- Email: [email protected]
Realumination of dealuminated HZSM-5 zeolites by acid treatment: a reexamination
Microporous and Mesoporous Materials 46 (2001) 177±184
www.elsevier.com/locate/micromeso
Realumination of dealuminated HZSM-5 zeolites by acid treatment: a reexamination Anna Omegna, Mohamed Haouas, Andreas Kogelbauer 1, Roel Prins * Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH), Universitatstrasse 6, CH-8092 Zurich, Switzerland Received 27 November 2000; received in revised form 2 March 2001; accepted 2 March 2001
Abstract The feasibility of reinsertion of non-framework aluminum into the lattice of dealuminated HZSM-5 zeolites upon acid treatment was investigated by powder X-ray diraction, nitrogen adsorption, elemental analysis, solid-state NMR and Fourier transform infrared (FTIR) spectroscopy. XRD and nitrogen adsorption revealed that no structural degradation of the HZSM-5 zeolite matrix took place during dealumination and acid treatment. FTIR and multinuclear NMR spectroscopy showed that hydrothermal treatment was more eective in dealumination than the mere calcination treatment. In the dealuminated samples, a fraction of the tetrahedral aluminum was present as extra-lattice aluminum. No spectroscopic evidence of reinsertion of aluminum into the framework was observed after acid treatment. On the contrary, in addition to leaching of the octahedral extra-framework aluminum, part of the framework aluminum was extracted as well. A fraction of the tetrahedral aluminum did not belong to the zeolite lattice. Elemental analysis revealed that silicon was present in the mother liquor of the acid treatments. It is proposed that an amorphous phase of silico-aluminate, with aluminum in tetrahedral coordination, is formed as a result of the reaction between this silicon and the aluminum in the extracted solution. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Realumination; Acid treatment; Leaching; Silica±alumina; Tetrahedral aluminum
1. Introduction Acid±base properties of zeolites clearly depend on the content of aluminum in the framework. The Si/Al ratio can be adjusted either during the crystallization or via post-synthesis modi®cation. Dealumination, the removal of framework aluminum from the zeolite lattice, is a well known pro* Corresponding author. Tel.: +41-1-6325490; fax: +41-16321162. E-mail address: [email protected] (R. Prins). 1 Present address: Department of Chemical Engineering and Chemical Technology; Imperial College of Science, Technology and Medicine, London SW7 2BY, UK.
cedure for stabilizing zeolites and for creating mesopores, which help to overcome diusional problems in the micropores of the zeolites [1]. Since the activity of zeolites for acid-catalyzed reactions is directly related to the number of framework aluminum atoms, the possibility of reinserting aluminum into the partially dealuminated, mesoporous framework has become a topic of increasing interest for the production of stable and highly active catalysts. As early as 1980, Breck and Skeels reported the reinsertion of aluminum into the lattice of a thermally decomposed NH4 Y zeolite upon treatment with an aqueous NaOH solution [2]. Since then, several attempts were made to reincorporate aluminum into the framework
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 2 8 1 - 5
178
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
of zeolites by alkaline treatment [3±7], following the concept that the application of conditions, similar to those under which hydrothermal synthesis occurs, may lead to the reinsertion of aluminum during post-synthetic treatment as well. However, the published reports leave many questions open regarding the actual incorporation of aluminum into the lattice, because more recent studies demonstrated that tetrahedral aluminum can be present as extra-lattice aluminum in the form of silica±alumina [8±11]. Sano et al. recently reported that acid treatment was eective in inducing the reinsertion of nonframework aluminum into the lattice of dealuminated HZSM-5 zeolites [12,13]. Their conclusion was based on 27 Al MAS NMR and Fourier transform infrared (FTIR) spectroscopic studies of acid treated HZSM-5 zeolites dealuminated by calcination or steaming. Acid treatment after dealumination is a common procedure for leaching the extra-framework aluminum (EFAL) species occluded in the pores of zeolites, thereby helping to minimize diusion problems during the catalytic reaction. On the other hand, the direct treatment of Al-rich zeolites with mineral acids is a well known dealumination procedure. Since a contradiction seemed to exist, a careful re-evaluation of the procedure proposed by Sano et al. [12,13] was deemed necessary. In this study samples were prepared following the procedure described by Sano et al. [12,13] and were characterized using a combination of powder X-ray diraction, elemental analysis, nitrogen adsorption, 27 Al, 1 H, 29 Si MAS NMR as well as FTIR spectroscopy. 2. Experimental 2.1. Materials Zeolite ZSM-5 PZ-2/40 (CU Chemie Uetikon) was calcined in ¯owing air at 550°C for 5 h to remove the template. HZSM-5 was then prepared by subsequent three-fold ion exchange with 1 M ammonium nitrate solution under re¯ux, followed by washing with deionized water and calcination of the NH4 -ZSM-5 at 550°C for 5 h. Thermal dealumination of the HZSM-5 was carried out in a
mue furnace at 600°C for 48 h (to give sample HZSM-5(T)). Hydrothermal dealumination was performed in a stainless steel reactor tube at 500°C for 6 h with a steam ¯ow rate of 33 ml water/h (to give sample HZSM-5(H)). One gram of the dealuminated HZSM-5(T) and HZSM-5(H) was treated with 100 ml of 2 M hydrochloric acid at 100°C for 120 h. The resulting materials were ®ltered, washed with deionized water, dried at 100°C and then calcined at 380°C for 8 h (to give samples HZSM-5(T)HCl and HZSM-5(H)HCl , respectively). Amorphous silica±alumina (13% Al2 O3 , Shell) was used as reference material. 2.2. Characterization Solid-state MAS NMR spectra were recorded on a Bruker AMX400 spectrometer operating at a static ®eld of 9.4 T. 27 Al MAS NMR measurements were performed at a resonance frequency of 104.26 MHz. Spectra were recorded at a spinning rate of 7 kHz, a pulse length of 1.2 ls p=12 and a delay time of 1 s. All zeolites were fully hydrated in a desiccator over a saturated aqueous NaCl solution for one week prior to the measurements. For quantitative evaluation, all samples were weighed, and the spectra were calibrated by measuring a known amount of (NH4 )Al(SO4 )2 12H2 O under identical conditions [14]. 1 H MAS NMR measurements were performed at a resonance frequency of 400.13 MHz. Spectra of HZSM-5, HZSM-5(T) and HZSM-5(T)HCl were recorded in 4-mm rotors at a spinning rate of 10 kHz, pulse length of 3.5 ls p=3 and delay time of 10 s, while the spectra of HZSM-5(H) and HZSM-5(HHCl were recorded in 7-mm rotors at a spinning rate of 7 kHz, pulse length of 4 ls p=4 and delay time of 10 s. All zeolites had been previously dehydrated under vacuum at 350°C for 4 h. For a quantitative comparison, all samples were weighed, and the spectra were calibrated by measuring a known amount of 1,1,1,3,3,3-hexa¯uoro-2-propanol under identical conditions [14]. 29 Si MAS NMR measurements were performed at a resonance frequency of 79.49 MHz. High power proton decoupling was applied during acquisition. The spinning rate was 5 kHz, the pulse length 1.5 ls p=6 and the delay time 10 s. The 29 Si NMR
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
179
ICP-OES agree within 3.5%. Analysis of the solutions after acid treatment was done by ICP-OES.
