on polymer substrates has increasingly grown in food packaging, (8) Jha, P.; Mason, L. W.; Way, J. D. J. Membr. Sci. 2006, 272, 125–136. (9) Senthilkumar, U.; Reddy, B. S. R. J. Membr. Sci. 2007, 292, 72–79. (10) Shi, Y.; Burns, C. M.; Feng, X. J. Membr. Sci. 2006, 282, 115–123. (11) Pinnau, I.; He, Z. J. Membr. Sci. 2004, 244, 227–233. (12) Merkel, T. C.; Bondar, V. I.; Nagai, K.; Freeman, B. D.; Pinnau, I. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 415–434. (13) Singh, A.; Freeman, B. D.; Pinnau, I. J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 289–301. (14) Al-Masri, M.; Kricheldorf, H. R.; Fritsch, D. Macromolecules 1999, 32, 7853–7858. (15) Wind, J. D.; Staudt-Bickel, C.; Paul, D. R.; Koros, W. J. Macromolecules 2003, 36, 1882–1888. (16) Tiemblo, P.; Guzman, J.; Riande, E.; Mijangos, C.; Reinecke, H. Macromolecules 2002, 35, 420–424. (17) Staudt-Bickel, C.; Koros, W. J. J. Membr. Sci. 1999, 155, 145–154.

Published on Web 12/08/2009

Langmuir 2010, 26(6), 4392–4399

Fu et al.

medical device industries, microelectronics, and gas separation membranes.18,19 Published results illustrated that SiOx films reduce oxygen and water permeation rates and provide high gas selectivity.20 The surface of PDMS can be transformed to SiOx by a variety of techniques. Pyrolytic degradation of PDMS has been found to produce silicon oxide films. However, this procedure requires high temperature (>800 °C), and the ceramic yield is usually low. The high temperature requirement and cost may be prohibitive for these processes for industrial applications. It has therefore been of interest to develop SiOx fabrication processes which can be applied at low temperatures. Maruyama and Tago21 proposed silicon oxide thin films prepared from the precursor silicon tetraacetate by a direct photochemical vapor deposition method. Thin films were obtained in the presence of oxygen gas at a substrate temperature of above 110 °C. Ferguson et al. reported the treatment of poly(dimethylsiloxane) with an oxygen plasma that produced a silicate layer directly.22 Once again, chemical vapor deposition and plasma require special environment conditions and may involve toxic siloxane precursors. An alternative approach that uses reactive oxygen and UV radiation is UVozone treatment. In comparison with chemical vapor deposition and plasma systems, UV-ozone equipment is relatively inexpensive. Mirely and Koberstein presented a simpler UV-ozone roomtemperature process.23,24 Langmuir-Blodgett monolayers of end-functional poly(dimethylsiloxane) were converted to SiOx by exposure to UV light in the presence of atmosphere oxygen at ambient conditions. Ozone is generated in situ from atmospheric oxygen by exposure to 184.7 nm UV light. Ozone absorbs 254.7 nm UV light and subsequently photodissociates into molecular oxygen and atomic oxygen.25 Atomic oxygen and ozone can react strongly with organic materials. Adsorption of UV light by polymer can lead to the formation of radicals. The organic radicals react with atomic oxygen or ozone to form volatile fragments such as CO2 and H2O, which can desorb from the surface. As such, UV-ozone treatment has been used effectively to remove thin layers of organic materials. For PDMS, UVozone treatment is effective in removing the organic portion of the PDMS while the siloxane component is converted to SiOx. Ouyang et al. reported that a thin silicon oxide layer can be created on the surfaces of a variety of siloxane polymers by UVozone treatment.26 Different siloxane polymers exhibit rates of photoinduced conversion that can be related to details of their chemical structures. Extending these results by the same authors, a thin SiOx surface was formed on porous Nylon membranes coated with a cross-linked polysiloxane by UV-ozone treatment. The resultant composite membranes exhibited good gas separation performance.27 Because the microstructure of the surface treatment layer is critical to gas separation performance of UV-ozone treated PDMS membranes, surface microstructure analysis therefore becomes vital to the understanding of gas transport problems. Positron annihilation spectroscopy (PAS) is a special nondestruc(18) Erlat, A. G.; Spontak, R. J.; Clarke, R. P.; Robinson, T. C.; Haaland, P. D.; Tropsha, Y.; Harvey, N. G.; Vogler, E. A. J. Phys. Chem. B 1999, 103, 6047–6055. (19) Sharma, J.; Berry, D. H.; Composto, R. J.; Dai, H.-L. J. Mater. Res. 1999, 14, 990–994. (20) Tropsha, Y. G.; Harvey, N. G. J. Phys. Chem. B 1997, 101, 2259–2266. (21) Maruyama, T.; Tago, T. Thin Solid Films 1993, 232, 201–203. (22) Ferguson, G. S.; Chaudhury, M. K.; Biebuyck, H. A.; Whitesides, G. M. Macromolecules 1993, 26, 5870–5875. (23) Mirley, C. L.; Koberstein, J. T. Langmuir 1995, 11, 1049–1052. (24) Mirley, C. L.; Koberstein, J. T. Langmuir 1995, 11, 2837–2839. (25) Egitto, F. D.; Matienzo, L. J. J. Mater. Sci. 2006, 41, 6362. (26) Ouyang, M.; Yuan, C; Muisener, R. J.; Boulares, A.; Koberstein, J. T. Chem. Mater. 2000, 12, 1591–1596. (27) Ouyang, M.; Muisener, R. J.; Boulares, A.; Koberstein, J. T. J. Membr. Sci. 2000, 177, 177–187.

