- Email: [email protected]
Study of mesoporous silica films by positron annihilation based on a slow positron beam: Effects of preparation conditions on pore size and open porosity
Chemical Physics 331 (2007) 213–218 www.elsevier.com/locate/chemphys
Study of mesoporous silica films by positron annihilation based on a slow positron beam: Effects of preparation conditions on pore size and open porosity Chunqing He *, Ryoichi Suzuki, Toshiyuki Ohdaira, Nagayasu Oshima, Atsushi Kinomura, Makoto Muramatsu, Yoshinori Kobayashi * National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8565 and 8568, Japan Received 2 June 2006; accepted 23 October 2006 Available online 10 November 2006
Abstract Positron annihilation spectroscopy (PAS) based on an intense pulsed slow positron beam was applied to the study of mesoporous silica films, synthesized using tetraethyl orthosilicate (TEOS) as the network precursor and a triblock copolymer (EO106PO70EO106) as the structure-directing agent. With positron annihilation lifetime spectroscopy (PALS), pore sizes were obtained from ortho-positronium (o-Ps) lifetimes of the films capped with a 20 nm thick SiO2 layer. Influences of preparation conditions such as heating, TEOS vapor infiltration and precursor solution ageing on the pore size were studied. Moreover, the effect of ageing of the precursor solution on film pore interconnectivity/open porosity was investigated through lifetime–energy correlation measurements by observing intrinsic annihilation of o-Ps diffused out from the uncapped film surface. Ó 2006 Elsevier B.V. All rights reserved. PACS: 78.70.Bj Keywords: Positronium; Slow positron beam; Micropore; Mesopore; Copolymer; Silica film
1. Introduction Mesoporous materials have attracted much attention because they can be used not only in separation, catalysis, encapsulation, chemical sensing, but also as low-dielectric constant (low-k) films and optical coatings. These applications in most cases require well defined pores with narrow size distribution in addition to high stability and processability [1–6]. Various techniques such as gas adsorption porosimetry, small angle X-ray scattering (SAXS), small angle neutron scattering (SANS), transmission electron microscopy (TEM), ellipsometric porosimetry (EP) and positron annihilation spectroscopy (PAS) have been used *
Corresponding authors. Tel.: +81 29 861 4622; fax: +81 29 861 4622. E-mail addresses: [email protected] (C. He), [email protected] (Y. Kobayashi). 0301-0104/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chemphys.2006.10.016
to characterize the porous materials [6]. Of these, PAS possesses significant and unique advantages. First, one can measure the pore size down to a few angstrom and pore size distribution with positron annihilation lifetime spectroscopy (PALS). Secondly, the use of variable energy positron beams enables depth-profiling of thin films, since positrons can be injected into desired depths of a few nanometer to 1lm by adjusting the positron energy. Because of these advantages, PAS is often used to study mesoporous films and low-k dielectrics [7–14]. PAS relies on that the positron, the anti-particle of an electron, has unique annihilation properties in condensed matter. In porous materials, positronium (Ps), the bound state of a positron and an electron, forms either as a singlet state (para-Ps, p-Ps) or a triplet state (ortho-Ps, o-Ps). The ratio of the formation probability of p-Ps to o-Ps is 1:3. The intrinsic lifetime of p-Ps via 2c-annihilation is 0.125 ns,
214
C. He et al. / Chemical Physics 331 (2007) 213–218
while that of o-Ps via 3c-annihilation in a vacuum is 142 ns. However, when localized in a nanopore, o-Ps undergoes 2c pick-up annihilation and its lifetime is shortened to a few – a few tens nanoseconds depending on the pore size. This makes it possible to estimate the pore size from the measured o-Ps lifetime [7,11,12,15–21]. In the presence of interconnected, open pores, o-Ps may escape through the pores into a vacuum and then undergoes 3c-annihilation, which provides information on pore interconnectivity/open porosity. Mesoporous silica films can be synthesized by removal of polymer templates (surfactants) from mesostructural silica composite films prepared by the sol–gel processes [5,22]. The pore size, its distribution and open porosity are influenced by the molecular weight [22,23] and hydrophilicity of the surfactants [23,24], precursor sol and post-treatments of the deposited films [22,25]. Understanding the relationship between the pore structures and preparation conditions is very important for synthesizing mesoporous silica films with desired pores. In the present work, PAS was applied to study the pore size and pore interconnectivity/ open porosity of the mesoporous silica films synthesized under various conditions, using tetraethyl orthosilicate (TEOS) as the network precursor and a triblock copolymer EO106PO70EO106 (BASF surfactant, Pluronic F127) as the structure-directing agent. The effects of such preparation conditions as calcination, TEOS vapor infiltration and precursor ageing on the mesopore characteristics are discussed. 2. Experimental 2.1. Preparation of mesoporous thin silica films The mesoporous silica films were deposited by dip-coating [26] on a polished silicon (1 0 0) wafer. The coating solution was prepared by the addition of an ethanol solution of a PEO–PPO–PEO triblock copolymer (BASF Pluronic F127, EO106PO70EO106, with a molecular weight of 12600 g/mol) to silica sol–gel made by an acid-catalyzed process [25]. First, the silica sol was prepared by stirring the mixture of calculated amounts of TEOS (4.16 g), EtOH (3.5 ml), H2O and HCl (molar ratio = 1:3:8:0.003) at about 70 °C for 90 min. After aging at about 60 °C for 25 min, the triblock copolymer solution (1.512 g EO106PO70EO106: 32.5 ml EtOH) was added with stirring. The final molar ratio of TEOS:EO106PO70EO106:EtOH:H2O:HCl was 1:0.006:30:8:0.003. The film thickness was controlled to a few hundreds nm by adjusting the dip-coating rate. After the precursor solution aged at RT for 25 days, films A and B were deposited and dried overnight in air before calcination. Films C and D were deposited after 45 days and dried in air for a few days. After drying, the deposited films were placed in a tube furnace under Ar flow. After they were maintained in Ar for more than 30 min, the furnace temperature was raised, and then the films were calcined at 450 °C for 3 h under Ar gas flow
before being cooled slowly. After calcination, the mass of the films decreased by 52% due to the removal of the solvent and polymer template. The effect of pretreatment on the pore properties was studied with films subjected to different thermal treatments and a film subjected to TEOS vapor infiltration. Film A was heated with an increasing rate of 15 °C/min and calcined at 450 °C for 3 h. Films B and C were treated under Ar gas flow around 90 °C for 3 h and heated with an increasing rate of 1.5 °C/min before calcination at 450 °C for 3 h. Film D was subjected to TEOS vapor infiltration around 90 °C for 3 h before calcination. To prevent Ps diffusion out of the surface, some films were capped with a 20 nm thick SiO2 layer using electron beam deposition. Thicknesses of the calcined films were 300–440 nm and their refractive indices were around 1.27 as measured by an ellipsometer. The refractive indices were lower than the value of 1.46 for nonporous SiO2, indicating introduction of porosity in the films after calcination. The total porosity of the calcined films was estimated to be about 40%. The film preparation conditions are listed in Table 1. 2.2. PALS and lifetime–energy correlation measurement An experimental setup based on a variable energy intense pulsed slow positron beam at the National Institute of Advanced Industrial of Science and Technology was used to record positron annihilation lifetime spectra and lifetime–energy correlation data [27]. The positron annihilation signals were detected by a truncated BaF2 scintillator coupled to a photo-multiplier tube (PMT, Hamamatsu R2083Q). The output of the PMT was divided; one signal was connected to a constant fraction discriminator (CFD) and the other to a spectroscopy amplifier. The energy window was set to accept both 2-c and 3-c annihilation events. The positron annihilation lifetime spectrum was recorded by measuring the time difference between the CFD signal and the timing signal of the pulsing system using a timeto-amplitude converter (TAC) and analog-to-digitalconverter (ADC). Another ADC was used to record the c-ray energy data from the spectroscopy amplifier. To record lifetime–energy correlation spectra, coincident lifetime and energy signals were stored in a two-dimensional multi-channel analyzer (2D-MCA). Before the PAS experiments, all films were baked at 300 °C for 30 min in order to remove absorbed water.
