H. Chon, S.-K. Ihm and Y.S. Uh (Editors) Progress in Zeolite and Microporous Materials
Studies in Surface Science and Catalysis, Vol. 105 9 1997 Elsevier Science B.V. All rights reserved.
93
Synthesis of titanium-containing mesoporous molecular sieves with a cubic structure K. A. Koyano and T. Tatsumi Engineering Research Institute, Faculty of Engineering, The University of Tokyo, Yayoi, Tokyo 113, Japan 1. INTRODUCTION Recently a new family of mesoporous molecular sieves named M41S was discovered by the researchers at Mobil [1, 2]. The M41S family is classified into several members: MCM-41 (hexagonal), MCM-48 (cubic) and other species. These materials have uniform pore, and the pore size are able to be tailored in the range 16/~ to 100/~ through the choice of surfactant, auxiliary chemicals and reaction conditions. Because of the favorable uniformity and size of the pore, synthesis and utilization of mesoporous materials have been investigated by numbers of researchers. Some circumstantial reports about the synthesis have been made [3-5]. Adsorption property and catalytic activity of silica-based and Al-containing MCM-41 have also been reported [6, 7]. Ti-, V-, B-, Fe- and Mn-substituted MCM-41 and Ti-substituted hexagonal mesoporous silica (Ti-HMS) have also been synthesized [8-15]. These Ti- and V-substituted mesoporous molecular sieves pioneered the potential to oxidize bulky molecules which cannot enter into the micropores of zeolites such as TS-1, TS-2 and Ti-beta. MCM-48, characterized by a three-dimensional channel system, may have several advantages over MCM-41 with a one-dimensional channel system when applied to catalytic reactions: for instance, the three-dimensional pore system should be more resistant to blockage by extraneous materials than the one-dimensional pore system. Here we report the synthesis of Ti-MCM-48 and its use as a catalyst for epoxidation of alkenes and unsaturated alcohols. The effects of gel composition and the gel preparation method on the structure of mesoporous materials are also reported. 2. EXPERIMENTAL
Mesoporous materials were synthesized under hydrothermal conditions at 373 K in a static Teflon bottle for 10 days. The procedures of gel preparation were as follows. For the preparation of pure silica mesoporous materials, an aqueous solution of cetyltrimethylammonium chloride / hydroxide (CTMACI/OH, CI/OH - 70/30) was added dropwise to tetraethyl orthosilicate (TEOS) under vigorous stirring at 278 K. After stirred for 1 h, the mixture was heated at 358 K for 4 h to remove the ethanol produced in the hydrolysis of TEOS. For the synthesis of Ti-containing mesoporous materials, two types of hydrolysis method were employed. Ti-MCM-48(1) was prepared by a one-stage hydrolysis method: TEOS and tetrabutyl orthotitanate (TBOT) were hydrolyzed simultaneously after being mixed for 30 minutes at 298 K. Ti-MCM-48(2) was prepared by a two-stage method : To a 44% solution
94 of TEOS in propan-2-ol, CTMAOH in methanol and water (water / TEOS (molar) = 2) were added to partly hydrolyze TEOS at 278 K. After 1 h, a propan-2-ol solution of TBOT was added to this resultant mixture very slowly under vigorous stirring. The.mixture was then stirred for 1 h, when the aqueous solution of CTMAC1 was finally added. When water-glass (SiO 2 = 28 - 30%, Na20 = 9 - 10%) was used as the Si source, sulfuric acid was added to the mixture in order to adjust the pH to 11.6 and CTMAC1 was used as the template. The molar compositions of the gels subjected to hydrothermal synthesis (373 K, 10 days) were as follows : SiO 2. xTiO 2 9 yNa20 9 CTMA 9 zH20, where 0 _< x _<0.02, y = 0 (TEOS) or 0.29 0.35 (water-glass), and 46.5 _
3. RESULTS AND DISCUSSION 3.1. Time course of hydrothermal synthesis Before we attempt to introduce Ti into the framework of MCM-48, we investigated the optimal conditions for the synthesis of pure silica MCM-48. Since it was found that the incorporation of titanium into zeolite frameworks was negatively affected by the presence of alkali cations, TEOS was used as the Si source. In the preparation of MCM-48, we conducted the hydrothermal synthesis in a Teflon
(211)
(420) .,..~ r~
(c) (B)
(A) 2
4
6
20 / degree
8
10
Figure 1. Change of XRD pattern with time in hydrothermal synthesis of MCM-48. (A) 2.5 days, (B) 6 days, (C) 10 days.
