Changes in Silicon Elastomeric Surface Properties under Stretching

Institut de Chimie des Surfaces et Interfaces, I.C.S.I., C.N.R.S., UPR 9069, 15, Rue Jean Starcky, 68057. Mulhouse Cedex, France, Institut Charles Sad...
0 downloads 0 Views 1MB Size
Download PDF
13136

Langmuir 2007, 23, 13136-13145

Changes in Silicon Elastomeric Surface Properties under Stretching Induced by Three Surface Treatments V. Roucoules,*,† A. Ponche,† A. Geissler,† F. Siffer,† L. Vidal,† S. Ollivier,† M. F. Vallat,† P. Marie,‡ J. C. Voegel,§ P. Schaaf,‡ and J. Hemmerle´§ Institut de Chimie des Surfaces et Interfaces, I.C.S.I., C.N.R.S., UPR 9069, 15, Rue Jean Starcky, 68057 Mulhouse Cedex, France, Institut Charles Sadron, I.C.S., C.N.R.S., UPR 9069, 6, Rue Boussingault, 67083 Strasbourg Cedex, France, and Institut National de la Sante et de la Recherche Medicale, Unite 595, 11, Rue Humann, 67085 Strasbourg Cedex, France ReceiVed May 18, 2007. In Final Form: September 21, 2007 Poly(dimethylsiloxane) (PDMS) substrates are used in many applications where the substrates need to be elongated and various treatments are used to regulate their surface properties. In this article, we compare the effect of three of such treatments, namely, UV irradiation, water plasma, and plasma polymerization, both from a molecular and from a macroscopic point of view. We focus our attention in particular on the behavior of the treated surfaces under mechanical stretching. UV irradiation induces the substitution of methyl groups by hydroxyl and acid groups, water plasma leads to a silicate-like layer, and plasma polymerization causes the formation of an organic thin film with a major content of anhydride and acid groups. Stretching induces cracks on the surface both for silicate-like layers and for plasma polymer thin coatings. This is not the case for the UV irradiated PDMS substrates. We then analyzed the chemical composition of these cracks. In the case of water plasma, the cracks reveal native PDMS. In the case of plasma polymerization, the cracks reveal modified PDMS. The contact angles of plasma polymer and UV treated surfaces vary only very slightly under stretching, whereas large variations are observed for water plasma treatments. The small variation in the contact angle values observed on the plasma polymer thin film under stretching even when cracks appear on the surface are explained by the specific chemistry of the PDMS in the cracks. We find that it is very different from native PDMS and that its structure is somewhere between Si(O2) and Si(O3). This is, to our knowledge, the first study where different surface treatments of PDMS are compared for films under stretching.

Introduction Cross-linked siloxane elastomers are among the most widely used class of polymers that have been the subject of numerous studies.1-4 The reason for this widespread use is related to an attractive combination of materials properties that include a hydrophobic surface, high electrical resistance, low toxicity, and flexibility.5,6 However, when these materials are involved in microfluidic devices or used for medical applications, the advantages announced previously become problematic.7-10 In the literature, various approaches have been used to modify poly(dimethylsiloxane) (PDMS) surfaces, including (i) chemical methods, such as adsorption of specific molecules,11-14 silaniza* To whom correspondence should be addressed. E-mail: Vincent. [email protected]. † Institut de Chimie des Surfaces et Interfaces. ‡ Institut Charles Sadron. § Institut National de la Sante et de la Recherche Medicale. (1) Hillborg, H.; Gedde, U. W. IEEE Trans. Dielectr. Electr. Insul. 1999, 6, 703-717. (2) Pike, J. K.; Ho, T.; Wynne, K. J. Chem. Mater. 1996, 8, 856-860. (3) Ikada, Y. Biomaterials 1994, 15, 725-736. (4) Kim, J.; Chaudhury, M. K.; Owen, M. J. IEEE Trans. Dielectr. Electr. Insul. 1999, 6, 695-702. (5) Gorur, R.; Cherney, E.; Burnham, J. Outdoor Insulation. 8208 short courses notes, College of Engineering and Applied Science, Arizona State University, 1998. (6) Campbell, B. Dow Corning Corp. Paper, 1996. (7) Niwano, M.; Kinashi, K.; Saito, K.; Miyamoto, N.; Honma, K. J. Electrochem. Soc. 1994, 141, 1556-1561. (8) Joubert, O.; Hollinger, G.; Fiori, C.; Devine, R. A. B.; Paniez, P.; Pantel, R. J. Appl. Phys. 1991, 69, 6647-6651. (9) Koch, M.; Evans, A.; Brunnschweiler, A. Microfluidic Technology and Applications; Baldock Research Studies Press: Hertfordshire, U.K., 2000. (10) Donzel, C.; Geissler, M.; Bernard, A.; Wolf, H.; Michel, B.; Hilborn, J.; Delamarche, E. AdV. Mater. 2001, 13, 1164-1167. (11) Lenz, P.; Ajo-Franklin, C. M.; Boxer, S. G. Langmuir 2004, 20, 1109211099. (12) Ro, K. W.; Chang, W. J.; Kim, H.; Koo, Y. M.; Hahn, J. H. Electrophoresis 2003, 24, 3253-3259.

