- Email: [email protected]
Structural and chemical influences on the MTO reaction: a comparison of chabazite and SAPO-34 as MTO catalysts
Microporous and Mesoporous Materials 29 (1999) 185–190
Structural and chemical influences on the MTO reaction: a comparison of chabazite and SAPO-34 as MTO catalysts Ivar M. Dahl, Helle Mostad *, Duncan Akporiaye, Rune Wendelbo SINTEF Applied Chemistry, P.O. Box 124, Blindern, N-0314 Oslo, Norway Received 31 March 1998; received in revised form 16 October 1998; accepted 19 October 1998
Abstract A comparison of isostructural H-SAPO-34 and H-chabazite in high space-velocity experiments of the methanol to olefins (MTO) reaction has been performed. The initial selectivity was shown to be similar for both catalysts, and also to be independent of acid site density. However, the deactivation rate was critically dependent on the acid site density. Thus, low acid-site density samples exhibited longer lifetimes at low space velocity conditions than their high acid-site density analogues. © 1999 Elsevier Science B.V. All rights reserved. Keywords: CHA; MTO; Reaction mechanism; SAPO
1. Introduction Since Mobil discovered the MTO/MTG reaction over zeolite-type catalysts, several zeolites and SAPO materials have been examined as catalysts in this reaction [1–3]. Selectivity patterns are clearly determined by geometrical restrictions. Hydrogen transfer and the production of aromatics vary widely for different materials. Comparisons have been made for different SAPOs [1] and for isostructural SAPOs and zeolites [2,3] in the MTO reaction. A ‘‘hydrocarbon pool’’ mechanism has previously been proposed for the MTO reaction over SAPO-34 [4–6 ]. This model implies a dynamic situation in which large carbonaceous species build * Corresponding author. Tel: +47 22 067692; Fax: +47 22 067350. E-mail address: [email protected]
up inside the cages of the structure. These carbonaceous compounds are continuously adding reactants and splitting off products. Test experiments on SAPO and zeolite materials have previously been performed at low feed rates, resulting in complete conversion of the methanol/dimethyl ether feedstock [1–4]. More than 50% conversion of methanol has been demonstrated at space velocities above 300 h−1 over SAPO-34 for a short period [7]. Thus, at space velocities below 1 h−1, a large surplus of activity exists in addition to what is necessary to achieve a 95% conversion of methanol. This excess activity will result in secondary reactions, which slowly brings the product composition towards chemical equilibrium. Thus, to determine the primary selectivity pattern of the MTO reaction, it is important that the investigations be performed at conditions where the secondary reactions are of minor importance. To minimize the influence of secondary
1387-1811/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S1 3 8 7 -1 8 1 1 ( 9 8 ) 0 0 33 0 - 8
186
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
reactions, a sufficiently high space velocity has to be used in the comparison of the isostructural SAPO-34 and chabazite in the MTO reaction. This is the purpose of the present work. 2. Experimental 2.1. Synthesis Chabazite (CHAB AS ) was synthesized according to Zones and Van Nordstrand [8,9] using the recrystallisation of zeolite Y with adamantyl trimethyl ammonium iodide. The reagents used were NaOH (Aldrich), sodium silicate ( Kebo) and LZ-Y62 zeolite Y (Si/Al=2.47) as the alumina source. A typical gel oxide ratio of Al O : 2 3 80SiO :25Na O:620H O:3.5R was employed (R= 2 2 2 adamantyl trimethyl ammonium iodide= ATMAI ). The NaOH was first dissolved in a specified amount of water. To this solution was added zeolite Y and then sodium silicate. After stirring, the ATMAI was carefully added to obtain a relatively homogeneous mixture. The whole mixture was transferred to 50 ml autoclaves rotating axially at 15 rpm in heated ovens. Crystallization was carried out at 150°C for 120 h. After the synthesis of another batch of CHAB AS, this was dealuminated twice to obtain a high-Si sample (CHAB DA2). Approximately 1 g of a synthetic high-silica material was calcined at 550°C to remove the organic amine before carrying out a double ion-exchange treatment with 2 M ammonium nitrate. This material was dried overnight at 100°C before steam treatment at 973 K for 16 h. Repetition of the ion-exchange/drying/steaming procedure was followed by a final wash with hydrochloric acid. The ‘‘normal’’-silica H-SAPO-34 sample (SAPO-34 NS) was prepared according to the procedure of Lok et al. [10]. The low-silica H-SAPO-34 sample (SAPO-34 LS ) was synthesized by Professor Lillerud and his group at the University of Oslo [11]. 2.2. Characterization X-ray powder diffractograms ( XRD) were obtained using a Siemens D5000 instrument with Ni-filtered Cu Ka radiation.