Langmuir 2010, 26(6), 4392–4399

Article

Figure 1. Chemical structure of the cross-linking agent.

tive evaluation technique for materials characterization which uses the positron.28 PAS was developed as a useful tool to probe the microscopic properties of polymeric materials during the recent decades. One of the great successes in this line of research is the direct determination of polymer free-volume and hole properties at an atomic scale (0.2-2 nm). Recent investigations in polymer materials using PAS have been reported and discussed mainly in the average bulk properties using the conventional PAL technique.29-34 Current positron spectroscopy uses a variable monoenergy positron beam (from the electronvolts range to several kiloelectronvolts)35 coupled with PAL and momentum density measurements, and it is capable of probing defect profiles from the surface, interfaces, and to the bulk. In this work, we use the method of positron annihilation radiation, Doppler broadening of energy spectra, coupled with a slow positron beam to examine the free volume profile of surface treatment layer. In this work, we employ the UV-ozone conversion method to create silicon oxide surface layers on dense PDMS membranes and develop high performance gas separation membranes. The physical characteristics, chemical composition, and permeability of these PDMS membranes are determined as a function of UVozone exposure time and cross-linking agent content. Gas permeation measurements show that gas barrier properties increase significantly with exposure to UV-ozone and that selectivity greater than 5 can be attained for the O2/N2 system.

Experimental Section Membrane Preparation. Silicon rubber (PDMS), crosslinking agent (SYL-OFF 7048), and catalyst (SYL-OFF 4000) were obtained from the Dow Corning Corporation. The chemical structure of the cross-linking agent is shown in Figure 1. Proper cross-linking agent and catalyst were added into the PDMS prepolymer and stirred for 1 h at room temperature. The resultant solution was cast onto Teflon substrates. Samples were cured at 120 °C for 12 h, before being exposed to UV-ozone. The thickness of the resultant membrane was about 180 μm. A low-pressure mercury-quartz lamp, generating UV emissions at 185 and 254 nm (140 mW cm-2), was used as UV source. The ozone is simultaneously generated during the UV exposure period from atmospheric oxygen. The samples were placed approximately 5 cm from the lamp. The mechanism of the photochemical oxidation of PDMS to silicon oxide, as shown in Figure 2, has been discussed in the literature.36 In this work, a particular membrane is denoted by a letter-number code. The former two (28) Jean, Y. C.; Mallon, P. E.; Schrader, D. M. Principles and Applications of Positron and Positronium Chemistry; World Sci. Press: Singapore, 2003. (29) Consolati, G.; Pegoraro, M.; Quasso, F.; Severini, F. Polymer 2001, 42, 1265–1269. (30) Nagel, C.; Gunther-Schade, K.; Fritsch, D.; Strunskus, T.; Faupel, F. Macromolecules 2002, 35, 2071–2077. (31) Sharma, A.; Thampi, S. P.; Suggala, S. V.; Bhattacharya, P. K. Langmuir 2004, 20, 4708–4714. (32) Satyanarayana, S. V.; Bhattacharya, P. K. J. Membr. Sci. 2004, 238, 103– 115. (33) Kruse, J.; Kanzow, J.; Ratzke, K.; Faupel, F.; Heuchel, M.; Frahn, J.; Hofmann, D. Macromolecules 2005, 38, 9638–9643. (34) Wang, X.-Y.; Willmore, F. T.; Raharjo, R. D.; Wang, X.; Freeman, B. D.; Hill, A. J.; Sanchez, I. C. J. Phys. Chem. B 2006, 110, 16685–16693. (35) Coleman, P. Positron Beams and Their Applications; World Sci. Press: Singapore, 2000. (36) Graubner, V.-M.; Jordan, R.; Nuyken, O.; Schnyder, B.; Lippert, T.; K€otz, R.; Wokaun, A. Macromolecules 2004, 37, 5936–5943.

