Preparation of micropatterned poly(silsesquioxane) thin films using adamantylphenol molecules

Thin Solid Films 515 (2007) 4603 – 4608 www.elsevier.com/locate/tsf

Preparation of micropatterned poly(silsesquioxane) thin films using adamantylphenol molecules Bong Jun Cha a,⁎, Young-Jin Choi a , Baekjin Kim b , Hern Kim a a

Department of Environmental Engineering and Biotechnology, Myongji University, San 38-2, Nam-dong, Yongin-si, Kyonggi-Do, 449-728, Korea b Korea Institute of Industrial Technology, 35-3, Hongcheon-ri, Ipjang-myeon, Cheonan-si, Chungcheongnam-do, 330-825, Korea Received 2 January 2006; received in revised form 26 October 2006; accepted 20 November 2006 Available online 28 December 2006

Abstract Poly(methyl silsesquioxane) (PMSSQ) thin films with micropatterned surface structures have been prepared by use of adamantylphenol molecules as a photo-thermal sensitive moiety together with UV lithographic technique and mild heating process. Initially, PMSSQ films form positive patterns of micronscale due to the film densification triggered by photooxidation and photopolymerization of doped moieties within UV exposed region. With thermal treatment, negative patterns from these pre-patterned films are formed by the difference of polycondensation rates between non-exposed regions and irradiated areas without additional developing or etching steps. This structural transformation of PMSSQ thin films was investigated using Fourier-transformed infrared spectroscopy, ultraviolet visible spectroscopy, X-ray photoelectron spectroscopy, Auger electron spectroscopy, and field emission scanning electron microscopy. © 2006 Elsevier B.V. All rights reserved. Keywords: Silsesquioxane; Photooxidation; Structural properties; Surface structure

1. Introduction Polysilsesquioxane (PSSQ) materials represented by empirical formula (RSiO3/2)n recently receive an increasing attention in both fundamental and applied science. Potential application for these materials includes use in nanocomposites as reinforcing fillers, low dielectric (low-k) insulators in microelectronic devices, and micro-optical waveguides [1–3]. In addition to such promising fields, patterning of PSSQ has considerably been interested for their practical applications to copper interconnect of low-dielectric materials, anti-reflective films, and surface relief gratings [4–6]. Since it is well known that PSSQ can be crosslinked or polymerized by heat or irradiation source such as X-ray and electron beam, selective irradiation or heating within the confined film surface can make the formation of diverse patterned structures possible [7–9]. Patterning of nanoporous poly(methyl silsesquioxane) (PMSSQ) films was also introduced using a typical UV lithographic technique combined with aid of photoacid/base ⁎ Corresponding author. Tel.: +82 31 330 6824; fax: +82 31 336 6336. E-mail address: [email protected] (B.J. Cha). 0040-6090/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2006.11.043

generators and ozone generation [10–12]. Together with these versatile methods to prepare patterned PMSSQ structures, there is still a need to control patterned structures using simpler efficient process. Moreover, for successful implementation of PMSSQ as low-k materials into advanced integration devices, the development of technique without etching or developing step is required to fulfill the practical advantages such as a direct formation of via or trench. Adamantylphenol grafted to PMSSQ matrix was previously employed to prepare the nanoporous PMSSQ thin films using only thermal treatment for low-k application. As was previously mentioned, the decomposition temperature of adamantylphenol was observed to shift to higher temperature compared to pristine adamantylphenol due to the chemical bonds of adamantylphenol into the PMSSQ matrix [13,14]. This suggests that the introduction of covalently bonded adamantylphenol is not suitable for realizing the diverse patterned structure under the condition of mild heating. In present study, we described a simple approach for micropatterning, where positive patterns formed on PMSSQ films are transformed to negatively patterned structures without additional developing or etching steps. This method employs adamantylphenol molecules