Table 1 Film preparation conditions Film
Aging time (day)
Treatment
Heating rate (°C/min)
A B C D
25 25 45 45
90 °C 3 h 90 °C 3 h With VI, 90 °C 3 h
15 1.5 1.5 1.5
VI means TEOS vapor infiltration. All films were finally calcined at 450 °C for 3 h in order to remove the template.
C. He et al. / Chemical Physics 331 (2007) 213–218
215
The positron energy was fixed at 2 keV so that almost all positrons were injected into the mesoporous films. Due to the bunching of some positrons that remained after chopping, a number of weak satellite peaks appeared on the background of a positron lifetime spectrum [28]. Therefore, Kapton was used as a reference to obtain the background with the satellite peaks, to be subtracted from the lifetime spectra of the mesoporous silica films. Details of the positron annihilation system can be found elsewhere [29,30]. 3. Results and discussion 3.1. Positron annihilation lifetime spectra and pore sizes of porous films Positron annihilation lifetime spectra for uncapped and capped films are displayed in Fig. 1. It is seen that the spectra of the uncapped films differ from those of the capped ones; the o-Ps intensities for the latter films are significantly higher, because the 20 nm thick capping layer inhibits o-Ps diffusion out of the films. In our setup, the detector is placed behind the sample and in between them there is a c-ray collimator with 1 cm opening to reduce the background. In the case of an uncapped film with open porosity, some o-Ps atoms are emitted from the surface, fly over different lengths and undergo 3c-annihilation in a vacuum. The farther away from the film o-Ps decays, the less annihilation c-rays are detected. Therefore, o-Ps annihilation in a vacuum provides not only lower detection probability (in comparison with 2c/3c-annihilation inside the film) but also a highly distorted lifetime spectrum for an uncapped porous film with open porosity. Positron lifetime spectra of zeolites [14] and hypercrosslinked polystyrenes [31] have been reasonably decomposed into five components. We applied CONTIN [32,33] analysis to the lifetime spectrum of a capped silica film and found five well-separated dispersions [34]. Therefore, all the positron lifetime spectra for our capped films were resolved into five components using the PATFIT program [35], by fixing the first component s1 = 0.125 ns (p-Ps annihilation). The resolved second component s2 0.400 ns,
Fig. 1. Positron lifetime spectra for capped and uncapped films A, B, C and D.
invariant from one film to another, is due to the free positron annihilation. The lifetimes and relative intensities of three long-lived o-Ps components are listed in Table 2: s3 2.0 ns, which is often observed in nonporous amorphous SiO2, is attributed to o-Ps annihilation in the microvoids in the amorphous SiO2 matrix; s4 8.0 ns is probably due to o-Ps in micropores; the longest component, s5 60–70 ns is the o-Ps lifetime in mesopores. The relationship between the o-Ps lifetime and effective pore radius R 6 1 nm is given by the semi-empirical quantum-mechanical model of Tao and Eldrup [15–17] 1 R 1 2pR þ sin so-Ps ¼ s0 1 ; ð1Þ R þ dR 2p R þ dR where s0 is the spin averaged annihilation lifetime (0.5 ns), dR is an empirical parameter that describes the thickness of
Table 2 Positronium lifetimes and their intensities, diameters/side lengths of microvoids and micro/mesopores for different films Film
s3 (ns) I3 (%)
s4 (ns) I4 (%)
s5 (ns) I5 (%)
Microvoid 2Ra (nm)
Micropore 2Ra (nm)
Mesoporeb Dt/Dc (nm)
A
2.12 ± 0.03 7.21 ± 0.03 1.86 ± 0.02 7.15 ± 0.07 2.09 ± 0.02 8.28 ± 0.05 2.10 ± 0.02 7.13 ± 0.03
8.68 ± 0.07 5.25 ± 0.04 7.02 ± 0.04 4.94 ± 0.03 8.57 ± 0.07 4.06 ± 0.03 8.95 ± 0.06 4.29 ± 0.03
68.2 ± 0.1 22.05 ± 0.04 63.7 ± 0.1 23.35 ± 0.04 69.8 ± 0.1 23.32 ± 0.05 73.6 ± 0.1 26.54 ± 0.04
0.59
1.24
3.89/5.26
0.57
1.12
3.56/4.76
0.59
1.23
4.02/5.44
0.59
1.26
4.34/5.93
B C D
Errors for the lifetimes and intensities indicate statistical uncertainties computed by the PATFIT program. The errors are not shown for the deduced pore sizes because they are quite small compared with the listed pore sizes. a Microvoid and micropore sizes were estimated from s3 and s4 according to the TE model assuming a spherical pore (2R). b Mesopore sizes were obtained from s5 according to the RTE model assuming both an infinite square tubular pore (Dt) and a cubic pore (Dc).
216
C. He et al. / Chemical Physics 331 (2007) 213–218
the electron layer effectively seen by the positron in Ps and ˚ by fitting Eq. (1) to Ps has been determined to be 1.656 A lifetimes in materials with known pore sizes. For larger pores (R > 1 nm), the intrinsic annihilation and classical particle nature of Ps have to be taken into account. Recently, the Tao–Eldrup model (TE model) has been extended to account for large pores by Shantarovich [18], Goworek et al. [19,20], Gidley et al. [7,11,12] and Ito et al. [21]. Plots of lifetime vs. pore size computed based on the TE model and the extended rectangular Tao–Eldrup model (RTE model, developed by Gidley et al.) are shown in Fig. 2. The latter model by Gidley et al. has been frequently applied to determine the mesopore sizes in porous materials and low-k films [36–38], because pore sizes can be simply computed from the measured o-Ps lifetimes. We calculated the sizes of microvoids and micropores from the o-Ps lifetimes according to the TE model. For mesopores, we used the RTE model. The sizes of the microvoids and micropores estimated from s3 and s4 are 0.6 and 1.2 nm in diameter, respectively. However, the mesopore size was found to depend on the preparation condition and vary from 4.8 to 5.9 nm or 3.6 to 4.3 nm. The two pore size ranges refer to the values obtained under assumptions of cubic and tubular pores, respectively. The mesopore sizes in our films templated by F127 derived by PALS are comparable to those of mesoporous silica films synthesized using the same template evaluated by gas adsorption [22,39] and those of low-k films prepared from TEOS/F127 with a similar composition (1:0.005) estimated by PALS [13]. 3.2. Effects of thermal treatment and TEOS vapor infiltration on the mesopore size The contraction of sol–gel synthesized silica films, depending on their thermal stability and treatment conditions, is often observed after calcination [40]. The estimated pore sizes of films A and B are 3.9 (tubular pore)/5.3 (cubic pore) nm and 3.6/4.8 nm, respectively. These films were subjected to different thermal treatments before being calcined at 450 °C for 3 h. Film A was heated from room temperature to 450 °C with a rate of 15 °C/min and then
Fig. 2. Ps lifetime vs. pore diameter according to the classical TE model and pore side length according to the extended RTE model. The pore side length does not include the thickness of the electron layer.