tion,15 atom transfer radical polymerization,16 and chemical vapor deposition17 or (ii) physical methods, such as exposure to energy sources as ultraviolet light,18-20 corona discharges,21,22 radioinduced graft polymerization,23,24 glow discharge plasma,25-28 and continuous discharge plasma polymerization.29 Recently, modified PDMS substrates have been used in the design of new types of surface responsive materials.30 They consist of layerby-layer deposition of polyanions and polycations (i.e., polyelectrolyte multilayer deposition) on maleic anhydride pulsed plasma polymer modified PDMS substrates.31,32 Under mechan(13) Wu, D. P.; Luo, Y.; Zhou, X.; Dai, Z.; Lin, B. Electrophoresis 2005, 26, 211-218. (14) Park, J.; Hammond, P. T. Macromolecules 2005, 38, 10542-10550. (15) Hellmich, W.; Regtmeier, D. T. T.; Ros, R.; Anselmetti, D.; Ros, A. Langmuir 2005, 21, 7551-7557. (16) Xiao, D. Q.; Van Le, T.; Wirth, M. J. Anal. Chem. 2004, 76, 2055-2061. (17) Lahann, J.; Balcells, M.; Lu, H.; Rodon, T.; Jensen, K. F.; Langer, R. Anal. Chem. 2003, 75, 2117-2122. (18) Schnyder, B.; Lippert, T.; Ko¨tz, R.; Wokaun, A.; Graubner, V. M.; Nuyken, O. Surf. Sci. 2003, 532-535, 1067-1071. (19) Ola`h, A.; Hillborg, H.; Vansco, G. J. Appl. Surf. Sci. 2005, 239, 410-423. (20) Graubner, V.-M.; Jordan, R.; Nuyken, O.; Schnyder, B.; Lippert, T.; Kotz, R.; Wokaun, A. Macromolecules 2004, 37, 5936-5943. (21) Hillborg, H.; Gedde, U. W. Polymer 1998, 19, 1991-1998. (22) Pocius, A. V.; Kinning, D. J.; Yarusso, D. J. Thakkar, B.; Mangipudi, V. S.; Turrell, M. Plast. Eng. 1997, 53, 31-36. (23) Bae, W. S.; Convertine, A. J.; McCormick, C. L.; Urban, M. W. Langmuir 2007, 23, 667-672. (24) Kim, H.; Urban, M. W. Langmuir 1996, 12, 1047-1050. (25) Bodas, D.; Khan-Malek, C. Microelectron. Eng. 2006, 83, 1277-1279. (26) Ginn, B. T.; Steinbock, O. Langmuir 2003, 19, 8117-8118. (27) Lawton, R. A.; Price, C. R.; Runge, F, A.; Doherty, W. J., III; Saavedra, S. S. Colloids Surf., A 2005, 253, 213-215. (28) Chen, I. J.; Lindner, E. Langmuir 2007, 23, 3118-3122. (29) Barbier, V.; Tatoulian, M.; Li, H.; Arefi-Khonsari, F.; Ajdari, A.; Tabeling, P. Langmuir 2006, 22, 5230-5232. (30) Hemmerle´, J.; Roucoules, V.; Fleith, G.; Nardin, M.; Ball, V.; Lavalle, P.; Marie, P.; Voegel, J. C.; Schaaf, P. Langmuir 2005, 21, 10328-10331. (31) Ryan, M. E.; Hynes, A. M.; Badyal, J. P. S. Chem. Mater. 1996, 8, 37-42.

10.1021/la701460f CCC: $37.00 © 2007 American Chemical Society Published on Web 11/14/2007

Changes in Silicon Elastomeric Surface Properties

ical stimulation (elongation/retraction cycles), these surfaces change their hydrophobicity. Once again, the tailored PDMS substrates with polar groups and a high surface energy have been of great value: the plasma polymer charged groups allowed the polyelectrolyte multilayers to be uniquely deposited by strong electrostatic interactions. However, no further studies were carried out to elucidate the mechanisms involved during the cyclic mechanical process. In particular, the functionalized PDMS substrates must support the effects due to the elongation/retraction cycles. In this context, it will be helpful to compare the behavior under stretching of maleic anhydride pulsed plasma polymer functionalized PDMS substrates with the behavior under stretching of PDMS substrates modified by two commonly used physical methods as (i) UV light irradiation (λ ) 254 nm) and (ii) water plasma. All these treatments create oxygen-containing polar groups at the PDMS surface. UV irradiation at a wavelength of 254 nm generates atomic oxygen, a very strong oxidizing agent. Atomic oxygen is generated in ambient air by the combination of photochemical processes.33 This technique has been used to produce silicate layers, but it suffers from reproducibility. Radio frequency glow discharge plasma has been widely used for the surface modification of various materials. The process is friendly to the environment and induces little damage to the intrinsic properties of the substrates. The plasma contains a complex mixture of high-energy photons, electrons, ions, radicals, and excited species.34 It has been demonstrated that water plasma can be used to obtain hydrophilic polymer surfaces but that the characteristics of the surfaces gradually change during aging and the surfaces recover their hydrophobicity. Although H2O plasma treatment may not alter the surface topography of a substrate in most cases, it has been found to be useful for surface etching by H2O plasma under certain conditions.34 This finding can be both an advantage and a hindrance following the desired final properties required. Plasma polymerization is also a simple one-step dry surface procedure that proceeds at room temperature.35 Most previous studies on plasma polymerization focused on continuous discharge. In this discharge mode, many active species are created by high-energy electrons and ions, which serve as the subsequent reactive species. Therefore, the polymer films are made of a high degree cross-linked structure.36 To meet with the increasing demand for surface functionalization of solid substrates, a pulsed discharge technique has emerged in plasma polymerization.37 This technique entails modulating the electrical discharge on the millisecond to microsecond time scale. In the case of gaseous precursors containing polymerizable carbon-carbon double bonds, pulsed plasma functionalization is comprised of two distinct reaction regimes corresponding to the plasma duty cycle on and off periods, where the on-time generates active sites in the gas phase and at the surface, while conventional polymerization occurs during the subsequent off-time.31,37 This gives rise to extremely high levels of structural retention and incorporation of specific functional groups at the surface. Maleic anhydride plasma polymer thin films are of particular interest (32) Roucoules, V.; Fail, C. A.; Schofield, W. C. E.; Teare, D. O. H.; Badyal, J. P. S. Langmuir 2005, 21, 1412-1415. (33) Ye, H.; Gu, Z.; Gracias, D. H. Langmuir 2006, 22, 1863-1868. (34) Parker, J. L.; Cho, D. L.; Claesson, P. M. J.Phys. Chem. 1989, 93, 61216125. (35) Boening, H. V. Fundamentals of Plasma Chemistry and Technology; Technomic Publishing Company, Inc.: 1988; p 75. (36) Yasuda, H.; Plasma Polymerization; Academic Press: London, 1985. (37) Siffer, F.; Ponche, A.; Fioux, P.; Schultz, J.; Roucoules, V. Anal. Chim. Acta 2005, 539, 289-299.