The MAS NMR spectra were obtained using a Varian VXR 30 S spectrometer operating at a proton resonance frequency of 300 MHz. The instrument was equipped with both Jakobsen and Doty MAS probes. The conditions for the acquisition of the 29Si MAS NMR spectra were as follows: 29Si frequency=59.6 MHz, sweep width= 14 000 Hz, pulse width=8 ms (90° pulse, 8.3 ms), repetition time=5 s, acquisition time=1 s, number of scans=1000, MAS spinning speed=4.5 kHz, reference to Me Si. 4 Scanning electron micrographs (SEM ) were taken using a JEOL JSM-840 instrument. The dry powder catalyst samples were mounted on a tape and coated with a thin layer of Au. Chemical analyses of the samples were carried out using an electron microprobe on a Cameca instrument (model Camebax microbeam) equipped with three wavelength-dispersive spectrometers and a quantitative LINK energy dispersive analyzer ( EDA). TPD measurements were performed on an AMI-1 instrument from Altamira Instruments Inc. 200 mg of catalyst with a particle size of 0.2–0.5 mm was used. TPD of ammonia was carried out according to the following procedure. (1) Pretreatment with 30 cm3 min−1 He flow up to 600°C, saturation of the catalyst surface with 30 cm3 min−1 10 vol.% ammonia in a He flow at 200°C, flushing with 30 cm3min−1 He flow at 200°C for 30 min. (2) TPD at increasing temperature from 200 to 600°C at 20°C min−1 (30 cm3 min−1 He flow) and then standby for 15 min at 600°C. The high space-velocity MTO reaction experiments were performed in a fixed bed of 30 mg of calcined catalyst with a 0.2–0.4 mm particle size. A He flow of 220 cm3 min−1 was maintained with a mass flow controller. The MeOH content in the feed gas was 9.1 vol.%, and thus a WHSV of about 70 g MeOH per g of catalyst per hour could be calculated. The reaction temperature was 400°C. In order to follow the reaction at short time intervals, the reactor effluent was analyzed on an on-line Fisons Sensolab mass spectrometer (MS). The mass spectrometric conditions were electron impact at 70 eV, 275°C and 4–100 amu. The mass spectrometric analyses of the reactor effluent were
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
performed by monitoring selected specific masses (m/e), i.e. propene=42 and butene=56, measured relative to the m/e value of He (m/e=4). As propene is the main reaction product at the chosen conditions and the selectivity to propene is similar throughout a run [4], the intensity of the m/e=42 ion signal was used as a measure of the product formation rate. The GC analyses were performed on a Hewlett Packard 5890 instrument equipped with a 50 m Carboplot column and an FID detector.
3. Results and discussion 3.1. Catalytic materials In order to compare SAPO-34 and chabazite materials with equal acid site densities in the MTO reaction, the synthesis of chabazites with sufficiently low amount of Al (i.e. Si/Al ratios in the region 7–30) is important. Cartlidge et al. [3] have synthesized chabazites with Si/Al ratios of 3–4, and performed further dealumination by ionexchange/steaming/acid washing to obtain a Si/Al ratio of 10.5. Yuen et al. [2] have synthesized and examined the MTO behavior of chabazites with claimed Si/Al ratios of 5–30. We were unable to synthesize chabazite with a higher Si/Al ratio than 8.5 with the method of Yuen et al. [2], but this ratio may be increased to 37 by the subsequent steam dealumination/ion-exchange procedure [12]. The properties of the materials tested are summarized in Table 1. As can be seen in Table 1, the catalyst samples obtained have about the same crystal size. In order
187
to compare the acid site density of the SAPOs and the chabazites, there is a need for a common parameter for SAPOs and zeolites. Thus, the TAS (tetrahedral atoms perr acid site) parameter is advantageous to the Si/Al ratio of the zeolites (the Si/Al ratios were equal to 7 and 25 for CHAB AS and CHAB DA2, respectively) and the (Al+P)/Si ratio of the SAPOs (the (Al+P)/Si ratios were equal to 10 and 30 for SAPO-34 NS and SAPO-34 LS, respectively). Furthermore, the calculated acid site densities from the 29Si NMR data of the chabazites were about twice the acid site densities calculated by NH TPD. This is probably due to 3 the presence of non-zeolitic SiO [12,13] in the 2 high-silica chabazites synthesized according to the present procedure. Thus, the TAS numbers (from NMR) are typical of the acid site density in the zeolite phase, while the NH TPD results represent 3 the total number of acid sites in a sample. A much better agreement was observed between the calculated and measured acid site density of the SAPO-34s, indicating that these samples consist of nearly pure SAPO-34.
