Colloids and Surfaces A: Physicochemical and Engineering Aspects 179 (2001) 133 – 138 www.elsevier.nl/locate/colsurfa
Synthesis of microporous titano-alumino-silicate ETAS-10 with different framework aluminum contents Z. Lin a, J. Rocha a,*, A. Ferreira a, M.W. Anderson b b
a Department of Chemistry, Uni6ersity of A6eiro, 3810 A6eiro, Portugal Department of Chemistry, UMIST, PO Box 88, Manchester M60 1QD, UK
Abstract Optimized synthesis conditions for obtaining highly crystalline microporous titano-alumino-silicate ETAS-10 materials with different framework aluminum contents (Al/Ti molar ratio 0.1 – 0.48) have been reported. Samples with low Al contents (Al/Ti 0.10) have been prepared using anatase and pseudoboehmite as, respectively, the Ti and Al sources. Good quality ETAS-10 materials with Al framework contents above 0.35 have been obtained using a modification of the synthesis procedure that involves drying and calcining the parent gel prior to the hydrothermal treatment. Changing the K and Na cation composition of the parent gel allows the fine-tuning of the framework Al content. It has been found that 29Si MAS NMR spectroscopy provides the best means of assessing the framework Al content of ETAS-10 materials. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Titano-alumino-silicate; ETAS-10; Microporous materials; Zeolites; Synthesis; Isomorphous substitution
1. Introduction ETS-10 is a microporous framework zeolitetype titanosilicate, possessing a three dimensional 12-ring pore system, and presenting considerable potential for being used as a catalyst, desiccant and ion-exchanger [1 – 4]. In ETS-10, the TiO6 octahedra are connected to each other by two opposite corners forming a chain which is surrounded by the corner-sharing SiO4 tetrahedra. The acidic properties of ETS-10 may be introduced by forming bridging hydroxyl groups such * Corresponding author. Tel.: +351-234-370730; fax: + 351-234-370084. E-mail address:
[email protected] (J. Rocha).
as Ti–(OH)–Si. In order to improve further these properties, aluminum may be inserted in the framework of ETS-10 in non-stoichiometric amounts, affording materials denominated ETAS10. The preparation of ETAS-10 has already been reported [5–7]. However, in our earlier work we were mainly concerned with the structural characterization of ETAS-10 and, hence, little reference was given to the synthesis procedure. Here, we wish to expound on this work and report detailed and optimized synthesis conditions for obtaining highly crystalline ETAS-10 materials with different framework aluminum contents. In particular, we describe a novel method for preparing good quality ETAS-10 samples with a high Al content.
0927-7757/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S0927-7757(00)00648-8
134
Z. Lin et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 133–138
All samples were characterized by bulk chemical analysis, powder X-ray diffraction (XRD), scanning electron microscopy (SEM), 29Si and 27Al magic-angle spinning (MAS) nuclear magnetic resonance (NMR), Raman spectroscopy, thermogravimetric analysis (TGA).
2. Experimental The hydrothermal synthesis of ETAS-10 was carried out under autogenous pressure without stirring. The typical synthesis gel was prepared as follows.
2.1. Titanium trichloride route An alkaline solution was obtained by mixing 37.88 g water, 4.84 g sodium hydroxide (Merck), 3.68 g sodium chloride (Aldrich), 2.54 g potassium chloride (Aldrich) and 4.60 g potassium fluoride (Aldrich). 1.80 g sodium aluminate (41 w/w% Na2O, 54 w/w% Al2O3, Riedel-de Haen) was added to this solution followed by 28.25 g titanium trichloride aqueous solution (1.9 M in 2.0 M HCl, Aldrich). 57.07 g sodium silicate aqueous solution (27 w/w% SiO2, 8 w/w% Na2O, BDH) was added last and the gel was thoroughly stirred. 0.6 g ETS-10 seed crystals were added to the final gel and mixed with it carefully.