Langmuir, Vol. 23, No. 26, 2007 13137

because of their reactivity and polarity.30 But, it is still difficult to obtain such plasma polymers since even at relatively low input energies, the anhydride groups are lost and the deposited thin film contains mostly dissociation products rather than the anhydride groups. In recent work, we have shown that under optimal conditions, the retention of the maleic anhydride group could reach a maximum value of 32%.37 In this study, we used UV irradiation (λ ) 254 nm), water plasma, and maleic anhydride pulsed plasma polymerization to modify the surface of a commercial PDMS substrate involved in surface responsive material applications. SEM, AFM, XPS, and contact angle measurements were used as investigating tools, providing a better knowledge of the performance of the thin layers deposited onto the silicon substrate when the latter is stretched. Experimental Procedures Materials. Silicone substrates were molded by Statice Sante´ SAS using MED-4750 from NuSil Silicone Technology LCC. It is a two-part silicone elastomer that undergoes a platinum catalyzed addition reaction between PDMS chains carrying vinyl CdC bonds and a dimethyl methyl hydrogen siloxane copolymer. The blend also contains 30% of amorphous silica. The reaction can be accelerated by heat. PDMS films of 0.5 mm thickness were obtained by molding in a heated press between glass plates (durometer: 50; maximum of elongation: 1000%; tear strength: 44.1 kN/m; and tensile strength: 1450 psi). After removal of the plates, the silicone film was protected by PET sheets for storage before treatment. The transmission infrared spectrum of the film shows no peak corresponding to SiH remaining functional groups, indicating that the reaction is complete. About 3% of extractable material was measured after swelling at equilibrium in cyclohexane. Short chains of PDMS were identified, but all films were used as such after the molding step without further preparation. UV Irradiation. UV irradiation was carried out by exposure for 5 h to a short wavelength mercury lamp with a line emission at λ ) 254 nm. The PDMS sheet was placed at 25 mm from the source. At this distance, the intensity at the sample was 4500 µW/cm2. Plasma Chamber. Water plasma treatments and plasma polymer deposition were carried out in the same reactor. The plasma reactor consists of an electrodeless cylindrical glass reactor (6 cm diameter, 680 cm3 volume, base pressure 5 × 10-4 mbar, and leak rate better than 1.0 × 10-10 kg s-1) enclosed in a Faraday cage. The chamber was fitted with a gas inlet, a Pirani pressure gauge, a two-stage rotary pump (Edwards) connected to a liquid nitrogen cold trap, and an externally wound copper coil (4 mm diameter, 5 turns). All joints were grease-free. An L-C matching network (Dressler, VM 1500 W--ICP) was used to match the output impedance of a 13.56 MHz radio frequency (rf) power supply (Dressler, Cesar 133) to the partially ionized gas load by minimizing the standing wave ratio of the transmitted power. During electrical pulsing, the pulse shape was monitored with an oscilloscope, and the average power 〈P〉 delivered to the system was calculated using the following expression: 〈P〉 ) Pp[ton/(ton + toff)], where Pp is the average continuous wave power output and ton/(ton + toff) is defined as the duty cycle. Prior to each experiment, the reactor was cleaned by scrubbing with detergent, rinsing in propan-2-ol, oven drying, followed by a 30 min highpower (60 W) air plasma treatment. The system was then vented to air, and a PDMS sheet (6 cm × 1.5 cm) was placed in the chamber (8 cm from the gas inlet) followed by evacuation back down to initial pressure. Water Plasma. In the case of water plasma experiments, ultrapurified water was loaded into a stoppered glass gas delivery tube. Subsequently, water vapor was introduced into the reaction chamber at a constant pressure of 0.2 mbar and with a flow rate of approximately 1.2 × 10-9 kg s-1. At this stage, the plasma was ignited with a peak power of 60 W and run for 5 min. Then, the rf generator was switched off, and the system was vented up to atmospheric pressure.

13138 Langmuir, Vol. 23, No. 26, 2007 Plasma Polymerization. In the case of plasma polymerization experiments, maleic anhydride (Prolabo, 99.5% purity) was ground into a fine powder and loaded into a stoppered glass gas delivery tube. Subsequently, maleic anhydride vapor was introduced into the reaction chamber at a constant pressure of 0.2 mbar and with a flow rate of approximately 1.6 × 10-9 kg s-1. At this stage, the plasma was ignited and run for 30 min. The optimum deposition conditions correspond to power output 5 W, pulse on-time 25 µs, and off-time 1200 µs. These parameters were previously optimized on the basis of a full factorial design and a central composite approach.37 Upon completion of deposition, the rf generator was switched off, and the monomer feed was allowed to continue to flow through the system for a further 2 min prior to venting up to atmospheric pressure. Elongation Step. The differently modified PDMS sheets were elongated and analyzed immediately after treatment. Stretching experiments were carried out with a homemade stretching device. PDMS sheets were elongated up to 2 times (100%) their initial length, and the elongation rate corresponded to 24 µm/s. SEM. The SEM observations were performed by using a FEI environmental microscope (Quanta 400 model) working at 30 keV. The films were observed using high vacuum mode. AFM. AFM images were realized with a Dimension 3000 scanning probe microscope (Digital Instruments). A silicon cantilever was used for all measurements. The spring constant of the cantilever was 20-100 N/m. Typically, the surface morphology of 5 µm × 5 µm areas near the center of each sample was observed in the tapping mode of the scanning probe microscope. Contact Angle Measurements. Contact angles were measured on a Digidrop apparatus. Three pure water droplets were deposited on the surface, and two independent measurements were performed on each droplet. Each contact angle value was thus the average of six values. XPS. XPS analysis of the functionalized surfaces utilized a LEYBOLD LHS11 instrument equipped with an non-monochromatized Mg KR1,2 X-ray source (1253.6 eV) and a concentric hemispherical analyzer. Photoemitted electrons were collected at a takeoff angle of 90° from the substrate, with electron detection in the constant analyzer energy mode (CAE, pass energy 20 eV). Core level spectra were fitted using a Gaussian/Lorentzian product function with equal fwhm using CASAXPS software. The Gaussian/ Lorentzian ratios were taken as 30% for all peaks. The binding energy of the CHx component in the C1s region was set to 285.0 eV and used for referencing. Instrumental sensitivity (multiplication) factors for C(1s)/O(1s)/N(1s)/Al(2p) were taken, respectively, as equal to 1.0:2.86:1.77:0.57. Infrared Spectroscopy. Attenuated total reflection Fourier transform infrared (ATR-FTIR) spectra were recorded with a Brucker IFS 66 FTIR spectrometer equipped with an ATR accessory from Specac. The spectrometer was equipped with a nitrogen cooled MCT detector. Spectra were the result of the sum of 256 scans and were recorded in the range of 4000-650 cm-1 with a spectral resolution of 4 cm-1. A single reflection Ge crystal (refractive index ∼4.0) with a 65° beam incidence was used. The spectrum of a freshly cleaned Ge crystal was used as the background. Thickness Measurements. Thickness measurements of the plasma polymer thin film deposited on a silicon wafer were made by ellipsometry and AFM. The ellipsometric measurements were performed on a phase modulation Multiskop from Physik Instrumente (Model M-033k001) at 632.8 nm (HeNe laser). The cross-section of the laser beam was about 1 mm2. Measurements in air were performed at different positions (at least five) of the samples to check the uniformity of the films. The AFM measurements were performed on silicon wafers by masking part of the surface before polymer plasma treatment. After treatment, a height profile was determined. The profile shows the deposition of a smooth film with a clear and sharp edge. This allows us to easily determine the film thickness.

