Active Sites in Working Bifunctional GaH-TON Aromatization Catalysts

It is generally agreed that Ga sites are involved in dehydrogenation reaction steps and that the zeolite acid sites catalyze cracking, oligomerization...
0 downloads 0 Views 100KB Size
Download PDF
J. Phys. Chem. B 2006, 110, 18473-18480

18473

Active Sites in Working Bifunctional GaH-TON Aromatization Catalysts: Kinetic Evaluation Dmitry B. Lukyanov* and Tanya Vazhnova Catalysis and Reaction Engineering Group, Department of Chemical Engineering, UniVersity of Bath, Bath BA2 7AY, United Kingdom ReceiVed: May 30, 2006; In Final Form: July 24, 2006

The conversion of light alkanes to high value aromatics proceeds with a high selectivity over bifunctional, gallium (Ga) containing zeolite catalysts. It is generally agreed that Ga sites are involved in dehydrogenation reaction steps and that the zeolite acid sites catalyze cracking, oligomerization, and cyclization reactions. However, understanding of the precise roles of the acid and Ga sites in the reaction mechanisms is significantly hampered since the number of these sites in working catalysts is not known. This paper describes a kinetic approach to evaluation of the acid and Ga active sites in working Ga containing TON zeolite catalysts that relies on the analysis of the rates of formation of the primary products of a n-butane aromatization reaction. Our results show that the rate of ethane formation at low n-butane conversions can be used as a quantitative estimate of acidity in working bifunctional zeolite catalysts and demonstrate, for the first time, a significant decrease in the number of Brønsted acid sites in the Ga containing catalysts under reaction conditions: around 47 and 79% for the catalysts with Ga loading of 1.5 and 2.5 wt %, respectively. We conclude that the reduction in acidity is associated with the formation of catalytically active Ga+ ions and obtain estimates for the number and steady-state turnover activity of the acid and Ga active sites in n-butane transformation. We anticipate that our work will facilitate understanding of the precise roles of the acid and Ga sites in the mechanisms of alkane aromatization and, as a far-reaching implication, will prompt wider use of detailed kinetic studies for the evaluation of active sites in working catalysts.

Introduction The conversion of light alkanes, such as ethane, propane, or n-butane, to high value aromatic hydrocarbons is of considerable industrial importance and forms the basis for a few commercial processes (e.g., UOP/BP Cyclar process,1 Mobil M-2 forming process,2 and IFP/SALUTEC Aroforming process3). This reaction is also interesting from a scientific viewpoint since it represents an example of a complex, multistage reaction catalyzed by bifunctional catalysts that contain active sites of two different types. Many studies have focused on the use of medium-pore MFI zeolites as catalysts for alkane aromatization (according to the International Zeolite Association (IZA), the structure code MFI refers to ZSM-5) and have demonstrated that this reaction can be represented schematically as a sequence of three stages.4-9 With acidic H-MFI catalysts, the first stage, alkane conversion into alkenes, proceeds via two routes: (i) protolytic cracking of C-C and C-H bonds in the alkane molecules and (ii) hydrogen transfer (HT) between the feed alkane and the product alkenes.5-7,10,11 The second stage, alkene interconversion, involves alkene isomerization, oligomerization, and cracking steps.4,9,12,13 The third stage, alkene aromatization, proceeds via a sequence of cyclization and hydrogen transfer steps.6,13-17 According to the stoichiometry of the bimolecular hydrogen transfer aromatization mechanisms, formation of hydrogendeficient hydrocarbons such as aromatics is balanced by formation of hydrogen-rich products such as alkanes.13-17 These alkanes together with the lower alkanes produced in the primary, * Corresponding author. Tel.: +44 1225 383329; fax: +44 1225 385713; e-mail:[email protected].

protolytic cracking steps decrease the maximum obtainable aromatics yield. However, this yield can be increased essentially by using bifunctional, metal containing MFI catalysts instead of the purely acidic H-MFI zeolites. Catalysts with different metals have been investigated extensively (mainly, with gallium, zinc, and platinum), and the reviews of these studies indicate the Ga containing zeolites as the most efficient catalysts for aromatization reactions.3,18-21 It is generally agreed that Ga active sites are involved in dehydrogenation reaction steps and that zeolite Brønsted acid sites (BAS) catalyze the cracking, oligomerization, and cyclization reaction steps. However, the precise roles of the acid and Ga active sites in the reaction mechanisms are not clear (for thorough consideration of this subject, see the papers by Kwak et al.,8 Price et al.,22 and Biscardi and Iglesia23). Further progress in understanding of the reaction chemistry is significantly hampered due to the lack of information on the number of the acid and Ga active sites in working catalysts. Generation of such information is a challenge since, as it follows from the literature,24-32 the numbers of the acid and Ga sites, as well as the nature of the latter, may change under reaction conditions due to the reduction of Ga3+ ions that are initially present in the catalysts. Studies of the catalysts prereduced in H2 suggest24-31 that this process involves H2 and the zeolite BAS, as was first proposed by Price and co-workers:25,26

Ga2O3 + 2H2 + 2H+Z- f 2Ga+Z- + 3H2O

(1)

where Z- is an anionic zeolite site. Using in situ Ga K-edge X-ray absorption spectroscopy, Meitzner et al.32 have demonstrated that, indeed, Ga3+ ions that were initially present in the

10.1021/jp063333c CCC: $33.50 © 2006 American Chemical Society Published on Web 08/30/2006