signal of TMS was used as the chemical shift reference. Deconvolution of the NMR spectra was performed using the Bruker software W I N F I T . FTIR spectra were obtained at 4 cm 1 resolution with a Mattson Galaxy spectrometer equipped with an MCT detector. Prior to the investigation, all the samples were pressed into self-supporting wafers (10±15 mg/cm2 ) and evacuated at 350°C for 6 h under a residual pressure of 10 6 Pa. For quantitative comparison spectra were normalized using the integrated intensities of the Si±O overtone vibrations. XRD powder diraction patterns from 5° to 60° 2h were obtained on a Siemens D5000 diractometer using CuKa radiation k Nitrogen adsorption measurements 1:5406 A. were carried out at 196°C on a Micromeritics ASAP 2010 instrument using a conventional volumetric technique. Prior to analysis the samples were outgassed at 400°C for 4 h at pressures below 1:33 10 3 Pa. The Si/Al ratios were determined by means of laser ablation, inductively coupled plasma mass spectrometry (LA-ICP-MS) [15]. An excimer laser (193 nm) (MicroLas GmbH) was attached to an ELAN 6100 (Perkin Elmer) mass spectrometer for direct solid sampling. The zeolites were pressed into self-supporting pellets and placed in an airtight ablation cell. The laser sampled 80 lm of the zeolite (10 Hz, 120 mJ, 5 replicates per sample), and the produced aerosol was transported into the mass spectrometer [16]. The concentration ratio was determined by comparison with the reference glass 610 from NIST as the external calibration standard. Two zeolite samples were digested and analyzed as solutions by ICP optical emission spectroscopy (ICP-OES). The data from the direct solid-sampling technique (LA-ICP-MS) and from
3. Results and discussion The results of the chemical analysis and nitrogen adsorption are summarized in Table 1. Elemental analysis of the calcined and steamed samples revealed little change in the bulk composition as compared to the parent HZSM-5. The original bulk Si/Al ratio of 19.1 was maintained after calcination (19.3) and increased slightly after steaming (20.1). On the contrary, analysis of the acid-treated samples indicated that the bulk composition of the zeolite changed during the treatment with acid. The Si/Al ratio of the calcined sample increased to 25.3 and that of the steamed sample to 38.3, indicating that leaching of aluminum occurred during the acid treatment. Analysis of the solutions of the acid treatment revealed that dissolution of silicon took place as well. Of the original silicon content of the calcined and steamed zeolites, 2.2 wt.% and 3.8 wt.%, respectively, were dissolved during the acid treatment. Nitrogen adsorption measurements showed that a secondary pore system did not form after the various treatments. The micropore volume remained unchanged, and the pore volume distribution of the meso- and macropore region (2±100 nm), calculated according to the BJH method, showed that there was no change in the textural properties. The BET surface area of the parent material HZSM-5 was not appreciably modi®ed as a result of the subsequent treatments. X-ray powder diraction patterns are shown in Fig. 1. They correspond to those expected for
Table 1 Nomenclature and properties of the HZSM-5 zeolites Sample
Treatment
Si/Al (0:1)
BET surface area 5 m2 =g
Micropore volume (cm3 =g)
HZSM-5 HZSM-5(T) HZSM-5(T)HCl HZSM-5(H) HZSM-5(H)HCl
Calcination, 600°C, 48 h 2 M HCl, 100°C, 120 h Steaming, 500°C, 6 h, 33 ml/min 2 M HCl, 100°C, 120 h
19.1 19.3 25.3 20.1 38.3
429 406 416 381 400
0.16 0.15 0.15 0.15 0.16
180
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
Fig. 1. XRD patterns of (a) HZSM-5, (b) HZSM-5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM-5(H)HCl .
samples of MFI topology. The diractograms indicate that crystallinity was retained after every single treatment. Fig. 2 shows a quantitative comparison of the 27 Al MAS NMR spectra of the HZSM-5 zeolites. The spectrum of the parent HZSM-5 (Fig. 2a) is characterized by an intense and sharp signal at 53 ppm due to aluminum species in tetrahedral coordination (AlIV ). An additional weak signal is observed at 0 ppm, assigned to octahedrally co-
Fig. 2. 27 Al MAS NMR spectra of (a) HZSM-5, (b) HZSM5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H), (e) HZSM-5(H)HCl and (f) silica±alumina. Asterisks indicate spinning side bands.
ordinated EFAl species (AlVI ), either present in the original samples or formed during the preliminary thermal treatments of the ZSM-5. After thermal (Fig. 2b) or hydrothermal treatment (Fig. 2d), the 27 Al tetrahedral aluminum signal decreased in intensity, indicating that removal of aluminum from the lattice occurred. Table 2 reports the results of the quantitative analysis of the deconvoluted spectra. They show that only a limited extent of tetrahedral aluminum was extracted from the framework in the case of the calcined material HZSM-5(T), while almost 50% loss was observed for the hydrothermally treated HZSM-5(H) sample. These results agree with the known resistance of HZSM-5 to dealumination and with the fact that the degree of dealumination is greater in the presence of steam [17]. At the same time, the intensity of the peak at 0 ppm increased and a feature appeared at 30 ppm. The latter signal has been assigned to highly distorted tetrahedral [18] as well as to ®ve-coordinated aluminum species [19,20] (because of the ambiguity of the assignment, we will refer to this signal as Al30 ppm ). It is worthwhile noticing that the dealuminated samples HZSM-5(T) and HZSM-5(H) do not contain ``NMR invisible'' aluminum, as it is inferred from the fact that the total aluminum content did not change after dealumination (see Table 2). The spectra after acid treatment of the samples dealuminated by calcination or steaming are shown in Fig. 2c and e, respectively. In contrast to the results reported by Sano et al. [12], acid treatment of the dealuminated zeolites did not lead to the recovery of the intensity of the tetrahedral aluminum signal. On the contrary, a quantitative comparison of the areas of the signal at 53 ppm before and after acid treatment (see Table 2) showed a decrease in the amount of tetrahedrally coordinated aluminum species. This result indicates that the acid treatment did not result in the reinsertion of non-framework aluminum into the lattice, but rather to the further dissolution of structural aluminum. Our 27 Al MAS NMR results clearly show that the main eect of the acid treatment is the leaching of EFAl species. This is demonstrated by the decrease in the amount of the species responsible for the signal at 0 ppm on HZSM-5(T)HCl and HZSM-5(H)HCl (see Table 2 and Fig. 2c and e).