Results and Discussion AFM and SEM Analysis. The average roughness (Ra) of the films estimated by scanning a 5 µm × 5 µm area is listed in Table

Roucoules et al. Table 1. Average Roughness of PDMS Substrates Ra (nm) unstretched PDMS PDMS/UV irradiation PDMS/water plasma PDMS/plasma polymerization

7.9 8.3 2.2 13.7

1. The surface of the virgin PDMS is relatively smooth (7.9 nm), which is consistent with the surface of glass used for molding. An obvious difference between the surface morphologies of the virgin and the surface morphologies of the modified PDMS was obtained. A notable result is that UV irradiation induced no change in the surface roughness (Ra ) 8.3 nm), while water plasma and plasma polymerization let to a decrease (down to 2.2 nm) or an increase (up to 13.7 nm) in the average roughness, respectively. The corresponding phase images are also presented in Figures S1-S4 of the Supporting Information. In the tapping mode, the measurement of the difference between the phase angle of the excitation signal and the phase angle of the cantilever response was used to map compositional variations such as stiffness, hardness, and viscoelasticity on the sample surface.38 The PDMS specimens used exhibited phase images with large features that could result from heterogeneous filler aggregates (see Supporting Information Figure S1). Similar aggregates are still observable after UV irradiations (see Supporting Information Figure S2). The contrast in the phase images of the PDMS surface after water plasma or plasma polymerization is completely different. The phase image of the water plasma treated PDMS substrate shows globules in a very flat surface (see Supporting Information Figure S3). The phase image of the plasma polymer treated PDMS corresponds typically to the phase image obtained after maleic anhydride plasma polymer deposition (see Supporting Information Figure S4).37 Figures 1-3 show SEM images of modified PDMS while a strain was applied. The arrows show the direction of elongation. The first remarkable characteristic is that the UV irradiated PDMS substrate (Figure 1) showed no sign of surface cracking, while water plasma coatings (Figure 2) and plasma polymer coatings (Figure 3) revealed fractures at the surface. These cracks were previously observed by Hillborg et al.39 for PDMS substrates treated by corona discharges as well as by Efimenko et al.40 for a UV/ozone exposed elongated PDMS substrate. This behavior is explained by the composite, which is formed by the superposition of a stiff layer at the top of a flexible substrate. The fragmentation normal to the loading direction depends on the transfer of the tensile stress of the substrate to the surface layer as well as on the tensile properties and thickness of that layer. According to these authors, the cracks were attributed to the fragmentation of the silica-like layer formed, which depends on the conditions of treatment and on the degree of cross-linking of the PDMS. The length of the cracks for the given strain (100%) that is observed in this work is clearly smaller for the water plasma treated samples. Although the distribution of the crack length is less regular than that observed by Hillborg et al.,39 the crack length is of the same order of magnitude (about 5 µm) for these physical treatments (corona or plasma), whereas it is of the order of about 15-20 µm for the plasma polymer treatment. Moreover, plasma polymer thin films show cracks perpendicular (38) Maiti, M.; Bhowmick, A. K. Polymer 2006, 47, 6156-6166. (39) Hillborg, H.; Sandelin, M.; Gedde, U. W. Polymer 2001, 42, 7349-7362. (40) Efimenko, K.; Rackaitis, M.; Manias, E.; Vaziri, A.; Mahadevan, L.; Genzer, J. Nat. Mater. 2005, 4, 293-297.

Changes in Silicon Elastomeric Surface Properties

Figure 1. SEM images of UV treated and stretched PDMS at two different scales.

Langmuir, Vol. 23, No. 26, 2007 13139

Figure 2. SEM images of water plasma treated and stretched PDMS at two different scales.

to the direction of the strain (Figure 3). In water plasma coatings, most of the cracks are perpendicular to the direction of the loading, and additional ones are oriented at 45° to the direction of the strain (Figure 2). Therefore, the dark bands show the PDMS substrate made visible due to opening cracks in the coatings upon stretching. It should be noted here that all the specimens were subjected to minor deformations by the preparation procedure prior to the SEM studies. In spite of precautions, random cracks could appear in the film. Hillborg et al.41 reported that cracking of the oxidized layer may occur spontaneously: the formation of a surface layer rich in SiOx involves a significant reduction in specific volume, and this volume change leads to a buildup of tensile stresses in the film and the formation of surface cracks. In the meantime, all the surfaces are folded and wavy: a wavy pattern or buckles superimpose on the fragments in a direction perpendicular to the loading direction. These waves result from the different shrinkage ratios of the substrate and surface layer.