2.2. Anatase route A total of 15.09 g sodium silicate aqueous solution (27 w/w% SiO2, 8 w/w% Na2O, BDH), 1.31 g potassium fluoride (Aldrich), 1.50 g potassium chloride (Aldrich), 5.22 g sodium chloride (Aldrich) and 20.09 g water were mixed to obtain a gel. 1.05 g anatase (Merck) was added into the gel followed by 0.55 g pseudoboehmite (67.01 w/w% Al2O3, Condea). In both routes, the titanium – aluminum–silicate gel was sealed in Teflon-lined autoclaves and treated at 200 – 230°C for 1 – 3 days. In the synthesis using a calcined gel precursor, the gel obtained in the titanium trichloride route was dried over night at 130°C. After being ground, the dried gel was calcined at 500°C for 1.5 h and mixed with a
basic solution (2.5 w/w% sodium hydroxide) and ETS-10 seeds. The weight ratio of basic solution to powder was 2. The mixture was then autoclaved at 200°C for 1 day. The resulting crystals were filtered using paper MN 640w from Macherey-Nagel and washed at room temperature with distilled water, and dried at 90–120°C. Powder XRD data were collected on a Rigaku diffractometer using CuKa radiation filtered by Ni. SEM images were recorded on a Hitachi S-4100 microscope. 29Si and 27Al MAS NMR spectra were recorded at 79.49 and 104.3 MHz, respectively, on a (9.4 T) Bruker MSL 400 P spectrometer. The 29Si MAS NMR spectra were recorded using 40° pulses, a spinning rate of 5 kHz and 40–60 s recycle delays. Chemical shifts are quoted in ppm from TMS. The framework Al/Ti and Si/Al ratios of samples were calculated from the 29Si MAS NMR spectrum in a similar way to zeolites according to our previous work [7]. Single-quantum (1Q) 27Al MAS NMR spectra were recorded using very short, 0.6 ms (10°) and powerful radiofrequency pulses with a 0.5 s recycle delay and a spinning rate of 15 kHz. The quintuple-quantum (5Q) 27Al MAS NMR spectra were recorded with radio-frequency magnetic field amplitude of 166 kHz, 1 s recycle delay and a spinning rate of 14 kHz. To produce pure absorption line-shapes in the MQ MAS spectra the optimum conditions for excitation and transfer of the MQ coherences using a simple two-pulse sequence were used [8]. A total of 256 points were acquired in the t1 dimension in increments of 5 ms. The chemical shifts are quoted in ppm from Al(H2O)36 + . TGA curves were measured on a Shimadzu TGA-50 with a 5° C min − 1 heating rate. The Raman spectra were measured at room temperature with a Renishaw imaging microscope 2000, using a 25 mW Spectra Physics 127 HeNe excitation laser (632.8 nm), a resolution of 0.8– 1.6 cm − 1.
3. Results and discussion ETAS-10 samples with Al/Ti ratio up to 0.35 can be prepared by the titanium trichloride route and using an appropriate gel composition (Table
Z. Lin et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 133–138
1). The pH of the parent gel (after 1:100 dilution with water) is 10.4 – 10.6. The aluminum contents of ETAS-10 materials calculated from the 29Si MAS NMR spectra [7] are given in Table 2. The Al content of ETAS-10 may be fine-tuned by slightly changing the K:(Na + K) molar ratio of the starting gel. For example, for sample A2 when this ratio increases from 0.17 to 0.24, the Al/Ti molar ratio of ETAS-10 materials decreases from Table 1 Compositions of the parent gels for preparing ETAS-10 Samplea
Na2O
K2O
SiO2
TiO2
Al2O3
A A1 A2 A3 A4
4.88 4.76 4.56 3.81 4.61
1.63 1.05 0.92 1.49 0.91
5.16 5.32 5.95 5.90 5.94
1.00 1.00 1.00 1.00 1.00
0.25 0.11 0.22 0.30 0.23
H2O 127 120 122 121 143
a Sample A is prepared using pseudoboehmite and anatase as aluminum and titanium sources while sources for samples A1–A4 are sodium aluminate and titanium trichloride.
Table 2 Atomic ratios of ETAS-10 samples obtained from deconvolution of the 29Si MAS NMR spectra Sample
Al/Ti
Si/Al
A A1 A2 A3 A4
0.10 0.16 0.26 0.35 0.48
47.1 31.0 18.7 13.6 9.7
Fig. 1. Powder XRD patterns of ETS-10 and selected ETAS10 samples (A2 and A4).