DOI: 10.1021/la903445x

4393

Article

Fu et al.

Figure 2. Mechanism of the photochemical oxidation of PDMS. numbers identify the cross-linking agent content (parts of hundred resin (phr)); the later three numbers identify the UV-ozone exposure time. For example, PDMS30060 denotes that a PDMS membrane contains 30 phr cross-linking agent and 60 min exposure to UV-ozone. Membrane Characterization. The attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra of nonUV-ozone exposure and UV-ozone exposure PDMS membranes were measured using a Perkin-Elmer FTIR spectrometer (PerkinElmer Cetus Instruments, Norwalk, CT) in the wavenumber range 600-4000 cm-1. Contact angles of water on the membrane were measured by using a face contact angle meter CA-D type (Kyowa Interface Science Co. Ltd.). Deionized water was used as the probe liquid. All measurements were taken at ambient temperature. The reported values are usually the average of at least 10 measurements, taken at different locations on the surface. The surface topography of the sample was characterized by atomic force microscopy (Digital Instrument, DI5000). A scanning electron microscope (model S-3000, Hitachi) was used to observe surfaces of the PDMS membranes. Chemical compositions of UV-ozone exposure layers were analyzed qualitatively using an energy-dispersive X-ray (EDX) microanalysis system. Surface compositions of the membranes were also measured using a Thermo Fisher Scientific Thermo Scientific K-Alpha X-ray photoelectron spectrometer (XPS), equipped with a monochromatic Al KR X-ray source (1486.6 eV). The S parameters of Doppler broadening energy spectra (DBES) were measured as a function of positron implantation energy (depth) at room temperature using the slow positron beam (0-30 keV) at the membrane center of Chung Yuan University. The DBES spectra were recorded as a function of positron energy from 100 eV to 30 keV. The DBES spectra were measured using an HP Ge detector (EG&G Ortec. with 35% efficiency and energy resolution of 1.5 at 511 keV peak) at a counting rate of approximately 2500 cps. The total number of counts for each DBES spectrum was 2  106. The obtained DBES spectra were characterized by an S parameter, defined as the ratio of integrated counts between energy 510.3 and 511.7 keV (S width) to total counts after the background was properly subtracted. Since the S parameter represents the relative value of the low momentum part of positronelectron annihilation radiation, it is sensitive to the change of the positron and positronium (Ps) states due to microstructural changes. When the positron and Ps are localized in a hole or free volume with a finite size, the observed S parameter is a measure of the momentum broadening according to the uncertainty principle: a larger hole results in a larger S parameter value and the amount of parapositronium (singlet state). The S parameter has been successfully used in detecting the free-volume depth profile in polymeric systems.37,38 Positron annihilation lifetime experiments were performed using a variable monoenergy 30 keV positron beam. We also used the conventional PAL method to measure the free-volume properties in the bulk membrane samples. The PAL data contain quantitative information on the free-volume properties in polymeric systems from the surface, interfaces, and to the bulk. The lifetime resolution was (37) Jean, Y. C.; Zhang, R.; Cao, H.; Yuan, J.-P.; Huang, C.-M.; Nielsen, B.; Asoka-Kumar, P. Phys. Rev. B 1997, 56, 8459–8462. (38) Li, Y.; Zhang, R.; Chen, H.; Zhang, J.; Suzuki, R.; Ohdaira, T.; Feldstein, M. M.; Jean, Y. C. Biomacromolecules 2003, 4, 1856–1864.

