J. Phys. Chem. B 2007, 111, 12119-12123
Solid-State
13C
12119
NMR Studies on Organic-Inorganic Hybrid Zeolites
Katsutoshi Yamamoto,†,‡ Yasuyuki Sakata,§ and Takashi Tatsumi*,|,⊥ Department of Applied Chemistry, Graduate School of Engineering, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan, Yokkaichi Laboratory, Mitsubishi Chemical Group Science and Technology Research Centre, Inc., Yokkaichi 510-8530, Japan, and DiVision of Materials Science & Engineering, Graduate School of Engineering, Yokohama National UniVersity, 79-5 Tokiwadai, Hodogaya-ku, Yokohama 240-8501, Japan ReceiVed: July 30, 2007
A novel type of organic-inorganic hybrid zeolite with organic lattice (ZOL) is studied in detail by solidstate 13C magic angle spinning nuclear magnetic resonance (MAS NMR). The 13C MAS NMR measurements employing several pulse sequences quantitatively demonstrate that methylene groups are really incorporated in the framework, although they are partially cleaved into methyl groups. The organic species in ZOL materials are open for adsorbates, which is evidenced by the 13C MAS NMR measurements for an n-hexane-adsorbing ZOL material. This finding strongly suggests that organic moieties are incorporated as a zeolite framework, indicating that ZOL is not a physical mixture of a carbon-containing amorphous aggregate and a conventional zeolite but a true organic-inorganic hybrid zeolite.
Introduction The organic-inorganic hybridization of silica-based porous materials has been attracting much attention because it may provide an inorganic matrix with new functions and properties. Shea et al. synthesized an organic-inorganic hybrid porous silsesquioxane from an organosilane having a bulky bridging organic group between two silicon atoms.1 Bulky organic groups such as phenylene and biphenylene groups act as a spacer to open a micropore in a silica matrix. A lot of porous silsesquioxane materials with various organic groups were successfully synthesized by the sol-gel method, and later similar materials were also synthesized by Corriu et al.2 In 1999, similar types of organosilanes having bridging organic groups were employed by several research groups to synthesize hexagonally ordered mesoporous silica materials templated by surfactant micelles.3-5 In these hybrid mesoporous materials, bridging organic groups such as ethylene and vinylene groups are embedded in the meso-structured pore wall. Because the insertion of bridging organic groups does not form structural defects in the pore wall, such materials can incorporate large amounts of organic groups without deteriorating the structural order compared with mesoporous silica materials modified with terminal organic groups. They have proved to have well-ordered structures3 and those framework organic groups can be further functionalized.4 After these reports, a large number of hybrid mesoporous silica materials having various bridging organic groups, structures, and characteristic properties were successfully synthesized.6-8 * To whom correspondence should be addressed. Tel: +81-45-924-5238. Fax: +81-45-924-5282. E-mail:
[email protected]. † The University of Tokyo. ‡ Current address: Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980-8577, Japan. § Mitsubishi Chemical Group Science and Technology Research Centre, Inc. || Yokohama National University. ⊥ Current address: Chemical Resources Laboratory, Tokyo Institute of Technology, Yokohama 226-8503, Japan.
In the field of zeolite hybridization, Jones et al. innovatively functionalized a *BEA-type zeolite with a terminal organic group for the application to a shape-selective acid catalyst.9 This approach using an organosilane with a pendant organic group as a part of silicon source succeeded in adding new functions to a microporous inorganic matrix, although the presence of terminal organic groups inevitably gives structural defects. In this regard, the strategy of methylene-substitution for a lattice oxygen atom could be favorable to synthesizing a zeolite material having a large amount of organic groups. From this standpoint, we have synthesized a novel inorganic-organic hybrid zeolite (ZOL), in which a methylene group is incorporated as a lattice.10 This hybrid zeolite material is synthesized from an organosilane with a bridging methylene group between two silicon atoms (Si-CH2-Si) to be substituted for a siloxane bridge (Si-O-Si). In the course of this study, however, it was revealed that the Si-CH2-Si linkage can be cleaved under the hydrothermal synthesis conditions. In this work, ZOL materials are examined in detail by 13C magic angle spinning nuclear magnetic resonance (MAS NMR) in order to quantitatively prove the presence of a methylene framework in this new hybrid zeolite material. Experimental Section Materials. Two types of ZOL materials, an LTA-type aluminosilicate zeolite ZOL-A and an MFI-type pure silica zeolite ZOL-2, were used in this study. These ZOL materials were synthesized by using bis(triethoxysilyl)methane (BTESM, Azmax), having a bridging methylene group between two silicon atoms, as a silicon source. As a control, an LTA-type zeolite modified with methyl groups, designated as Me-A, was synthesized by employing the equimolar mixture of TEOS (TCI) and methyltriethoxysilane (MTES, TCI) as a silicon source. The detailed synthesis procedures for these materials are described in previous papers.10 As evidenced by the XRD patterns in the precedent paper,10c both ZOL materials have well-crystallized zeolitic structures.