95 bottle at 373 K after keeping the mother liquids at 358 K for 4 hours to remove alcohols. The composition of reaction mixture was TEOS / CTMA / H20 = 1 / 1 / 46.5. As shown in Figure 1, mesoporous structure was formed after 2.5-day hydrothermal treatment. The typical pattern of the cubic structure with one strong peak and 7 weak ones in the 20 range of 2-7 ~ distinctly appeared after 10 days. This finding indicates that 10-day hydrothermal treatment is required to synthesize MCM-48 with the cubic structure, using TEOS as the Si source under our conditions. It is to be noted for the synthesis of MCM-41 under the same hydrothermal conditions except for the composition of reaction mixture ( TEOS / dodecyltrimethylammonium (DTMA) / H20 = 0.75 / 1 / 84 ), the formation of hexagonal structure also required 10 days while mesoporous structure was formed before the hydrothermal treatment. 3.2. Effect of surfactant concentration
Vartuli et al. have reported that the surfactant/silica molar ratio (Surf/Si) is a critical variable in the formation of M41S materials; cubic mesoporous material MCM-48 was formed when the Surf/Si ratio was adjusted to 1-1.5 [5]. We attempted the synthesis of mesoporous materials by varying the concentration of the surfactant with the Surf/Si ratio kept constant at 1.0. At 20.0 wt% of CTMAC1/OH (70/30) (H20/CTMA = 69.8), MCM-41 was exclusively obtained. The structure was transformed from hexagonal into cubic with increasing concentration of the surfactant (Table 1), in good agreement with the effect of concentration on the formation of a liquid-crystal phase for the CTMA system [16]; the cubic phase is favored over the hexagonal phase at higher concentrations for the chloride containing system. Obviously, the structure of the products was affected not only by Surf/Si but also by the surfactant concentration. Using TBOT as the Ti source, Ti-MCM-48-(1) and -(2) molecular sieves were successfully obtained from the gels with the following molar compositions: SiO 2 9 0.01TiO 2 9 0.7CTMACI 9 0.3CTMAOH 9 46.5I-I20. The increase in the Ti content to Si/Ti = 50 (two stage hydrolysis method) resulted in the formation of an ill-defined mesoporous material, which showed only one strong intensity peak at ca. 20 = 2 o. Table 1 Effect of concentration of surfactant and Si source on the structure of products Si source a
Si/Ti b
TEOS TEOS TEOS TEOS
0 0 100 50
H20/CTMA 69.8 c 46.5 c 46.5 c 46.Y
structure hexagonal cubic cubic ill-defined mesoporous
aTEOS: tetraethyl orthosilicate, Si/CTMA (cetyltrimethylammonium) = 1.0. bTi source : tetrabutyl orthotitanate, c CTMAC1/OH = 70/30. Hydrothermal synthesis conditions: 373 K; 10 days; under static condition. 3.3. Effect of Si source
Using colloidal silica as a Si source instead of TEOS, MCM-48 was obtained; however, the crystallinity was slightly lower than the product formed from TEOS. (Figure 2) When water-glass was used as the Si source, MCM-41 was obtained from the gels with the same Surf/Si and HzO/Surf ratios from which MCM-48 was obtained. There is an important
96
,i..o
(c) O
/
/
(A) !