4394 DOI: 10.1021/la903445x

Figure 3. ATR-FTIR spectra of silica, non-UV-ozone exposure, and UV-ozone exposure PDMS membranes. 450-800 ps for a counting rate of 100-500 cps. Each PAL spectrum contains 2  106 counts. The obtained PAL data were fitted into three lifetime components using the PATFIT program. Analyzed results of positron lifetimes (τ1, τ2, and τ3) and intensities (I1, I2, and I3) from PAL spectra are attributed to the positron and positronium annihilation in polymeric membrane materials. The shortest τ1 (0.125 ns) is from p-Ps annihilation, τ2 (0.45 ns) is from the positron annihilation, and τ3 is due to o-Ps annihilation. Since Ps is known to preferentially localize in defect sites, particularly in the free volume before annihilation takes place, parameters from o-Ps annihilation have been successfully used to obtain the electron properties and depth profiles of free volumes in thin film polymers. The o-Ps lifetime τ3 is on the order of 1-5 ns in polymeric materials, the so-called pick-off annihilation with electrons in molecules, and is used to calculate the mean free-volume radius R (A˚ to nm) based on an established semiempirical correlation equation from a spherical-cavity model.39 Gas Permeation Measurements. A gas permeation analyzer (Yanaco GTR10) was used to measure pure-gas permeability coefficients for the PDMS membranes to O2 and N2. The tests were carried out under isothermal conditions at 35 °C ((0.5 °C). The usual unit of permeability (P) is the barrer [10-10 (cm3 (STP) cm)/(cm2 s cm Hg)]. Ideal selectivities were calculated as the ratio of permeability coefficients: RA=B ¼

PA PB

ð3Þ

where PA and PB are the permeability coefficients of pure gases A and B.

Results and Discussion The gas transport behavior through a novel UV-ozone exposure PDMS membrane can be influenced by the intrinsic properties of the pristine membrane and the UV-ozone exposure conditions. The optimal cross-linking agent content and UVozone exposure time is envisaged to achieve the best gas separation performance of UV-ozone exposure PDMS membranes. Surface Characterization of PDMS Membranes. The observed degradation and modification of the PDMS membranes from the ATR-FTIR spectra as a function of UV-ozone exposure time is shown in Figure 3. FTIR bands at 780, 1020, and (39) Nakanishi, N.; Wang, S. J.; Jean, Y. C. In Positron Annihilation Studies of Fluids; Sharma, S. C., Ed.; World Scientific: Singapore 1988; p 292.

Langmuir 2010, 26(6), 4392–4399

Fu et al.

Article Table 2. Experimental (Determined by EDX) Surface Chemical Compositions of PDMS Membranes Exposed for Different Times membranes element

PDMS30000

PDMS30030

PDMS30060

C1s (%) O1s (%) Si2p (%) O/Si

44.12 27.48 28.40 0.97

38.05 33.18 28.77 1.15

33.21 37.62 29.10 1.29

Table 3. Experimental (Determined by XPS) Surface Chemical Compositions of PDMS Membranes Exposed for Different Times membranes

Figure 4. Variation of water contact angle as a function of UVozone exposure time for PDMS membranes containing 30 phr cross-linking agent. Table 1. Variation in Surface Roughness of PDMS Membranes Determined by AFM membrane

rms (nm)