18474 J. Phys. Chem. B, Vol. 110, No. 37, 2006 GaH-MFI catalyst were reduced during propane aromatization into catalytically active Ga+ species that existed under reaction conditions only. This finding, later confirmed by IR spectroscopic investigation of GaH-MFI catalysts under different conditions,31 reveals the critical need for in situ studies of the active sites in the Ga containing zeolite catalysts. Therefore, the aim of this work was to attempt quantitative evaluation of the BAS and Ga active sites in working Ga containing zeolite catalysts during n-butane aromatization. To achieve this goal, we developed a kinetic approach to the evaluation of active sites in working catalysts that relies on analysis of the rates of formation of certain reaction products that can be associated with the active sites of a particular type. This approach was applied in this work for analysis of the experimental data obtained with Ga free and Ga containing medium pore Theta-1 zeolite, which has a TON structure according to the official classification of the IZA (as a consequence, we are using the term TON to define this zeolite in our paper). The reason behind the choice of the TON zeolite is our recent finding that the introduction of Ga into this zeolite (Ga/Al ) 0.48) results in an essentially higher increase in the aromatics selectivity when compared with the corresponding increase observed with a MFI zeolite.33 The amount of gallium in two GaH-TON catalysts used in this work corresponds to a Ga/Al atomic ratio of 0.48 and 0.81 (see Experimental Procedures). This implies that, in principle, around 50 and 80% of all protons present in the parent H-TON zeolite can be replaced by Ga+ cations in two prepared Ga containing catalysts. Experimental Procedures A sample of H-Theta-1 (TON) zeolite was obtained from BP Chemicals. X-ray diffraction (XRD) analysis confirmed high crystallinity of this zeolite and the absence of other phases. Scanning electron microscopy (SEM) has shown that the zeolite crystallites have the shape of needles and rods with an average length and width of 1.5 and 0.2 µm, respectively. The Si/Al atomic ratio was 37; the Na content was below 0.01 wt %, and the Al content was 1.2 wt % (445 µmol g-1). This sample, defined as H-TON zeolite, was used as a catalyst without any modification. Two Ga containing zeolite catalysts were prepared by incipient wetness impregnation of the parent H-TON zeolite, using aqueous solutions of Ga(NO3)3 of different concentration. After impregnation, the catalyst samples were slowly dried at room temperature for about 48 h and then at ∼350 K for 24 h. The Ga content was 1.5 wt % (215 µmol g-1) in the catalyst defined as GaH-TON and 2.5 wt % (359 µmol g-1) in the catalyst defined as GaH-TON-A. For kinetic studies, the catalyst samples were pressed into disks, crashed, and sieved to obtain particle sizes in the range of 250-500 µm. Prior to the kinetic experiments, the catalyst samples were slowly heated (1 K/min) in the reactor under N2 flow (30 mL/min) to 803 K and kept at this temperature for 4 h before switching to the reaction. In the experiments with the prereduced Ga containing catalysts, activation of the catalyst samples in N2 flow at 803 K was followed by the reduction in a flow of 100% H2. First, the sample was cooled to 473 K (5 K/min) and N2 was replaced by H2 (30 mL/min). Then, the sample was heated (5 K/min) under H2 flow to 803 K, kept at this temperature for 1 h, and then purged by N2 (30 mL/min, 10 min) before switching to the reaction. The kinetic studies of n-butane transformation were carried out at atmospheric pressure in a continuous flow microreactor with 100% n-butane as feed. The reaction temperature was 803 K. The reaction mixture was analyzed by on-line GC using a

Lukyanov and Vazhnova Varian CP-3800 Gas Chromatograph, which was equipped with a molecular sieve 13X packed column, a thermal conductivity detector for analysis of H2, and a 25 m long PLOT Al2O3/KCl capillary column with a flame ionization detector for analysis of hydrocarbons. The catalyst activity in the primary reaction steps (n-butane dehydrogenation and protolytic cracking steps) was estimated by extrapolation of the rate data to zero n-butane conversion. Different levels of conversions were obtained by performing experiments at different values of contact time that was defined as WHSV-1 (h), where WHSV (h-1) was the weight hour space velocity. Time on stream (TOS) studies demonstrated no changes in n-butane conversion and product distribution between 15 and 60 min after the beginning of the reaction. Therefore, the experimental data for the steady-state experiments were obtained at TOS between 15 and 50 min. The initial conditions (TOS ) 15 min) were restored at the end of each experimental run (normally, two experiments) to verify the absence of catalyst deactivation. For investigation of the catalyst performance during the initial period of reaction, a multi-port valve was used that allowed rapid collection of up to 16 samples of the reaction mixture, their storage at 473 K, and subsequent GC analysis. FTIR spectra of the self-supported catalyst disks were collected at a resolution of 2 cm-1 using a Nicolet Magna 550 FTIR spectrometer and a purpose-built IR cell that allowed high temperature treatment of samples in situ.34 Adsorption of ammonia was carried out at 423 K to estimate the number of acid sites present in the parent H-TON catalyst, while the pyridine adsorption, which was also carried out at 423 K, was used to characterize Brønsted and Lewis acid sites in this catalyst. Prior to all FTIR experiments, the samples were activated under vacuum (10-5 mbar) at 673 K overnight. The detailed experimental procedure is described elsewhere.34,35 The experiments with ammonia and pyridine revealed that the number of all acid sites in the H-TON catalyst was 433 µmol g-1 and that the numbers of the Brønsted and Lewis acid sites were 390 and 43 µmol g-1, respectively. The IR spectra of the reduced Ga containing catalysts were collected after reduction of the catalyst samples by hydrogen at 803 K (the H2 pressure was 50 mbar and duration of the treatment was 1 h). Results and Discussion Approach to Evaluation of Active Sites. The underlying idea of our approach was to use the rates of n-butane transformation into stable primary reaction products, such as H2, methane (C1), and ethane (C2), for the quantitative evaluation of the acid and Ga active sites in working Ga containing zeolites. Indeed, it is known11,36,37 that these products are formed over zeolite Brønsted acid sites in n-butane protolytic cracking steps, as shown next:

n-C4H10 f H2 + n-C4H8

(2)

n-C4H10 f CH4 + C3H6

(3)

n-C4H10 f C2H6 + C2H4

(4)