10.1021/jp076046x CCC: $37.00 © 2007 American Chemical Society Published on Web 10/03/2007
12120 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Yamamoto et al. decay for Me-A (Figure 2a) linearly decreases in accordance with the fact that this material has only one kind of carbon species. In contrast, the curve for ZOL-A (Figure 2b) consists of two straight lines with different slopes, suggesting that it contains two kinds of carbon species having different T1c values. One with a shorter T1c decay must be the carbon atom of a methyl group judging from the similarly short T1c of MeA, and the other with a longer relaxation time would be that of a methylene group. Similarly, the plot for another ZOL material ZOL-2 (Figure 2c) also indicates the presence of both methyl and methylene groups in this material. These findings suggest that methylene groups in the starting reagent actually remain in ZOL materials, although they are partially cleaved to methyl groups during the hydrothermal synthesis. Even in ZOL-A(0.2), which is crystallized for a time as short as 7 days from the mixture of 20% BTESM and 80% TEOS (as Si molar ratio) as silicon sources, the cleavage of methylene species are confirmed by the presence of methyl species. The proportions of methyl and methylene groups (MCH3/MCH2) in the samples can be estimated from the curve fitting simulation according to the following equation.
Figure 1. (A) 29Si DD/MAS and (B) 13C CP/MAS NMR spectra of (a) ZOL-A and (b) ZOL-2. Chemical shifts are given in ppm from TMS. The number of scans is shown on each spectrum. 13C MAS NMR Measurements. 13C MAS NMR spectra were collected at room temperature with Chemagnetics CMX400 and Varian NMR SYSTEMS 400WB spectrometers operating at 100.563 and 100.549 MHz, respectively. A Chemagnetics MAS probe with a 7.5 mm (o.d.) rotor was used, and the sample spinning rate was 4 kHz. The π/2 pulse width was 5.0 µs. The 13C dipolar decoupling (DD)/MAS NMR spectrum was obtained with the pulse delay of 120 s and the strength of proton dipolar decoupling field of 50 kHz. The cross polarization (CP)/MAS measurements were made with the contact time of 1 ms. Dipolar dephasing measurements were done with a delay time of 100 µs. The Goldman-Shen pulse sequence11 is applied to the measurement for an n-hexane-adsorbing ZOL-A sample, which is prepared by treating Ca2+-exchanged ZOL-A with n-hexane vapor at 298 K overnight.