PDMS30000 PDMS30060 PDMS30120 PDMS30240

0.656 0.884 1.468 2.355

1250 cm-1 are characteristic of PDMS bands reported in transmission FTIR studies. The bands observed at 780 and 1250 cm-1 correspond, respectively, to the Si-C rocking vibrations and the -CH3 deformation in the Si(CH3)2 groups. The intensity of the absorption bands at 780 and 1250 cm-1 were reduced considerably by oxidation of the PDMS and decreased with increasing UV-ozone exposure time. In addition, the broad and strong bands with maxima at 1020 cm-1 correspond to the asymmetric Si-O-Si stretching vibrations. Upon UV-ozone exposure, Si-O-Si bands at 1020 cm-1 undergo significant changes in shape and frequency. The intensity decreases consistent with the expectation that the Si coordinated to two oxygen atoms in PDMS is being converted to Si coordinated with four oxygen atoms as in SiO2. The ATR-FTIR results confirm the oxidation of PDMS to form a thin surface film of SiOx. Figure 4 shows water contact angle as a function of UV-ozone exposure time for PDMS membranes. Reduction in the contact angle from a value greater than 100° to a value less than 60° is due to the presence of SiOx. Bettens et al. presented that the contact angle of water on a silica membrane was 52°.40 This result reconfirms the oxidation of PDMS to form a thin surface film of SiOx. Atomic force microscopy (AFM) was used to evaluate changes in surface topography that occur during the UV-ozone conversion process. AFM pictures and surface roughness results are shown in Figure 5 and Table 1, respectively. Surface roughness increases remarkably with treatment time. For PDMS, UV-ozone treatment is effective in removing the organic portion of the PDMS while the siloxane component is converted to silicon oxides. Hence, the evaporation of low molecular weight volatile fragments resulted in a surface roughness increase. Surface composition analysis of PDMS membranes, prior to and after various periods of UV-ozone exposure, was carried out by EDX and XPS. The probing depths of EDX and XPS were (40) Bettens, B.; Dekeyzer, S.; Van der Bruggen, B.; Degreve, J.; Vandecasteele, C. J. Phys. Chem. B 2005, 109, 5216–5222.

Langmuir 2010, 26(6), 4392–4399

element

PDMS30000

PDMS30030

PDMS30060

C1s (%) O1s (%) Si2p (%) O/Si siloxane (%) SiOx (%)

45.50 25.72 28.78 0.89 100 0 0 0

24.42 46.49 29.09 1.60 37.32 0 62.68 0

11.00 59.89 29.11 2.02 0 17.94 0 82.06

101.9 eV 102.8 eV 103.2 eV 103.6 eV

smaller than 2 μm and 10 nm, respectively. Combined, these two techniques can probe the siloxane conversion difference between the very near surface and deeper surface. Atomic compositions obtained from EDX are shown in Table 2. It is apparent that carbon composition decreases and oxygen composition increases with exposure time; moreover, the silicon composition almost keeps the same. Oxygen composition increments result from conversion of silicones to silicon oxides as well as oxidation of carbon-containing groups. The O/Si ratios were observed to increase from 0.97 to 1.29 after exposure in UV-ozone for 60 min. It indicates that UV-ozone exposure dose not completely remove PDMS fragments but modifies the membrane considerably. Surface composition analysis of PDMS membranes was also accomplished with XPS. XPS results are reported in Table 3 and illustrate the change trend of atomic compositions and O/Si ratios in a consistent manner as the results in Table 2. However, the ranges of these changes in Table 3 were more dramatic. For the PDMS30060 membrane, the O/Si ratio approximates the ideal ratio of SiO2 (O/Si = 2). Comparing Tables 3 and 2, we can conclude that the degree of silicones converted into silicon oxides in very near surface of the PDMS membrane is higher than that in the deeper surface. The binding energy for Si(2p) centered at 101.9 eV is attributed to silicon coordinated to two oxygen atoms. The non-UV-ozone exposure PDMS membrane had a Si(2p) binding energy of 101.9 eV (see Table 3), indicating the existence of a pure PDMS phase. The nature of changes in the silicon configuration during UV-ozone exposure can be monitored by analysis of the Si(2p) binding energy. After exposure to UVozone, the Si(2p) binding energy shifted slightly toward higher electron voltages than that of the pure PDMS phase, indicating the existence of partial SiOx domains. In the case of membrane PDMS30030, 62.7% siloxane converts to SiOx. The XPS Si(2p) spectra for the PDMS30060 membrane indicate that the siloxane essentially completes conversion to SiOx. The Si(2p) binding energy at 103.6 eV was found in amorphous SiO2, which suggests that 82% of the surface of PDMS30060 membrane is very nearly that of silicon dioxide. From the results in Tables 2 and 3, we can deduce that higher conversion rates need longer UV-ozone exposure time. It is generally considered that the conversion of siloxane during UV-ozone exposure may be a function of cross-linking agent content. Table 4, which shows the atomic compositions, O/Si and DOI: 10.1021/la903445x

4395

Article

Fu et al.