135
0.26 to 0.20. A powder XRD pattern and a SEM image of a select sample are shown in Figs. 1 and 2. The 29Si MAS NMR spectra of ETAS-10 materials are depicted in Fig. 3. The preparation of highly crystalline ETAS-10 samples with an Al/Ti molar ratio higher than 0.35 is difficult. We have found that in order to obtain materials of good quality it is necessary to dry and calcine the parent gel before carrying out the hydrothermal treatment. For example, sample A4 (Al:Ti = 0.48), gel composition given in Table 1) was synthesized in this way and its 29Si MAS NMR spectrum is depicted in Fig. 3. Samples A2 (Al:Ti= 0.26) and A4 were prepared from the similar parent gels but only in the synthesis of the latter was this gel dried and calcined before hydrothermal treatment. Clearly, gel calcination has a significant effect on the resulting ETAS-10 synthesis products and it is of interest to understand why this is. Consider the synthesis of ETAS-10 with (Al:Ti= 0.48). The powder XRD patterns of the parent gel (1) dried at 40, (2) dried at 130°C, and (3) dried at 130°C and calcined at 500°C for 1.5 h are similar, indicating the presence of an amorphous material containing some NaCl and KCl. However, their 27Al MAS NMR spectra (Fig. 4) are different. The solid dried at 40°C contains 80% six-coordinated Al (peak at 0 ppm) while in the other two solids, the Al atoms are essentially four-coordinated (peaks at 55 ppm). In addition, the calcination of the gel at 500°C broadens the NMR resonance, suggesting that the local environment of the Al tetrahedra are more distorted. Therefore, we speculate that gel calcination converts hexa-coordinated Al into highly reactive (distorted) four-coordinated Al and favors the synthesis process. In an effort to further optimize the synthesis procedure, alternative aluminum and titanium sources were tested. ETAS-10 with a low (0.10) Al/Ti molar ratio was prepared at 230°C for 3 days by the anatase route and using pseudoboehmite as the Al source (gel composition A in Table 1). No seeds were used. A mixture of ETAS-10, ETS-4 and unidentified aluminosilicates was obtained. However because the ETS-4 and aluminosilicates particles (\ 7 mm) are larger than the ETAS-10 crystals ( B 1 mm) it is possible to sepa-
136
Z. Lin et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 133–138
Fig. 2. Typical SEM images of ETAS-10 (sample A2).
rate them by paper filtration and to obtain almost pure ETAS-10. As it was found previously for the synthesis of ETS-10 from anatase [4], the rate-limiting step in the synthesis of ETAS-10 is the slow dissolution of anatase. Seeding does not accelerate the synthesis, indicating that this process is not controlled by nucleation. When the pseudoboehmite content in the parent gel was increased the ETAS-10 Al content remained low but the amount of contaminating impurities increased. The structures of ETS-10 and ETAS-10 are closely related. These materials display very similar powder XRD patterns (Fig. 1). In principle, increasing the Al content of the framework should cause an increase in the d-spacing because the Al–O bond length is longer than the Si –O bond length. Indeed, a small shift of all XRD reflections to higher d-values (24.63°/2u for ETS-10, 24.44°/2u for ETAS-10, Al:Ti= 0.48) is observed as the Al content increases. As revealed by SEM, the elongated octahedra ETAS-10 crystals are similar to ETS-10 [4] crystals with a particle size around 1–4 mm, depending on the synthesis procedure. The 29Si MAS NMR spectra of ETAS-10 with a range of Al substitutions (Fig. 3) are more complicated than the spectrum of ETS-10. Besides all the peaks also given by ETS-10, centered at −94.7 and −96.6 ppm [ascribed to Si(3Si, 1Ti) chemical environments] and at − 103.7 ppm [ascribed to Si(4Si) environments] ETAS-10 gives an additional broad peak 4 ppm down-field from the Si(3Si, 1Ti) signal [1,2,6,7]. This broad peak is ascribed to Si(2Si, 1Al, 1Ti) environments [6,7].
Hence, 29Si MAS NMR spectroscopy provides direct evidence for the aluminum substitution in ETS-10 and allows the distinction between samples with different Al/Ti (and Si/Al) molar ratios. In order to complement the previously reported data [6,7] showing that Al is introduced into only one of the possible ETS-10 silicon sites [the Si(4Si) site], we have now recorded 27Al 5Q MAS NMR spectra for selected samples. The spectrum in Fig. 5 displays a single 27Al NMR resonance, which is broadened by the presence of a distribution of quadrupole parameters and isotropic chemical shifts. The TGA curves of ETAS-10 samples (Fig. 6) with different aluminum contents reveal a weight
Fig. 3. 29Si MAS NMR spectra of ETAS-10 materials with the Al/Ti molar ratios indicated.