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
181
Table 2 Concentration of the dierent aluminum and proton species in the dierent environments on the zeolites HZSM-5 Zeolite
AlIV lmol=g 27 Al NMR
AlVI lmol=g 27 Al NMR
Al30 ppm lmol=g 27 Al NMR
Si(OH) lmol=g 1 H NMR
Si(OH)Al lmol=g 1 H NMR
HZSM-5 HZSM-5(T) HZSM-5(T)HCl HZSM-5(H) HZSM-5(H)HCl
406 386 348 217 179
21 33 19 69 27
± 38 81 111 95
256 197 237 118 221
420 323 288 83 72
This is not surprising since the acid treatment is usually applied to dissolve EFAl that is occluded in the pores of dealuminated zeolites. The extent of the dissolution was higher in the case of the steamed material. Contrary to what was reported by Sano et al. [13], HCl treatment of HZSM-5(T) led to an increase in the concentration of the species characterized by the signal at 30 ppm. Therefore, it would not be possible for realumination to occur through the reinsertion of these species. A slight decrease was observed in the case of the steamed material. In conclusion, the 27 Al MAS NMR characterization of the acid-treated samples showed the following. (i) Octahedral aluminum species were washed out, although not completely. (ii) An additional fraction of tetrahedral aluminum was removed from the lattice. (iii) In the case of the thermally dealuminated sample, some tetrahedrally coordinated aluminum probably changed into penta-coordinated or tetrahedrally distorted species. 1 H MAS NMR further supported the conclusion that non-framework aluminum was not reinserted into the lattice of dealuminated HZSM-5 zeolites during acid treatment. In Fig. 3 the 1 H MAS NMR spectra of the dierent samples are quantitatively compared. At least three signals are present in the spectrum of the parent HZSM-5 (see Fig. 4), with maxima at 1.3, 1.8 and 3.4 ppm. The peak at 1.3 ppm is due to protons of silanols, the signal at 3.4 ppm is due to bridged Si(OH)Al groups, and the component at 1.8 ppm is assigned to non-acidic hydroxyls which are bonded to EFAl species [21]. There is a broad feature between 4 and 8 ppm, which may be due to H-bonded silanols [22,23] and/or Si(OH)Al groups interacting with
Fig. 3. 1 H MAS NMR spectra of (a) HZSM-5, (b) HZSM5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM5(H)HCl .
Fig. 4. Computer deconvolution of the 1 H MAS NMR spectrum of HZSM-5: experimentally observed spectrum (Ð); resolved components ( ); ®t (- - -).
182
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
the framework oxygens via H-bonds [14,24]. Upon dealumination, the concentration of Brùnsted acidic protons declined remarkably (see Fig. 3b and d and Table 2). The extent of the decrease was much greater in the case of the steamed sample (Fig. 3d). This con®rms again that the degree of dealumination by hydrothermal treatment is higher compared to the mere thermal treatment. The area of the silanol peak at 1.3 ppm (Fig. 3b and d) diminished, suggesting that silicon migrates to the tetrahedral vacancies generated by the dealumination of the framework. In both dealuminated samples HZSM-5(T) and HZSM-5(H), the amount of tetrahedral aluminum species (27 Al NMR) exceeds that of the Brùnsted acid centers determined by 1 H NMR (see Table 2). For the HZSM-5(H) sample this dierence is quite large. Since every aluminum atom in the zeolite lattice is associated with a Si(OH)Al Brùnsted center, some tetrahedral aluminum in the dealuminated samples must be present as EFAlIV , either as aluminum (hydr)oxide or as silica±alumina. The formation of EFAlIV associated with an amorphous silica±alumina phase after dealumination has been described for zeolites Y [19,25] and ZSM-5 [8]. Silica±alumina possesses aluminum in tetrahedral environment, as revealed by the presence of a signal at 53 ppm in the 27 Al NMR spectrum (see Fig. 2f). On the basis of these considerations, we believe that the peak at 53 ppm on the 27 Al NMR spectrum of the dealuminated samples is due to the combination of the signals of framework aluminum and a new EFAlIV , possibly associated with an amorphous silica±alumina phase formed from the reaction between the extracted silicon and aluminum. Acid treatment after dealumination clearly led to a decrease of the concentration of the Brùnsted acid centers (Fig. 3c and e and Table 2), con®rming that even more structural aluminum was removed from the zeolite matrix. Also in the case of the acid-treated samples, the amount of tetrahedral aluminum exceeded the amount of bridged Si(OH)Al groups, thus revealing the formation of new EFAlIV species. At the same time, the amount of silanol groups (peak at 1.4 ppm) increased, indicating the possible formation of an amorphous phase of silica±alumina from the reaction between
the extracted aluminum and silicon in solution. The formation of silico-aluminates after washing was ®rst proposed by Lutz et al. for dealuminated zeolite Y treated with an alkaline solution [9,10]. The increase of the silanol population may also be due to the formation of a new silica phase. Acid treatment may dissolve alumina of the aluminumrich silica±alumina debris, thus leading to silica species. Another signal of weak intensity was present in the spectra of the acid-treated samples at 0 ppm and represents non-acidic unperturbed hydroxyl groups, which are bonded to EFAl species [14]. The corresponding FTIR spectra are quantitatively compared in Fig. 5. There are at least four bands at 3743, 3667, 3610 and 3550±3300 cm 1 in the FTIR spectrum of the parent HZSM-5 zeolite (Fig. 5a). The IR spectrum is the mirror image of the corresponding 1 H NMR spectrum. There is general agreement that the band at 3743 cm 1 corresponds to the O±H stretching mode of isolated silanols, while that at 3610 cm 1 is due to bridged Si(OH)Al groups [26±29]. The weak feature at 3667 cm 1 is assigned to hydroxyl groups attached to EFAl oxide species [30,31]. Finally, the broad absorption around 3550±3300 cm 1 was assigned to hydrogen-bonded silanols [22,23] and/ or bridged Brùnsted acid groups [14,24]. Because of the similarity between FTIR and 1 H NMR spectroscopy, it is not surprising that the evolution
Fig. 5. FTIR spectra in the OH stretching spectral region of (a) HZSM-5, (b) HZSM-5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM-5(H)HCl .
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
of the FTIR spectra pro®le of HZSM-5 upon subsequent treatments is consistent with the trend found in 1 H NMR. However, no quantitative analysis of the FTIR spectra was attempted, since IR intensities are strongly in¯uenced by the different environments (molar extinction coecients vary for dierent families of hydroxyl groups). On the basis of the comparison of the spectra of the acid-treated HZSM-5 samples, the following conclusions can be drawn. (i) Leaching of EFAl took place, as evidenced by the decrease of the band at 3667 cm 1 . (ii) The band at 3743 cm 1 increased, suggesting formation of structural defects and the formation of a new silica phase. (iii) The intensity of the band at 3660 cm 1 (Brùnsted acid sites) decreased slightly. The 29 Si MAS NMR spectra and their resolved lines of the parent HZSM-5, the dealuminated HZSM-5(T) and the acid-treated HZSM-5(T)HCl samples are shown in Fig. 6. Four peaks were distinguished and assigned as follows. The signals at 112.6 and 106.7 ppm were attributed to the resonance of the silicon atoms in the SiO4 tetrahedra surrounded by four (4Si, 0Al) and three (3Si, 1Al) silicon atoms, respectively. The peak at 103 ppm was assigned to silanols of (Si(OH)1 (±OSi)3 ) units, while the shoulder at 116.1 ppm was attributed to the presence of crystallographically inequivalent sites in the zeolite [32,33]. 29 Si MAS NMR cannot provide quantitative information about the framework transformation, since the three spectra exhibit very similar patterns and only show minor dierences in peak intensities. Therefore, only qualitative conclusions can be drawn regarding the feasibility of realumination. Thermal treatment (Fig. 6b) did not appreciably aect the area of the signal at 106.7 ppm associated to Brùnsted acid sites. As we showed above, a more severe treatment is necessary to remove a consistent amount of aluminum from the framework. Acid treatment after thermal dealumination led to the spectrum of Fig. 6c. If a realumination had occurred, the amount of framework aluminum species should have increased. However, no signi®cant change in the area of the signal at 106.7 ppm, associated with these species, was observed, con®rming that no reinsertion of aluminum into the framework has occurred.