0
2
4 6 20 ! degree
8
10
Figure 2. XRD patterns of as-synthesized mesoporous materials prepared from various Si sources. (A) TEOS, (B) colloidal silica, (C) sodium silicate.
difference between the Si sources; water-glass giving MCM-41 contains a large amount of Na, in contrast to TEOS and colloidal silica which give MCM-48. When a small amount of Na § (Na/Si molar ratio = 0.15) was added to TEOS together with the surfactant, the ill-defined mesostructure was formed. Even when Na § was added to the gel prepared from TEOS immediately before the hydrothermal treatment, we obtained poorly crystalline material. It has been shown that the addition of Na § dramatically changes the phase diagram of aqueous CTMA [ 17]. Thus it is conceivable that the presence of Na seriously affected the structure of the products. The details will be reported elsewhere. 3.4. Effect of pH
The pH is one of the important factors in the synthesis of all the kind of zeolites. We investigated the effect of pH by varying the ratio of anion of surfactant (CTMAC1/OH 70/30, 80/20, 90/10). As shown in Figure 3, the cubic structure was formed at the CTMAC1/OH = 70/30. With decreasing pH, the product structure was transformed from cubic into ill-defined
d .,..~ r~
(c)
O
(B) ~ ,-"-7~--4
,,
,(A)
6
8
20 / degree
Figure 3. XRD pattems of as-synthesized mesoporous materials prepared by variety of CTMACI/OH ratios. (A) 90/10, (B) 80/20, (C) 70/30.
97 mesostructure and finally amorphous. No solid products were obtained at the pH higher than 12. These results lead to the conclusion that the appropriate pH for the synthesis of MCM-48 is the range of 11 - 12.
3.5. Ti-eontaining MCM-4$ Titanium-containing mesoporous material with the cubic structure, Ti-MCM-48, has been synthesized by the addition of tetrabutyl orthotitanate as a Ti source. As shown in Figure 4, Ti-MCM-48(1) exhibited a very strong peak at d - 35.6 A and medium or weak peaks a t d = 31.0, 23.4, 21.8, 19.5, 18.6, 17.9 and 17.2/~. These eight peaks were indexed on a cubic unit cell with a - 87.2/~. The nitrogen physisorption isotherm was characteristic of mesoporous materials with uniform pore size. The average pore diameter determined by the Dolimore-Heal method was 25/~ and the BET surface area was ca. 1000 m2g1 . These data totally agree with those of pure silica MCM-48. We studied the pore structure of Ti-MCM-48 by high resolution transmission electron microscopy. Figure 5 exhibits a monoclinic type repeat pattern containing a 109.5 ~ angle, on the cubic [110] phase. Besides, typical regular two-dimensional patterns, such as [111] and [100], were observed, although most of the observed sheets show [111] or [110] patterns. These observations were in good agreement with those made by Mobil researchers for pure-silica MCM-48 [5]. From the XRD and TEM patterns, it has been revealed that the symmetry of Ti-MCM-48 was consistent with an I a 3 d space group, as was proposed by Schmidt et al. [18]. No differences in the TEM patterns were found between pure silica MCM-48 and Ti-MCM-48. As shown in Figure 6, solid-state 29Si MAS NMR spectrum of Ti-MCM-48 closely resemble that of pure silica MCM-48. These spectra are able to be separated into three broad peaks from Q4, Q3 and Q2 silicones by an appropriate choice of CP conditions. Though the peaks were rather broad, Q3 content (27%) roughly agreed with the N/Si molar ratio (= 0.30) observed for the as-synthesized pure silica materials [5]. This agreement in absolute numbers, .
.
.
.
.
.
.:
..
.. . . . . . .
...
..
:,::
,..
:...
:
.
.
.
.
.
r~
.~
(B___2____) (A) 0
I
I
I
I
2
4
6
8
10
-'~-- bU n m ~
.t ~ .~:, ~. - %~ . - %m ~.m-. ~' m. . . -9 ~
20 / degree
Figure 4. XRD patterns of MCM-41. (A) pure silica MCM-48, (B) Ti-MCM-48(1)
Figure 5. Transmission electron micrographs of Ti-MCM-48(1) showing the [110] projection.
