Preparation of Supported Mesoporous Thin Films Concerning

A novel, efficient and facile method for the template removal from mesoporous materials. Lu Chen ... Rafal M. Grudzien , Bogna E. Grabicka , Mietek Ja...
1 downloads 0 Views 135KB Size
Langmuir 2005, 21, 1171-1174

1171

Preparation of Supported Mesoporous Thin Films Concerning Template Removal by Supercritical Fluid Extraction L. Huang,*,† C. Poh,‡ S. C. Ng,§ K. Hidajat,‡ and S. Kawi‡ Institute of Chemical and Engineering Sciences, 1 Pesek Road, Jurong Island, Singapore 627833, Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 117576, and Department of Chemistry, National University of Singapore, Singapore 117543 Received September 24, 2004. In Final Form: December 29, 2004 Thin films of silicate MCM-41 and silicate MCM-48 have been prepared on porous ceramic supports by the hydrothermal method. A comparative study of template removal has been made on supported thin films and on powder. By supercritical fluid extraction (SFE) with CH3OH-modified CO2, at least 78% of the template can be removed from as-synthesized materials at 85 °C. X-ray diffraction (XRD) observations indicate that the resulting supported thin films after SFE are structurally stable and ordered with a weak pore contraction. The advantages of SFE over calcination in template removal are presented with a series of results obtained on supported thin films and on powder by XRD and N2 adsorption-desorption.

Introduction In the processes of preparation of molecular sieve thin films, removing organic templates via calcination at high temperatures (540-600 °C) usually results in distortion or cracking of thin films,1,2 although it is most efficient. As thin films are heated, they shrink in volume due to the condensation of silanol groups in the pores to form siloxane bonds. Once the films are attached to the substrate and unable to shrink in that direction, the reduction in volume is accommodated completely by a reduction in thickness. Therefore, low temperatures are required to eliminate organic templates from thin films. Among a number of low-temperature methods for eliminating organic templates from molecular sieves,3-7 supercritical fluid extraction (SFE) is efficient and advantageous in practice.8,9 Supercritical CO2 has widely been used as an extraction fluid, because of low point (31.1 °C and 72.8 bar), low cost, and low toxicity and reactivity. CO2 SFE has found a couple of successful applications in the quantitative extraction of organic templates from mesoporous molecular sieves.7,9,10 Here we report a study on the modified CO2 SFE of cetyltrimethylammonium bromide (CTMABr) from as* To whom correspondence should be addressed. E-mail: [email protected]. † Institute of Chemical and Engineering Sciences. ‡ Department of Chemical and Biomolecular Engineering, National University of Singapore. § Department of Chemistry, National University of Singapore. (1) Brinker, C. T.; Scherer, G. W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing; Academic Press: Boston, 1990. (2) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710. (3) Whitehurst, D. D. U.S. Patent 5,143,879, 1992. (4) Tanev, P. T.; Pinnavaia, T. J. Science 1995, 267, 865. (5) Keene, M. T. J.; Denoyed, R.; Llewellyn, Ph. L. Chem. Commun. 1998, 2203. (6) Hozumi, A.; Sugimura, H.; Hiraku, K.; Kameyama, T.; Takai, O. Chem. Mater. 2000, 12, 3842. (7) Kawi, S.; Lai, M. W. AIChE J. 2002,48, 1572. (8) Camel, V.; Tambute´, A.; Caude, M. J. Chromatogr. 1993, 642, 263. (9) Grieken, R. van; Calleja, G.; Stucky, G. D.; Melero, J. A.; Garcı´a, R. A.; Iglesias, J. Langmuir 2003, 19, 3966. (10) Lu, X.-B.; Zhang, W.-H.; Xiu, J.-H.; He, R.; Chen, L.-G.; Li, X. Ind. Eng. Chem. Res. 2003, 42, 653.