calcined at the same temperature for 3 h, while film B was dried at 90 °C for 3 h before being heated to 450 °C with a rate of 1.5 °C/min (Table 1). Thus, faster heating seems to be responsible for the larger pore size of film A. Two competitive processes play important roles in determining the pore size of the calcined films: reactions between Si–OH groups and relaxation of the pore walls. The pore walls surrounding the template molecules contract during condensation of the Si–OH groups. The condensation is not completed at room temperature and proceeds more effectively at higher temperatures. Slow heating after drying the deposited film at 90 °C enables the silanol groups of the silica walls to well relax and results in more contraction of pore walls after calcination. On the other hand, the rapid formation of rigid Si–O–Si pore wall network under fast heating hinders the further relaxation of Si–OH groups before their complete condensation. This can account for the larger pore size of film A in comparison with film B. It has been reported that mesoporous silica thin films prepared under TEOS vapor infiltration (VI) [41,42] have higher structural stability than those prepared without VI. In order to investigate the effect of TEOS vapor infiltration, film D was treated under TEOS vapor around 90 °C for 3 h before further being elevated to a high temperature of 450 °C for the removal of the template. In Table 2, the size of mesopores in the film with VI (D) is found to be 4.3/5.9 nm, which is larger than the pore size of the film without VI (C), i.e., 4.0/5.4 nm. CONTIN analysis was applied to the positron lifetime spectra of films C and D. The o-Ps lifetime distributions in the mesopores are shown in Fig. 3. The film without VI (C) has an o-Ps lifetime distribution centered around 65 ns with a full width at half maximum (FWHM) of 12 ns. On the other hand, the o-Ps lifetime distribution of the film with VI (D) is shifted to 67 ns and has a FWHM of 18 ns. Assuming cubic pores, the pore sizes of the film with and without TEOS VI are distributed over 4.2– 6.6 nm and 3.9–7.9 nm, respectively. These results indicate
Fig. 3. Lifetime distributions of the longest o-Ps components in silica capped films, which were treated with and without TEOS vapor infiltration before calcination. Cubic pore sizes calculated from o-Ps lifetimes are indicated on the upper axis.
C. He et al. / Chemical Physics 331 (2007) 213–218
217
that the mesopore size distribution in the film with VI is slightly less uniform than that without VI. The structural stability of a silica network in as-deposited film may be low because a number of Si–OH groups remain unreacted. The density of the silica walls surrounding the templates may be low, resulting in considerable structural contraction of pore walls during condensation of the silanol groups at high temperatures. However, during VI-treatment TEOS molecules penetrate the silica walls and react with the residual Si–OH groups. The silica walls of the VI treated mesoporous silica film have higher density, structural stability and less contract under the same calcination condition. Thus, the larger pores in film D are attributed to improved stiffness of the pore walls. 3.3. Effect of ageing of precursor solution on pore interconnectivity/open porosity Ps can escape from the uncapped film surface and annihilate via 3c-annihilation if the pores are interconnected and open to a vacuum. To study the pore interconnectivity/open porosity, we measured lifetime–energy spectra for the mesoporous films. Positron annihilation c-ray spectra in the annihilation time region of 28.8–340 ns for uncapped and capped films are shown in Fig. 4. The spectra are almost the same for all the capped films (Fig. 4(a)). Careful inspection enables us to find that the intensity of 3-c annihilation is higher for a film with larger pores as expected. However, in
Fig. 5. Normalized positron annihilation c-ray spectra in different annihilation time regions for uncapped porous silica films (a) B and (b) C obtained by lifetime–energy correlation measurement. The incident positron energy is 2 keV.
Fig. 4(b), the contribution of 3c-annihilation centered around 0.350 MeV for uncapped films C and D are much higher than those of uncapped films A and B. Further, in spite of different thermal treatments, the c-ray spectra for films B and C are essentially the same as films A and D, respectively. The c-ray spectra in different positron annihilation time regions for uncapped films B and C are plotted in Fig. 5. We do not present the spectra of uncapped films A and D because they are almost identical to B and C, respectively. At times shorter than 2.4 ns, a distinct 2c-annihilation peak around 0.511 MeV is observed for films B and C. The 3c fraction is enhanced drastically in the longer positron ages for film C (and also D, data not shown). In the time region from 48.4 to 72.3 ns, the 2c-annihilation characteristic peak is clearly observable for film B (and A, data not shown), but not for film C (and D, data not shown). The above results show that the fraction of o-Ps that escapes from films C/D is higher, indicative of their much higher pore interconnectivity/open porosity than films A/ B. A possible reason is that highly connected-micelles of the template are formed in the precursor solution after longer ageing, resulting in well-developed, interconnected open pores after removal of the template. 4. Conclusions
Fig. 4. Normalized positron annihilation c-ray spectra for (a) capped and (b) uncapped films in the annihilation time region from 28.8 to 340 ns obtained by lifetime–energy correlation measurements. The incident positron energy is 2.0 keV.