3.2. Catalytic testing The relative intensities of the product-specific MS traces are described in Fig. 1. The propene selectivity was observed to remain fairly constant during an MTO run. Thus, the ratio between the intensity of the mass number typical for propene (m/e=42) and the mass number of the inert gas He (m/e=4) was used as a relative measure of the production rate of propene (r42 in Figs. 1–3). Moreover, the same ratio was used as a relative
Table 1 Characteristics of the catalysts tested Catalyst
TASa,b
Crystal size (mm)
Calculated acid site densityb (mmol g−1) Measured acidity from NH TPD (mmol g−1) 3
CHAB AS CHAB DA2 SAPO-34 NS SAPO-34 LS
8 26 11 31
0.2–2 <1 0.5–1 0.2–0.8
2.1 0.6 1.5 0.5
0.9 0.3 1.3 0.5
aTAS=tetrahedral atoms per potentially acid site, i.e. (Si+Al )/Al for chabazite and (Si+Al+P)/Si for SAPO-34. bCalculated from 29Si NMR analysis for the chabazites [13] and microprobe analysis for the SAPO-34 samples.
188
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Fig. 1. The intensity of the m/e=42 ion (relative to the m/e=4 ion) as a function of time after the admission of methanol to the catalysts.
measure of the total conversion of methanol to hydrocarbons. With the exception of the chabazite sample with the highest Al content, an activation period was observed for all the other catalysts (Fig. 1). This activation period is in agreement with a ‘‘hydrocarbon pool’’ mechanism of the MTO reaction over SAPO-34, and may represent the time needed to build up the active intermediates. A comparison of the amounts of products from the four catalysts with time on-stream is given in Fig. 2. The amounts of hydrocarbons produced were calculated from the GC analysis. The percentage of hydrocarbons (as C equivalents) in the 1 product stream after the condensation of water was used. The selectivity of the various catalysts was observed to be fairly constant during a run (precision of about 2%) ( Table 2). The differences in selectivity observed between chabazite and SAPO-34 mayt be significant. It seems that chabazite has a somewhat higher ethene selectivity than SAPO-34 and a correspondingly lower propene selectivity. In our experience, however, similar differences in selectivity (5%) may also be found in SAPO-34 samples of different origin. In these short contact-time experiments we do not observe any significant amount of propane (detection limit of 1% because of incomplete resolution of the propane and propene peaks). This is in contrast to the results of Yuen et al. [2], which find nearly 20% propane selectivity for a chabazite with Si/Al=4.5 at an LHSV of 0.27 h−1. With this
high contact time, secondary reactions determine the selectivity pattern. The most striking feature in Fig. 2 is the difference in the deactivation rate behavior of the catalysts. The maximum conversion obtained seems to be correlated to the total acidity of the catalyst. For both catalyst samples with the highest content of potential acid sites (SAPO-34 NS and CHAB AS ), a high conversion was observed in the beginning of the experiment with a rather steep deactivation profile of the curves. On the other hand, the conversion over the two samples with the lowest number of acid sites (SAPO-34 LS and CHAB DA2) was on a lower level, but the observed deactivation rates were reduced. It appears from the results presented in Fig. 2 that the differences in deactivation rates depend more on the density of acid sites than on the differences in chemical composition of SAPO0-34 and chabazite. If the deactivation of the catalysts obeys firstorder kinetics in the number of acid sites, linear semi-logarithmic plots would be expected at sufficiently low conversions, reaching an asymptotic value at full conversion. Semi-logarithmic plots of the ratio between the MS m/e=42 and m/e=4, which is a relative measure of the oxygenate conversion with reaction time of the four investigated catalysts, is illustrated in Fig. 3. For a first-order deactivation reaction, straight lines in these plots at below 20% conversion are expected, leveling off at higher conversions. The observed curvature of the plots in Fig. 3 indicates higher-order deactivation rates in site concentrations at high activities. This is particularly significant for the chabazite sample with the highest acid site density. Moreover, this deactivation rate is not a simple function of the activity, as the samples with a higher initial activity have a larger deactivation rate at a comparable activity. Thus, at sufficiently long times on-stream, the catalysts with the highest densities of acid sites have a lower activity than their low acid-site density counterparts. Consequently, ranking the catalytic properties of these materials must be performed with great care, as the ranking will depend heavily on which test conditions and ranking criteria are chosen. We can see two possible causes for the rapid deactivation at high acid site density. (1) If the
189
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Fig. 2. Deactivation of SAPO-34 NS, SAPO-34 LS, CHAB AS and CHAB DA2 measured from the gas chromatographic data as a percentage of hydrocarbons in the product stream.