made thin films of MCM-41 and MCM-48 on porous ceramic supports. This work includes relevant quantitative extraction analysis and structural characterization (Xray diffraction (XRD) and N2 adsorption-desorption) using powder samples besides an XRD study on supported thin films. To our knowledge, this contribution includes for the first time the use of SFE for the removal of surfactant templates from mesoporous thin films. Experimental Section Sodium silicate solution (25.5-28.5% SiO2, 7.5-8.5% Na2O) and CTMABr (98-101%) were purchased from Merck. Tetraethyl orthosilicate (TEOS, 98%) and tetraethylammonium hydroxide solution (20%) were obtained from Aldrich and from Sigma. Porous ceramic disks with a pore diameter of 0.5 µm were supplied by Coors Tek. Silicate MCM-41 powder was synthesized as described in ref 11. After hydrothermal reaction, the solid was filtered off, washed with deionized water until free of Br-, and dried under vacuum at room temperature. Supported silicate MCM-41 thin films were synthesized using a molar composition of gel of 1.0 TEOS/0.48 CTMABr/0.5 NaOH/60 H2O. NaOH (1.057 g) and 55 mL of deionized water were added to 8.81 g of CTMABr. After the solution had been stirred for 10 min, 11.4 mL of TEOS was poured to form a gel. The gel solution was stirred for 1 h 45 min and then dripped onto porous ceramic disks and spin coated. The disks following spin coating were transferred into a polypropylene bottle together with the gel solution. The system was heated in an oven at 100 °C for 4 days under autogenerated pressure. The resulting supported films were rinsed with deionized water and dried at room temperature under vacuum. Silicate MCM-48 powder and supported silicate MCM-48 thin films were synthesized using a molar composition of gel of 1.0 TEOS/0.65 CTMABr/0.50 NaOH/62 H2O. NaOH (1.098 g) and 56 mL of deionized water were added to 12.00 g of CTMABr. After stirring for 10 min, 11.2 mL of TEOS was poured to form a gel. The gel solution was stirred for 1.5 h and then dripped onto porous ceramic disks and spin coated. The disks following spin coating were transferred into a polypropylene bottle together with the gel solution. The system was heated in an oven at 100 °C for 4 days under autogenerated pressure. The resulting powder and supported films were washed with deionized water until free of Br- and dried at room temperature under vacuum. The disks covered with silica films were subjected to some polishing with a spatula. (11) Huang, L.; Wu, J. C.; Kawi, S. J. Mol. Catal. A 2003, 206, 371.

10.1021/la047621g CCC: $30.25 © 2005 American Chemical Society Published on Web 01/19/2005

1172

Langmuir, Vol. 21, No. 4, 2005

Letters

Table 1. Template Removal Results and Properties of Mesoporous Powder Samples SFE efficiency (%)

material as-synth MCM-41a as-synth MCM-41b as-synth MCM-48a as-synth MCM-48b cured MCM-48a

78

mean pore diam (Å)

total pore volume (cm3/g)

BET surface area (m2/g)

structure after template removal

29.2 30.5

0.81 1.29

1332 1651

26.3 23.6

0.81 0.82

1225 1411

stable stable destroyed stable stable

93 82

a Template removal by SFE under the conditions of 85 °C, 90 bar, CH3OH/CO2 ) 0.1/1.0 mL/min, and 3 h. b Template removal by calcination at 540 °C.

Figure 2. N2 adsorption-desorption isotherms of (a) assynthesized MCM-41 powder after 6 h of template SFE; (b) as-synthesized MCM-41 powder after calcination.

Figure 1. XRD patterns of (a) as-synthesized MCM-41 powder; (b) as-synthesized MCM-41 powder after 6 h of template SFE; (c) as-synthesized MCM-41 powder after calcination. Template removal was done with SFE or calcination. SFE was conducted on a Jasco SFE apparatus. In virtue of the study of SFE conditions,12 we chose 85 °C, 90 bar, 3 h, and 0.1/1.0 mL/min as the most appropriate extraction temperature, extraction pressure, extraction time, and CH3OH/CO2 ratio, respectively. Calcination was carried out with programmed heating from 23 to 540 °C at 0. 2 °C/min followed by 24 h of isothermic heating at 540 °C. The thickness of supported thin films was measured using scanning electron microscopy (SEM) on a JSM-5600 LV apparatus. The template amount was determined by thermogravimetry (TG) on a Shimadzu DTG-50 thermogravimetric analyzer to estimate the SFE efficiency. XRD was performed on a Shimadzu XRD-6000 spectrometer. N2 adsorption-desorption was measured on a Quantachrome Autosorb-1 analyzer.

Figure 3. Pore size distributions of (a) as-synthesized MCM41 powder after 6 h of template SFE; (b) as-synthesized MCM41 powder after calcination.