4604

B.J. Cha et al. / Thin Solid Films 515 (2007) 4603–4608

without covalent bond as photo-thermal sensitive moiety together with UV irradiation as well as mild thermal treatment. In addition, the detailed mechanism for formation of patterned structures was discussed. 2. Experimental details 2.1. Preparation of PMSSQ and adamantylphenol All the reagents used in present study were purchased from Aldrich and were used as received unless otherwise specified. In this study, PMSSQ polymers were synthesized with 1,2bis-trimethoxysilylethane and methyltrimethoxysilane by carefully controlled hydrolysis-condensation reaction. A detailed synthesis procedure of PMSSQ polymer was previously published as well as adamantylphenol moiety [13,14]. The prepared PMSSQ and adamantylphenol characterized by 1H nuclear magnetic resonance (1H NMR) measurement (Bruker, AVANCE 500 MHz) are in a good agreement with the results reported previously [13,14]. From 1H NMR spectrum, the SiOH content in polymer was 19.3 mol% and number average molecular weight of polymer measured by Gel Permeation Chromatography was 7400. 2.2. Preparation of thin films As a typical example, PMSSQ thin films containing adamantylphenol were prepared by spin-coating from polymer solutions dissolved in methylisobutylketone (MIBK) with concentration of 30 wt.% adamantylphenol. The silicon wafer pre-treated with piranha solution (H2SO4/H2O2 = 7/3 v/v) was spun at 2000 rpm for 30 s to give film thickness ranging from 550 nm to 650 nm. Subsequently coated films were dried under vacuum for 3 days to remove residual MIBK at 50 °C. The dried films were irradiated through copper grid (100 or 400 mesh, GLIDER) with UV light (wavelength; 254 nm) of 3.9 mW/cm2 for exposure time of 12 h in air, as can be seen Fig. 1. Some nonirradiated films were thermally treated as well as irradiated films with home-made furnace under N2 flow at 300 °C for 2 h at a heating rate of 3 °C/min, respectively. 2.3. Characterization The Fourier-transformed infrared spectroscopic (FT-IR) measurements were performed on a JASCO FT/IR 200 spectrometer. Ten accumulations were signal-averaged at a resolution of 4 cm− 1. Baseline corrected infrared spectra were obtained for films on silicon wafer in absorbance mode at room temperature. The ultraviolet-visible (UV-VIS) spectra were collected for film coated on quartz plate using Perkin-Elmer UV-VIS spectrometer (model 2287) with an operational wavelength range of 190–400 nm. Auger electron depth profiles were obtained on a Physical Electronics scanning Auger microprobe systems (model 660) by etching the film surface with an argon ion beam (2 keV, 20 μA/cm2). X-ray photoelectron spectra (XPS) were collected on a SIGMA PROBE (ThermoVG, UK) equipped with a monochromatic Al-

Fig. 1. Schematic procedure to prepare the PMSSQ films with positive and negative patterns.

Kα X-ray source (1486.6 eV). Spectra were recorded for a takeoff angle of 90° and the peaks were corrected to the C 1s peak at 284.6 eV. Optical microscopic measurements were carried out using a Nikon OPTIPHOT-2POL in reflection mode. Atomic force microscopy (AFM) was performed using a Digital Instruments Nanoscope IIIa in a tapping mode with a Si3N4 tip. The field emission scanning electron microscopy (FE-SEM) (JEOL 6330F) was employed to investigate the surface texture of the thin films. Spin coated films were cut into small pieces and fixed on a sampling holder to analyze the surface structure. In order to minimize the film damage due to the electron beam and also to obtain clear images, gold was sputtered onto the thin films. 3. Results and discussion Previous studies presented the feasibility of employing covalently bonded adamantylphenol as a porogen (pore generating material) in the preparation of nanoporous low-k films [13,14]. When adamantylphenol was chemically incorporated into the PMSSQ matrix, the decomposition temperature of adamantylphenol significantly increased to 450 °C, which is higher than that of pristine adamantylphenol without the presence of a covalent bond. This suggests that chemically linked adamantylphenol is a good pore forming agent leading to the formation of nanoporous films. In order to form the porous films, adamantylphenol should be decomposed after the vitrification of the PMSSQ matrix. On the other hand, adamantylphenol without chemical bond can decompose below 300 °C, which indicates that there are no nanopores in the film owing to the sublimation of adamantylphenol before the condensation of PMSSQ. In this study, the adamantylphenol molecules without the chemical bond are used as photothermally active dopant because it can sublime at relatively low