Results and Discussion Figure 1 exhibits the 29Si DD/MAS and 13C CP/MAS NMR spectra of ZOL-A and ZOL-2. In the 29Si NMR spectra, resonance peaks attributable to organically modified silicon species are observed from around -60 to -70 ppm. For 13C NMR spectra, only one resonance peak is observed at around 0 ppm for both samples. This peak should be assigned to carbon species directly bonded to a silicon atom, which indicates the successful incorporation of organically functionalized silicon species into the product as well as the absence of organic structure-directing agents in these materials. Our previous papers10 have revealed that the Si-CH2-Si linkage can be cleaved into Si-OH and CH3-Si by the attack of hydroxyl anions, which implies that not only methylene groups but also methyl groups could be involved in these ZOL materials. However, on the basis of usual CP/MAS NMR spectra, methylene and methyl groups cannot be differentiated because their resonance peaks appear so close. Therefore, in the following experiments several pulse sequences are employed to quantitatively evaluate the amount of framework methylene groups. Figure 2 exhibits the semilogarithm plots of the 13C spinlattice relaxation (T1c) decay for carbon species in ZOL-A, ZOL-2, and Me-A using Torchia’s pulse sequence.12 The T1c
I ) MCH3exp(-τ/T1CH3) + MCH2exp(-τ/T1CH2) where I is the relative intensities of resonance peaks, T1CH3 is the relaxation time for carbon species in the methyl group, and T1CH2 is the relaxation time for carbon species in methylene group. As a control, an amorphous aggregate is prepared by hydrolyzing a mixture of 67% BTESM and 33% MTES in acid conditions. The T1c decay of this standard material also consists of two lines having different slopes (Figure 2e) to prove the presence of methylene and methyl species. The curve-fitting simulation based on the equation above estimates the CH2/CH3 ratio at 71:29, which demonstrates the validity of this method for evaluating the ratio of methylene and methyl species. Table 1 exhibits the 13C relaxation time and the molar ratio of methyl and methylene species in ZOL materials together with their hydrothermal synthesis conditions. According to the curve fitting, the percentage of methylene species to the total organic species contained in ZOL-A is estimated at 61%. Because methylene groups could be cleaved to methyl groups with increasing synthesis time, a higher percentage of methylene species (71%) remains in ZOL-A(0.2). On the other hand, ZOL-2 has 45% of methylene species, which is smaller than that of ZOL-A, presumably because it is crystallized at higher synthesis temperature. The relaxation of carbon species in methyl groups in ZOL-A completes within 15 s (≈5 × T1c). Therefore, by the measurement using Torchia’s pulse sequence with the relaxation delay time of 15 s, only the resonance peak attributable to methylene species can be obtained. Figure 3a shows thus obtained spectrum of ZOL-A exhibiting a resonance peak at around 1 ppm. On the other hand, the resonance peak derived only from carbon species in methyl groups is obtained by the measurement using the dipolar dephasing pulse sequence. Empirically, due to the dipole-dipole interaction between 13C and 1H, resonance peaks of methylene and methine groups almost completely diminish within 100 µs. Therefore, the 13C NMR spectrum using the dipolar dephasing pulse sequence with the delay time of 100 µs should exhibit a resonance peak only for a methyl group. In such a measurement for ZOL-A (Figure 3b), the resonance peak of methyl species is observed at around -2 ppm, being close to but appreciably different from that of methylene species.
13C
NMR of Organic-Inorganic Hybrid Zeolites
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12121
Figure 2. 13C T1c relaxation decay curves of (a) Me-A, (b) ZOL-A, (c) ZOL-2, (d) ZOL-A(0.2), and (e) an amorphous aggregate prepared from a mixture of 67% BTESM and 33% MTES.
TABLE 1: Quantitative Analyses of Carbon Species in ZOL Materials through the Curve-Fitting Simulation of 13C T1c Relaxation Decay hydrothermal conditions
ZOL-A ZOL-A(0.2) ZOL-2 a
C relaxation time (s)
molar ratio (%)
silicon sourcea
temp (°C)
time (days)
T1CH3
50% MTES + 50% TEOS 100% BTESM 20% BTESM + 80% TEOS 100% BTESM
100
3
3.7
100 100
22 7
2.8 0.9
23.4 10.8
39 29
61 71
140
20
4.0
59.3
55
45
material Me-A
13
T1CH2
MCH3
MCH2
100
As a silicon molar ratio.
The 13C DD/MAS NMR spectrum of ZOL-A (Figure 4), whose peak area should reflect the abundance ratio of all the carbon species involved, shows one broad peak at around -1 ppm. This broad resonance peak would be composed of the peaks for methylene and methyl species having different chemical shifts. Therefore, through the deconvolution of the spectrum based on the resonance peaks assigned as methylene and methyl species as described above, the molar ratio of Si-
CH2-Si/CH3-Si is estimated at 58:42, which is similar to the value obtained above through the curve-fitting simulation for the T1c relaxation decay (Table 1). This method would be more precise than the curve fitting of T1c decay described before. However, for a ZOL material having a long relaxation time, such as ZOL-2, it might be practically difficult to use this method because the signal acquired after the complete relaxation of a certain species
12122 J. Phys. Chem. B, Vol. 111, No. 42, 2007
Figure 3. 13C MAS NMR spectra of ZOL-A for (a) a methylene group using Torchia’s pulse sequence with the relaxation delay time of 15 s and (b) a methyl group using the dipolar dephasing pulse sequence with the delay time of 100 µs. Chemical shifts are given in ppm from TMS.