Results and Discussion In Table 1 are presented the template SFE efficiencies and some properties of powder samples. In the case of MCM-41, 78% of CTMABr was extracted from assynthesized powder. Figure 1 gives the XRD patterns before and after template removal on as-synthesized MCM-41 powder. The diffraction peaks after SFE were similar to those after template removal by calcination, with only slight upward shift in position. This implies that as-synthesized MCM-41 powder is stable under SFE conditions and quite ordered after template SFE in structure. N2 adsorption-desorption indicated that after SFE as-synthesized MCM-41 powder had a type IV adsorption-desorption isotherm (Figure 2) that was accompanied by a pore size distribution around 29.2 Å (Figure 3). Its capillary condensation step appeared at lower relative pressures than those at which the capillary condensation step of as-synthesized MCM-41 powder after template removal by calcinations arose. Compared to the case of as-synthesized MCM-41 powder after calcination, (12) Huang, L.; Poh, C.; Ng, S. C.; Hidajat, K.; Kawi, S. Talanta, in press.

Figure 4. XRD patterns of (a) as-synthesized MCM-48 powder; (b) as-synthesized MCM-48 powder after 3 h of template SFE; (c) cured MCM-48 powder; (d) cured MCM-48 powder after 6 h of template SFE; (e) as-synthesized MCM-48 powder after calcination.

its capillary condensation step height and mean pore diameter were small. These may be due to irregular pore filling with the remaining template in an important amount (22%). Consistently, SFE-treated MCM-41 powder had smaller total pore volumes and Brunauer-EmmettTeller (BET) surface areas than calcined MCM-41 powder. The results illustrate that SFE is effective for the template removal from MCM-41 without the structural degradation of MCM-41. In the case of MCM-48, 93% of CTMABr was extracted from as-synthesized powder. Figure 4 presents the XRD patterns before and after template removal on MCM-48 powder. The diffraction peaks disappeared after SFE on as-synthesized powder, which was indicative of the destruction of MCM-48 structure under SFE conditions.

Letters

Langmuir, Vol. 21, No. 4, 2005 1173

Figure 5. N2 adsorption-desorption isotherms of (a) cured MCM-48 powder after 6 h of template SFE; (b) as-synthesized MCM-48 powder after calcination.

Figure 7. XRD patterns (a) of an as-synthesized MCM-41 thin film on a porous ceramic support; (b) after 6 h of template SFE on (a); (c) after calcination following (b).

Figure 6. Pore size distributions of (a) cured MCM-48 powder after 6 h of template SFE; (b) as-synthesized MCM-48 powder after calcination.

Figure 8. XRD patterns (a) of an as-synthesized MCM-48 thin film on a porous ceramic support; (b) after 6 h of template SFE on (a); (c) after calcination following (b).

To improve the structural stability by strengthening crosslinking of the silicate framework, as-synthesized MCM48 powder was cured at 160 °C under vacuum overnight. Cured powder gave a template SFE efficiency of 82% and exhibited more intense diffraction peaks after SFE than those of as-synthesized powder and before SFE. The diffraction peaks were comparable to those of assynthesized powder after calcination in intensity. Meanwhile curing as-synthesized powder caused a larger upward shift of diffraction peaks in position, that is attributed to more severe pore contraction of MCM-48. These results show that cured MCM-48 powder can remain structurally stable during the SFE even though its structure shrinks markedly. According to N2 adsorptiondesorption, the adsorption-desorption isotherm of cured powder after SFE showed a capillary condensation step at lower relative pressures than those of as-synthesized powder after calcination (Figure 5), and its mean pore diameter (23.6 Å) was smaller than that of as-synthesized powder after calcination (26.3 Å) (Figure 6). This confirms that more serious pore contraction arises with cured MCM48 powder, in accordance with XRD data. After SFE, cured MCM-48 powder had a high total pore volume and a high BET surface area, that were comparable to those of assynthesized MCM-48 powder after calcination. Figures 7 and 8 show the XRD patterns before and after template removal on porous ceramic supports. In the case of MCM-41, the as-synthesized supported thin film exhibited a set of diffraction peaks at 2.12, 3.68, 4.26, and 5.58° shown in Figure 7a, which are well characteristic of MCM-41. This clearly accounts for the success in the synthesis of the MCM-41 thin film on a porous ceramic support using a CTMABr-TEOS gel by hydrothermal reaction. After SFE on this supported thin film, the