B.J. Cha et al. / Thin Solid Films 515 (2007) 4603–4608

4605

temperature of 200 °C and absorb UV light below 300 nm. This feature makes patterning of PMSSQ thin films possible upon heating or exposure under optimal conditions, when doped with adamantylphenol. Fig. 2(a) shows a FE-SEM image of patterned surface structure formed on PMSSQ thin film after UV irradiation for 12 h through copper grid (400 mesh). From the fact that UV irradiated regions (squares) are brighter than non-exposed areas (lines), it can be expected that micropatterns with square lines of 30 μm were formed without additional steps such as etching or developing. Since pattern formation may result from changes in film thickness between the exposed and protected areas, patterned surface structures were further investigated using an atomic force microscope. As shown in Fig. 2(b), average depth of exposed areas was lower than that of protected lines with the relative difference of 48 ± 6 nm, suggesting that pattern formation is mainly due to the difference of film thickness between irradiated and non-irradiated regions. Additionally, pattern contrast by such thickness change could be further

Fig. 3. (a) UV-vis and (b) FT-IR spectra of the thin films before and after UV irradiation.

Fig. 2. (a) FE-SEM micrographs of PMSSQ film surface formed after UV irradiation and (b) AFM image of the positive patterned surface structure. Scale bar denotes 100 μm.

controlled with adamantylphenol content and irradiation time. It is noted that these surface patterns were not formed through irradiation under vacuum or for films without adamantylphenol. Thus, pattern formation would be related to photooxidation of adamantylphenol group in matrix. To understand the origin of pattern formation in detail, the chemical structure of PMSSQ thin films was investigated using UV and FT-IR spectroscopy. Fig. 3(a) shows the change of photoelectronic structures of films before and after irradiation, respectively. After exposure, a band at 274 nm due to π–π⁎ transition of adamantylphenol significantly decreases whereas a long tail band above 300 nm increases clearly. It can be also noted that there is no such absorbance change for film without adamantylphenol. It was previously reported that a similar photochemical reaction occurred for the degradation of phenol by UV photolysis [15]. Accordingly, the reduction of such band intensity can be understood in terms of photodecomposition of

4606

B.J. Cha et al. / Thin Solid Films 515 (2007) 4603–4608

adamantylphenol moieties within film concomitant with photooxidation. Similarly, the increment of long tail band may be related to the formation of polymerized adamantylphenol products accompanied with photooxidation of adamantylphenol, as case of photopolymerization of methacrylate [6]. Indeed, it has been well known that organic molecules interacting with UV light give rise to various reactive species such as free radicals and excited molecules through reaction with atomic oxygen. In this study, a possible mechanism involves first the dissociation of the conjugated C_C bonds of the phenyl rings that react subsequently with atomic oxygen to give highly oxidized and polymerized carbons. Photooxidation and photopolymerization of adamantylphenol are further evidenced by infrared spectra, as shown in Fig. 3(b). IR band of C–H and C_C stretching vibration at 2910 and 1517 cm − 1 reflects the expected adamantylphenol groups. On the other hand, an IR band seen at around 3340 cm− 1 corresponds to the presence of silanol groups or Si–OH stretching band as well as adsorbed water. A strong broad band appeared from 1000 to 1200 cm− 1 is attributed to the presence of Si–O–Si groups on PMSSQ backbone. In addition, peaks due to Si–C and Si–OH on the PMSSQ were also detected at 1275 and 902 cm− 1, respectively. FT-IR spectrum after irradiation reveals that adamantylphenol components were decomposed photochemically, as indicated by the reduction of band intensity at 2910 and 1517 cm− 1. In contrast, a new and broad band appears at 1718 cm− 1, which may be considered as an evidence of the formation of polymerized and oxidized adamantylphenol products mainly containing carbonyl groups. With irradiation, the peak maxima at about 3340 cm− 1 largely shifted to 3390 cm− 1 with a distinct band at 3747 cm− 1 A similar chemical transformation was observed to occur for the silica films in which UV/ozone treatment increases silica condensation among vicinal silanol groups together with the occurrence of isolated silanol groups assigned at about 3700 cm− 1 without the reduction of peak intensity at around 3340 cm− 1 [16]. It is well known that UV treatment

Fig. 4. Auger depth profiles of the Si, O, and C for irradiated and non-irradiated films. Bold and thin lines denote the irradiated and non-irradiated films, respectively.