Z. Lin et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 133–138
137
with increasing aluminum content. Titanosilicate ETS-4 and purely titaneous synthetic nenadkevichite [10] display the Ti–O–Ti vibration at 775 and 725 cm − 1, respectively. The 725 cm − 1 band may also contain some contribution from the SiO4 deformation mode. Indeed, research on AM-6, a vanadium silicate analogue of ETS-10, reveals that most Raman bands shift to higher frequency, confirming that the ETS-10 bands are associated with Ti modes. A weak band at −725 cm − 1 in the spectrum of AM-6 indicates that there is also
Fig. 4. 27Al MAS NMR spectra of materials obtained in the ETAS-10 (sample A4) synthesis process which involves drying and calcining the parent gel prior to the hydrothermal treatment.
Fig. 6. TGA curves of ETAS-10 samples with the Al/Ti molar ratios indicated.
Fig. 5. A4).
27
Al 5Q MAS NMR spectrum of ETAS-10 (sample
loss of 12–14 w/w%, which increases with increasing Al content. Raman spectra of selected ETAS-10 samples are shown in Fig. 7. No new band appears when the Al framework content increases. The spectra of both ETS-10 [9] and low aluminum content (Al:Ti= 0.10) samples are very similar and are dominated by one sharp peak at − 725 cm − 1. This peak broadens and shifts to higher frequency
Fig. 7. Raman spectra of selected ETAS-10 with the Al/Ti molar ratios indicated.
138
Z. Lin et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 179 (2001) 133–138
some contribution of the SiO4 modes for the main Raman peak of ETS-10 [9]. It is possible that in ETAS-10 the AlO4 stretching vibrations couple with the deformation of the SiO4 groups giving rise to a broadening effect, instead of producing new vibrational modes. As in zeolites, two faint bands above 900 cm − 1 may be due to Si – O stretch vibrations. The Raman spectrum of sample A does not show any band at 144 cm − 1, indicating that the material is free from anatase.
4. Conclusions Highly crystalline ETAS-10 with Al/Ti molar ratios between 0.10 and 0.48 were synthesized. Good quality samples with Al framework contents above 0.35 could only be obtained using a modification of the synthesis procedure, which involves drying and calcining the parent gel before the hydrothermal treatment. Low Al content ETAS-10 materials are best prepared using anatase as the Ti source and pseudoboehmite as the Al source. Changing the K and Na cations composition of the parent gel allows the fine-tuning of the Al/Ti molar ratio.
.
Acknowledgements The Portuguese team thanks PRAXIS XXI and FEDER for financial support.
References [1] M.W. Anderson, O. Terasaki, T. Ohsuna, A. Philippou, S.P. Mackay, A. Ferreira, J. Rocha, S. Lidin, Nature 367 (1994) 347. [2] M.W. Anderson, O. Terasaki, T. Ohsuna, A. Philippou, S.P. Mackay, A. Ferreira, J. Rocha, S. Lidin, Phil. Mag. B 71 (1995) 813. [3] S.M. Kuznicki, US Patent, 4 853 202, August 1, 1989. [4] J. Rocha, A. Ferreira, Z. Lin, M.W. Anderson, Microporous Mesoporous Mater. 23 (1998) 253. [5] S.M. Kuznicki, R.J. Madon, G.S. Koermer, K.A. Thrush, Euro Patent, 0 405 978 A1, 02, 01, 1991. [6] M.W. Anderson, A. Philippou, Z. Lin, A. Ferreira, J. Rocha, Angew Chem. Int. Ed. Engl. 34 (1995) 1003. [7] M.W. Anderson, J. Rocha, Z. Lin, A. Philippou, I. Orion, A. Ferreira, Microporous Mater. 6 (1996) 195. [8] C. Fernandez, J.P. Amoureux, J.M. Chezeau, L. Delmotte, H. Kessler, Microporous Mater. 6 (1996) 331. [9] J. Rocha, P. Branda˜o, Z. Lin, M.W. Anderson, V. Alfredsson, O. Terasaki, Angew Chem. Int. Ed. Engl. 36 (1997) 100. [10] J. Rocha, P. Branda˜o, Z. Lin, A.P. Esculcas, A. Ferreira, M.W. Anderson, J. Phys. Chem. 100 (1996) 14978.