diffraction peaks slightly shifted upward in position and decreased in intensity. The presence of the spectrum in Figure 7b significantly indicates that the structure of MCM-41 thin film basically remains stable and ordered after template removal via SFE. When the supported thin film was further subjected to calcination at 540 °C for eliminating the remaining template, the diffraction peaks strongly broadened and shifted upward as seen in Figure 7c. This suggests that template removal from the supported MCM-41 thin film via high-temperature calcination is accompanied by serious pore contraction and pore uniformity diminution in the thin film structure. It is clear from the observed XRD pattern in Figure 7a that the resultant supported MCM-41 thin film displays (100), (110), (200), and (210) reflections as is the case for the MCM-41 powder. The presence of the (110) and (210) reflections as well as the (100) and (200) ones for the MCM41 thin film suggests that the mesoporous channel axis is oriented randomly rather than parallel to the plane of the film. It is of significance to compare the variations of structural properties of as-synthesized MCM-41 powder and as-synthesized supported MCM-41 thin film during the processes of template removal. In the powder form, as-synthesized MCM-41 shows weak pore contraction and almost unchanged pore uniformity, based on the fact that much the same XRD peak width and the weak XRD peak upward shift are observed, even after calcination. In the supported thin film form, as-synthesized MCM-41 presents strong pore contraction and decreased pore uniformity at high temperatures, since the XRD peaks shift upward greatly and the XRD peak width increases tremendously after calcination. It is evident that the substrate exerts a critical influence on the structural

1174

Langmuir, Vol. 21, No. 4, 2005

stability of MCM-41 thin film. It seems that the interaction between the substrate and the MCM-41 thin film is so strong that the thin film cannot retain the initial pore uniformity once the structural shrinkage takes place as the template removal proceeds at high temperatures. To avoid such structural distortion, SFE can replace calcination to remove the template at mild temperatures, as indicated above. In the case of MCM-48, the as-synthesized supported thin film exhibited a set of diffraction peaks at 2.18, 2.54, 4.06, and 4.28° as shown in Figure 8a, which are attributed to MCM-48. This proves that the MCM-48 thin film is successfully synthesized on a porous ceramic support with a conventional surfactant-silicate gel by hydrothermal reaction. Surprisingly, SFE on this supported thin film did not cause the damage of mesoporous structure, differing from the case of as-synthesized MCM-48 powder. The diffraction peaks of MCM-48 with intensities comparable to those in Figure 8a were still observed, shifting upward a little in position as seen in Figure 8b. This significantly demonstrates that the supported MCM-48 thin film is structurally stable and ordered after template removal via SFE. After the remaining template had been burned off via calcination, the diffraction peaks increased greatly in intensity and continued to shift upward in position as seen in Figure 8c. No matter how the template was removed from the as-synthesized supported MCM48 thin film, the diffraction peak width remained substantially unchanged. This fact implies that the uniform pore sizes in the structure of supported MCM-48 thin films can be retained during the template removal. In contrast to what happens to as-synthesized MCM48 powder, as-synthesized MCM-48 thin films acquire a tremendously enhanced mesoporous structural ordering on porous ceramic supports according to the XRD peak intensity. They are stabilized via attachment to the substrate so as to avoid the structural collapse under SFE conditions. The substrate undoubtedly plays a beneficial role in the formation of stable and ordered MCM-48 thin films. Apparently, the presence of the substrate not only does not result in a larger pore contraction but produces

Letters

uniform MCM-48 mesopores after template removal, in view of the XRD peak position and width. Despite the template removal by calcination at 540 °C, the obtained supported MCM-48 thin film still displays a sharp principal diffraction peak and a peak upward shift equivalent to that in the case of as-synthesized MCM-48 powder. The equivalent upward shifts suggest that the thin film and powder have equivalent pore contractions. The influence of the substrate on the MCM-48 thin film appears different from that on the MCM-41 thin film. Such a consequence may be due to the difference in the natures of interactions of the substrate with the MCM-48 thin film and with the MCM-41 thin film. The observations of cross-sections of polished supported thin films by SEM indicated that the supported thin films of MCM-41 and MCM-48 were 20 µm or so in thickness. Conclusions This paper has dealt with the preparation and structural stability of silicate MCM-41 and silicate MCM-48 thin films on porous ceramic supports with the involvement of template removal via CH3OH-modified CO2 SFE. The supported MCM-41 and MCM-48 thin films are synthesized by the hydrothermal method from TEOS using CTMABr as the template. At least 78% of the template can be eliminated from these mesoporous materials with SFE at 85 °C, which is satisfactory. SFE enables production of uniform mesopores in the structures of supported thin films with a weak structural shrinkage. By contrast, calcination at 540 °C for template removal causes the structural damage of supported thin films. Supported MCM-41 thin films lose pore uniformity with a strong pore contraction after calcination, possibly due to the strong attachment of the MCM-41 thin film to the substrate. Supported MCM-48 thin films display a more marked pore contraction after calcination than after SFE. Acknowledgment. This project was supported by ABB Lummus Global Inc. We thank Professor Frits M. Dautzenberg for valuable discussion. LA047621G