Figure 4. 13C DD/MAS NMR spectrum of ZOL-A with the pulse delay of 120 s and 4000 scans (thick line) and a simulated spectrum composed of the spectra for methylene and methyl species (thin lines). Chemical shifts are given in ppm from TMS.
becomes too weak. Considering the similarity in the estimation results for ZOL-A, the curve-fitting simulation based on the T1c decay would satisfactorily quantify the framework methylene species in ZOL materials having long relaxation time. The 13C CP/MAS NMR spectra of the n-hexane-adsorbing ZOL-A sample were acquired using the Goldman-Shen pulse sequence11 with various spin-diffusion times. In this pulse sequence, all the 1H nuclei are initially magnetized and relaxed. After an appropriate relaxation time, mobile 1H in adsorbed n-hexane molecules, which has longer relaxation time, remains magnetized selectively. Then the magnetization is spin-diffused to rigid 1H involved in organic species of ZOL-A during the spin-diffusion time. The closer the distance between two 1H is, the faster the magnetization diffuses. Then the magnetization is transferred to 13C via cross-polarization to detect the resonance. Figure 5 exhibits relative peak areas of thus obtained resonance peak for framework organic species against that for adsorbed n-hexane after several spin-diffusion time periods. This figure clearly shows that the relative peak area is quickly increased with increasing spin-diffusion time. Figure 6A
Yamamoto et al.
Figure 5. Relative peak area of framework organic species against adsorbed n-hexane based on 13C CP/MAS NMR spectra using the Goldman-Shen pulse sequence for (a) ZOL-A and (b) an amorphous aggregate prepared from BTESM.
Figure 6. (a) Usual 13C CP/MAS NMR spectra (4096 scans) and (b) 13 C CP/MAS spectra using the Goldman-Shen pulse sequence (10 000 scans) with the spin-diffusion time of 50 ms for n-hexane-adsorbing (A) ZOL-A and (B) an amorphous aggregate from BTESM. Numbers indicated above peaks represent relative peak area, where the total area for n-hexane peaks is normalized to 100.
compares the spectrum having the spin-diffusion time of 50 ms with the usual CP/MAS spectrum, in which all the carbon species are magnetized. In these two spectra, the ratios of the peak area attributable to the framework organic species to that of n-hexane are almost the same. Consequently, these NMR measurements indicate that all the organic moieties in ZOL-A are accessible for n-hexane so that the magnetization of 1H can be spin-diffused and quickly equilibrated. As a control, an amorphous aggregate having similar organic content is prepared by hydrolyzing and polymerizing BTESM in alkaline conditions at room temperature, and the same series of NMR measurements is conducted after the n-hexane adsorption under the same conditions. In this case, the relative peak area of the framework organic species is quite small (Figure 5b) compared with the case of ZOL-A, and even after 50 ms the peak area ratio is also much smaller than that observed in the usual CP/MAS spectrum (Figure 6B). In the amorphous aggregate prepared from BTESM, some organic groups may be buried inside the aggregate, being inaccessible to n-hexane. Therefore, the spin-diffusion from adsorbed n-hexane to such
13C
NMR of Organic-Inorganic Hybrid Zeolites
buried carbon species is impossible or at least very slow. As a result, the peak area cannot reach to the value of the usual CP/ MAS spectrum. These experimental facts reasonably suggest that organic moieties in ZOL-A are exposed to the surface. That is, ZOL-A is not a physical mixture of a carbon-containing amorphous aggregate and conventional zeolite A but a genuine organicinorganic hybrid zeolite material having framework organic groups that are accessible to adsorbate molecules. Conclusion The detailed 13C MAS NMR studies have revealed that a methylene framework is actually contained in ZOL materials, although they are partly hydrolyzed to give rise to methyl species. The proportion of methylene and methyl species involved in ZOL-A is quantitatively checked by two methods, and both methods show a similar result that Si-CH2-Si/CH3Si is about 60:40. The measurements using the Goldman-Shen pulse sequence demonstrate a strong interaction between organic moieties involved in ZOL-A and n-hexane molecules adsorbed on ZOL-A to suggest that organic moieties in ZOL materials are open and easily accessible for adsorbate molecules. This finding strongly indicates that organic moieties are not buried in amorphous aggregate but involved as an open framework of ZOL material, supporting that ZOL material is a real organicinorganic hybrid zeolite material. Acknowledgment. The authors thank Mitsubishi Chemical Corp. for financial and technical support. This work was partly supported by Core Research for Evolutional Science and Technology (CREST) of JST Corp. to T.T.