Mesoporous silica films were synthesized using a triblock copolymer as the structure director. Positron annihilation lifetime spectra were well resolved into five
218
C. He et al. / Chemical Physics 331 (2007) 213–218
components, including three long-lived o-Ps lifetimes. Microvoids and micropores were found to be 0.6 and 1.2 nm in diameter, respectively, and almost independent of pre-treatments. However, pre-treatments on the asdeposited films influenced the mesopore sizes and pore interconnectivity/open porosity of the calcined films. Fast heating resulted in slightly larger mesopores. The pores in a calcined film with TEOS vapor infiltration were observed larger than those in the film without VI. A possible reason is that TEOS molecules penetrated the silica network, reacted with some residual Si–OH groups, thereby reduced the contraction of the silica walls during their condensation. The positron annihilation lifetime–energy correlation results indicated that longer ageing of the precursor solution led to a considerable increment in the pore interconnectivity/open porosity in the mesoporous silica films. Acknowledgements This work was supported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan. One of the authors (C. He) would like to thank the Japan Society for the Promotion of Science (JSPS) for the financial support. Drs. K. Ito and T. Oka are appreciated for their assistance of experiments. References [1] M.E. Davis, Nature 417 (2002) 813. [2] P. Yang, T. Deng, D. Zhao, P. Feng, D. Pine, B.F. Chmelka, G.M. Whitesides, G.D. Stucky, Science 282 (1998) 2244. [3] J.Y. Ying, C.P. Mehnert, M.S. Wong, Angew. Chem., Int. Ed. 38 (1999) 56. [4] X. He, D. Antonelli, Angew. Chem., Int. Ed. 41 (2002) 214. [5] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. [6] K. Maex, M.R. Baklanov, D. Shamiryan, F. Iacopi, S.H. Brongersma, Z.S. Yanovitskaya, J. Appl. Phys. 39 (2003) 8793. [7] D.W. Gidley, W.E. Frieze, T.L. Dull, J. Sun, A.F. Yee, C.V. Nguyen, D.Y. Yoon, Appl. Phys. Lett. 76 (2000) 1282. [8] J. Xu, J. Moxom, S. Yang, R. Suzuki, T. Ohdaira, Chem. Phys. Lett. 364 (2002) 309. [9] M.P. Petkov, M.H. Weber, K.G. Lynn, K.P. Rodbell, S.A. Cohen, J. Appl. Phys. 86 (1999) 3104. [10] Y. Kobayashi, W. Zheng, T.B. Chang, K. Hirata, R. Suzuki, T. Ohdaira, K. Ito, J. Appl. Phys. 91 (2002) 1704. [11] D.W. Gidley, W.E. Frieze, T.L. Dull, A.F. Yee, E.T. Ryan, H.M. Ho, Phys. Rev. B 60 (8) (1999) R5157. [12] T.L. Dull, W.E. Frieze, D.W. Gidley, J. Sun, A.F. Yee, J. Phys. Chem. 105 (2001) 4657. [13] A. van Veen, R. Escobar, H. Schut, S.W.H. Eijt, C.V. Falub, A.R. Balkenende, F.K. de Theije, Mater. Sci. Eng. B 102 (2003) 2.
[14] R. Zaleski, J. Wawryszczuk, A. Borowka, J. Goworek, T. Goworek, Micropor. Mesopor. Mater. 62 (2003) 47. [15] S.J. Tao, J. Chem. Phys. 56 (1972) 5499. [16] M. Eldrup, D. Lightbody, J.N. Sherwood, Chem. Phys. 63 (1981) 51. [17] H. Nakanishi, S.J. Wang, Y.C. Jean, in: S.C. Sharma (Ed.), Positron Annihilation Studies of Fluids, World Scientific, Singapore, 1988, pp. 292–298. [18] V.P. Shantarovich, J. Radioanal. Nucl. Chem. 210 (1996) 357. [19] T. Goworek, K. Ciesielski, B. Jasin´ska, J. Wawryszczuk, Chem. Phys. 230 (1998) 305. [20] T. Goworek, K. Ciesielski, B. Jasin´ska, J. Wawryszczuk, Radiat. Phys. Chem. 58 (2000) 719. [21] K. Ito, H. Nakanishi, Y. Ujihira, J. Phys. Chem. B 103 (1999) 4555. [22] D. Zhao, Q. Huo, J. Feng, B.F. Chmella, G.D. Stucky, J. Am. Chem. Soc. 120 (1998) 6024. [23] C. He, M. Muramatsu, T. Ohdaira, N. Oshima, A. Kinomura, R. Suzuki, Y. Kobayashi, Radiat. Phys. Chem. (2006) in press, doi:10.1016/j.radphyschem.2006.03.036. [24] C. Yu, J. Fan, B. Tian, G.D. Stucky, D. Zhao, J. Phys. Chem. B 107 (2003) 13368. [25] X. Li, J.C. Birnbaum, R.E. Williford, G.E. Fryxell, C.A. Coyle, G.C. Dunham, S. Baskaran, Chem. Commun. (2003) 2054. [26] Y. Lu, R. Ganguli, C.A. Drewien, M.T. Anderson, J.C. Brinker, W. Gong, Y. Guo, H. Soyez, B. Dunn, M.H. Huang, J.I. Zink, Nature 389 (1997) 364. [27] R. Suzuki, T. Ohdaira, Y. Kobayashi, K. Ito, Y. Shioya, T. Ishimaru, Raidat. Phys. Chem. 68 (2003) 339. [28] C. He, E. Hamada, N. Djourelov, T. Suzuki, H. Kobayashi, K. Kondo, Y. Ito, Nucl. Instrum. Methods Phys. Res., Sect. B 211 (2003) 571. [29] R. Suzuki, T. Ohdaira, A. Uedono, Y. Kobayashi, Appl. Surf. Sci. 194 (2002) 89. [30] R. Suzuki, T. Ohdaira, Y. Kobayashi, K. Ito, R.S. Ru, Y. Shioya, H. Ichikawa, H. Hosomi, K. Ishikiriyama, H. Shirataki, S. Matsuno, J. Xu, Mater. Sci. Forum 445–446 (2004) 224. [31] C. He, T. Suzuki, V.P. Shantarovich, K. Kondo, Y. Ito, Chem. Phys. 303 (2004) 219. [32] S.W. Provencher, Comput. Phys. Commun. 27 (1982) 213. [33] R.B. Gregory, Nucl. Instrum. Methods Phys. Res., Sect. A 302 (1991) 496. [34] C. He, M. Muramatsu, T. Ohdaira, A. Kinomura, R. Suzuki, K. Ito, Y. Kobayashi, Appl. Surf. Sci. 252 (2006) 3221. [35] P. Kirkegaard, M. Eldrup, O.E. Mogensen, N.J. Pedersen, Comput. Phys. Commun. 23 (1981) 307. [36] T. Ohdaira, R. Suzuki, H. Shirataki, S. Matsuno, Mater. Sci. Forum 445–446 (2004) 334. [37] J.H. Yim, Y.Y. Lyu, H.D. Jeong, S.A. Song, I.S. Hwang, J.H. Lee, A.K. Mah, S.C. Chang, J.G. Park, Y.F. Hu, J.N. Sun, D.W. Gidley, Adv. Funct. Mater. 13 (2003) 382. [38] S. Li, J. Sun, Z. Li, H. Peng, D.W. Gidley, E. Todd Ryan, Y. Yan, J. Phys. Chem. B 108 (2004) 11689. [39] L. Pei, K.I. Kurumada, M. Tanigaki, M. Hori, K. Susa, J. Mater. Sci. 39 (2004) 663. [40] M. Ogawa, N. Masukawa, Micropor. Mesopor. Mater. 38 (2000) 35. [41] N. Nishiyama, S. Tanaka, Y. Egashira, Y. Oku, K. Ueyama, Chem. Mater. 15 (2003) 1006. [42] S. Tanaka, N. Nishiyama, Y. Egashira, Y. Oku, K. Ueyama, Micropor. Mesopor. Mater. 63 (2003) 105.