On the other hand, it is well-established3,19,22,23,38 that Ga active sites participate in the alkane dehydrogenation steps (eq 2). Hence, we anticipated that comparison of the rates of H2 formation over parent (H-form) and Ga containing zeolites would provide information on the Ga active sites, while comparison of the rates of C1 and C2 formation would allow evaluation of the zeolite acid sites. We decided to test our approach using an aromatization reaction of n-butane and not

Working Bifunctional GaH-TON Aromatization Catalyst

J. Phys. Chem. B, Vol. 110, No. 37, 2006 18475

Figure 1. Effect of contact time on n-butane conversion over (O) H-TON and (2) GaH-TON catalysts. Contact time is defined as WHSV-1, where WHSV (h-1) is the weight hour space velocity. The experimental data were obtained at 803 K with 100% n-butane as feed.

of propane since in the latter case Ga species were found to have noticeable activity in propane cracking into methane and ethene.7,38 Evaluation of Acidity in Working Ga Containing TON Catalysts. Our recent study of n-butane aromatization over the GaH-TON catalyst (1.5 wt % Ga), which is also used in this work, has demonstrated an evolution of the aromatization activity of this catalyst during initial periods of the reaction.33 The changes in the catalyst performance were completed in about 15 min, and after that time, the catalyst demonstrated stable operation. It has been suggested that the observed transient catalyst behavior was associated with the formation of catalytically active Ga+ species in accordance with the chemistry described by eq 1. If this is the case, then the acidity of the GaH-TON catalyst would be reduced significantly during the first 15 min of the reaction. Further changes in the catalyst acidity as well as in the Ga active sites at longer TOS are unlikely, as it is clear from the stable catalyst performance that was observed after the initial transient period of the reaction.33 Therefore, the evaluation of the acidity and Ga active sites in the working GaH-TON catalyst under steady-state reaction conditions was done in this work using the experimental data collected at TOS between 15 and 50 min (see Experimental Procedures). Modification of the H-TON zeolite with gallium (1.5 wt %) has only a slight effect on the overall catalyst activity in the n-butane aromatization reaction (Figure 1) but considerably affects the reaction product distribution. Figure 2 compares the formation of the primary reaction products (H2, C1, and C2) on the H-TON and GaH-TON catalysts at low n-butane conversions. The data on H2 formation indicate the enhanced activity of the latter catalyst in the dehydrogenation of n-butane, while consideration of C1 and C2 formation suggests that the protolytic cracking activity could be suppressed in the presence of gallium. These suggested changes can explain qualitatively the observed weak effect of gallium on the overall catalyst activity in n-butane transformation. Figure 3 reveals a significant effect of Ga on the aromatics formation. The parent H-TON zeolite exhibits low aromatization activity, as one would expect to observe with this zeolite. Indeed, the 1-D TON channel system39,40 does not provide enough space for bimolecular hydrogen transfer reactions that lead to aromatics formation on the acidic zeolites.14,15,17 As a consequence, the aromatization process in the H-TON zeolite is severely spatially constrained. The introduction of Ga into the H-TON zeolite results in the significant increase in aromatics concentration (Figure 3), verifying that the Ga active sites dehydrogenate the intermediate species involved in the aromatization process, thus replacing the bimolecular HT

Figure 2. Concentrations of (A) hydrogen, (B) methane, and (C) ethane as functions of n-butane conversion over (b) H-TON and (2) GaHTON catalysts. The experimental data were obtained at 803 K with 100% n-butane as feed.

Figure 3. Concentration of aromatics as a function of n-butane conversion over (b) H-TON and (2) GaH-TON catalysts. The experimental data were obtained at 803 K with 100% n-butane as feed.

reactions that form a part of the aromatization process over zeolite BAS. The results considered previously show that both the acid and the Ga active sites are involved in n-butane conversion into aromatics. The next question to be answered concerned the numbers of these sites in the working Ga containing catalyst. According to our hypothesis, this vital information could be acquired on the basis of the rates of formation of H2, C1, and C2 in the primary reaction steps. Hence, we calculated the formation rates of these products at different conversions using the experimental data shown in Figure 2 and the rate equations for a differential reactor41 (i.e., Ri ) (WHSV × Ci)/Mi, where

18476 J. Phys. Chem. B, Vol. 110, No. 37, 2006

Lukyanov and Vazhnova

Figure 5. IR spectra of H-TON and GaH-TON catalysts. Spectra a-c correspond to H-TON, as-prepared GaH-TON, and prereduced GaHTON catalysts, respectively. The band at 3603 cm-1 is assigned to Brønsted acid sites associated with framework aluminum (tSi(OH)Alt), while that at 3744 cm-1 is assigned to nonacidic terminal silanol groups (tSi(OH)).42

Figure 4. Effect of n-butane conversion on the rates of formation of (A) hydrogen, (B) methane, and (C) ethane over (b) H-TON and (2) GaH-TON catalysts. The experimental data were obtained at 803 K with 100% n-butane as feed.