The Poisson ratio of PDMS is about equal to 0.5, whereas that of the surface layer is much smaller. Therefore, under tensile loading, compression stresses appear in the surface layer due to lower lateral contraction as compared to the substrate.42 The width of the waves is almost constant over the entire area of the photographs but differs slightly following the treatment used. UV irradiated PDMS substrates show only waves parallel to the tension direction (Figure 1). The wave crests lay parallel to the loading direction in all cases. We also used AFM to probe the fractured surface on a stretched samples (see Supporting Information Figures S5-S8). To better visualize these defects, the samples were subjected to crosssectional analysis. A cross-sectional line was drawn over the scanned image. The cross-sectional profiles along this line are shown in Figure 4. We would expect that the difference in height between the top surface and the trough of the waves should correspond to the thickness of the coatings.

(41) Hillborg, H.; Ankner, J. F.; Gedde, U. W.; Smith, G. D.; Yasuda, H. K.; Wikstro¨m, K. Polymer 2000, 41, 6851-6863.

(42) Volynskii, A. L.; Bazhenov, S.; Lebedeva, O. V.; Bakeev, N. F. J. Mater. Sci. 2000, 35, 547-554.

13140 Langmuir, Vol. 23, No. 26, 2007

Figure 3. SEM images of plasma polymer treated and stretched PDMS at two different scales.

Under the same plasma polymer deposition conditions on silicon wafers, we formed a uniform film that has a well-defined thickness equal to 22 ( 2 nm (this value has been obtained by ellipsometry and AFM analysis (see Supporting Information Figure S9). One can expect a similar thickness on our PDMS substrate (see Supporting Information Figure S10). In the case of the water plasma treatment, Hillborg et al.41 have reported conversion of PDMS to an SiOx film surface having a thickness of 130-160 nm after 3 min of treatment using rf plasma modification techniques in comparable conditions. One can also expect a similar thickness on our PDMS substrate. The depth of the cracks is on the order of 600-800 nm, and they seem to be slightly deeper in water plasma treated samples than in plasma polymer treated ones. According to the estimated values of the thicknesses, it is difficult to imagine that the final thickness of the plasma polymer will be higher than 100 nm. The difference in height can not only be due to cracks in the plasma polymer thin film but due to spreading of the fractures in an underlying layer as well. This result could be explained by the presence of a PDMS sublayer

Roucoules et al.

weakened by additional cross-links due to the plasma treatment process (i.e., water plasma and plasma polymerization). Then, the exposed PDMS in the cracks should be different from the native PDMS used. Again, this point will be confirmed by XPS analysis and contact angle measurements. Infrared Analysis. The chemical changes of the PDMS surfaces exposed to the three treatments were studied by ATRFTIR spectroscopy. The information depth of the ATR technique is between 0.5 and 2 µm depending on the wavenumber. The reference spectrum of commercial PDMS is in accordance with previous published data (see Supporting Information Figure S11).43 The bands originating from the symmetric and asymmetric stretching vibrations of the methyl groups (νas,s(CH3)) are centered at 2963 cm-1 (A) and 2906 cm-1 (B), respectively. The corresponding deformation vibration (δ(CH3)) modes are located at 1446 cm-1 (C), 1412 cm-1 (D), and 1259 cm-1 (E). The asymmetric stretching vibrations of the Si-O-Si group (νas(SiOSi)) appear between 930 and 1200 cm-1 (F), the methyl rocking (Fr(CH3)) and Si-C stretching vibrations (ν(SiC)) are centered around 795 cm-1 (G), and the Si-O bending vibrations (δ(SiO)) are located at about 702 cm-1 (H). After treatment, the presence of the background PDMS infrared bands confirmed that degradation had only taken place near the extreme surface (see Supporting Information Figure S11). When the PDMS was exposed to UV irradiation, the infrared spectrum exhibited two small peaks centered at 1722 cm-1 (L) and 1702 cm-1 (M) corresponding to the appearance of carboxylic acid groups. At the same time, a slight shift of νas(CH3) (B) and νs(CH3) (A) toward higher wavenumbers and a broad absorption band centered at 3400 cm-1 (N) (see Supporting Information Figure S13) were observed. These changes are the result of the substitution of methyl groups by oxygen units and hydroxyl groups. Similar results were observed after water plasma, indicating the presence of hydroxyl groups in the silicate-like layer, while no infrared signal was observed in the region of 1600-2000 cm-1. Subsequent degradation of the polymer was also indicated by a shift toward higher wavenumbers of the νas(SiOSi) band (F) due to an increase of the number of oxygens at the silicon (see Supporting Information Figure S14). According to ref 43, a shoulder should appear around 1150 cm-1 belonging to the outof-phase oxygen motion associated with the Si(O4) configuration. In our case, this band was not detectable. An explanation could be to the thickness of the SiOx layer, which is lower than the effective sensing depth of the ATR technique including, therefore, the response of the underlying PDMS. In previous work, infrared analysis of the deposited maleic anhydride pulsed plasma polymer films onto silicon wafers confirmed a high degree of anhydride group incorporation.37 The following characteristic infrared absorption features of cyclic anhydride groups were identified: asymmetric and symmetric CdO stretching (1860 and 1796 cm-1), cyclic conjugated anhydride group stretching (1241-1196 cm-1), C-O-C stretching vibrations (1097-1062 cm-1), and cyclic unconjugated anhydride group stretching (964-906 cm-1). When the same polymer thin film was deposited onto commercial PDMS, only asymmetric and symmetric CdO (J and K) stretching bands were observed because of the presence of the background PDMS infrared bands (see Supporting Information Figures S11 and S12). Here again, the results confirmed the high degree of retention of the maleic anhydride group in the deposited plasma polymer thin film. A slight infrared signal at 1722 cm-1 (L) was also (43) Graubner, V. M.; Jordan, R.; Nuyken, O.; Schnyder, B.; Lippert, T.; Ko¨tz, R.; Wokaun, A. Macromolecules 2004, 37, 5936-5943.