Figure 5. AFM pictures of PDMS30000 (a), PDMS 30060 (b), PDMS30120 (c), and PDMS30240 (d). Table 4. Experimental (Determined by EDX) Surface Chemical Compositions of PDMS Membranes Containing Different Cross-Linking Agent Content membranes element

PDMS30000

PDMS10060

PDMS30060

PDMS40060

PDMS50060

C (%) O (%) Si (%) O/Si O/C

44.12 27.48 28.40 0.97 0.62

36.83 34.35 28.82 1.19 0.93

33.21 37.62 29.10 1.29 1.13

43.48 32.81 23.71 1.38 0.75

29.44 41.36 29.20 1.42 1.40

O/C atomic ratios for UV-ozone exposure membranes, as a function of the cross-linking agent content, indicates that carbon composition decreases and oxygen composition increases with cross-linking agent content. The higher amount of oxygen at the UV-ozone treated surface is mostly the result of silicones converting to silicon oxides as well as oxidation of carbon-containing groups as discussed previously. O/Si and O/C atomic ratios shift to a higher value after 60 min of UV-ozone exposure, suggesting that higher cross-linking agent content seemed favorable for silicones to convert to silicon oxides. It is clear that the siloxane on the PDMS surface is not completely converted to siliconoxides (O/Si < 2). In order to examine very near surface composition, XPS was used. Table 5 shows the very near surface 4396 DOI: 10.1021/la903445x

composition changes of UV-ozone exposure PDMS membranes with different cross-linking agent content. For UV-ozone exposure membranes, O/Si atomic ratios approximate to 2 as the crosslinking agent content is higher than 10 phr. It is indicated that the cross-linking agent content will not affect the siloxane conversion ability of PDMS membranes. However, the cross-linking agent content will affect the SiO2 fraction on the membrane surface, and PDMS membrane content of 30 phr cross-linking agent has the highest SiO2 conversion. Membrane Gas Separation Performance and Free Volume. To investigate the improvement in gas separation performance of PDMS membranes, the gas permeation properties of UV-ozone exposure membranes were measured. The O2 permeability of Langmuir 2010, 26(6), 4392–4399

Fu et al.

Article

Table 5. Experimental (Determined by XPS) Surface Chemical Compositions of PDMS Membranes Containing Different Cross-Linking Agent Content membranes element C (%) O (%) Si (%) O/Si O/C siloxane (%) SiOx (%)

BE (eV)

101.9 102.8 103.2 103.6

PDMS30000

PDMS10060

PDMS30060

PDMS40060

PDMS50060

45.50 25.72 28.78 0.89 0.57 100 0 0 0

13.84 57.58 28.58 2.01 4.16 0 37.94 0 62.06

11.00 59.89 29.11 2.02 5.44 0 17.94 0 82.06

10.11 60.45 29.44 2.05 5.98 0 9.00 67.02 23.98

9.82 60.84 29.34 2.07 6.20 0 0 68.57 31.43

Figure 6. Variation of permeability and selectivity as a function of UV-ozone exposure time for PDMS membranes containing 30 phr cross-linking agent.

PDMS30000 membrane was 777 barrer and the O2/N2 selectivity was 1.9 as shown in Figure 6. Besides, the O2 permeability and O2/N2 selectivity of the silica membrane were 0.51 barrer and 7.28, respectively.41 The data presented in Figure 6 show that, for increasing UV-ozone exposure time, the permeability of oxygen decreases first and then increases as the exposure time exceeds 60 min. However, as the UV-ozone exposure time increased, the variation of O2/N2 selectivity is the opposite of permeability. The decrease in permeability can be attributed to the composition and free volume change due to the UV-ozone treatment. That is, as the SiOx forms, it becomes a barrier that deduces free volume size. This result indicates that the surface of a SiOx layer acts as a barrier film to the penetration of gas molecules and leads to an increase in selectivity. For exposure times exceeding 60 min, the oxygen permeability increased but the selectivity decreased. Figure 7 illustrates the results obtained by SEM. There are some cracks formed on the PDMS30090 membrane surface, suggesting less effective gas barrier properties were found for long time UV-ozone exposure PDMS membranes. Figure 8 presents S parameters as a function of positron energy for the PDMS membranes. The S parameter of the non-UVozone exposure PDMS membrane (PDMS30000) increases rapidly and then reaches a plateau. The plateau presents that the property of free volume is homogeneous in the non-UV-ozone exposure membrane. However, as the positron energy increased, the rate of increase in the S parameter of PDMS30060 considerably increased. This means that the size of free volume in the UVozone exposure layer increases from surface to bulk. S parameter curves superposed at 27 keV present that the thickness of the SiOx (41) Molyneux, P. J. Membr. Sci. 2008, 320, 42–56.