Fig. 5. Si 2p core level spectra of irradiated and non-irradiated films.

resulted in the formation of an enhanced network structure accompanied by the change of peak intensity at 3340 cm− 1 which corresponds to the presence of the silanol groups [17]. However, the observed IR spectrum indicates that there was no the clear reduction of peak intensity at 3340 cm− 1 corresponding to silanols after UV irradiation, although the peak intensity at about 1125 cm− 1 corresponding to Si–O–Si backbone slightly increased. In addition, it was reported that UV treatment leads to the formation of Si–OH groups for MSQ type films [18,19]. Thus, this spectral variation can be explained in terms of the significant change of interactions among adamantylphenol and silanol groups by the photodecomposition and photooxidation of adamantylphenol. Since densification of film can be strongly correlated to the chemical composition of films with photooxidation, the change of film composition with irradiation was investigated using Auger electron spectroscopy (AES). Fig. 4 shows AES depth profiling of films before and after UV irradiation, respectively. UV irradiated films relative to non-irradiation clearly show the significant increase of atomic concentrations of oxygen accompanied with the reduction of carbon species, while silicon concentration is nearly constant through total film thickness. From the change of each atomic concentration averaged through the bulk of the film, the relative atomic compositions represented by C/Si and O/Si were estimated before and after irradiation, respectively. The calculated results show that a relative oxygen content increased by 19.1% from 1.52, which is similar to PMSSQ composition presented by RSiO1.5, to 1.81. In contrast, carbon concentration lowers from 1.94 to 1.48 by 21.7% after irradiation. As can be expected from oxidation combined with degradation of carbon species, these results imply that atomic oxygen was incorporated into PMSSQ film, perhaps, by mechanism such as substitution of carbon by oxygen in accompanied with the formation of enhanced Si–O siloxane bonds. As a result, film density increases, as reflected by the enhancement of Si–O–Si band intensity at about 1100 cm− 1. Additionally, the oscillatory profiles observed for

B.J. Cha et al. / Thin Solid Films 515 (2007) 4603–4608

4607

irradiated film is assumed to result from interference effects between the incident beam and that of reflected from the silicon substrate, as previously reported [20]. The film densification is straightforwardly elucidated from XPS analysis, as shown in Fig. 5. In Si 2p core level spectra of film before irradiation, two peaks corresponded to 101.4 and 102.5 eV was assigned to Si–C and O–Si–C, respectively [21,22]. After irradiation, it is evident that peak maximum at 101.4 eV was slightly shifted to higher binding energy together with the occurrence of a broad band to higher binding energy. This behavior is interpreted in terms of densification of films resulting from the formation of oxygen enriched environment by photooxidation. This selective densification relative to that in non-exposed areas leads to the reduction of film thickness in normal direction to surface. Therefore, pattern formation is due to the lowering of film thickness originated from film densification concomitant with photooxidation of adamantylphenol groups within irradiated area. The effect of thermal treatment on pre-patterned film structure was investigated to realize the controllable patterning

Fig. 7. (a) UV-vis and (b) FT-IR spectra of the irradiated and non-irradiated PMSSQ films after thermal treatment at 300 °C for 2 h.