reaction is diffusion-limited in the products, a highactivity catalyst will have a high internal concentration of olefins, which will increase the rate for higher-order reactions, resulting in aromatic coke formation. At comparable activity levels, the sample with a high acid-site density will have more coke, and accordingly greater diffusion resistance and higher internal olefin concentration, which accounts for the higher deactivation rate. (2) Assuming that the reaction intermediate is locked
in one cage of the structure, the number of reaction intermediates with access to more than one acid site increases rapidly with the acid site density. If two or more acid sites can act in a concerted way to effect hydrogen transfer reactions and coke formation, this may lead to the observed effect. Depending on which of the four oxygen atoms around an Al position the proton is associated with, the acid site of the chabazite structure may in principle be active in three cages. For a random
Table 2 Olefin selectivity Catalyst
SAPO-34 NS
SAPO-34 NS
SAPO-34 LS
SAPO-34 LS
CHAB AS
CHAB AS
CHAB DA2
CHAB DA2
TOS (min) C= 2 C= 3 C= 4
13 35 48 17
133 30 50 19
18 32 50 17
135 35 48 16
7 40 40 17
62 36 40 24
7 35 44 19
123 37 42 20
190
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Acknowledgements The authors wish to express their gratitude to Terje Fuglerud and Steinar Kvisle for valuable discussions, and to Anne Andersen for skillful technical assistance in the synthesis of the chabazites. We also wish to thank Professor Karl Petter Lillerud for supplying the SAPO-34 LS sample. Financial support of this work by the Norwegian Research Council is gratefully acknowledged.
Fig. 3. Data from Fig. 1 in a logarithmic plot. %: SAPO-34 NS, n: SAPO-34 LS, 6: CHAB AS, #: CHAB DA2.
incorporation of acid sites, there is a maximum of cages with access to only one acid site in catalyst materials with a low Si content ( TAS of 36).
4. Conclusion Chabazite and SAPO-34 have very similar properties as MTO catalysts. The acid-site density is the most important parameter for the deactivation behavior. There are also smaller differences between SAPO-34 and chabazite which may be due to their chemical compositions. There is clearly an optimum acid density of these catalyst materials, although this optimum will depend on the process conditions chosen.
References [1] S.M. Yang, S.I. Wang, C.C. Huang, in: A. Holmen, K.J. Jens, S. Kolboe (Eds.), Natural Gas Conversion, Studies in Surface Science and Catalalysis, vol. 61, Elsevier, Amsterdam, 1991, p. 429. [2] L.-T. Yuen, S.I. Zones, T.V. Harris, E.J. Gallegos, A. Auroux, Microporous Mater. 2 (1994) 105. [3] S. Cartlidge, R. Patel, in: P.A. Jacobs, R.A. van Santen ( Eds.), Zeolites: Facts, Figures, Future, Studies in Surface Science and Catalysis, vol. 49, Elsevier, Amsterdam, 1989, p. 1151. [4] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329. [5] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458. [6 ] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304. [7] D. Chen, K. Moljord, T. Fuglerud, A. Holmen, Microporous Mesoporous Mater. 29 (1999) 191 (this issue) [8] S.I. Zones, R.A. Van Nordstrand, Zeolites 8 (1988) 166. [9] S.I. Zones, US Patent 4 665 110, 1987. [10] B.M. Lok, C.A. Messina, R.L. Patton, R.T. Gajek, T.R. Cannan, E.M. Flanigen, US Patent 4 440 871, 1984. [11] E.N. Halvorsen, PhD thesis, University of Oslo, 1996. [12] D.E. Akporiaye et al., in preparation. [13] D.E. Akporiaye, I.M. Dahl, H.B. Mostad, R. Wendelbo, J. Phys. Chem. C 100 (1996) 4148–4153.