Changes in Silicon Elastomeric Surface Properties

Langmuir, Vol. 23, No. 26, 2007 13141

Figure 4. Section analyses of stretched (a) water plasma and (b) plasma polymer modified PDMS substrate. Table 2. XPS and Contact Angle Analysis of PDMS Substrates XPS %O

% Si

contact angles ((2°)

Unstretched PDMS 46.2 19.2 PDMS/UV irradiation 37.5 28.1 PDMS/water plasma 19.5 38.5 PDMS/plasma polymerization 69.6 22.3

34.6 34.4 42.0 8.1

115 95 62 58

Stretched PDMS 45.7 PDMS/UV irradiation 45.4 PDMS/water plasma 30.6 PDMS/plasma polymerization 59.5

35.5 30.4 37.4 17.3

116 90 124 59

%C

18.8 24.2 32.0 23.2

observed, corresponding to the carboxylic acid contained in the plasma polymer thin film.31,37 The PDMS substrates also have been analyzed after stretching. But unfortunately, except for an overall decrease of plasma thin film absorption bands, no significant changes were observed in the infrared features. XPS Analysis. The effect of the treatments on the atomic surface composition of PDMS before and after stretching was assessed by XPS. The probed depth by XPS was ∼7-10 nm. The results are summarized in Table 2 (see also Supporting Information Table S1). The theoretical atomic composition of PDMS based on the repeat unit is 50% C, 25% O, and 25% Si. In the case of the studied PDMS used, the carbon content was 46 atom %, which is close to the theoretical value of 50 atom %. However, a higher silicon content (34.6% instead of 25.0%) was observed at the expense of oxygen (19.2% instead of 25.0%). Other authors19,44,45 have reported for commercial silicone substrates compositions that are not in agreement with the theoretical composition with a higher content in oxygen. This (44) Hillborg, H.; Tomczak, N.; Ola`h, A.; Scho¨nherr, H.; Vancso, G. J. Langmuir 2004, 20, 785-794. (45) Kim, J.; Chaudhury, M. K.; Owen, M. J. J. Colloid Interface Sci. 2006, 293, 364-375.

deviation was related tentatively to the presence of silica fillers, although other studies reported that the thickness analyzed by XPS cannot sense the fillers.46 In our case, the atomic concentration of Si is higher than expected. Slight differences were observed for the extracted sample by cyclohexane showing that the short free chains that are in the network cannot be responsible for the result that is obtained. One could therefore attribute the difference to the molding conditions against glass or to the presence of the copolymer that is carrying the SiH functional groups, which is not known. The analysis of the C(1s) envelope indicated two types of carbon functionalities (see Supporting Information Figure S15 and Table S1): 64% hydrocarbon (CHx ∼285.0 eV) and 36% carbon singly bonded to a silicon atom (C-Si- ∼284.4 eV). The hydrocarbon CHx was found to be the predominant carbon center in the C(1s) envelope, with smaller amounts of siliconated functionalities. We considered that part of this carbon signal was due to adventitious carbon present on any XPS sample exposed previously to air. The high-resolution spectra for Si(2p) (see Supporting Information Figure S16 and Table S1) showed only one peak at 102.1 eV, and the high-resolution spectra for the O(1s) orbital peaks (see Supporting Information Figure S17 and Table S1) showed only one peak at 531.9 eV. Both peaks correspond to PDMS, and their position corresponds quite well to the values of the literature.47 After UV irradiation, there is only a small increase in roughness (Table 2). XPS analysis of UV irradiated PDMS substrates indicates three types of carbon functionalities on the C(1s) envelope (see Supporting Information Figure S15 and Table S1): 22.4% carbon singly bonded to a silicon (C-Si ∼284.4 eV), 68.7% hydrocarbon (CHx ∼285.0 eV), and 8.9% carbon singly bonded to oxygen (-C-O ∼287.0 eV). The Si(2p) (46) Efimenko, K.; Wallace, W. E.; Genzer, J. J. Colloid Interface Sci. 2002, 254, 306-315. (47) Moulder, J. F.; Stickle, P. E.; Sobol, P. E.; Bomben, K. D. In Handbook of X-ray Photoelectron Spectroscopy; Chastain, J., Ed.; PerkinElmer Corp.: Eden Praire, MN, 1992; p 22.

13142 Langmuir, Vol. 23, No. 26, 2007

envelope (see Supporting Information Figure S16 and Table S1) showed only one functionality but a significant shift toward a higher binding energy at 102.9 eV corresponding to a Si(O3) configuration.44 The O(1s) envelope (see Supporting Information Figure S16 and Table S1) showed only one peak at 532.7 eV. This treatment leads to an increase in the number of oxygen bonds to silicon at the expense of the carbon concentration as can be seen in Table 2. An oxidative cross-linking process seems to be effective in the probed depth by XPS. This may correspond to a different chemical environment of the Si and O atoms that can appear with the substitution of methyl groups by hydroxyl groups and oxygen, respectively. After water plasma exposure, XPS analysis of the PDMS substrates indicated three types of carbon functionalities on the C(1s) envelope (see Supporting Information Figure S15 and Table S1): 91.8% carbon singly bonded to a silicon (C-Si ∼284.4 eV), 5.3% carbon singly bonded to oxygen (-C-O ∼286.9 eV), and 2.9% carbon triply bonded to oxygen (OsCdO ∼289.1 eV). No aliphatic carbon component (CHx) was detected. The Si(2p) envelope (see Supporting Information Figure S16) shows a shoulder at 103.4 eV, which corresponds to the formation of a silica-like structure (SiOx) in the surface region. This peak position agrees quite well with literature values of 103.3-103.7 eV for the Si(O4) configuration, in which Si is linked to four neighboring oxygen atoms.47 But, a contribution from PDMS at 102.0 eV remains visible. According to Hillborg et al.,44 two main reasons could explain this result: the first is the limitation of the conversion process to the silicate layer due to the presence of imperfect domains in the converted layers or to the migration of free siloxanes into the oxidized region. The second could be the thickness of the SiOx layer, which can be lower than the effective sampling depth of XPS (∼7-10 nm). The corresponding peak of the O(1s) transition (see Supporting Information Figure S17) shifts from 532.4 eV for untreated PDMS to 532.7 eV for water plasma treated PDMS. Again, this result corresponds quite well to the data found in the literature.47 The mean value of the interband energy between the O(1s) orbital peak and the Si(2p) orbital peak is 429.2 eV, and the value found in this study was 429.3 eV. Finally, the PDMS substrates were treated by maleic anhydride pulsed plasma polymerization. XPS analysis indicated also five types of carbon functionalities on the C(1s) envelope (see Supporting Information Figure S15): 54.7% hydrocarbon (CHx ∼285.0 eV), 14.0% carbon singly bonded to an anhydride group (CsC(O)dOs ∼285.7 eV), 10.0% carbon singly bonded to oxygen (-C-O ∼286.6 eV), 0.7% carbon doubly bonded to oxygen (OsCsO/sCdO ∼287.9 eV), and 20.5% anhydride groups (OdCsOsCdO ∼289.4 eV). Previous works37 show typical XPS analysis of the maleic anhydride pulsed plasma polymer thin film deposited onto silicon wafers, indicating five types of carbon functionality on the C(1s) envelope: 27.6% hydrocarbon (CHx ∼285.0 eV), 30.1% carbon singly bonded to an anhydride group (CsC(O)dOs ∼285.7 eV), 7.5% carbon singly bonded to oxygen (-C-O ∼286.6 eV), 4.7% carbon doubly bonded to oxygen (OsCsO/sCdO ∼287.9 eV), and 30.1% anhydride groups (OdCsOsCdO ∼289.4 eV) in agreement with a high degree of retention of maleic anhydride groups in the plasma polymer thin film. If these results are compared with the ones obtained on the silicone treated substrate in the same conditions, the hydrocarbon CHx was found to be the predominant carbon center in the C(1s) envelope, with smaller amounts of oxygenated functionalities. Also, a decrease in the relative importance of the component corresponding to the anhydride groups was observed (30.1-20.5%). Although the