Langmuir 2010, 26(6), 4392–4399

Figure 7. Surface SEM images of PDMS30000 (a), PDMS 30060 (b), and PDMS30090 (c).

conversion layer is smaller than 6 μm. The ratio of the thickness of the SiOx layer to that of the overall membrane was about 0.033. In DOI: 10.1021/la903445x

4397

Article

Fu et al.

Figure 10. Variation of permeability and selectivity as a function of cross-linking agent content for 60 min UV-ozone exposure PDMS membrane.

Figure 8. Variation of S parameters as a function of positron energy: effect of UV-ozone exposure time.

Figure 11. Variation of S parameters as a function of positron Figure 9. Variation of permeability and selectivity as a function of cross-linking agent content for non-UV-ozone exposure PDMS membrane.

addition, the S parameter was higher for membrane PDMS30000 compared to membrane PDMS30060. This result can explain the fact that the conversion by UV-ozone exposure creates a smaller free-volume size surface layer, which acted as a barrier to gas transport. To aid in understanding the contributions of the cross-linking agent in the membrane, the gas permeation properties of non-UVozone exposure and UV-ozone exposure membranes were measured. The data presented in Figure 9 show that, for increasing cross-linking agent content, the permeability of oxygen slightly decreases in the non-UV-ozone exposure PDMS membranes. It would be expected that increasing the cross-linking agent content leads to an increased degree of cross-linking of the membrane. This increase in degree of cross-linking leads probably to a decrease in polymer free volume. With this premise, it is not surprising that permeability slightly decreased with the crosslinking agent content. Because of the rubbery and high freevolume nature of PDMS membranes, the effect of cross-linking agent content on oxygen permeation rate is similar to nitrogen. Therefore, the selectivity for all of the non-UV-ozone exposure PDMS membranes is almost the same (Figure 9). Figure 10 presents permeability and selectivity as a function of cross-linking agent content for the UV-ozone exposure PDMS membranes. 4398 DOI: 10.1021/la903445x

energy: effect of cross-linking agent content.

The O2/N2 selectivities increase rapidly up to a cross-linking agent content of about 30 phr and then decrease at higher content. In the case of a cross-linking agent content of 30 phr, the membrane showed the highest fraction of silicon oxide, as shown in Table 5, that was converted from the siloxane. This result suggests the fraction of silicon oxide in the PDMS membrane significantly influenced the selectivity of the membrane. In the present study, the highest O2/N2 selectivity is about 5.4. A comparable result was observed by Ouyang et al. in their work with PDMS membranes.27 Park et al. have investigated the effect UV-ozone exposure on the conversion of the PDMS layer in poly(imide siloxane) films.42 The oxygen permeability was 2.16 barrer for the UV-ozone exposure poly(imide siloxane) film with 5.4 O2/N2 selectivity. In this work, the oxygen permeability of the membrane possessing 5.4 O2/N2 selectivity was 127 barrer. The comparison shows that the UV-ozone exposure PDMS membranes possess better gas separation properties than their copolymers, and that formation of a surface SiOx layer by suitable cross-linking agent content provides improved selectivity and better barrier properties. Figure 11 shows the S parameter variation of the UV-ozone exposure PDMS membranes with different cross-linking agent content as a function of positron energy. The values of the S parameter for membrane PDMS30060 was lower than that for (42) Park, H. B.; Han, D. W.; Lee, Y. M. Chem. Mater. 2003, 15, 2346–2353.