Structural and chemical influences on the MTO reaction: a comparison of chabazite and SAPO-34 as MTO catalysts Ivar M. Dahl, Helle Mostad *, Duncan Akporiaye, Rune Wendelbo SINTEF Applied Chemistry, P.O. Box 124, Blindern, N-0314 Oslo, Norway Received 31 March 1998; received in revised form 16 October 1998; accepted 19 October 1998
Abstract A comparison of isostructural H-SAPO-34 and H-chabazite in high space-velocity experiments of the methanol to olefins (MTO) reaction has been performed. The initial selectivity was shown to be similar for both catalysts, and also to be independent of acid site density. However, the deactivation rate was critically dependent on the acid site density. Thus, low acid-site density samples exhibited longer lifetimes at low space velocity conditions than their high acid-site density analogues. © 1999 Elsevier Science B.V. All rights reserved. Keywords: CHA; MTO; Reaction mechanism; SAPO
1. Introduction Since Mobil discovered the MTO/MTG reaction over zeolite-type catalysts, several zeolites and SAPO materials have been examined as catalysts in this reaction [1–3]. Selectivity patterns are clearly determined by geometrical restrictions. Hydrogen transfer and the production of aromatics vary widely for different materials. Comparisons have been made for different SAPOs [1] and for isostructural SAPOs and zeolites [2,3] in the MTO reaction. A ‘‘hydrocarbon pool’’ mechanism has previously been proposed for the MTO reaction over SAPO-34 [4–6 ]. This model implies a dynamic situation in which large carbonaceous species build * Corresponding author. Tel: +47 22 067692; Fax: +47 22 067350. E-mail address: [email protected]
up inside the cages of the structure. These carbonaceous compounds are continuously adding reactants and splitting off products. Test experiments on SAPO and zeolite materials have previously been performed at low feed rates, resulting in complete conversion of the methanol/dimethyl ether feedstock [1–4]. More than 50% conversion of methanol has been demonstrated at space velocities above 300 h−1 over SAPO-34 for a short period [7]. Thus, at space velocities below 1 h−1, a large surplus of activity exists in addition to what is necessary to achieve a 95% conversion of methanol. This excess activity will result in secondary reactions, which slowly brings the product composition towards chemical equilibrium. Thus, to determine the primary selectivity pattern of the MTO reaction, it is important that the investigations be performed at conditions where the secondary reactions are of minor importance. To minimize the influence of secondary
1387-1811/99/$ – see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S1 3 8 7 -1 8 1 1 ( 9 8 ) 0 0 33 0 - 8
186
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
reactions, a sufficiently high space velocity has to be used in the comparison of the isostructural SAPO-34 and chabazite in the MTO reaction. This is the purpose of the present work. 2. Experimental 2.1. Synthesis Chabazite (CHAB AS ) was synthesized according to Zones and Van Nordstrand [8,9] using the recrystallisation of zeolite Y with adamantyl trimethyl ammonium iodide. The reagents used were NaOH (Aldrich), sodium silicate ( Kebo) and LZ-Y62 zeolite Y (Si/Al=2.47) as the alumina source. A typical gel oxide ratio of Al O : 2 3 80SiO :25Na O:620H O:3.5R was employed (R= 2 2 2 adamantyl trimethyl ammonium iodide= ATMAI ). The NaOH was first dissolved in a specified amount of water. To this solution was added zeolite Y and then sodium silicate. After stirring, the ATMAI was carefully added to obtain a relatively homogeneous mixture. The whole mixture was transferred to 50 ml autoclaves rotating axially at 15 rpm in heated ovens. Crystallization was carried out at 150°C for 120 h. After the synthesis of another batch of CHAB AS, this was dealuminated twice to obtain a high-Si sample (CHAB DA2). Approximately 1 g of a synthetic high-silica material was calcined at 550°C to remove the organic amine before carrying out a double ion-exchange treatment with 2 M ammonium nitrate. This material was dried overnight at 100°C before steam treatment at 973 K for 16 h. Repetition of the ion-exchange/drying/steaming procedure was followed by a final wash with hydrochloric acid. The ‘‘normal’’-silica H-SAPO-34 sample (SAPO-34 NS) was prepared according to the procedure of Lok et al. [10]. The low-silica H-SAPO-34 sample (SAPO-34 LS ) was synthesized by Professor Lillerud and his group at the University of Oslo [11]. 2.2. Characterization X-ray powder diffractograms ( XRD) were obtained using a Siemens D5000 instrument with Ni-filtered Cu Ka radiation.