Roucoules et al.

plasma polymer treated PDMS shows continuous deposited polymer thin film, an unexpected Si(2p) signal was detected. The corresponding high-resolution spectra (see Supporting Information Figure S16) showed only one peak at 102.2 eV, indicating the contribution of the PDMS. This result is surprising because a continuous and homogeneous thin film of maleic anhydride plasma polymer with a thickness of approximately 20 nm was expected.37 Recently, similar results were observed, and no explanation was given.29 The discussions for precise interpretation are still under progress. Nevertheless, the presence of anhydride groups is confirmed by two peaks in the highresolution spectra of O(1s) corresponding to sCdO at 532.4 eV and to OdCsOsCdO at 533.9 eV for the plasma polymer (see Supporting Information Figure S17). To facilitate a straightforward examination of the effects of the different treatments, the C/Si and O/Si atomic ratios were calculated (see Supporting Information Figure S18). It is clear that the three treatments produced an increase in the oxygen concentration given by a higher ratio of O/Si than virgin PDMS. The highest oxygen-containing group (i.e., polar groups) concentration was obtained after plasma polymer deposition. However, water plasma and UV irradiation produced a decrease in the C/Si ratio due to the formation of silicate-like layers or to the substitution of methyl groups by hydroxyl groups and oxygen, while plasma polymerization produced an increase in the C/Si ratio due to the organic thin film deposition. The elongated PDMS substrates were also analyzed by XPS (see Table 2 and Supporting Information Table S1). The stretched UV irradiated PDMS substrate indicated three types of carbon functionalities on the C(1s) envelope: 4.0% carbon singly bonded to a silicon (C-Si ∼284.4 eV), 88.5% hydrocarbon (CHx ∼285.0 eV), and 7.5% carbon singly bonded to oxygen (-C-O ∼286.6 eV) (Figure 5 d). As compared to the results obtained before stretching, the C-Si concentration decreases at the expense of the CHx concentration. No change in the Si(2p) and O(1s) envelopes was observed. This point will be confirmed by contact angle measurements and will be discussed later. The C(1s) envelope of the stretched water plasma modified PDMS substrate after stretching indicated three types of carbon functionalities: 89.0% carbon singly bonded to a silicon (-CSi ∼284.4 eV), 7.9% carbon singly bonded to oxygen (-C-O ∼286.8 eV), and 3.1% carbon triply bonded to oxygen (Os CdO ∼288.5 eV) (Figure 5c). The Si(2p) envelope shows two peaks: 46.8% of the contribution of PDMS at 102.0 eV and 53.2% of 103.6 eV for the silicate-like structure (see Supporting Information Figure S19). There is an increase of the concentration of the silicium coming from the PDMS substrate (26.5-46.8%) upon stretching. This change is due to the cracks opening in the silicate-like layer leading to the contribution of PDMS (∼102.0 eV), which is very similar to that of the native PDMS (∼102.1 eV). The interband energy between the O(1s) orbital peak and the Si(2p) orbital peak is 429.0 eV, confirming the Si(O4) configuration at the top of the coating layer. Finally, the C(1s) envelope of the stretched plasma polymer modified PDMS substrate indicated five types of carbon functionalities: 54.1% hydrocarbon (CHx ∼285.0 eV), 12.8% carbon singly bonded to an anhydride group (CsC(O)dOs ∼285.7 eV), 15.2% carbon singly bonded to oxygen (-C-O ∼286.6 eV), 2.5% carbon doubly bonded to oxygen (OsCs O/sCdO ∼287.9 eV), and 15.4% anhydride groups (OdCs OsCdO ∼289.4 eV) (Figure 5b). As compared to the results obtained before stretching, the main difference is the increase of the C-O functionalities at the expense of the OdCsOs

Changes in Silicon Elastomeric Surface Properties

Langmuir, Vol. 23, No. 26, 2007 13143

Figure 5. High-resolution XPS spectra of the C(1s) peak for stretched (a) untreated, (b) plasma polymer, (c) water plasma, and (d) UV treated PDMS.