Fig. 6. (a) FE-SEM micrographs of PMSSQ film surface formed after thermal treatment at 300 °C for 2 h with irradiated film and (b) AFM image of the negative patterned surface structure. Scale bar denotes 100 μm.

for PMSSQ films. Fig. 6(a) shows a FE-SEM image of the irradiated film treated thermally at 300 °C under N2. UV irradiated regions (squares) are darker than non-exposed areas (lines) as opposed to previous films prepared by only UV irradiation. As confirmed by AFM image of Fig. 6(b), film thickness within non-irradiated areas was also lower than that of exposed areas with relative difference of 126 ± 19 nm. Thus, such inversed contrast is caused by the difference of film thickness between non-irradiated and irradiated region. To understand a mechanism about the inversed pattern formation further, both chemical structures of non-irradiated and irradiated films were investigated after heat treatment with UV and FT-IR spectroscopy. As illustrated in photoelectronic spectra of Fig. 7(a), there is no UV absorbance for non-irradiated film. In contrast, UV spectrum of irradiated film still showed featureless absorbance in spite of thermal treatment, indicating that the oxidized products remained within film. Thus, no absorbance in non-irradiated film can be understood in terms of sublimation of

4608

B.J. Cha et al. / Thin Solid Films 515 (2007) 4603–4608

adamantylphenol below 300 °C. These results are further elucidated by infrared spectroscopic measurements, as shown in Fig. 7(b). Relative to the spectrum of non-irradiated film before treatment in Fig. 3(b), FT-IR spectra of the treated films demonstrates that peak at 2910 cm− 1 due to C–H stretching vibration of adamantylphenol significantly decreased, as expected by sublimation of adamantylphenol. In contrast, peaks corresponding to polymerized adamantylphenol products at 2910 and 1720 cm− 1 still survived in irradiated film, reflecting that photopolymerized and oxidized adamantylphenol components have higher thermal stability than that of intact adamantylphenol. The effect of thermal treatment on each film can have an influence on chemical structure of PMSSQ matrix. As can be seen in Fig. 7(b), the intensity of the vibrational band assigned to the ladder-like Si–O stretching at around 1040 cm− 1 clearly increases after treatment in addition to the significant reduction of band intensity due to the cage-like Si–O stretching vibration at 1125 cm− 1. This molecular transformation is also accompanied with the removal of silanol groups, as indicated by the disappearance of band intensity at 3400 cm− 1 in Fig. 3(b). It has been well known that this spectroscopic trend implies a structural rearrangement from a cage-like structure to a ladderlike network structure due to polycondensation like film densification [23–26]. In contrast, such rearrangement reaction significantly attenuates for irradiated film, as shown in Fig. 7(b). As a result, the peak ratio (0.70) of the cage to ladder Si–O bonds in the non-irradiated film is lower than the value (0.82) for the irradiated film, indicating that the retainment of oxidized adamantylphenol products may limit further condensation and rearrangement reaction. Thus, the occurrence of inversed patterns with thermal treatment can be explained in term of the difference between the high polycondensation rate concomitant with sublimation of intact adamantylphenol in nonexposed region and rate retarded by the subsistence of oxidized and polymerized products in exposed areas, respectively. 4. Conclusion In present study, micropatterned surface structures were formed on PMSSQ thin film by employing photo-thermal sensitive adamantylphenol molecules together with UV irradiation and mild heat treatment. By irradiation, positive patterns were built up due to the film densification triggered by the photooxidation and polymerization of adamantylphenol. Prepatterned films were also transformed to film with negative patterns after thermal treatment. Such change of patterning structure could be explained in terms of difference of polycondensation rates in both non-irradiated and exposed regions. This facile photothermal patterning techniques is believed to be very useful in the design of patterned functional and microarrayed structures to introduce sensing, optical, and electrical components.