The MAS NMR spectra were obtained using a Varian VXR 30 S spectrometer operating at a proton resonance frequency of 300 MHz. The instrument was equipped with both Jakobsen and Doty MAS probes. The conditions for the acquisition of the 29Si MAS NMR spectra were as follows: 29Si frequency=59.6 MHz, sweep width= 14 000 Hz, pulse width=8 ms (90° pulse, 8.3 ms), repetition time=5 s, acquisition time=1 s, number of scans=1000, MAS spinning speed=4.5 kHz, reference to Me Si. 4 Scanning electron micrographs (SEM ) were taken using a JEOL JSM-840 instrument. The dry powder catalyst samples were mounted on a tape and coated with a thin layer of Au. Chemical analyses of the samples were carried out using an electron microprobe on a Cameca instrument (model Camebax microbeam) equipped with three wavelength-dispersive spectrometers and a quantitative LINK energy dispersive analyzer ( EDA). TPD measurements were performed on an AMI-1 instrument from Altamira Instruments Inc. 200 mg of catalyst with a particle size of 0.2–0.5 mm was used. TPD of ammonia was carried out according to the following procedure. (1) Pretreatment with 30 cm3 min−1 He flow up to 600°C, saturation of the catalyst surface with 30 cm3 min−1 10 vol.% ammonia in a He flow at 200°C, flushing with 30 cm3min−1 He flow at 200°C for 30 min. (2) TPD at increasing temperature from 200 to 600°C at 20°C min−1 (30 cm3 min−1 He flow) and then standby for 15 min at 600°C. The high space-velocity MTO reaction experiments were performed in a fixed bed of 30 mg of calcined catalyst with a 0.2–0.4 mm particle size. A He flow of 220 cm3 min−1 was maintained with a mass flow controller. The MeOH content in the feed gas was 9.1 vol.%, and thus a WHSV of about 70 g MeOH per g of catalyst per hour could be calculated. The reaction temperature was 400°C. In order to follow the reaction at short time intervals, the reactor effluent was analyzed on an on-line Fisons Sensolab mass spectrometer (MS). The mass spectrometric conditions were electron impact at 70 eV, 275°C and 4–100 amu. The mass spectrometric analyses of the reactor effluent were
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
performed by monitoring selected specific masses (m/e), i.e. propene=42 and butene=56, measured relative to the m/e value of He (m/e=4). As propene is the main reaction product at the chosen conditions and the selectivity to propene is similar throughout a run [4], the intensity of the m/e=42 ion signal was used as a measure of the product formation rate. The GC analyses were performed on a Hewlett Packard 5890 instrument equipped with a 50 m Carboplot column and an FID detector.
3. Results and discussion 3.1. Catalytic materials In order to compare SAPO-34 and chabazite materials with equal acid site densities in the MTO reaction, the synthesis of chabazites with sufficiently low amount of Al (i.e. Si/Al ratios in the region 7–30) is important. Cartlidge et al. [3] have synthesized chabazites with Si/Al ratios of 3–4, and performed further dealumination by ionexchange/steaming/acid washing to obtain a Si/Al ratio of 10.5. Yuen et al. [2] have synthesized and examined the MTO behavior of chabazites with claimed Si/Al ratios of 5–30. We were unable to synthesize chabazite with a higher Si/Al ratio than 8.5 with the method of Yuen et al. [2], but this ratio may be increased to 37 by the subsequent steam dealumination/ion-exchange procedure [12]. The properties of the materials tested are summarized in Table 1. As can be seen in Table 1, the catalyst samples obtained have about the same crystal size. In order
187
to compare the acid site density of the SAPOs and the chabazites, there is a need for a common parameter for SAPOs and zeolites. Thus, the TAS (tetrahedral atoms perr acid site) parameter is advantageous to the Si/Al ratio of the zeolites (the Si/Al ratios were equal to 7 and 25 for CHAB AS and CHAB DA2, respectively) and the (Al+P)/Si ratio of the SAPOs (the (Al+P)/Si ratios were equal to 10 and 30 for SAPO-34 NS and SAPO-34 LS, respectively). Furthermore, the calculated acid site densities from the 29Si NMR data of the chabazites were about twice the acid site densities calculated by NH TPD. This is probably due to 3 the presence of non-zeolitic SiO [12,13] in the 2 high-silica chabazites synthesized according to the present procedure. Thus, the TAS numbers (from NMR) are typical of the acid site density in the zeolite phase, while the NH TPD results represent 3 the total number of acid sites in a sample. A much better agreement was observed between the calculated and measured acid site density of the SAPO-34s, indicating that these samples consist of nearly pure SAPO-34.