TABLE 1: Rates of Formation of Primary Products during n-Butane Conversion over H-TON and GaH-TON Catalysts catalyst

H2 rate (mmol g-1 h-1)

C1 rate (mmol g-1 h-1)

C2 rate (mmol g-1 h-1)

H-TONa GaH-TONa acid sitesb Ga sitesb

15.0 42.1 8.06 34.04

18.7 13.4 10.04 3.36

22.7 12.2 12.2 0

a Rate values are based on the data shown in Figure 4. b Rate values are calculated (see Activities of the Acid and Ga Sites for an explanation).

Ri is the rate of formation of the ith product (mol g-1 h-1), and Mi and Ci are the molecular weight (g mol-1) and mass fraction of this product in the reaction mixture, respectively). The calculated rates are shown in Figure 4, and their extrapolation to zero n-butane conversion provides quantitative estimates for the rates of formation of H2, C1, and C2 in the primary reaction steps over the H-TON and GaH-TON catalysts. The obtained estimates (Table 1) demonstrate that insertion of Ga into the H-TON zeolite results in enhanced dehydrogenation activity and reduced cracking activity of the catalyst. On the basis of the literature data,3,7,19,24-32 we have concluded that the observed enhancement of the dehydrogenation activity is due to gallium, which is probably reduced to Ga+ species under reaction conditions. The observed decrease in the cracking activity (Table 1) can be explained (i) by the decrease in the number of the acid sites or/and (ii) by the decrease in the acid strength of these sites by Ga species present in the GaH-TON catalyst. In both

cases, one would expect the same decrease in the rates of C1 and C2 formation (assuming that both products are formed on the same active sites). However, comparison of the rate values obtained for the H-TON and GaH-TON catalysts (Table 1) shows that the decrease in the C1 formation rate (1.4 times) is essentially smaller than the decrease in the rate of C2 formation (1.9 times). Such a difference can be understood if C2 is formed over acid sites only, while C1 formation takes place over both the acid and the Ga active sites. The latter assumption is strongly supported by the kinetic studies of propane transformation over Ga2O3 and GaH-MFI catalysts that demonstrated that the Ga species were active in the formation of methane from propane.7,38 Consequently, based on the assumption that ethane is produced on the acid sites only, we arrive at conclusion that the acidity of the working GaH-TON catalyst should be considerably lower (1.9 times) than the acidity of the parent H-TON catalyst. To check this conclusion, we collected the IR spectra of the H-TON and GaH-TON catalysts after their calcination in a vacuum. These spectra (Figure 5, spectra a and b) reveal that there is practically no difference in the number of the Brønsted acid sites in these two catalysts (the intensities of the 3603 cm-1 band, which is associated with the zeolite BAS,42 are very similar). This finding, however, does not prove that the numbers of BAS in these catalysts are similar under reaction conditions. Indeed, it is generally agreed28,30-32 that catalyst thermal treatment leads to the decomposition of gallium nitrate and the formation of Ga3+ species that are present as Ga2O3 clusters on the outer surface of zeolite crystallites. As a result, Ga species do not migrate into zeolite channels after catalyst calcination, and little or no ion exchange with protons is observed (this explains very similar intensities of the 3603 cm-1 band in spectra a and b, Figure 5). However, as discussed in the Introduction, the Ga3+ species are likely to be reduced under reaction conditions (in the presence of H2), and this reduction should lead to the consumption of BAS (see eq 1). Therefore, we reduced the GaH-TON catalyst in the IR-cell in H2 at 803 K and obtained the IR spectrum of the reduced GaH-TON zeolite (Figure 5, spectrum c). The intensity of the 3603 cm-1 band in the reduced catalyst is about 53% of the intensity of this band in the parent zeolite. This value practically coincides with the difference in the C2 formation rate (53.7%) that was established for the GaH-TON and H-TON catalysts (Table 1). Thus, we can conclude that the observed decrease in the steady-state catalyst cracking activity is due to the decrease in the number of BAS in the working GaH-TON catalyst that is about 47% for the catalyst used in this work (1.5 wt % Ga). This conclusion implies that the rate of ethane formation over

Working Bifunctional GaH-TON Aromatization Catalyst

Figure 6. Effect of time on stream on the rates of ethane formation over (b) H-TON, (2) as-prepared GaH-TON, and (0) prereduced in H2 GaH-TON catalysts. The experiments were carried out under the same reaction conditions: WHSV ) 190 h-1, T ) 803 K, and 100% n-butane as feed. The conversion of n-butane was in the range of 1.82.2% with all three catalysts.