Study of mesoporous silica films by positron annihilation based on a slow positron beam: Effects of preparation conditions on pore size and open porosity Chunqing He *, Ryoichi Suzuki, Toshiyuki Ohdaira, Nagayasu Oshima, Atsushi Kinomura, Makoto Muramatsu, Yoshinori Kobayashi * National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8565 and 8568, Japan Received 2 June 2006; accepted 23 October 2006 Available online 10 November 2006
Abstract Positron annihilation spectroscopy (PAS) based on an intense pulsed slow positron beam was applied to the study of mesoporous silica films, synthesized using tetraethyl orthosilicate (TEOS) as the network precursor and a triblock copolymer (EO106PO70EO106) as the structure-directing agent. With positron annihilation lifetime spectroscopy (PALS), pore sizes were obtained from ortho-positronium (o-Ps) lifetimes of the films capped with a 20 nm thick SiO2 layer. Influences of preparation conditions such as heating, TEOS vapor infiltration and precursor solution ageing on the pore size were studied. Moreover, the effect of ageing of the precursor solution on film pore interconnectivity/open porosity was investigated through lifetime–energy correlation measurements by observing intrinsic annihilation of o-Ps diffused out from the uncapped film surface. Ó 2006 Elsevier B.V. All rights reserved. PACS: 78.70.Bj Keywords: Positronium; Slow positron beam; Micropore; Mesopore; Copolymer; Silica film
1. Introduction Mesoporous materials have attracted much attention because they can be used not only in separation, catalysis, encapsulation, chemical sensing, but also as low-dielectric constant (low-k) films and optical coatings. These applications in most cases require well defined pores with narrow size distribution in addition to high stability and processability [1–6]. Various techniques such as gas adsorption porosimetry, small angle X-ray scattering (SAXS), small angle neutron scattering (SANS), transmission electron microscopy (TEM), ellipsometric porosimetry (EP) and positron annihilation spectroscopy (PAS) have been used *
Corresponding authors. Tel.: +81 29 861 4622; fax: +81 29 861 4622. E-mail addresses: [email protected] (C. He), [email protected] (Y. Kobayashi). 0301-0104/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chemphys.2006.10.016
to characterize the porous materials [6]. Of these, PAS possesses significant and unique advantages. First, one can measure the pore size down to a few angstrom and pore size distribution with positron annihilation lifetime spectroscopy (PALS). Secondly, the use of variable energy positron beams enables depth-profiling of thin films, since positrons can be injected into desired depths of a few nanometer to 1lm by adjusting the positron energy. Because of these advantages, PAS is often used to study mesoporous films and low-k dielectrics [7–14]. PAS relies on that the positron, the anti-particle of an electron, has unique annihilation properties in condensed matter. In porous materials, positronium (Ps), the bound state of a positron and an electron, forms either as a singlet state (para-Ps, p-Ps) or a triplet state (ortho-Ps, o-Ps). The ratio of the formation probability of p-Ps to o-Ps is 1:3. The intrinsic lifetime of p-Ps via 2c-annihilation is 0.125 ns,
214
C. He et al. / Chemical Physics 331 (2007) 213–218
while that of o-Ps via 3c-annihilation in a vacuum is 142 ns. However, when localized in a nanopore, o-Ps undergoes 2c pick-up annihilation and its lifetime is shortened to a few – a few tens nanoseconds depending on the pore size. This makes it possible to estimate the pore size from the measured o-Ps lifetime [7,11,12,15–21]. In the presence of interconnected, open pores, o-Ps may escape through the pores into a vacuum and then undergoes 3c-annihilation, which provides information on pore interconnectivity/open porosity. Mesoporous silica films can be synthesized by removal of polymer templates (surfactants) from mesostructural silica composite films prepared by the sol–gel processes [5,22]. The pore size, its distribution and open porosity are influenced by the molecular weight [22,23] and hydrophilicity of the surfactants [23,24], precursor sol and post-treatments of the deposited films [22,25]. Understanding the relationship between the pore structures and preparation conditions is very important for synthesizing mesoporous silica films with desired pores. In the present work, PAS was applied to study the pore size and pore interconnectivity/ open porosity of the mesoporous silica films synthesized under various conditions, using tetraethyl orthosilicate (TEOS) as the network precursor and a triblock copolymer EO106PO70EO106 (BASF surfactant, Pluronic F127) as the structure-directing agent. The effects of such preparation conditions as calcination, TEOS vapor infiltration and precursor ageing on the mesopore characteristics are discussed. 2. Experimental 2.1. Preparation of mesoporous thin silica films The mesoporous silica films were deposited by dip-coating [26] on a polished silicon (1 0 0) wafer. The coating solution was prepared by the addition of an ethanol solution of a PEO–PPO–PEO triblock copolymer (BASF Pluronic F127, EO106PO70EO106, with a molecular weight of 12600 g/mol) to silica sol–gel made by an acid-catalyzed process [25]. First, the silica sol was prepared by stirring the mixture of calculated amounts of TEOS (4.16 g), EtOH (3.5 ml), H2O and HCl (molar ratio = 1:3:8:0.003) at about 70 °C for 90 min. After aging at about 60 °C for 25 min, the triblock copolymer solution (1.512 g EO106PO70EO106: 32.5 ml EtOH) was added with stirring. The final molar ratio of TEOS:EO106PO70EO106:EtOH:H2O:HCl was 1:0.006:30:8:0.003. The film thickness was controlled to a few hundreds nm by adjusting the dip-coating rate. After the precursor solution aged at RT for 25 days, films A and B were deposited and dried overnight in air before calcination. Films C and D were deposited after 45 days and dried in air for a few days. After drying, the deposited films were placed in a tube furnace under Ar flow. After they were maintained in Ar for more than 30 min, the furnace temperature was raised, and then the films were calcined at 450 °C for 3 h under Ar gas flow
before being cooled slowly. After calcination, the mass of the films decreased by 52% due to the removal of the solvent and polymer template. The effect of pretreatment on the pore properties was studied with films subjected to different thermal treatments and a film subjected to TEOS vapor infiltration. Film A was heated with an increasing rate of 15 °C/min and calcined at 450 °C for 3 h. Films B and C were treated under Ar gas flow around 90 °C for 3 h and heated with an increasing rate of 1.5 °C/min before calcination at 450 °C for 3 h. Film D was subjected to TEOS vapor infiltration around 90 °C for 3 h before calcination. To prevent Ps diffusion out of the surface, some films were capped with a 20 nm thick SiO2 layer using electron beam deposition. Thicknesses of the calcined films were 300–440 nm and their refractive indices were around 1.27 as measured by an ellipsometer. The refractive indices were lower than the value of 1.46 for nonporous SiO2, indicating introduction of porosity in the films after calcination. The total porosity of the calcined films was estimated to be about 40%. The film preparation conditions are listed in Table 1. 2.2. PALS and lifetime–energy correlation measurement An experimental setup based on a variable energy intense pulsed slow positron beam at the National Institute of Advanced Industrial of Science and Technology was used to record positron annihilation lifetime spectra and lifetime–energy correlation data [27]. The positron annihilation signals were detected by a truncated BaF2 scintillator coupled to a photo-multiplier tube (PMT, Hamamatsu R2083Q). The output of the PMT was divided; one signal was connected to a constant fraction discriminator (CFD) and the other to a spectroscopy amplifier. The energy window was set to accept both 2-c and 3-c annihilation events. The positron annihilation lifetime spectrum was recorded by measuring the time difference between the CFD signal and the timing signal of the pulsing system using a timeto-amplitude converter (TAC) and analog-to-digitalconverter (ADC). Another ADC was used to record the c-ray energy data from the spectroscopy amplifier. To record lifetime–energy correlation spectra, coincident lifetime and energy signals were stored in a two-dimensional multi-channel analyzer (2D-MCA). Before the PAS experiments, all films were baked at 300 °C for 30 min in order to remove absorbed water.