3.2. Catalytic testing The relative intensities of the product-specific MS traces are described in Fig. 1. The propene selectivity was observed to remain fairly constant during an MTO run. Thus, the ratio between the intensity of the mass number typical for propene (m/e=42) and the mass number of the inert gas He (m/e=4) was used as a relative measure of the production rate of propene (r42 in Figs. 1–3). Moreover, the same ratio was used as a relative
Table 1 Characteristics of the catalysts tested Catalyst
TASa,b
Crystal size (mm)
Calculated acid site densityb (mmol g−1) Measured acidity from NH TPD (mmol g−1) 3
CHAB AS CHAB DA2 SAPO-34 NS SAPO-34 LS
8 26 11 31
0.2–2 <1 0.5–1 0.2–0.8
2.1 0.6 1.5 0.5
0.9 0.3 1.3 0.5
aTAS=tetrahedral atoms per potentially acid site, i.e. (Si+Al )/Al for chabazite and (Si+Al+P)/Si for SAPO-34. bCalculated from 29Si NMR analysis for the chabazites [13] and microprobe analysis for the SAPO-34 samples.
188
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Fig. 1. The intensity of the m/e=42 ion (relative to the m/e=4 ion) as a function of time after the admission of methanol to the catalysts.
measure of the total conversion of methanol to hydrocarbons. With the exception of the chabazite sample with the highest Al content, an activation period was observed for all the other catalysts (Fig. 1). This activation period is in agreement with a ‘‘hydrocarbon pool’’ mechanism of the MTO reaction over SAPO-34, and may represent the time needed to build up the active intermediates. A comparison of the amounts of products from the four catalysts with time on-stream is given in Fig. 2. The amounts of hydrocarbons produced were calculated from the GC analysis. The percentage of hydrocarbons (as C equivalents) in the 1 product stream after the condensation of water was used. The selectivity of the various catalysts was observed to be fairly constant during a run (precision of about 2%) ( Table 2). The differences in selectivity observed between chabazite and SAPO-34 mayt be significant. It seems that chabazite has a somewhat higher ethene selectivity than SAPO-34 and a correspondingly lower propene selectivity. In our experience, however, similar differences in selectivity (5%) may also be found in SAPO-34 samples of different origin. In these short contact-time experiments we do not observe any significant amount of propane (detection limit of 1% because of incomplete resolution of the propane and propene peaks). This is in contrast to the results of Yuen et al. [2], which find nearly 20% propane selectivity for a chabazite with Si/Al=4.5 at an LHSV of 0.27 h−1. With this
high contact time, secondary reactions determine the selectivity pattern. The most striking feature in Fig. 2 is the difference in the deactivation rate behavior of the catalysts. The maximum conversion obtained seems to be correlated to the total acidity of the catalyst. For both catalyst samples with the highest content of potential acid sites (SAPO-34 NS and CHAB AS ), a high conversion was observed in the beginning of the experiment with a rather steep deactivation profile of the curves. On the other hand, the conversion over the two samples with the lowest number of acid sites (SAPO-34 LS and CHAB DA2) was on a lower level, but the observed deactivation rates were reduced. It appears from the results presented in Fig. 2 that the differences in deactivation rates depend more on the density of acid sites than on the differences in chemical composition of SAPO0-34 and chabazite. If the deactivation of the catalysts obeys firstorder kinetics in the number of acid sites, linear semi-logarithmic plots would be expected at sufficiently low conversions, reaching an asymptotic value at full conversion. Semi-logarithmic plots of the ratio between the MS m/e=42 and m/e=4, which is a relative measure of the oxygenate conversion with reaction time of the four investigated catalysts, is illustrated in Fig. 3. For a first-order deactivation reaction, straight lines in these plots at below 20% conversion are expected, leveling off at higher conversions. The observed curvature of the plots in Fig. 3 indicates higher-order deactivation rates in site concentrations at high activities. This is particularly significant for the chabazite sample with the highest acid site density. Moreover, this deactivation rate is not a simple function of the activity, as the samples with a higher initial activity have a larger deactivation rate at a comparable activity. Thus, at sufficiently long times on-stream, the catalysts with the highest densities of acid sites have a lower activity than their low acid-site density counterparts. Consequently, ranking the catalytic properties of these materials must be performed with great care, as the ranking will depend heavily on which test conditions and ranking criteria are chosen. We can see two possible causes for the rapid deactivation at high acid site density. (1) If the
189
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Fig. 2. Deactivation of SAPO-34 NS, SAPO-34 LS, CHAB AS and CHAB DA2 measured from the gas chromatographic data as a percentage of hydrocarbons in the product stream.