J. Phys. Chem. B, Vol. 111, No. 42, 2007 12123 References and Notes (1) (a) Shea, K. J.; Loy, D. A.; Webster, O. W. Chem. Mater. 1989, 1, 572. (b) Loy, D. A.; Shea, K. J. Chem. ReV. 1995, 95, 1431. (c) Shea, K. J.; Loy, D. A. Chem. Mater. 2001, 13, 3306. (2) Corriu, R. J. P.; Moreau, J. J. E.; Thepot, P.; Man, M. W. C. Chem. Mater. 1992, 4, 1217. (3) Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc. 1999, 121, 9611. (4) Melde, B. J.; Holland, B. T.; Blanford, C. F.; Stein, A. Chem. Mater. 1999, 11, 3302. (5) Asefa, T.; MacLachlan, M. J.; Coombs, N.; Ozin, G. A. Nature 1999, 402, 867. (6) Lu, Y.; Fan, H.; Doke, N.; Loy, D. A.; Assink, R. A.; LaVan, D. A.; Brinker, C. J. J. Am. Chem. Soc. 2000, 122, 5258. (7) (a) Guan, S.; Inagaki, S.; Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc. 2000, 122, 5660. (b) Inagaki, S.; Guan, S.; Ohsuna, T.; Terasaki, O. Nature 2002, 416, 304. (c) Yang, Q. H.; Kapoor, M. P.; Inagaki, S. J. Am. Chem. Soc. 2002, 124, 9694. (d) Goto, Y.; Inagaki, S. Chem. Commum. 2002, 2410. (8) (a) Ishii, C. Y.; Asefa, T.; Coombs, N.; MacLachlan, M. J.; Ozin, G. A. Chem. Commum. 1999, 2539. (b) Asefa, T.; MacLachlan, M. J.; Grondey, H.; Coombs, N.; Ozin, G. A. Angew. Chem., Int. Ed. 2000, 39, 1808. (c) Asefa, T.; Ishii, C. Y.; MacLachlan, M. J.; Ozin, G. A. J. Mater. Chem. 2000, 10, 1751. (d) Kuroki, M.; Asefa, T.; Whitnal, W.; Kruk, M.; Yoshina-Ishii, C.; Jaroniec, M.; Ozin, G. A. J. Am. Chem. Soc. 2002, 124, 13886. (9) (a) Jones, C. W.; Tsuji, K.; Davis, M. E. Nature 1998, 393, 52. (b) Tsuji, K.; Jones, C. W.; Davis, M. E. Microporous Mesoporous Mater. 1999, 29, 339. (c) Jones, C. W.; Tsuji, K.; Davis, M. E. Microporous Mesoporous Mater. 1999, 33, 223. (d) Jones, C. W.; Tsapatsis, M.; Okubo, T.; Davis, M. E. Microporous Mesoporous Mater. 2001, 42, 21. (10) (a) Yamamoto, K.; Takahashi, Y.; Tatsumi, T. Stud. Surf. Sci. Catal. 2001, 135, 299. (b) Yamamoto, K.; Sakata, Y.; Nohara, Y.; Takahashi, Y.; Tatsumi, T. Science 2003, 300, 470. (c) Yamamoto, Y.; Nohara, Y.; Domon, Y.; Takahashi, K.; Sakata, Y.; Ple´vert, J.; Tatsumi, T. Chem. Mater. 2005, 17, 3913. (11) Goldman, M.; Shen, L. Phys. ReV. 1966, 144, 321. (12) Torchia, D. A. J. Magn. Reson. 1978, 30, 613.