Table 1 Film preparation conditions Film
Aging time (day)
Treatment
Heating rate (°C/min)
A B C D
25 25 45 45
90 °C 3 h 90 °C 3 h With VI, 90 °C 3 h
15 1.5 1.5 1.5
VI means TEOS vapor infiltration. All films were finally calcined at 450 °C for 3 h in order to remove the template.
C. He et al. / Chemical Physics 331 (2007) 213–218
215
The positron energy was fixed at 2 keV so that almost all positrons were injected into the mesoporous films. Due to the bunching of some positrons that remained after chopping, a number of weak satellite peaks appeared on the background of a positron lifetime spectrum [28]. Therefore, Kapton was used as a reference to obtain the background with the satellite peaks, to be subtracted from the lifetime spectra of the mesoporous silica films. Details of the positron annihilation system can be found elsewhere [29,30]. 3. Results and discussion 3.1. Positron annihilation lifetime spectra and pore sizes of porous films Positron annihilation lifetime spectra for uncapped and capped films are displayed in Fig. 1. It is seen that the spectra of the uncapped films differ from those of the capped ones; the o-Ps intensities for the latter films are significantly higher, because the 20 nm thick capping layer inhibits o-Ps diffusion out of the films. In our setup, the detector is placed behind the sample and in between them there is a c-ray collimator with 1 cm opening to reduce the background. In the case of an uncapped film with open porosity, some o-Ps atoms are emitted from the surface, fly over different lengths and undergo 3c-annihilation in a vacuum. The farther away from the film o-Ps decays, the less annihilation c-rays are detected. Therefore, o-Ps annihilation in a vacuum provides not only lower detection probability (in comparison with 2c/3c-annihilation inside the film) but also a highly distorted lifetime spectrum for an uncapped porous film with open porosity. Positron lifetime spectra of zeolites [14] and hypercrosslinked polystyrenes [31] have been reasonably decomposed into five components. We applied CONTIN [32,33] analysis to the lifetime spectrum of a capped silica film and found five well-separated dispersions [34]. Therefore, all the positron lifetime spectra for our capped films were resolved into five components using the PATFIT program [35], by fixing the first component s1 = 0.125 ns (p-Ps annihilation). The resolved second component s2 0.400 ns,
Fig. 1. Positron lifetime spectra for capped and uncapped films A, B, C and D.
invariant from one film to another, is due to the free positron annihilation. The lifetimes and relative intensities of three long-lived o-Ps components are listed in Table 2: s3 2.0 ns, which is often observed in nonporous amorphous SiO2, is attributed to o-Ps annihilation in the microvoids in the amorphous SiO2 matrix; s4 8.0 ns is probably due to o-Ps in micropores; the longest component, s5 60–70 ns is the o-Ps lifetime in mesopores. The relationship between the o-Ps lifetime and effective pore radius R 6 1 nm is given by the semi-empirical quantum-mechanical model of Tao and Eldrup [15–17] 1 R 1 2pR þ sin so-Ps ¼ s0 1 ; ð1Þ R þ dR 2p R þ dR where s0 is the spin averaged annihilation lifetime (0.5 ns), dR is an empirical parameter that describes the thickness of
Table 2 Positronium lifetimes and their intensities, diameters/side lengths of microvoids and micro/mesopores for different films Film
s3 (ns) I3 (%)
s4 (ns) I4 (%)
s5 (ns) I5 (%)
Microvoid 2Ra (nm)
Micropore 2Ra (nm)
Mesoporeb Dt/Dc (nm)
A
2.12 ± 0.03 7.21 ± 0.03 1.86 ± 0.02 7.15 ± 0.07 2.09 ± 0.02 8.28 ± 0.05 2.10 ± 0.02 7.13 ± 0.03
8.68 ± 0.07 5.25 ± 0.04 7.02 ± 0.04 4.94 ± 0.03 8.57 ± 0.07 4.06 ± 0.03 8.95 ± 0.06 4.29 ± 0.03
68.2 ± 0.1 22.05 ± 0.04 63.7 ± 0.1 23.35 ± 0.04 69.8 ± 0.1 23.32 ± 0.05 73.6 ± 0.1 26.54 ± 0.04
0.59
1.24
3.89/5.26
0.57
1.12
3.56/4.76
0.59
1.23
4.02/5.44
0.59
1.26
4.34/5.93
B C D
Errors for the lifetimes and intensities indicate statistical uncertainties computed by the PATFIT program. The errors are not shown for the deduced pore sizes because they are quite small compared with the listed pore sizes. a Microvoid and micropore sizes were estimated from s3 and s4 according to the TE model assuming a spherical pore (2R). b Mesopore sizes were obtained from s5 according to the RTE model assuming both an infinite square tubular pore (Dt) and a cubic pore (Dc).
216
C. He et al. / Chemical Physics 331 (2007) 213–218
the electron layer effectively seen by the positron in Ps and ˚ by fitting Eq. (1) to Ps has been determined to be 1.656 A lifetimes in materials with known pore sizes. For larger pores (R > 1 nm), the intrinsic annihilation and classical particle nature of Ps have to be taken into account. Recently, the Tao–Eldrup model (TE model) has been extended to account for large pores by Shantarovich [18], Goworek et al. [19,20], Gidley et al. [7,11,12] and Ito et al. [21]. Plots of lifetime vs. pore size computed based on the TE model and the extended rectangular Tao–Eldrup model (RTE model, developed by Gidley et al.) are shown in Fig. 2. The latter model by Gidley et al. has been frequently applied to determine the mesopore sizes in porous materials and low-k films [36–38], because pore sizes can be simply computed from the measured o-Ps lifetimes. We calculated the sizes of microvoids and micropores from the o-Ps lifetimes according to the TE model. For mesopores, we used the RTE model. The sizes of the microvoids and micropores estimated from s3 and s4 are 0.6 and 1.2 nm in diameter, respectively. However, the mesopore size was found to depend on the preparation condition and vary from 4.8 to 5.9 nm or 3.6 to 4.3 nm. The two pore size ranges refer to the values obtained under assumptions of cubic and tubular pores, respectively. The mesopore sizes in our films templated by F127 derived by PALS are comparable to those of mesoporous silica films synthesized using the same template evaluated by gas adsorption [22,39] and those of low-k films prepared from TEOS/F127 with a similar composition (1:0.005) estimated by PALS [13]. 3.2. Effects of thermal treatment and TEOS vapor infiltration on the mesopore size The contraction of sol–gel synthesized silica films, depending on their thermal stability and treatment conditions, is often observed after calcination [40]. The estimated pore sizes of films A and B are 3.9 (tubular pore)/5.3 (cubic pore) nm and 3.6/4.8 nm, respectively. These films were subjected to different thermal treatments before being calcined at 450 °C for 3 h. Film A was heated from room temperature to 450 °C with a rate of 15 °C/min and then
Fig. 2. Ps lifetime vs. pore diameter according to the classical TE model and pore side length according to the extended RTE model. The pore side length does not include the thickness of the electron layer.