the GaH-TON catalyst prereduced in H2 should not change under reaction conditions (acidity is already decreased during H2 pretreatment) and should be the same as the steady-state rate of ethane formation over the as-prepared GaH-TON catalyst. The kinetic experiments performed during the initial period of n-butane reaction (Figure 6) verify this hypothesis and reveal that, indeed, the reaction conditions and H2 pretreatment lead to the same steady-state cracking activity of the GaH-TON catalyst. Figure 6 shows that the initial cracking activity of the as-prepared GaH-TON catalyst (TOS f 0) is very similar to the cracking activity of the H-TON catalyst, as one would expect on the basis of the IR spectra of these two catalysts (Figure 5, spectra a and b). The observed decrease in the rate of ethane formation over the as-prepared GaH-TON catalyst (Figure 6) indicates that the reduction of the catalyst acidity takes place under reaction conditions and that this process is completed in about 15 min after the beginning of the reaction. It is worth noting that the reduction in the ethane formation rate is around 46% (Figure 6), which is practically the same as the reduction in the number of BAS (47%) that was determined on the basis of the IR data (Figure 5, spectra a and c). The results presented above strongly suggest that the rate of ethane formation during n-butane conversion over Ga containing zeolite catalysts is related directly to the number of Brønsted acid sites in these catalysts under reaction conditions. To verify this suggestion, we performed IR and kinetic studies with a GaH-TON-A catalyst that contained different amounts of gallium (2.5 wt %) and was expected to possess less BAS under reaction conditions and display a lower activity in ethane formation when compared with the GaH-TON catalyst (1.5 wt % Ga). The comparison of the intensities of the 3603 cm-1 band in the IR spectra of H-TON and as-prepared GaH-TON-A catalysts (Figure 7, spectra a and b) demonstrates very little difference (∼10%) in the number of BAS present in these catalysts. This result agrees with the IR data shown in Figure 5 for the GaHTON catalyst (1.5 wt %) and the generally accepted view28,30-32 that most of the Ga species in the calcined catalysts are present as Ga2O3 clusters on the outer surface of zeolite crystallites. The decreased intensity of the 3744 cm-1 band in GaH-TON-A catalyst can be explained, in our opinion, by the formation of weak hydrogen bonds between terminal silanol groups (tSi(OH)) and Ga2O3 clusters. Indeed, the intensity of the 3744 cm-1 band is practically restored after treatment of this catalyst with H2 (Figure 7) that leads to the reduction of Ga2O3 and migration of Ga+ species in the zeolite channels. As a result of this process, Ga+ ions replace protons (see eq 1), thus reducing

J. Phys. Chem. B, Vol. 110, No. 37, 2006 18477

Figure 7. IR spectra of H-TON and GaH-TON-A catalysts. Spectra a-c correspond to H-TON, as-prepared GaH-TON-A, and prereduced GaH-TON-A catalysts, respectively.

Figure 8. Effect of time on stream on the rates of ethane formation over (b) H-TON, (2) as-prepared GaH-TON-A, and (0) prereduced in H2 GaH-TON-A catalysts. The experiments were carried out under the same reaction conditions: WHSV ) 190 h-1, T ) 803 K, and 100% n-butane as feed. The conversion of n-butane was in the range of 1.5-1.9% with all three catalysts.

the acidity of the Ga containing catalyst. Quantitative comparison of the intensities of the 3603 cm-1 band in the IR spectra of H-TON and reduced GaH-TON-A catalysts (Figure 7, spectra a and c) reveals that the Brønsted acidity of the reduced GaHTON-A catalyst is about 22% of the acidity of the parent H-TON catalyst. Thus, the described IR study suggests that a significant decrease in the rate of ethane formation should be observed during the initial period of n-butane conversion over the GaHTON-A catalyst. Figure 8 shows the effect of TOS on the rate of ethane formation during n-butane conversion over H-TON, as-prepared GaH-TON-A, and prereduced in H2 GaH-TON-A catalysts. Qualitatively, these results are very similar to the results observed with the GaH-TON catalyst (Figure 6) and confirm that the reaction conditions and H2 pretreatment lead to the same steady-state cracking activity of Ga containing TON catalysts. However, quantitative consideration of the data in Figure 8 shows that the decrease in the steady-state rate of ethane formation is significantly larger for the GaH-TON-A catalyst than for the GaH-TON one. Comparison of the steady-state rates of ethane formation over the H-TON and GaH-TON-A catalysts (Figure 8) reveals that the reduction in this rate is around 79%, which is nearly the same as the decrease in the number of BAS (78%) that was determined on the basis of the IR data (Figure 7, spectra a and c). On the basis of the results described above, we can conclude that measurements of the steady-state rate of ethane formation at low n-butane conversions over Ga containing zeolite catalysts provide quantitative estimation of the number of Brønsted acid sites in these catalysts under reaction conditions. Moreover, it appears that the proposed method should allow the evaluation of acidity in working bifunctional zeolite catalysts that contain other metals (e.g., zinc or platinum).