CdO groups. Moreover, the Si(2p) envelope shows only one peak at 102.5 eV (see Supporting Information Figure S19). There is a shift of 0.3 eV to the higher binding energies as compared to the results obtained before stretching. This peak is due to the crack openings revealing PDMS. But, this peak does not correspond to native PDMS. According to the results in the literature, the exposed PDMS in the cracks corresponds to a configuration somewhere between Si(O2) and Si(O3). This point is also confirmed by the shift to a higher binding energy (532.4 to 532.6 eV) of the peak in the high-resolution spectra of O(1s) corresponding to SisEnDashO/sCdO from PDMS. These results could be explained by a mixture of Si(CH3)2O, Si(O,CH3)O, and SiO2 species.41 Contact Angle Measurements. Water contact angle measurements were carried out to determine the hydrophilicity of the PDMS surfaces and the effect of the different treatments on surface wettability upon stretching. Water droplets probe the outermost 5-10 Å. As expected, the water contact angle of untreated PDMS decreased after UV irradiation and more markedly after plasma polymerization and water plasma (see Table 2). As observed by Efimenko et al.,46 the absence of ozone during the UV treatment is much less efficient for increasing the surface hydrophilicity of the PDMS substrate. Complete spreading by water is not observed after plasma treatment as shown for other treatments such as UV/ozone46 or corona discharge.39 However, the remarkable result is that the plasma techniques (i.e., plasma polymerization and water plasma) induced lower contact angle values in much shorter times (10 and 30 min, respectively). Then, the modified PDMS substrates were switched back and forth between 20 and 100% elongation relative to the unstretched length. The variation of the surface properties was followed by

Figure 6. Evolution of the contact angle during elongation/retraction cycles.

contact angle measurements (Figure 6). No change in the contact angle value was observed on the native PDMS (results no shown). For the UV irradiated PDMS substrate, the contact angles exhibited small changes during the elongation/retraction cycles. Typically, the contact angle switched from 92 ( 2 to 99 ( 2°. Under loading conditions, the anisotropy shape of the droplet that was observed reveals the slight anisotropic character of the surface of the substrate (slight wave formation or orientation of functional groups, as shown by Holland-Moritz et al. on polyethylene substrates48). (48) Holland-Moritz, L. E.; Koenig, K.; van Werden, K. Makromol. Chem. 1981, 182, 651-655.

13144 Langmuir, Vol. 23, No. 26, 2007

Roucoules et al.

the cracks close (Figure 7b), and the contact angle value drops to 74°. Finally, in the case of the plasma polymer modified PDMS substrates, the contact angle switches reversibly back and forth from 75 ( 2 to 72 ( 2° for the two elongation states. The reversible state is reached only after two cycles of elongation/retraction. The two first cycles correspond to the early stage of the formation of the cracks. During this period, the width of the cracks is very small as compared to the macroscopic characteristic size of the water drop used during the contact angle measurements. The cracks do not influence the water contact angle value at least during the first cycle. In spite of the presence of cracks at the surface, the contact angle exhibits very small changes during the elongation/retraction cycles. The exposed PDMS in the cracks corresponds to a configuration between Si(O2) and Si(O3), which is more hydrophilic than the native PDMS. Moreover, the variations in contact angle values for the two elongation states are in the same range in the case of the UV irradiated PDMS substrate, where no cracks are observable. Because the cracks in the plasma polymer treated substrates are most likely deeper than in the water plasma treated substrates, it can be assumed that the effect of the roughness on the contact angle value is of the second order and that the main influence is due to the change in surface properties on the PDMS, which is exposed upon stretching.

Conclusion

Figure 7. Height and phase images of water plasma modified PDMS (a) after 2.0 elongation and (b) after switching back to 1.2 elongation.

In the case of water plasma modified PDMS substrates, the contact angle switched reversibly back and forth from 74 ( 2 to 125 ( 2° for the two elongation states. More than 10 elongation/ retracting cycles were performed without altering the reversibility. The aforementioned behavior of the water droplet is due to the opening/closing of cracks during the mechanical stretching. Before stretching, the silicate-like layer exhibited a contact angle value of 62 ( 2°. We have shown previously that the cracks appear in the silicate-like layer upon elongation, exposing native PDMS. This makes the surface more hydrophobic (the 65° contact angle value changes to 125°) due to (i) the hydrophobicity of the native PDMS (the contact angle value is 115°) and (ii) the increase in the surface roughness (Table 1 and Figure 7-a). When the same substrate is switched back to 20% elongation,

We studied the performance of three surface treatments onto a commercial silicon substrate when the latter was stretched. The UV irradiation induces the substitution of methyl groups by hydroxyl groups and oxygen on the PDMS substrate, and the contact angle values remain very high (∼95°) after 5 h of treatment. No cracks are shown at the surface under stretching. The 5 min of water plasma treatment produces the formation of silicate-like layers at the PDMS surface, inducing smaller contact angle values (∼62°), and 30 min of maleic anhydride plasma polymer deposition leads to a continuous organic thin film at the PDMS surface with a contact angle value of (∼58°). Both silicatelike layer and plasma polymer coating induce cracks under stretching. But, the influence of the cracks on the contact angle during the elongation/retraction cycles is much smaller when the PDMS is treated by plasma polymerization (only 3° of variation). Even more, the variations in the contact angle values are in the same range as those observed on the UV irradiated PDMS substrate, where no cracks appear under stretching. This result is due to the specific chemistry of PDMS revealed in the plasma polymer cracks, which is very different from the chemistry of native PDMS. In previous work, polyelectrolyte multilayers were deposited on a maleic anhydride plasma polymer modified PDMS substrate.30 These substrates were also elongated up to 120% relative to their initial length and thus switched back and forth between 0 and 120% elongation. For example, in the case of a PEI-(Naf-PAH)4-Naf-(PAH-PAA)2 film, the contact angle switches reversibly back and forth from 60 ( 5 to 101 ( 5° for the two elongation states. Comparing these results (∼41°) and the variations obtained in this study with the maleic anhydride plasma polymer thin film alone (∼3°), the effect of the polyelectrolyte multilayer is certain. These results are encouraging, and pulsed plasma polymerization seems to be a useful candidate in all adhesion problems dealing with surface responsive materials. Furthermore, one of the inherent advantages of plasma polymerization as compared to other techniques includes the fact that the densities of the functional groups can be tailored by

Changes in Silicon Elastomeric Surface Properties

programming the duty cycle. Moreover, maleic anhydride pulsed plasma polymer surfaces could be modified by introducing different groups in its structure (via aminolysis reaction) to regulate its final chemical properties. We are now working on possibilities to eliminate all the cracks in the plasma polymer coating under stretching. This point will be the subject of a future paper.

Langmuir, Vol. 23, No. 26, 2007 13145

Acknowledgment. This work was supported by grants from the region of Alsace (France). Supporting Information Available: Additional figures as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org. LA701460F