calcined at the same temperature for 3 h, while film B was dried at 90 °C for 3 h before being heated to 450 °C with a rate of 1.5 °C/min (Table 1). Thus, faster heating seems to be responsible for the larger pore size of film A. Two competitive processes play important roles in determining the pore size of the calcined films: reactions between Si–OH groups and relaxation of the pore walls. The pore walls surrounding the template molecules contract during condensation of the Si–OH groups. The condensation is not completed at room temperature and proceeds more effectively at higher temperatures. Slow heating after drying the deposited film at 90 °C enables the silanol groups of the silica walls to well relax and results in more contraction of pore walls after calcination. On the other hand, the rapid formation of rigid Si–O–Si pore wall network under fast heating hinders the further relaxation of Si–OH groups before their complete condensation. This can account for the larger pore size of film A in comparison with film B. It has been reported that mesoporous silica thin films prepared under TEOS vapor infiltration (VI) [41,42] have higher structural stability than those prepared without VI. In order to investigate the effect of TEOS vapor infiltration, film D was treated under TEOS vapor around 90 °C for 3 h before further being elevated to a high temperature of 450 °C for the removal of the template. In Table 2, the size of mesopores in the film with VI (D) is found to be 4.3/5.9 nm, which is larger than the pore size of the film without VI (C), i.e., 4.0/5.4 nm. CONTIN analysis was applied to the positron lifetime spectra of films C and D. The o-Ps lifetime distributions in the mesopores are shown in Fig. 3. The film without VI (C) has an o-Ps lifetime distribution centered around 65 ns with a full width at half maximum (FWHM) of 12 ns. On the other hand, the o-Ps lifetime distribution of the film with VI (D) is shifted to 67 ns and has a FWHM of 18 ns. Assuming cubic pores, the pore sizes of the film with and without TEOS VI are distributed over 4.2– 6.6 nm and 3.9–7.9 nm, respectively. These results indicate
Fig. 3. Lifetime distributions of the longest o-Ps components in silica capped films, which were treated with and without TEOS vapor infiltration before calcination. Cubic pore sizes calculated from o-Ps lifetimes are indicated on the upper axis.
C. He et al. / Chemical Physics 331 (2007) 213–218
217
that the mesopore size distribution in the film with VI is slightly less uniform than that without VI. The structural stability of a silica network in as-deposited film may be low because a number of Si–OH groups remain unreacted. The density of the silica walls surrounding the templates may be low, resulting in considerable structural contraction of pore walls during condensation of the silanol groups at high temperatures. However, during VI-treatment TEOS molecules penetrate the silica walls and react with the residual Si–OH groups. The silica walls of the VI treated mesoporous silica film have higher density, structural stability and less contract under the same calcination condition. Thus, the larger pores in film D are attributed to improved stiffness of the pore walls. 3.3. Effect of ageing of precursor solution on pore interconnectivity/open porosity Ps can escape from the uncapped film surface and annihilate via 3c-annihilation if the pores are interconnected and open to a vacuum. To study the pore interconnectivity/open porosity, we measured lifetime–energy spectra for the mesoporous films. Positron annihilation c-ray spectra in the annihilation time region of 28.8–340 ns for uncapped and capped films are shown in Fig. 4. The spectra are almost the same for all the capped films (Fig. 4(a)). Careful inspection enables us to find that the intensity of 3-c annihilation is higher for a film with larger pores as expected. However, in
Fig. 5. Normalized positron annihilation c-ray spectra in different annihilation time regions for uncapped porous silica films (a) B and (b) C obtained by lifetime–energy correlation measurement. The incident positron energy is 2 keV.
Fig. 4(b), the contribution of 3c-annihilation centered around 0.350 MeV for uncapped films C and D are much higher than those of uncapped films A and B. Further, in spite of different thermal treatments, the c-ray spectra for films B and C are essentially the same as films A and D, respectively. The c-ray spectra in different positron annihilation time regions for uncapped films B and C are plotted in Fig. 5. We do not present the spectra of uncapped films A and D because they are almost identical to B and C, respectively. At times shorter than 2.4 ns, a distinct 2c-annihilation peak around 0.511 MeV is observed for films B and C. The 3c fraction is enhanced drastically in the longer positron ages for film C (and also D, data not shown). In the time region from 48.4 to 72.3 ns, the 2c-annihilation characteristic peak is clearly observable for film B (and A, data not shown), but not for film C (and D, data not shown). The above results show that the fraction of o-Ps that escapes from films C/D is higher, indicative of their much higher pore interconnectivity/open porosity than films A/ B. A possible reason is that highly connected-micelles of the template are formed in the precursor solution after longer ageing, resulting in well-developed, interconnected open pores after removal of the template. 4. Conclusions
Fig. 4. Normalized positron annihilation c-ray spectra for (a) capped and (b) uncapped films in the annihilation time region from 28.8 to 340 ns obtained by lifetime–energy correlation measurements. The incident positron energy is 2.0 keV.