Acknowledgements This work was supported by 2005 Research Fund of Myongji University. Authors thank Mr. U. Cho for helping AFM measurements. References [1] R.H. Baney, M. Itoh, A. Sakakibara, T. Suzuki, Chem. Rev. 95 (1995) 1409. [2] Y. Ni, S. Zheng, Chem. Mater. 16 (2004) 5141. [3] P. Yang, G. Wirnsberger, H.C. Huang, S.R. Cordero, M.D. McGehee, B. Scott, T. Deng, G.M. Whitesides, B.F. Chmelka, S.K. Buratto, G.D. Stucky, Science 287 (2000) 465. [4] K. Su, D.R. Bujalski, K. Eguchi, G.V. Gordon, D.-L. Ou, P. Chevalier, S. Hu, R.P. Boisvert, Chem. Mater. 17 (2005) 2520. [5] H.-C. Kim, J.B. Wilds, C.R. Kreller, W. Volksen, P.J. Brock, V.Y. Lee, T. Magbitang, J.L. Hedrick, C.J. Hawker, R.D. Miller, Adv. Mater. 14 (2002) 1637. [6] D. Blanc, S. Pelissier, K. Saravanamuttu, S.I. Najafi, M.P. Andrews, Adv. Mater. 11 (1999) 1508. [7] T.C. Chang, T.M. Tsai, P.T. Liu, Y.S. Mor, C.W. Chen, J.T. Sheu, T.Y. Tseng, Electrochem. Solid-State Lett. 6 (2003) G69. [8] D.P. Mancini, K.A. Gehoski, E. Ainley, K.J. Nordquist, D.J. Resnick, T.C. Bailey, S.V. Sreenivasan, J.G. Ekerdt, C.G. Willson, J. Vac. Sci. Technol., B 20 (2002) 2896. [9] T. Kikkawa, T. Nagahara, H. Matsuo, Appl. Phys. Lett. 78 (2001) 2557. [10] H.-C. Kim, G. Wallraff, C.R. Kreller, S. Angelos, V.Y. Lee, W. Volksen, R.D. Miller, Nano Lett. 4 (2004) 1169. [11] H.-C. Kim, C.R. Kreller, K.A. Tran, V. Sisodiya, S. Angelos, G. Wallraff, S. Swanson, R.D. Miller, Chem. Mater. 16 (2004) 4267. [12] Y.Y. Lyu, J.H. Yim, Y. Byun, J.M. Kim, J.K. Jeon, Thin Solid Films 496 (2006) 526. [13] B.J. Cha, S. Kim, K. Char, Macromol. Res. 13 (2005) 257. [14] B.J. Cha, S. Kim, K. Char, J.-K. Lee, D.Y. Yoon, H.-W. Rhee, Chem. Mater. 18 (2006) 378. [15] B. Roig, C. Gonzalez, O. Thomas, Spectrochim. Acta, Part A 59 (2003) 303. [16] T. Clark, J.D. Ruiz, H. Fan, C.J. Brinker, B.I. Swanson, A.N. Parikh, Chem. Mater. 12 (2000) 3879. [17] F. Iacopi, Y. Travaly, B. Eyckens, T. Abell, E.P. Guyer, D.M. Gage, R.H. Dauskardt, T. Sajavaara, K. Houthoofd, P. Grobet, P. Jacobs, K. Maex, J. Appl. Phys. 99 (2006) 053511. [18] C. Waldfried, Q. Han, O. Escorcia, A. Margolis, R. Albano, I. Berry, Proceeding of IEEE International Interconnect Technology Conference, San Francisco, U.S.A., June 3–5, 2002, 2002, p. 226. [19] R.C. Hedden, C. Waldfried, H.-J. Lee, O. Escoricia, J. Electrochem. Soc. 151 (2004) F178. [20] O. Joubert, G. Hollinger, C. Fiori, R.A.B. Devine, P. Paniez, R. Pantel, J. Appl. Phys. 69 (1991) 6647. [21] J. Ma, L. Shi, Y. Shi, S. Luo, J. Xu, J. Appl. Polym. Sci. 85 (2002) 1077. [22] W.K. Choi, T.Y. Ong, L.S. Tan, F.C. Loh, K.L. Tan, J. Appl. Phys. 83 (1998) 4968. [23] W.-C. Liu, C.-C. Yang, W.-C. Chen, B.-T. Dai, M.-S. Tsai, J. Non-Cryst. Solids 311 (2002) 233. [24] C.T. Chua, G. Sarkar, X. Hu, J. Electrochem. Soc. 145 (1998) 4007. [25] Q.R. Huang, W. Volksen, E. Huang, M. Toney, C.W. Frank, R.D. Miller, Chem. Mater. 14 (2002) 3676. [26] L.H. Lee, W.C. Chen, W.C. Liu, J. Polym. Sci., Polym. Chem. Ed. 40 (2002) 1560.