183
Fig. 6. 29 Si MAS NMR spectra and decomposition into the dierent component lines: (a) HZSM-5, (b) HZSM-5(T) and (c) HZSM-5(T)HCl .
4. Conclusions The results of our characterization are in clear contrast to the proposal by Sano et al. [12,13] regarding the feasibility of the reinsertion of nonframework aluminum into vacant framework sites by treatment with an aqueous solution of HCl. Investigation by FTIR and multinuclear NMR spectroscopy of the samples prepared under the same conditions did not show any evidence of reinsertion of aluminum into the lattice. The extent of the dealumination of the zeolite HZSM-5 under hydrothermal conditions was found to be much greater than under dry conditions. As a result of the dealumination, the extracted framework
184
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
tetrahedral aluminum changed into the following non-framework species: (i) hexa-coordinated, associated to the signal at 0 ppm in the 27 Al NMR; (ii) highly distorted tetrahedral or penta-coordinated (signal at 30 ppm); (iii) tetrahedral, either associated to aluminum (hydr)oxide or to an amorphous phase of silica±alumina. The latter species gives a signal at 53 ppm in the 27 Al NMR spectrum, which overlaps with the signal of the framework tetrahedral aluminum. Upon acid treatment of the dealuminated materials, leaching of extra-lattice octahedral aluminum was observed, as expected. Moreover, an additional fraction of structural aluminum was extracted. Part of this extracted aluminum probably changed into penta-coordinated or tetrahedrally distorted aluminum species (in the case of the HZSM-5 dealuminated by steaming), and part reacted with the extracted silicon in solution to form further amounts of amorphous silica±alumina. Some silica phase could also be present in the acid-treated samples, originating from the silica±alumina debris after dissolution of the alumina. Acknowledgements The authors thank Prof. D.G unther for carrying out the elemental analysis. References [1] L. Bertea, H.W. Kouwenhoven, R. Prins, Appl. Catal. A: Gen. 129 (1995) 229. [2] D.W. Breck, G.W. Skeels, in: L.V.C. Rees (Ed.), Proceedings 5th International Zeolite Conference, Heyden, London, 1980, p. 335. [3] X. Liu, J. Klinowski, J.M. Thomas, J. Chem. Soc., Chem. Commun. (1986) 582. [4] H. Hamadan, B. Sulikowski, J. Klinowski, J. Phys. Chem. 93 (1989) 350. [5] Z. Zhang, X. Liu, Y. Xu, R. Xu, Zeolites 11 (1991) 232. [6] V. Calsavara, E. Falabella Sousa-Aguiar, N.R.C. Fernandes Machado, Zeolites 17 (1996) 340.
[7] B. Sulikowski, J. Datka, B. Gil, J. Ptaszynski, J. Klinowski, J. Phys. Chem. B 101 (1997) 6929. [8] Y.C. Long, M.Y. Jin, Y.J. Sun, T.L. Wu, L.P. Wang, L. Fei, J. Chem. Soc., Faraday Trans. 92 (1996) 1647. [9] W. Lutz, W. Gessner, D. M uller, Zeolites 19 (1997) 209. [10] W. Lutz, W. Gessner, R. Bertram, I. Pitsch, R. Fricke, Micropor. Mater. 12 (1997) 131. [11] L. Heeribout, R. Vincent, P. Batamack, C. DoremieuxMorin, J. Fraissard, Catal. Lett. 53 (1998) 23. [12] T. Sano, R. Dtadenuma, Z. Wang, K. Soga, Chem. Commun. (1997) 1945. [13] T. Sano, Y. Uno, Z.B. Wang, C.H. Ahn, K. Soga, Micropor. Mesopor. Mater. 31 (1999) 89. [14] M. M uller, G. Harvey, R. Prins, Micropor. Mesopor. Mater. 34 (2000) 281. [15] D. G unther, S.E. Jackson, H.P. Longerich, Spectrochim. Acta B 54 (1999) 381. [16] D. G unther, R. Frischknecht, C.A. Heinrich, J. Anal. At. Spectrom. 12 (1997) 939. [17] S.M. Campbell, D.M. Bibby, J.M. Coddington, R.F. Howe, R.H. Meinhold, J. Catal. 161 (1996) 338. [18] A. Samoson, E. Lippmaa, G. Engelhardt, U. Lohse, H.G. Jerschkewitz, Chem. Phys. Lett. 134 (1987) 589. [19] J. Sanz, V. Fornes, A. Corma, J. Chem. Soc., Faraday Trans. 84 (1988) 3113. [20] J.P. Jilson, G.C. Edwards, A.W. Peters, K. Rajagopalan, R.F. Wormsbecker, T.G. Roberie, M.P. Shatlock, J. Chem. Soc., Chem. Commun. (1987) 91. [21] M. Hunger, Catal. Rev. Sci. Engng. 39 (1997) 345. [22] V.L. Zholobenko, L.M. Kustov, V.Y. Borovkov, V.B. Kazansky, Zeolites 8 (1988) 175. [23] E. Bourgeat-Lami, P. Massiani, P. Espiau, F.D. Renzo, F. Fajula, Appl. Catal. 72 (1991) 139. [24] C. Paze, A. Zecchina, S. Spera, A. Cosma, E. Merlo, G. Span o, G. Girotti, Phys. Chem. Chem. Phys. 1 (1999) 2627. [25] A. Janin, M. Maache, J.C. Lavalley, J.F. Joly, F. Raatz, N. Szydlowski, Zeolites 11 (1991) 391. [26] G. Qin, L. Zheng, Y. Xie, C. Wu, J. Catal. 95 (1985) 609. [27] J. Dwyer, Stud. Surf. Sci. Catal. 37 (1988) 333. [28] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Petrini, G. Leofanti, M. Padovan, C. Otero Arean, J. Chem. Soc., Faraday Trans. 88 (1992) 2959. [29] H. Kn ozinger, S. Huber, J. Chem. Soc., Faraday Trans. 94 (1998) 2047. [30] L.M. Kustov, V.B. Kazansky, S. Beran, L. Kubelkova, P. Jiru, J. Phys. Chem. 91 (1987) 5247. [31] A. Zecchina, C. Otero Arean, Chem. Soc., Rev. 25 (1996) 187. [32] E. Brunner, H. Ernst, D. Freude, J. Catal. 127 (1991) 34. [33] W. Zhang, X. Bao, X. Guo, X. Wang, Catal. Lett. 60 (1999) 89.