18478 J. Phys. Chem. B, Vol. 110, No. 37, 2006 Activities of the Acid and Ga Sites. In the following paragraphs, we consider the absolute activities of the acid and Ga sites in n-butane transformation. Such information may facilitate better understanding of the precise mechanisms of alkane aromatization reactions. It is also of general interest since, in principle, it allows comparison of the activities of the similar catalytic sites in different environments (e.g., BAS and Ga active sites in different zeolites, etc.). Let us first consider n-butane transformation over the H-TON catalyst that contains acid sites only. The rates of the primary reaction steps (eqs 2-4) are collected in Table 1, which shows that the total rate of n-butane conversion in these steps over the H-TON zeolite is 56.4 mmol g-1 h-1. This zeolite has a Si/Al ratio of 37 and contains 390 µmol g-1 of Brønsted acid sites (this value was determined in the IR experiments using ammonia and pyridine as probe molecules, see Experimental Procedures). According to the previous detailed studies of highly siliceous (Si/Al > 15) H-MFI zeolites,43-46 it was reasonable to assume the same catalytic properties of all BAS present in the H-TON zeolite used in this work. On the basis of this assumption, the turnover activity of an individual BAS in n-butane transformation was calculated and found to be equal to 145 molecules h-1 site-1 at 803 K. We also calculated the turnover number for BAS in H-MFI zeolite (Si/Al ) 35) for 803 K using the rate data on n-butane cracking at 773 K and activation energies reported by Narbeshuber et al.37 Comparison of this number, 184 molecules h-1 site-1, with the turnover number found for H-TON zeolite reveals very similar catalytic properties of the Brønsted acid sites in the H-MFI and H-TON zeolites. To estimate the turnover number for an individual Ga active site in n-butane transformation, we have assumed that generation of one Ga active site results in the consumption of one BAS. This assumption is based on the stoichiometry of eq 1 that is supported by the results reported by different authors22,31,47 as well as by the data obtained with two catalysts in this work. Thus, we consider Ga+ ions as Ga sites that are catalytically active in n-butane transformation. This view is in line with the recent experimental31,48 and theoretical49 studies of various Ga species in MFI catalysts (Ga+, GaH2+, GaH+, and GaO+) that provide strong evidence in favor of Ga+ ions as the active sites for alkane dehydrogenation under steady-state reaction conditions. As discussed earlier in this paper, the difference in the rates of formation of ethane over the H-TON and GaH-TON catalysts (see Table 1) corresponds to the difference in the numbers of BAS present in these catalysts under reaction conditions. For our experiments, this means that 53.7% of BAS that were present in the H-TON catalyst are still present in the working GaH-TON catalyst (1.5 wt % Ga). Using this number and the rates of H2 and C1 formation over the H-TON catalyst (100% of BAS), we have calculated the rates of formation of these two products over acid sites present in the working GaH-TON catalyst. Then, the rates of H2 and C1 formation over the Ga active sites in the working GaH-TON catalyst were determined as the difference between the observed total rates and the rates of formation of these products over acid sites (the calculated rates are shown in Table 1). Further, based on the stoichiometry of the Ga reduction process and the number of the remaining BAS in the working GaH-TON catalyst (209 µmol g-1), we determined the number of the Ga active sites in this catalyst (181 µmol g-1). Using this number and the rates of H2 and C1 formation over the Ga active sites in the working GaH-TON catalyst (Table 1), we calculated the turnover activity of the

Lukyanov and Vazhnova TABLE 2: Rates of Formation of Primary Products during n-Butane Conversion over H-TON and GaH-TON-A Catalysts catalyst

H2 rate (mmol g-1 h-1)

C1 rate (mmol g-1 h-1)

C2 rate (mmol g-1 h-1)

H-TONa GaH-TON-Ab acid sitesc Ga sitesc

15.0 32.9 3.17 29.73

18.7 6.8 3.95 2.85

22.7 4.8 4.8 0

a Rate values for the H-TON catalyst are taken from Table 1. b Rate values for the GaH-TON-A catalyst are determined under steady-state conditions at n-butane conversion of 1.5% (WHSV ) 190 h-1). c Rate values are calculated using the same procedure as described above for Table 1.

individual Ga active site in n-butane transformation and found it to be equal to 206 molecules h-1 site-1 at 803 K. Hence, we have established that the individual Ga active site is about 1.4 times more active in n-butane transformation than the individual BAS in the GaH-TON catalyst that contains 1.5 wt % Ga. It is worth noting that this comparison has been done for the first time since no information on the number of Ga active sites in the working catalysts and, consequently, on their turnover activity in alkane transformation was available until now. Analysis of the literature shows that there are two essentially different views on the mechanisms of the catalytic action of the acid and Ga active sites in alkane reactions. One group of the authors5,6,38,53,54 have proposed that these sites operate independently, as in a classical bifunctional catalysis. According to this viewpoint, the activation of alkane takes place both at acid and at Ga active sites, and this means that the turnover activities of these sites should not change with the gallium loading (assuming that the nature of the Ga active sites does not change). Other authors23,28,32,50,51,55 have suggested that the Ga active sites may work in concert with BAS, the latter being responsible for alkane activation. In this case, the turnover activity of the Ga sites can depend on the number of BAS that, in turn, is a function of the gallium content in the working catalyst (as is shown in this paper). Therefore, it was of interest to compare the turnover activities of the acid and Ga sites in two Ga containing catalysts used in this work. Table 2 shows the steady-state rates of formation of the primary reaction products of n-butane conversion over the H-TON and GaH-TON-A catalysts. The rates of formation of these products (H2, C1, and C2) over acid and Ga active sites were calculated using the same procedure as described previously for Table 1 and assuming that 21.1% of BAS, which was present in H-TON catalyst, is still present in the working GaHTON-A catalyst (21.1% is based on the ratio between the rates of ethane formation over GaH-TON-A and H-TON catalysts). On the basis of the stoichiometry of the Ga reduction process (eq 1) and the number of the remaining BAS in the working GaH-TON-A catalyst (82 µmol g-1), we calculated the number of the Ga active sites in this catalyst (308 µmol g-1). Using this number and the rates of H2 and C1 formation over Ga active sites in the working GaH-TON-A catalyst (32.58 mmol g-1 h-1, Table 2), we determined that the turnover activity of the individual Ga active site in n-butane transformation over this catalyst was 106 molecules h-1 site-1 at 803 K. This activity is about 2 times lower than the activity of the individual Ga site in the GaH-TON catalyst (206 molecules h-1 site-1). This result challenges the view that acid and Ga active sites operate independently but agrees well with the mechanism, suggested by Iglesia and co-workers,23,32,55 where Ga active sites act as a porthole for the exit of H-adatoms generated from alkane by BAS. Indeed, in accordance with this mechanism, the lower