98 Q3 (27%) and the N/Si ratio (30%) may indicate that Q3 is associated with the surfactant. The UV-VIS spectra for the Ti-MCM-48 samples are shown in Figure 7. The band at 220 nm has been assigned to isolated framework titanium in tetrahedral coordination [19]. TiMCM-48(1) showed a broad shoulder at ca. 270 nm attributed to extraframework titanium [20]. The anatase band, which occurs at 312 nm, seems to be superimposed on this band. In contrast, Ti-MCM-48(2) proved to be essentially free of extraframework titanium. These results indicate that the two-stage hydrolysis method is favorable for the isomorphous substitution of Ti for Si. This is interpreted in terms of the prerequisite condition to the efficient incorporation of Ti in the zeolite framework observed for TS-I: the rate of hydrolysis of Ti alkoxide should match that of Si alkoxide [21 ].
3
(B) (A) ! . . . . . . .
. . . . .
. . . .
. . . .
'-14o
ppm from TMS Figure 6. 29Si MAS NMR spectra of MCM-48. (A) pure silica MCM-48, (B) Ti-MCM-48(1).
200
!
400 600 Wavelength / nm
800
Figure 7. UV-VIS spectra of Ti-MCM-48. (A) Ti-MCM-48(2), (B) Ti-MCM-48(1).
3.6. Catalytic activity The oxidation of cyclododecene (c&/trans = 75/25) was performed using Ti-MCM-48, Ti-MCM-41, TS-1 and amorphous TiO2-SiO 2 with H202 or TBHP as the oxidant at 323 K. Ti-MCM-41 was synthesized at Surf/Si = 0.6, H20/CTMA = 75, and Si/Ti = 80 by the two stage method. As shown in Table 2, cyclododecene was epoxidized to cyclododecene oxide with either TBHP or H202 on Ti-MCM-48 and Ti-MCM-41. Preference for the epoxidation of trans-isomer was obtained for both oxidants. No other products were detected. The high activity of Ti-MCM-48 materials compared to Ti-MCM-41 may be due to the three-dimensional pore structure of Ti-MCM-48; Ti-MCM-48(2) was slightly more active than Ti-MCM-48(1). Amorphous TiO2-SiO 2 showed a very low activity in the epoxidation using H202. Compared with Ti-MCM-48(1), TiO2-SiO 2 showed a much more substantial difference in activity with H20 2 and TBHP, probably owing to severe inhibition of the catalyst by water [22]. No products were obtained with TS-1 probably as a consequence of the inability of cyclododecene to diffuse into the pores of TS-1 (5.6 x 5.3/~). Ti-MCM-48 was also applicable for the epoxidation of other alkenes. In the oxidation of pent-2-en-l-ol (10 mmol) using Ti-MCM-48(2) (100 mg) with H20 2 (60 mmol), oxidation
99 products were obtained in 17% yield (turnover number: 95); selectivities for epoxidation of the double bond and oxidation of the alcoholic group to aldehyde were 46 and 54%, respectively. The high reactivity of pent-2-en-l-ol compared to cyclododecene may be ascribed to the enhancement effect of the OH group [23]. Table 2 Epoxidation of cyclododecene on titanium-containing materials
Catalyst Ti-MCM-41(2) Ti-MCM-48(1) Ti-MCM-48(2) TS-1 TiO2-SiO 2
Si/Ti 71 80 83 79 85
Turnover Number (mol/mol-Ti) Oxidant H20 2 TBHP 0.73 (54/46)" 1.50 (59/41) ~ 1.61 (58/42)" 0 0.057 (-)"
--7.1(34/66)" --m 2.6 (30/70)"
"cis/trans ratio of products. Reaction conditions: 50 mg catalyst; 20 mmol cyclododecene (cis/trans = 75/25); 24 mmol oxidant; 323 K; 2 h. 4. CONCLUSION Titanium-containing mesoporous material with a cubic structure, Ti-MCM-48, has been successfully synthesized by means of using tetraethyl orthosilicate as a Si source under appropriate reaction conditions. Ti-MCM-48 has been found to show activity for epoxidation of bulky alkenes with H202. as well as tert-butyl hydroperoxide as an oxidant. The high catalytic activity of Ti-MCM-48 compared to Ti-MCM-41 may be related to its threedimensional channel system. 5. A C K N O W L E D G E M E N T S
The authors are grateful to Prof. S. Namba (The Nishi-Tokyo University) for measuring high resolution transmission electron micrographs, and also wish to thank Dr. S. Nakata (Chiyoda Corporation) for measuring 298i MAS NMR spectra. REFERENCES
1.