www.elsevier.com/locate/micromeso
Realumination of dealuminated HZSM-5 zeolites by acid treatment: a reexamination Anna Omegna, Mohamed Haouas, Andreas Kogelbauer 1, Roel Prins * Laboratory for Technical Chemistry, Swiss Federal Institute of Technology (ETH), Universitatstrasse 6, CH-8092 Zurich, Switzerland Received 27 November 2000; received in revised form 2 March 2001; accepted 2 March 2001
Abstract The feasibility of reinsertion of non-framework aluminum into the lattice of dealuminated HZSM-5 zeolites upon acid treatment was investigated by powder X-ray diraction, nitrogen adsorption, elemental analysis, solid-state NMR and Fourier transform infrared (FTIR) spectroscopy. XRD and nitrogen adsorption revealed that no structural degradation of the HZSM-5 zeolite matrix took place during dealumination and acid treatment. FTIR and multinuclear NMR spectroscopy showed that hydrothermal treatment was more eective in dealumination than the mere calcination treatment. In the dealuminated samples, a fraction of the tetrahedral aluminum was present as extra-lattice aluminum. No spectroscopic evidence of reinsertion of aluminum into the framework was observed after acid treatment. On the contrary, in addition to leaching of the octahedral extra-framework aluminum, part of the framework aluminum was extracted as well. A fraction of the tetrahedral aluminum did not belong to the zeolite lattice. Elemental analysis revealed that silicon was present in the mother liquor of the acid treatments. It is proposed that an amorphous phase of silico-aluminate, with aluminum in tetrahedral coordination, is formed as a result of the reaction between this silicon and the aluminum in the extracted solution. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Realumination; Acid treatment; Leaching; Silica±alumina; Tetrahedral aluminum
1. Introduction Acid±base properties of zeolites clearly depend on the content of aluminum in the framework. The Si/Al ratio can be adjusted either during the crystallization or via post-synthesis modi®cation. Dealumination, the removal of framework aluminum from the zeolite lattice, is a well known pro* Corresponding author. Tel.: +41-1-6325490; fax: +41-16321162. E-mail address: [email protected] (R. Prins). 1 Present address: Department of Chemical Engineering and Chemical Technology; Imperial College of Science, Technology and Medicine, London SW7 2BY, UK.
cedure for stabilizing zeolites and for creating mesopores, which help to overcome diusional problems in the micropores of the zeolites [1]. Since the activity of zeolites for acid-catalyzed reactions is directly related to the number of framework aluminum atoms, the possibility of reinserting aluminum into the partially dealuminated, mesoporous framework has become a topic of increasing interest for the production of stable and highly active catalysts. As early as 1980, Breck and Skeels reported the reinsertion of aluminum into the lattice of a thermally decomposed NH4 Y zeolite upon treatment with an aqueous NaOH solution [2]. Since then, several attempts were made to reincorporate aluminum into the framework
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 2 8 1 - 5
178
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
of zeolites by alkaline treatment [3±7], following the concept that the application of conditions, similar to those under which hydrothermal synthesis occurs, may lead to the reinsertion of aluminum during post-synthetic treatment as well. However, the published reports leave many questions open regarding the actual incorporation of aluminum into the lattice, because more recent studies demonstrated that tetrahedral aluminum can be present as extra-lattice aluminum in the form of silica±alumina [8±11]. Sano et al. recently reported that acid treatment was eective in inducing the reinsertion of nonframework aluminum into the lattice of dealuminated HZSM-5 zeolites [12,13]. Their conclusion was based on 27 Al MAS NMR and Fourier transform infrared (FTIR) spectroscopic studies of acid treated HZSM-5 zeolites dealuminated by calcination or steaming. Acid treatment after dealumination is a common procedure for leaching the extra-framework aluminum (EFAL) species occluded in the pores of zeolites, thereby helping to minimize diusion problems during the catalytic reaction. On the other hand, the direct treatment of Al-rich zeolites with mineral acids is a well known dealumination procedure. Since a contradiction seemed to exist, a careful re-evaluation of the procedure proposed by Sano et al. [12,13] was deemed necessary. In this study samples were prepared following the procedure described by Sano et al. [12,13] and were characterized using a combination of powder X-ray diraction, elemental analysis, nitrogen adsorption, 27 Al, 1 H, 29 Si MAS NMR as well as FTIR spectroscopy. 2. Experimental 2.1. Materials Zeolite ZSM-5 PZ-2/40 (CU Chemie Uetikon) was calcined in ¯owing air at 550°C for 5 h to remove the template. HZSM-5 was then prepared by subsequent three-fold ion exchange with 1 M ammonium nitrate solution under re¯ux, followed by washing with deionized water and calcination of the NH4 -ZSM-5 at 550°C for 5 h. Thermal dealumination of the HZSM-5 was carried out in a
mue furnace at 600°C for 48 h (to give sample HZSM-5(T)). Hydrothermal dealumination was performed in a stainless steel reactor tube at 500°C for 6 h with a steam ¯ow rate of 33 ml water/h (to give sample HZSM-5(H)). One gram of the dealuminated HZSM-5(T) and HZSM-5(H) was treated with 100 ml of 2 M hydrochloric acid at 100°C for 120 h. The resulting materials were ®ltered, washed with deionized water, dried at 100°C and then calcined at 380°C for 8 h (to give samples HZSM-5(T)HCl and HZSM-5(H)HCl , respectively). Amorphous silica±alumina (13% Al2 O3 , Shell) was used as reference material. 2.2. Characterization Solid-state MAS NMR spectra were recorded on a Bruker AMX400 spectrometer operating at a static ®eld of 9.4 T. 27 Al MAS NMR measurements were performed at a resonance frequency of 104.26 MHz. Spectra were recorded at a spinning rate of 7 kHz, a pulse length of 1.2 ls p=12 and a delay time of 1 s. All zeolites were fully hydrated in a desiccator over a saturated aqueous NaCl solution for one week prior to the measurements. For quantitative evaluation, all samples were weighed, and the spectra were calibrated by measuring a known amount of (NH4 )Al(SO4 )2 12H2 O under identical conditions [14]. 1 H MAS NMR measurements were performed at a resonance frequency of 400.13 MHz. Spectra of HZSM-5, HZSM-5(T) and HZSM-5(T)HCl were recorded in 4-mm rotors at a spinning rate of 10 kHz, pulse length of 3.5 ls p=3 and delay time of 10 s, while the spectra of HZSM-5(H) and HZSM-5(HHCl were recorded in 7-mm rotors at a spinning rate of 7 kHz, pulse length of 4 ls p=4 and delay time of 10 s. All zeolites had been previously dehydrated under vacuum at 350°C for 4 h. For a quantitative comparison, all samples were weighed, and the spectra were calibrated by measuring a known amount of 1,1,1,3,3,3-hexa¯uoro-2-propanol under identical conditions [14]. 29 Si MAS NMR measurements were performed at a resonance frequency of 79.49 MHz. High power proton decoupling was applied during acquisition. The spinning rate was 5 kHz, the pulse length 1.5 ls p=6 and the delay time 10 s. The 29 Si NMR
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
179
ICP-OES agree within 3.5%. Analysis of the solutions after acid treatment was done by ICP-OES.