Working Bifunctional GaH-TON Aromatization Catalyst turnover activity of Ga sites in the GaH-TON-A catalyst can be explained by the lower concentration of H-adatoms (hence, lower efficiency of Ga sites in their recombination) that is due to the lower number of BAS in this catalyst as compared to the GaH-TON one. This consideration also explains the wellknown3,19 poor efficiency of the Ga containing ZSM-5 catalysts with a high Si/Al ratio (low number of BAS) and allows us to predict that the highest turnover activity of the Ga active sites should be observed with zeolite catalysts with low Si/Al ratios and low Ga loading. Quantitative analysis of the catalytic performance of the Ga containing catalysts with different Si/ Al ratios and gallium loadings is required to verify this conception and generate a better, quantitative understanding of the exact catalytic action of Ga species in zeolite catalysts. Conclusion In this paper, we consider a detailed kinetic approach to the evaluation of the different active sites in working bifunctional zeolite catalysts that relies on analysis of the rates of formation of certain reaction products that can be associated with the active sites of a particular type. This approach develops further previous kinetic characterization of the acid sites in zeolites43,44,52 and, to the best of our knowledge, is used successfully for the first time for characterization of bifunctional catalysts. In the study described in this paper, we accomplished the kinetic evaluation of the acid and Ga active sites in working GaH-TON catalysts using the rates of formation of the primary products of the n-butane aromatization reaction. Analysis of these rates together with the IR spectra of the H-TON and GaHTON catalysts has demonstrated that the rate of ethane formation at low conversions of n-butane can be used for the quantitative characterization of the number of BAS present in working Ga containing TON catalysts. On the basis of this finding, we have shown that the number of BAS in the GaH-TON catalysts decreases considerably during the initial period of n-butane reaction and stabilizes after ∼15 min on stream. These results taken into consideration together with the results reported recently33 lead to the conclusion that the evolution of the GaHTON catalyst under reaction conditions is associated with the formation of catalytically active Ga+ species and related consumption of BAS according to the chemistry described by eq 1. As a consequence of this process, the introduction of Ga (1.5 wt %) into the H-TON zeolite has a very weak effect on the overall steady-state catalyst activity in n-butane conversion, as the gain in n-butane dehydrogenation activity is nearly compensated by the loss in n-butane protolytic cracking activity. For the steady-state reaction conditions, we have obtained, for the first time, quantitative estimates for the numbers of the BAS and Ga active sites in two working Ga containing catalysts (1.5 and 2.5 wt % Ga) as well as for the turnover activities of these sites in n-butane transformation. These new results provide further support to the mechanism, previously suggested by Iglesia and co-workers,23,32,55 where Ga active sites act as a porthole for the exit of H-adatoms that are generated from the alkane on BAS. It is anticipated that the study presented in this paper will facilitate understanding of the precise roles of the acid and Ga sites in the mechanisms of alkane aromatization and, as a farreaching implication, will prompt wider use of detailed kinetic studies for the evaluation of active sites in working catalysts. Acknowledgment. We thank Dr. W. J. Smith (BP Chemicals) for providing the zeolite sample. This work was supported by the Engineering and Physical Sciences Research Council of