C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli and J.S. Beck, Nature, 359
(1992) 710. 2.
3. 4.
J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T-U. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins and J.L. Schlenker, J. Am. Chem. Soc., 114 (1992) 10834. J.S. Beck, J.C. Vartuli, G.J. Kennedy, C.T. Kresge, W.J. Roth and S.E. Schramm,Chem. Mater., 6 (1994) 1816. J.C. Vartuli, C.T. Kresge, M.E. Leonowicz, A.S. Chu, S.B. McCullen, I.D. Johnson and E.W. Sheppard, Chem. Mater., 6 (1994) 2070.
100
10. 11. 12. 13. 14. 15. 16. 17. 18 19.
20. 21. 22. 23.
J.C. Vartuli, K.D. Schmitt, C.T. Kresge, W.J. Roth, M.E. Leonowicz, S.B. McCullen, S.D. Hellring, J.S. Beck, J.L. Schlenker, D.H. Olson and E.W. Sheppard, Chem. Mater., 6 (1994) 2317. P.J. Branton, P.G. Hall and K.S.W. Sing, J. Chem. Soc., Chem. Commun., (1993) 1257. A. Corma, V. Fornes, H. Garcia, M.A. Miranda and M.J. Sabater, J. Am. Chem. Soc., 116 (1994) 569. A. Corma, M.T. Navarro and J.P. Pariente, J. Chem. Soc., Chem. Commun., (1994) 147. K.M. Reddy, I. Moudrakovski and A. Sayari, J. Chem. Soc., Chem. Commun., (1994) 1059. P.T. Tanev, M. Chibwe and T.I. Pinnavaia, Nature, 368 (1994) 321. T. Blasco, M.T. Navarro and J.P. Pariente, J. Catal., 156 (1995) 65. S. Gontier and A. Tuel, Zeolites, 15 (1995) 601. A. Sayari, C. Danumah and I.L. Moudrakovski, Chem. Mater.,7 (1995) 813. Z.Y. Yuan, S.Q. Liu, T.H. Chen, J.Z. Wang and H.X. Li, J. Chem. Soc., Chem. Commun., (1995) 973. D. Zhao, and D. Goldfarb, J. Chem. Soc., Chem. Commun., (1995) 875. P.K. Vinson, J.R. Bellare, H.T. Davis, W.G. Miller and L.E. Scriven, J. Colloid Interface Sci., 142 (1991) 74. C.Y. Chen, S.L. Burkett, H.S. Li and M.E. Davis, Microporous Mater., 2 (1993) 27. R. Schmidt, M. Stocker, D. Akporiaye, E.H. Torstad and A. Olsen, Microporous Mater., 5(1995) 1. A. Zecchina, G. Spoto, S. Bordiga, A. Ferrero, G. Petrini, G. Leofanti and M. Padovan, Zeolite Chemistry and Catalysis, ed. P. A. Jacobs, N.I. Jaeger, L. Kubelkova and B. Wichterlova, Elsevier, Amsterdam, 1990, p. 251. Y.L. Kim, R.L. Riley, M.J. Huq, S. Salim, A.E. Le and T.E. Mallouk, Mat. Res. Soc. Symp. Proc. 233 (1991) 145. A. Thangaraj, R. Kumar, S.P. Mirajkar and P. Ratnasamy, .1. Catal., 130 (1991) 1. R.A. Sheldon, J. Mol. Cat., 7 (1980) 107. R. Kumar, G.C.G. Pais, B. Pandey and P. Kumar, J. Chem. Soc., Chem. Commun., (1995) 1315.