reaction is diffusion-limited in the products, a highactivity catalyst will have a high internal concentration of olefins, which will increase the rate for higher-order reactions, resulting in aromatic coke formation. At comparable activity levels, the sample with a high acid-site density will have more coke, and accordingly greater diffusion resistance and higher internal olefin concentration, which accounts for the higher deactivation rate. (2) Assuming that the reaction intermediate is locked
in one cage of the structure, the number of reaction intermediates with access to more than one acid site increases rapidly with the acid site density. If two or more acid sites can act in a concerted way to effect hydrogen transfer reactions and coke formation, this may lead to the observed effect. Depending on which of the four oxygen atoms around an Al position the proton is associated with, the acid site of the chabazite structure may in principle be active in three cages. For a random
Table 2 Olefin selectivity Catalyst
SAPO-34 NS
SAPO-34 NS
SAPO-34 LS
SAPO-34 LS
CHAB AS
CHAB AS
CHAB DA2
CHAB DA2
TOS (min) C= 2 C= 3 C= 4
13 35 48 17
133 30 50 19
18 32 50 17
135 35 48 16
7 40 40 17
62 36 40 24
7 35 44 19
123 37 42 20
190
I.M. Dahl et al. / Microporous and Mesoporous Materials 29 (1999) 185–190
Acknowledgements The authors wish to express their gratitude to Terje Fuglerud and Steinar Kvisle for valuable discussions, and to Anne Andersen for skillful technical assistance in the synthesis of the chabazites. We also wish to thank Professor Karl Petter Lillerud for supplying the SAPO-34 LS sample. Financial support of this work by the Norwegian Research Council is gratefully acknowledged.
Fig. 3. Data from Fig. 1 in a logarithmic plot. %: SAPO-34 NS, n: SAPO-34 LS, 6: CHAB AS, #: CHAB DA2.
incorporation of acid sites, there is a maximum of cages with access to only one acid site in catalyst materials with a low Si content ( TAS of 36).
4. Conclusion Chabazite and SAPO-34 have very similar properties as MTO catalysts. The acid-site density is the most important parameter for the deactivation behavior. There are also smaller differences between SAPO-34 and chabazite which may be due to their chemical compositions. There is clearly an optimum acid density of these catalyst materials, although this optimum will depend on the process conditions chosen.
References [1] S.M. Yang, S.I. Wang, C.C. Huang, in: A. Holmen, K.J. Jens, S. Kolboe (Eds.), Natural Gas Conversion, Studies in Surface Science and Catalalysis, vol. 61, Elsevier, Amsterdam, 1991, p. 429. [2] L.-T. Yuen, S.I. Zones, T.V. Harris, E.J. Gallegos, A. Auroux, Microporous Mater. 2 (1994) 105. [3] S. Cartlidge, R. Patel, in: P.A. Jacobs, R.A. van Santen ( Eds.), Zeolites: Facts, Figures, Future, Studies in Surface Science and Catalysis, vol. 49, Elsevier, Amsterdam, 1989, p. 1151. [4] I.M. Dahl, S. Kolboe, Catal. Lett. 20 (1993) 329. [5] I.M. Dahl, S. Kolboe, J. Catal. 149 (1994) 458. [6 ] I.M. Dahl, S. Kolboe, J. Catal. 161 (1996) 304. [7] D. Chen, K. Moljord, T. Fuglerud, A. Holmen, Microporous Mesoporous Mater. 29 (1999) 191 (this issue) [8] S.I. Zones, R.A. Van Nordstrand, Zeolites 8 (1988) 166. [9] S.I. Zones, US Patent 4 665 110, 1987. [10] B.M. Lok, C.A. Messina, R.L. Patton, R.T. Gajek, T.R. Cannan, E.M. Flanigen, US Patent 4 440 871, 1984. [11] E.N. Halvorsen, PhD thesis, University of Oslo, 1996. [12] D.E. Akporiaye et al., in preparation. [13] D.E. Akporiaye, I.M. Dahl, H.B. Mostad, R. Wendelbo, J. Phys. Chem. C 100 (1996) 4148–4153.