J. Phys. Chem. B, Vol. 110, No. 37, 2006 18479 the UK (Grants GR/A00096 and EP/C532554) and by the Department of Chemical Engineering (University of Bath) startup funds. References and Notes (1) Mowry, J. R.; Anderson, R. F.; Johnson, J. A. Oil Gas J. 1985, 83, (Dec 2), 128. (2) Chen, N. Y.; Yan, T. Y. Ind. Eng. Chem. Process Des. DeV. 1986, 25, 151. (3) Guisnet, M.; Gnep, N. S.; Alario, F. Appl. Catal. 1992, 89, 1. (4) Haag, W. O. In Proceedings of the 6th International Zeolite Conference; Olson, D. H., Bisio, A., Eds.; Butterworth: Surrey, UK, 1984; p 466. (5) Kitagawa, H.; Sendoda, Y.; Ono, Y. J. Catal. 1986, 101, 12. (6) Gnep, N. S.; Doyemet, J. Y.; Seco, A. M.; Ramoa Ribeiro, F.; Guisnet, M. Appl. Catal. 1987, 35, 93. (7) Lukyanov, D. B.; Gnep, N. S.; Guisnet, M. R. Ind. Eng. Chem. Res. 1995, 34, 516. (8) Kwak, B. S.; Sachtler, W. M. H.; Haag, W. O. J. Catal. 1994, 149, 465. (9) Biscardi, J. A.; Iglesia, E. J. Phys. Chem. B 1998, 102, 9284. (10) Haag, W. O.; Dessau, R. M. In Proceedings of the 8th International Congress on Catalysis; Dechema: Frankfurt-am-Main, Germany, 1984; Vol. 2, p 305. (11) Krannila, H.; Haag, W. O.; Gates, B. C. J. Catal. 1992, 135, 115. (12) Garwood, W. E. In ACS Symposium Series; Stucky, G. D., Dwyer, F. G., Eds.; American Chemical Society: Washington, DC, 1983; No. 218, p 383. (13) Quann, R. J.; Green, L. A.; Tabak, S. A.; Krambeck, F. J. Ind. Eng. Chem. Res. 1988, 27, 565. (14) Poutsma, M. L. In Zeolite Chemistry and Catalysis, ACS Monograph 171; Rabo, J. A., Ed.; American Chemical Society: Washington, DC, 1976; p 437. (15) Pines, H. Chemistry of Catalytic Hydrocarbon ConVersions; Academic Press: New York, 1981; Ch. 1. (16) Vedrine, J. C.; Dejaifve, P.; Garbowski, E. D.; Derouane, E. G. Stud. Surf. Sci. Catal. 1980, 5, 29. (17) Lukyanov, D. B.; Gnep, N. S.; Guisnet, M. R. Ind. Eng. Chem. Res. 1994, 33, 223. (18) Seddon, D. Catal. Today 1990, 6, 351. (19) Ono, Y. Catal. ReV. Sci. Eng. 1992, 34, 179. (20) Meriaudeau, P.; Naccache, C. Catal. ReV. Sci. Eng. 1997, 39, 5. (21) Fricke, R.; Kosslick, H.; Lischke, G.; Richter, M. Chem. ReV. 2000, 100, 2303. (22) Price, G. L.; Kanazirev, V.; Dooley, K. M.; Hart, V. I. J. Catal. 1998, 173, 17. (23) Biscardi, J. A.; Iglesia, E. J. Catal. 1999, 182, 117. (24) Kanazirev, V.; Price, G. L.; Dooley, K. M. J. Chem. Soc., Chem. Commun. 1990, 9, 712. (25) Price, G. L.; Kanazirev, V. J. Catal. 1990, 126, 267. (26) Dooley, K. M.; Chang, C.; Price, G. L. Appl. Catal., A 1992, 84, 17. (27) Meriaudeau, P.; Naccache, C. Appl. Catal. 1991, 73, 13. (28) Meriaudeau, P.; Naccache, C. J. Catal. 1995, 157, 283. (29) El-Malki, El-M.; van Santen, R. A.; Sachtler, W. M. H. J. Phys. Chem. B 1999, 103, 4611. (30) Nowak, I.; Quartararo, J.; Derouane, E. G.; Vedrine, J. C. Appl. Catal., A 2003, 251, 107. (31) Kazansky, V. B.; Subbotina, I. R.; van Santen, R. A.; Hensen, E. J. M. J. Catal. 2004, 227, 263. (32) Meitzner, G. D.; Iglesia, E.; Baumgartner, J. E.; Huang, E. S. J. Catal. 1993, 140, 209. (33) Lukyanov, D. B.; Vazhnova, T. Appl. Catal. A: Gen., submitted. (34) Zholobenko, V. L.; Makarova, M. A.; Dwyer, J. J. Phys. Chem. 1993, 97, 5962. (35) Lukyanov, D. B.; Zholobenko, V. L.; Dwyer, J.; Barri, S. A. I.; Smith, W. J. J. Phys. Chem. B 1999, 103, 197. (36) Ono, Y.; Kanae, K. J. Chem. Soc., Faraday Trans. 1991, 87, 663. (37) Narbeshuber, T. F.; Vinek, H.; Lercher, J. A. J. Catal. 1995, 157, 388. (38) Gnep, N. S.; Doyemet, J. Y.; Guisnet, M. J. Mol. Catal. 1988, 45, 281. (39) Barri, S. A. I.; Smith, G. W.; White, D.; Young D. Nature (London) 1984, 312, 533. (40) Kokotailo, G. T.; Schlenker, J. L.; Dwyer, F. G.; Valyocsik, E. W. Zeolites 1985, 5, 349. (41) Froment, G. F.; Bischoff, K. B. Chemical Reactor Analysis and Design, 2nd ed.; Wiley: New York, 1990; Ch. 2. (42) Borade, R. B.; Adnot, A.; Kaliaguine, S. Zeolites 1991, 11, 710. (43) Olson, D. H.; Haag, W. O.; Lago, R. M. J. Catal. 1980, 61, 390.

18480 J. Phys. Chem. B, Vol. 110, No. 37, 2006 (44) Haag, W. O.; Lago, R. M.; Weisz, P. B. Nature (London) 1984, 309, 589. (45) Lago, R. M.; Haag, W. O.; Mikovsky, R. J.; Olson, D. H.; Hellring, S. D.; Schmitt, K. D.; Kerr, G. T. In Proceedings of the 7th International Zeolite Conference; Murakami, Y., Iijima, A., Ward, J. W., Eds.; Kodansha: Tokyo, 1986; p 677. (46) Lukyanov, D. B. Zeolites 1991, 11, 325. (47) Kwak, B. S.; Sachtler, W. M. H. J. Catal. 1993, 141, 729. (48) Rane, N.; Overweg, A. R.; Kazansky, V. B.; van Santen, R. A.; Hensen, E. J. M. J. Catal. 2006, 239, 478. (49) Pidko, E. A.; Kazansky, V. B.; Hensen, E. J. M.; van Santen, R. A. J. Catal. 2006, 240, 73.

Lukyanov and Vazhnova (50) Meriaudeau, P.; Naccache, C. J. Mol. Catal. 1990, 59, 31. (51) Kwak, B. S.; Sachtler, W. M. H. J. Catal. 1994, 145, 456. (52) Lukyanov, D. B. J. Catal. 1994, 145, 54. (53) Kazansky, V. B.; Kustov, L. M.; Khodakov, A. Yu. In Zeolites: Facts, Figures, Future; Jacobs, P. A., van Santen, R. A., Eds.; Elsevier: Amsterdam, 1989; p 1173. (54) Gianetto, G.; Montes, A.; Gnep, N. S.; Florentino, A.; Cartraud, P.; Guisnet, M. J. Catal. 1993, 145, 86. (55) Iglesia, E.; Baumgartner, J. E.; Price, G. L. J. Catal. 1992, 134, 549.