signal of TMS was used as the chemical shift reference. Deconvolution of the NMR spectra was performed using the Bruker software W I N F I T . FTIR spectra were obtained at 4 cm 1 resolution with a Mattson Galaxy spectrometer equipped with an MCT detector. Prior to the investigation, all the samples were pressed into self-supporting wafers (10±15 mg/cm2 ) and evacuated at 350°C for 6 h under a residual pressure of 10 6 Pa. For quantitative comparison spectra were normalized using the integrated intensities of the Si±O overtone vibrations. XRD powder diraction patterns from 5° to 60° 2h were obtained on a Siemens D5000 diractometer using CuKa radiation k Nitrogen adsorption measurements 1:5406 A. were carried out at 196°C on a Micromeritics ASAP 2010 instrument using a conventional volumetric technique. Prior to analysis the samples were outgassed at 400°C for 4 h at pressures below 1:33 10 3 Pa. The Si/Al ratios were determined by means of laser ablation, inductively coupled plasma mass spectrometry (LA-ICP-MS) [15]. An excimer laser (193 nm) (MicroLas GmbH) was attached to an ELAN 6100 (Perkin Elmer) mass spectrometer for direct solid sampling. The zeolites were pressed into self-supporting pellets and placed in an airtight ablation cell. The laser sampled 80 lm of the zeolite (10 Hz, 120 mJ, 5 replicates per sample), and the produced aerosol was transported into the mass spectrometer [16]. The concentration ratio was determined by comparison with the reference glass 610 from NIST as the external calibration standard. Two zeolite samples were digested and analyzed as solutions by ICP optical emission spectroscopy (ICP-OES). The data from the direct solid-sampling technique (LA-ICP-MS) and from
3. Results and discussion The results of the chemical analysis and nitrogen adsorption are summarized in Table 1. Elemental analysis of the calcined and steamed samples revealed little change in the bulk composition as compared to the parent HZSM-5. The original bulk Si/Al ratio of 19.1 was maintained after calcination (19.3) and increased slightly after steaming (20.1). On the contrary, analysis of the acid-treated samples indicated that the bulk composition of the zeolite changed during the treatment with acid. The Si/Al ratio of the calcined sample increased to 25.3 and that of the steamed sample to 38.3, indicating that leaching of aluminum occurred during the acid treatment. Analysis of the solutions of the acid treatment revealed that dissolution of silicon took place as well. Of the original silicon content of the calcined and steamed zeolites, 2.2 wt.% and 3.8 wt.%, respectively, were dissolved during the acid treatment. Nitrogen adsorption measurements showed that a secondary pore system did not form after the various treatments. The micropore volume remained unchanged, and the pore volume distribution of the meso- and macropore region (2±100 nm), calculated according to the BJH method, showed that there was no change in the textural properties. The BET surface area of the parent material HZSM-5 was not appreciably modi®ed as a result of the subsequent treatments. X-ray powder diraction patterns are shown in Fig. 1. They correspond to those expected for
Table 1 Nomenclature and properties of the HZSM-5 zeolites Sample
Treatment
Si/Al (0:1)
BET surface area 5 m2 =g
Micropore volume (cm3 =g)
HZSM-5 HZSM-5(T) HZSM-5(T)HCl HZSM-5(H) HZSM-5(H)HCl
Calcination, 600°C, 48 h 2 M HCl, 100°C, 120 h Steaming, 500°C, 6 h, 33 ml/min 2 M HCl, 100°C, 120 h
19.1 19.3 25.3 20.1 38.3
429 406 416 381 400
0.16 0.15 0.15 0.15 0.16
180
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
Fig. 1. XRD patterns of (a) HZSM-5, (b) HZSM-5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM-5(H)HCl .
samples of MFI topology. The diractograms indicate that crystallinity was retained after every single treatment. Fig. 2 shows a quantitative comparison of the 27 Al MAS NMR spectra of the HZSM-5 zeolites. The spectrum of the parent HZSM-5 (Fig. 2a) is characterized by an intense and sharp signal at 53 ppm due to aluminum species in tetrahedral coordination (AlIV ). An additional weak signal is observed at 0 ppm, assigned to octahedrally co-
Fig. 2. 27 Al MAS NMR spectra of (a) HZSM-5, (b) HZSM5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H), (e) HZSM-5(H)HCl and (f) silica±alumina. Asterisks indicate spinning side bands.
ordinated EFAl species (AlVI ), either present in the original samples or formed during the preliminary thermal treatments of the ZSM-5. After thermal (Fig. 2b) or hydrothermal treatment (Fig. 2d), the 27 Al tetrahedral aluminum signal decreased in intensity, indicating that removal of aluminum from the lattice occurred. Table 2 reports the results of the quantitative analysis of the deconvoluted spectra. They show that only a limited extent of tetrahedral aluminum was extracted from the framework in the case of the calcined material HZSM-5(T), while almost 50% loss was observed for the hydrothermally treated HZSM-5(H) sample. These results agree with the known resistance of HZSM-5 to dealumination and with the fact that the degree of dealumination is greater in the presence of steam [17]. At the same time, the intensity of the peak at 0 ppm increased and a feature appeared at 30 ppm. The latter signal has been assigned to highly distorted tetrahedral [18] as well as to ®ve-coordinated aluminum species [19,20] (because of the ambiguity of the assignment, we will refer to this signal as Al30 ppm ). It is worthwhile noticing that the dealuminated samples HZSM-5(T) and HZSM-5(H) do not contain ``NMR invisible'' aluminum, as it is inferred from the fact that the total aluminum content did not change after dealumination (see Table 2). The spectra after acid treatment of the samples dealuminated by calcination or steaming are shown in Fig. 2c and e, respectively. In contrast to the results reported by Sano et al. [12], acid treatment of the dealuminated zeolites did not lead to the recovery of the intensity of the tetrahedral aluminum signal. On the contrary, a quantitative comparison of the areas of the signal at 53 ppm before and after acid treatment (see Table 2) showed a decrease in the amount of tetrahedrally coordinated aluminum species. This result indicates that the acid treatment did not result in the reinsertion of non-framework aluminum into the lattice, but rather to the further dissolution of structural aluminum. Our 27 Al MAS NMR results clearly show that the main eect of the acid treatment is the leaching of EFAl species. This is demonstrated by the decrease in the amount of the species responsible for the signal at 0 ppm on HZSM-5(T)HCl and HZSM-5(H)HCl (see Table 2 and Fig. 2c and e).