Mesoporous silica films were synthesized using a triblock copolymer as the structure director. Positron annihilation lifetime spectra were well resolved into five
218
C. He et al. / Chemical Physics 331 (2007) 213–218
components, including three long-lived o-Ps lifetimes. Microvoids and micropores were found to be 0.6 and 1.2 nm in diameter, respectively, and almost independent of pre-treatments. However, pre-treatments on the asdeposited films influenced the mesopore sizes and pore interconnectivity/open porosity of the calcined films. Fast heating resulted in slightly larger mesopores. The pores in a calcined film with TEOS vapor infiltration were observed larger than those in the film without VI. A possible reason is that TEOS molecules penetrated the silica network, reacted with some residual Si–OH groups, thereby reduced the contraction of the silica walls during their condensation. The positron annihilation lifetime–energy correlation results indicated that longer ageing of the precursor solution led to a considerable increment in the pore interconnectivity/open porosity in the mesoporous silica films. Acknowledgements This work was supported by the New Energy and Industrial Technology Development Organization (NEDO) of Japan. One of the authors (C. He) would like to thank the Japan Society for the Promotion of Science (JSPS) for the financial support. Drs. K. Ito and T. Oka are appreciated for their assistance of experiments. References [1] M.E. Davis, Nature 417 (2002) 813. [2] P. Yang, T. Deng, D. Zhao, P. Feng, D. Pine, B.F. Chmelka, G.M. Whitesides, G.D. Stucky, Science 282 (1998) 2244. [3] J.Y. Ying, C.P. Mehnert, M.S. Wong, Angew. Chem., Int. Ed. 38 (1999) 56. [4] X. He, D. Antonelli, Angew. Chem., Int. Ed. 41 (2002) 214. [5] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. [6] K. Maex, M.R. Baklanov, D. Shamiryan, F. Iacopi, S.H. Brongersma, Z.S. Yanovitskaya, J. Appl. Phys. 39 (2003) 8793. [7] D.W. Gidley, W.E. Frieze, T.L. Dull, J. Sun, A.F. Yee, C.V. Nguyen, D.Y. Yoon, Appl. Phys. Lett. 76 (2000) 1282. [8] J. Xu, J. Moxom, S. Yang, R. Suzuki, T. Ohdaira, Chem. Phys. Lett. 364 (2002) 309. [9] M.P. Petkov, M.H. Weber, K.G. Lynn, K.P. Rodbell, S.A. Cohen, J. Appl. Phys. 86 (1999) 3104. [10] Y. Kobayashi, W. Zheng, T.B. Chang, K. Hirata, R. Suzuki, T. Ohdaira, K. Ito, J. Appl. Phys. 91 (2002) 1704. [11] D.W. Gidley, W.E. Frieze, T.L. Dull, A.F. Yee, E.T. Ryan, H.M. Ho, Phys. Rev. B 60 (8) (1999) R5157. [12] T.L. Dull, W.E. Frieze, D.W. Gidley, J. Sun, A.F. Yee, J. Phys. Chem. 105 (2001) 4657. [13] A. van Veen, R. Escobar, H. Schut, S.W.H. Eijt, C.V. Falub, A.R. Balkenende, F.K. de Theije, Mater. Sci. Eng. B 102 (2003) 2.
[14] R. Zaleski, J. Wawryszczuk, A. Borowka, J. Goworek, T. Goworek, Micropor. Mesopor. Mater. 62 (2003) 47. [15] S.J. Tao, J. Chem. Phys. 56 (1972) 5499. [16] M. Eldrup, D. Lightbody, J.N. Sherwood, Chem. Phys. 63 (1981) 51. [17] H. Nakanishi, S.J. Wang, Y.C. Jean, in: S.C. Sharma (Ed.), Positron Annihilation Studies of Fluids, World Scientific, Singapore, 1988, pp. 292–298. [18] V.P. Shantarovich, J. Radioanal. Nucl. Chem. 210 (1996) 357. [19] T. Goworek, K. Ciesielski, B. Jasin´ska, J. Wawryszczuk, Chem. Phys. 230 (1998) 305. [20] T. Goworek, K. Ciesielski, B. Jasin´ska, J. Wawryszczuk, Radiat. Phys. Chem. 58 (2000) 719. [21] K. Ito, H. Nakanishi, Y. Ujihira, J. Phys. Chem. B 103 (1999) 4555. [22] D. Zhao, Q. Huo, J. Feng, B.F. Chmella, G.D. Stucky, J. Am. Chem. Soc. 120 (1998) 6024. [23] C. He, M. Muramatsu, T. Ohdaira, N. Oshima, A. Kinomura, R. Suzuki, Y. Kobayashi, Radiat. Phys. Chem. (2006) in press, doi:10.1016/j.radphyschem.2006.03.036. [24] C. Yu, J. Fan, B. Tian, G.D. Stucky, D. Zhao, J. Phys. Chem. B 107 (2003) 13368. [25] X. Li, J.C. Birnbaum, R.E. Williford, G.E. Fryxell, C.A. Coyle, G.C. Dunham, S. Baskaran, Chem. Commun. (2003) 2054. [26] Y. Lu, R. Ganguli, C.A. Drewien, M.T. Anderson, J.C. Brinker, W. Gong, Y. Guo, H. Soyez, B. Dunn, M.H. Huang, J.I. Zink, Nature 389 (1997) 364. [27] R. Suzuki, T. Ohdaira, Y. Kobayashi, K. Ito, Y. Shioya, T. Ishimaru, Raidat. Phys. Chem. 68 (2003) 339. [28] C. He, E. Hamada, N. Djourelov, T. Suzuki, H. Kobayashi, K. Kondo, Y. Ito, Nucl. Instrum. Methods Phys. Res., Sect. B 211 (2003) 571. [29] R. Suzuki, T. Ohdaira, A. Uedono, Y. Kobayashi, Appl. Surf. Sci. 194 (2002) 89. [30] R. Suzuki, T. Ohdaira, Y. Kobayashi, K. Ito, R.S. Ru, Y. Shioya, H. Ichikawa, H. Hosomi, K. Ishikiriyama, H. Shirataki, S. Matsuno, J. Xu, Mater. Sci. Forum 445–446 (2004) 224. [31] C. He, T. Suzuki, V.P. Shantarovich, K. Kondo, Y. Ito, Chem. Phys. 303 (2004) 219. [32] S.W. Provencher, Comput. Phys. Commun. 27 (1982) 213. [33] R.B. Gregory, Nucl. Instrum. Methods Phys. Res., Sect. A 302 (1991) 496. [34] C. He, M. Muramatsu, T. Ohdaira, A. Kinomura, R. Suzuki, K. Ito, Y. Kobayashi, Appl. Surf. Sci. 252 (2006) 3221. [35] P. Kirkegaard, M. Eldrup, O.E. Mogensen, N.J. Pedersen, Comput. Phys. Commun. 23 (1981) 307. [36] T. Ohdaira, R. Suzuki, H. Shirataki, S. Matsuno, Mater. Sci. Forum 445–446 (2004) 334. [37] J.H. Yim, Y.Y. Lyu, H.D. Jeong, S.A. Song, I.S. Hwang, J.H. Lee, A.K. Mah, S.C. Chang, J.G. Park, Y.F. Hu, J.N. Sun, D.W. Gidley, Adv. Funct. Mater. 13 (2003) 382. [38] S. Li, J. Sun, Z. Li, H. Peng, D.W. Gidley, E. Todd Ryan, Y. Yan, J. Phys. Chem. B 108 (2004) 11689. [39] L. Pei, K.I. Kurumada, M. Tanigaki, M. Hori, K. Susa, J. Mater. Sci. 39 (2004) 663. [40] M. Ogawa, N. Masukawa, Micropor. Mesopor. Mater. 38 (2000) 35. [41] N. Nishiyama, S. Tanaka, Y. Egashira, Y. Oku, K. Ueyama, Chem. Mater. 15 (2003) 1006. [42] S. Tanaka, N. Nishiyama, Y. Egashira, Y. Oku, K. Ueyama, Micropor. Mesopor. Mater. 63 (2003) 105.