Langmuir 2010, 26(6), 4392–4399

Fu et al.

Article

Figure 12. Variation of free-volume radius and intensity (I3) in PDMS30060 membrane as a function of positron energy.

Figure 13. Performance of UV-ozone exposure PDMS mem-

membrane PDMS10060. This observation is consistent with the change of selectivity in UV-ozone exposure PDMS membranes, shown in Figure 10, and indicates that high silicon oxide fraction or low free volume is responsible for the higher selectivity. For UV-ozone exposure PDMS membranes, there is a close relationship between free-volume properties on the membrane surface and gas separation properties. Understanding this relation is very important to tailor a UV-ozone exposure PDMS membrane for specific gas separation purposes. Slow positron beam is a unique and useful tool for investigating the free-volume depth profile on membrane surface layers. Figure 12 presents free volume size and number as a function of position incident energy. The free volume radii on UV-ozone exposure PDMS membrane surfaces were observed to increase from 2 to 3.5 A˚ as the positron incident energy increased from 1 to 10 keV. This is not surprising, since UV-ozone exposure PDMS exhibits almost complete SiO2 conversion in the very near surface but exhibits low SiO2 conversion in the deeper surface (Tables 4 and 5). The very near surface of the UV-ozone exposure layer has the smallest free-volume radius. It was proposed that an increase in selectivity and decrease in permeability observed after UV-ozone exposure was mainly due to shrinkage of the free-volume radius in the very near surface. The free-volume radius and I3 for the bulk non-UV-ozone exposure PDMS membrane were 3.98 A˚ and 19.7%, respectively. The higher free-volume radius of the bulk non-UV-ozone exposure PDMS membrane indicates that the thickness of the SiOx conversion layer should more than 2 μm. The values of I3 represented in Figure 12, the amount of free volume, varied in a very small range with positron incident energy. One can assume that the degree of siloxane converted to SiOx does not affect the amount of free volume on the surface of the UV-ozone exposure PDMS membrane. A comparison of O2/N2 separation performance for UV-ozone exposure PDMS membranes prepared in this work with other traditional polymeric materials used in gas separation membranes is shown in Figure 13. The solid line represents an “upper bound” proposed by Robeson above which no polymers are currently known to exit.43,44 The gas separation performance of UV-ozone

exposure PDMS membranes is just located on the upper bound; this result demonstrates that UV-ozone exposure PDMS membranes provide interesting gas separation properties and possess a potential for commercial application.

Langmuir 2010, 26(6), 4392–4399

branes.

Conclusions A thin SiOx layer can be created on the surfaces of a variety of cross-linked PDMS membranes by a UV-ozone induced photochemical conversion process. The majority of carbonaceous species are removed from the membrane surface, indicating the conversion of organic siloxane into inorganic SiOx. PDMS membranes with higher cross-linking agent content convert more rapidly while longer UV-ozone exposure time facilitates the siloxane conversion. We found that UV-ozone exposure increases the surface roughness and decreases the water contact angle of the PDMS membranes, which was due to the etching effect and the formation of SiOx. The O2 permeability of the cross-linked PDMS membrane before UV-ozone exposure was 777 barrer, and the O2/ N2 selectivity was 1.9; after UV-ozone exposure, the permeability decreased to 127 barrer while the selectivity increased to 5.4. These results demonstrate that UV-ozone exposure PDMS membranes provide interesting gas separation properties and that their barrier properties may be increased markedly by application of an UVozone exposure to create a surface SiOx layer. Slow positron beam results demonstrate the free volume changes which are due to UVozone exposure. The free-volume depth profile indicates that freevolume size increases as the distance below the surface increases; moreover, the effect of the degree of SiO2 conversion on the amount of free volume is not obvious. From the S parameter curve, it was possible to observe that the thickness of the SiOx conversion layer is smaller than 6 μm. Acknowledgment. The authors wish to sincerely thank the Ministry of Economic Affairs and the National Science Council of Taiwan, ROC, for financially supporting this project. (43) Robeson, L. M. J. Membr. Sci. 1991, 62, 165–185. (44) Robeson, L. M. J. Membr. Sci. 2008, 320, 390–400.

DOI: 10.1021/la903445x

4399