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
181
Table 2 Concentration of the dierent aluminum and proton species in the dierent environments on the zeolites HZSM-5 Zeolite
AlIV lmol=g 27 Al NMR
AlVI lmol=g 27 Al NMR
Al30 ppm lmol=g 27 Al NMR
Si(OH) lmol=g 1 H NMR
Si(OH)Al lmol=g 1 H NMR
HZSM-5 HZSM-5(T) HZSM-5(T)HCl HZSM-5(H) HZSM-5(H)HCl
406 386 348 217 179
21 33 19 69 27
± 38 81 111 95
256 197 237 118 221
420 323 288 83 72
This is not surprising since the acid treatment is usually applied to dissolve EFAl that is occluded in the pores of dealuminated zeolites. The extent of the dissolution was higher in the case of the steamed material. Contrary to what was reported by Sano et al. [13], HCl treatment of HZSM-5(T) led to an increase in the concentration of the species characterized by the signal at 30 ppm. Therefore, it would not be possible for realumination to occur through the reinsertion of these species. A slight decrease was observed in the case of the steamed material. In conclusion, the 27 Al MAS NMR characterization of the acid-treated samples showed the following. (i) Octahedral aluminum species were washed out, although not completely. (ii) An additional fraction of tetrahedral aluminum was removed from the lattice. (iii) In the case of the thermally dealuminated sample, some tetrahedrally coordinated aluminum probably changed into penta-coordinated or tetrahedrally distorted species. 1 H MAS NMR further supported the conclusion that non-framework aluminum was not reinserted into the lattice of dealuminated HZSM-5 zeolites during acid treatment. In Fig. 3 the 1 H MAS NMR spectra of the dierent samples are quantitatively compared. At least three signals are present in the spectrum of the parent HZSM-5 (see Fig. 4), with maxima at 1.3, 1.8 and 3.4 ppm. The peak at 1.3 ppm is due to protons of silanols, the signal at 3.4 ppm is due to bridged Si(OH)Al groups, and the component at 1.8 ppm is assigned to non-acidic hydroxyls which are bonded to EFAl species [21]. There is a broad feature between 4 and 8 ppm, which may be due to H-bonded silanols [22,23] and/or Si(OH)Al groups interacting with
Fig. 3. 1 H MAS NMR spectra of (a) HZSM-5, (b) HZSM5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM5(H)HCl .
Fig. 4. Computer deconvolution of the 1 H MAS NMR spectrum of HZSM-5: experimentally observed spectrum (Ð); resolved components ( ); ®t (- - -).
182
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
the framework oxygens via H-bonds [14,24]. Upon dealumination, the concentration of Brùnsted acidic protons declined remarkably (see Fig. 3b and d and Table 2). The extent of the decrease was much greater in the case of the steamed sample (Fig. 3d). This con®rms again that the degree of dealumination by hydrothermal treatment is higher compared to the mere thermal treatment. The area of the silanol peak at 1.3 ppm (Fig. 3b and d) diminished, suggesting that silicon migrates to the tetrahedral vacancies generated by the dealumination of the framework. In both dealuminated samples HZSM-5(T) and HZSM-5(H), the amount of tetrahedral aluminum species (27 Al NMR) exceeds that of the Brùnsted acid centers determined by 1 H NMR (see Table 2). For the HZSM-5(H) sample this dierence is quite large. Since every aluminum atom in the zeolite lattice is associated with a Si(OH)Al Brùnsted center, some tetrahedral aluminum in the dealuminated samples must be present as EFAlIV , either as aluminum (hydr)oxide or as silica±alumina. The formation of EFAlIV associated with an amorphous silica±alumina phase after dealumination has been described for zeolites Y [19,25] and ZSM-5 [8]. Silica±alumina possesses aluminum in tetrahedral environment, as revealed by the presence of a signal at 53 ppm in the 27 Al NMR spectrum (see Fig. 2f). On the basis of these considerations, we believe that the peak at 53 ppm on the 27 Al NMR spectrum of the dealuminated samples is due to the combination of the signals of framework aluminum and a new EFAlIV , possibly associated with an amorphous silica±alumina phase formed from the reaction between the extracted silicon and aluminum. Acid treatment after dealumination clearly led to a decrease of the concentration of the Brùnsted acid centers (Fig. 3c and e and Table 2), con®rming that even more structural aluminum was removed from the zeolite matrix. Also in the case of the acid-treated samples, the amount of tetrahedral aluminum exceeded the amount of bridged Si(OH)Al groups, thus revealing the formation of new EFAlIV species. At the same time, the amount of silanol groups (peak at 1.4 ppm) increased, indicating the possible formation of an amorphous phase of silica±alumina from the reaction between
the extracted aluminum and silicon in solution. The formation of silico-aluminates after washing was ®rst proposed by Lutz et al. for dealuminated zeolite Y treated with an alkaline solution [9,10]. The increase of the silanol population may also be due to the formation of a new silica phase. Acid treatment may dissolve alumina of the aluminumrich silica±alumina debris, thus leading to silica species. Another signal of weak intensity was present in the spectra of the acid-treated samples at 0 ppm and represents non-acidic unperturbed hydroxyl groups, which are bonded to EFAl species [14]. The corresponding FTIR spectra are quantitatively compared in Fig. 5. There are at least four bands at 3743, 3667, 3610 and 3550±3300 cm 1 in the FTIR spectrum of the parent HZSM-5 zeolite (Fig. 5a). The IR spectrum is the mirror image of the corresponding 1 H NMR spectrum. There is general agreement that the band at 3743 cm 1 corresponds to the O±H stretching mode of isolated silanols, while that at 3610 cm 1 is due to bridged Si(OH)Al groups [26±29]. The weak feature at 3667 cm 1 is assigned to hydroxyl groups attached to EFAl oxide species [30,31]. Finally, the broad absorption around 3550±3300 cm 1 was assigned to hydrogen-bonded silanols [22,23] and/ or bridged Brùnsted acid groups [14,24]. Because of the similarity between FTIR and 1 H NMR spectroscopy, it is not surprising that the evolution
Fig. 5. FTIR spectra in the OH stretching spectral region of (a) HZSM-5, (b) HZSM-5(T), (c) HZSM-5(T)HCl , (d) HZSM-5(H) and (e) HZSM-5(H)HCl .
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
of the FTIR spectra pro®le of HZSM-5 upon subsequent treatments is consistent with the trend found in 1 H NMR. However, no quantitative analysis of the FTIR spectra was attempted, since IR intensities are strongly in¯uenced by the different environments (molar extinction coecients vary for dierent families of hydroxyl groups). On the basis of the comparison of the spectra of the acid-treated HZSM-5 samples, the following conclusions can be drawn. (i) Leaching of EFAl took place, as evidenced by the decrease of the band at 3667 cm 1 . (ii) The band at 3743 cm 1 increased, suggesting formation of structural defects and the formation of a new silica phase. (iii) The intensity of the band at 3660 cm 1 (Brùnsted acid sites) decreased slightly. The 29 Si MAS NMR spectra and their resolved lines of the parent HZSM-5, the dealuminated HZSM-5(T) and the acid-treated HZSM-5(T)HCl samples are shown in Fig. 6. Four peaks were distinguished and assigned as follows. The signals at 112.6 and 106.7 ppm were attributed to the resonance of the silicon atoms in the SiO4 tetrahedra surrounded by four (4Si, 0Al) and three (3Si, 1Al) silicon atoms, respectively. The peak at 103 ppm was assigned to silanols of (Si(OH)1 (±OSi)3 ) units, while the shoulder at 116.1 ppm was attributed to the presence of crystallographically inequivalent sites in the zeolite [32,33]. 29 Si MAS NMR cannot provide quantitative information about the framework transformation, since the three spectra exhibit very similar patterns and only show minor dierences in peak intensities. Therefore, only qualitative conclusions can be drawn regarding the feasibility of realumination. Thermal treatment (Fig. 6b) did not appreciably aect the area of the signal at 106.7 ppm associated to Brùnsted acid sites. As we showed above, a more severe treatment is necessary to remove a consistent amount of aluminum from the framework. Acid treatment after thermal dealumination led to the spectrum of Fig. 6c. If a realumination had occurred, the amount of framework aluminum species should have increased. However, no signi®cant change in the area of the signal at 106.7 ppm, associated with these species, was observed, con®rming that no reinsertion of aluminum into the framework has occurred.
183
Fig. 6. 29 Si MAS NMR spectra and decomposition into the dierent component lines: (a) HZSM-5, (b) HZSM-5(T) and (c) HZSM-5(T)HCl .
4. Conclusions The results of our characterization are in clear contrast to the proposal by Sano et al. [12,13] regarding the feasibility of the reinsertion of nonframework aluminum into vacant framework sites by treatment with an aqueous solution of HCl. Investigation by FTIR and multinuclear NMR spectroscopy of the samples prepared under the same conditions did not show any evidence of reinsertion of aluminum into the lattice. The extent of the dealumination of the zeolite HZSM-5 under hydrothermal conditions was found to be much greater than under dry conditions. As a result of the dealumination, the extracted framework
184
A. Omegna et al. / Microporous and Mesoporous Materials 46 (2001) 177±184
tetrahedral aluminum changed into the following non-framework species: (i) hexa-coordinated, associated to the signal at 0 ppm in the 27 Al NMR; (ii) highly distorted tetrahedral or penta-coordinated (signal at 30 ppm); (iii) tetrahedral, either associated to aluminum (hydr)oxide or to an amorphous phase of silica±alumina. The latter species gives a signal at 53 ppm in the 27 Al NMR spectrum, which overlaps with the signal of the framework tetrahedral aluminum. Upon acid treatment of the dealuminated materials, leaching of extra-lattice octahedral aluminum was observed, as expected. Moreover, an additional fraction of structural aluminum was extracted. Part of this extracted aluminum probably changed into penta-coordinated or tetrahedrally distorted aluminum species (in the case of the HZSM-5 dealuminated by steaming), and part reacted with the extracted silicon in solution to form further amounts of amorphous silica±alumina. Some silica phase could also be present in the acid-treated samples, originating from the silica±alumina debris after dissolution of the alumina. Acknowledgements The authors thank Prof. D.G unther for carrying out the elemental analysis. References [1] L. Bertea, H.W. Kouwenhoven, R. Prins, Appl. Catal. A: Gen. 129 (1995) 229. [2] D.W. Breck, G.W. Skeels, in: L.V.C. Rees (Ed.), Proceedings 5th International Zeolite Conference, Heyden, London, 1980, p. 335. [3] X. Liu, J. Klinowski, J.M. Thomas, J. Chem. Soc., Chem. Commun. (1986) 582. [4] H. Hamadan, B. Sulikowski, J. Klinowski, J. Phys. Chem. 93 (1989) 350. [5] Z. Zhang, X. Liu, Y. Xu, R. Xu, Zeolites 11 (1991) 232. [6] V. Calsavara, E. Falabella Sousa-Aguiar, N.R.C. Fernandes Machado, Zeolites 17 (1996) 340.
[7] B. Sulikowski, J. Datka, B. Gil, J. Ptaszynski, J. Klinowski, J. Phys. Chem. B 101 (1997) 6929. [8] Y.C. Long, M.Y. Jin, Y.J. Sun, T.L. Wu, L.P. Wang, L. Fei, J. Chem. Soc., Faraday Trans. 92 (1996) 1647. [9] W. Lutz, W. Gessner, D. M uller, Zeolites 19 (1997) 209. [10] W. Lutz, W. Gessner, R. Bertram, I. Pitsch, R. Fricke, Micropor. Mater. 12 (1997) 131. [11] L. Heeribout, R. Vincent, P. Batamack, C. DoremieuxMorin, J. Fraissard, Catal. Lett. 53 (1998) 23. [12] T. Sano, R. Dtadenuma, Z. Wang, K. Soga, Chem. Commun. (1997) 1945. [13] T. Sano, Y. Uno, Z.B. Wang, C.H. Ahn, K. Soga, Micropor. Mesopor. Mater. 31 (1999) 89. [14] M. M uller, G. Harvey, R. Prins, Micropor. Mesopor. Mater. 34 (2000) 281. [15] D. G unther, S.E. Jackson, H.P. Longerich, Spectrochim. Acta B 54 (1999) 381. [16] D. G unther, R. Frischknecht, C.A. Heinrich, J. Anal. At. Spectrom. 12 (1997) 939. [17] S.M. Campbell, D.M. Bibby, J.M. Coddington, R.F. Howe, R.H. Meinhold, J. Catal. 161 (1996) 338. [18] A. Samoson, E. Lippmaa, G. Engelhardt, U. Lohse, H.G. Jerschkewitz, Chem. Phys. Lett. 134 (1987) 589. [19] J. Sanz, V. Fornes, A. Corma, J. Chem. Soc., Faraday Trans. 84 (1988) 3113. [20] J.P. Jilson, G.C. Edwards, A.W. Peters, K. Rajagopalan, R.F. Wormsbecker, T.G. Roberie, M.P. Shatlock, J. Chem. Soc., Chem. Commun. (1987) 91. [21] M. Hunger, Catal. Rev. Sci. Engng. 39 (1997) 345. [22] V.L. Zholobenko, L.M. Kustov, V.Y. Borovkov, V.B. Kazansky, Zeolites 8 (1988) 175. [23] E. Bourgeat-Lami, P. Massiani, P. Espiau, F.D. Renzo, F. Fajula, Appl. Catal. 72 (1991) 139. [24] C. Paze, A. Zecchina, S. Spera, A. Cosma, E. Merlo, G. Span o, G. Girotti, Phys. Chem. Chem. Phys. 1 (1999) 2627. [25] A. Janin, M. Maache, J.C. Lavalley, J.F. Joly, F. Raatz, N. Szydlowski, Zeolites 11 (1991) 391. [26] G. Qin, L. Zheng, Y. Xie, C. Wu, J. Catal. 95 (1985) 609. [27] J. Dwyer, Stud. Surf. Sci. Catal. 37 (1988) 333. [28] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Petrini, G. Leofanti, M. Padovan, C. Otero Arean, J. Chem. Soc., Faraday Trans. 88 (1992) 2959. [29] H. Kn ozinger, S. Huber, J. Chem. Soc., Faraday Trans. 94 (1998) 2047. [30] L.M. Kustov, V.B. Kazansky, S. Beran, L. Kubelkova, P. Jiru, J. Phys. Chem. 91 (1987) 5247. [31] A. Zecchina, C. Otero Arean, Chem. Soc., Rev. 25 (1996) 187. [32] E. Brunner, H. Ernst, D. Freude, J. Catal. 127 (1991) 34. [33] W. Zhang, X. Bao, X. Guo, X. Wang, Catal. Lett. 60 (1999) 89.