pubs.acs.org/Langmuir © 2009 American Chemical Society
Tantalum Clusters Supported on Silica-Alumina: Influence of Support Composition and Chemistry on Cluster Structure Junming Sun,† Miaofang Chi,‡ Rodrigo J. Lobo-Lapidus,† Shareghe Mehraeen,† Nigel D. Browning,† and Bruce C. Gates*,† † Department of Chemical Engineering and Materials Science, University of California, Davis, California 95616, and ‡Materials Science & Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831
Received April 12, 2009. Revised Manuscript Received June 6, 2009 Small cationic tantalum clusters were prepared on the surfaces of SiO2, silica-aluminas, and γ-Al2O3 supports by treating physisorbed pentabenzyltantalum at 523 K for 24 h in flowing H2. The rate of decomposition and the products formed in the decomposition of pentabenzyltantalum are dependent on the support composition. When the support was SiO2, the evolved products were mainly biphenyl and a small amount of toluene, indicating that the Ta-C bond in pentabenzyltantalum was activated. As the alumina content of the support increased, diphenylmethane, benzene, and ethylene were increasingly formed, and these products show that the activation of the C-C bonds linking the C atoms of methyl groups to the aromatic rings of the benzyl ligands was facilitated. Infrared spectra of the surface species and mass spectra of the effluents formed during the treatment show that the composition of the support had significant influence on the decomposition of pentabenzyltantalum, and the chemistry is inferred to be related to the electron-donor properties of the supports. Extended X-ray absorption fine structure (EXAFS) spectra recorded at the Ta LIII edge indicate the formation of clusters with a Ta-Ta first-shell coordination number of ∼3, and images obtained by scanning transmission electron microscopy (STEM) confirm the presence of such small clusters. X-ray absorption near edge structure (XANES) data indicate that the formal oxidation state of the tantalum in the clusters decreased from ∼3.0 to ∼2.6 as the support was changed from SiO2 to silica-aluminas to γ-Al2O3. The data suggest that the tantalum clusters were anchored to the supports via bridging O atoms. The EXAFS data show that the support composition had little influence on the cluster structure.
Introduction Clusters of the later transition metals that are highly dispersed on porous oxides and zeolites have been investigated extensively, because of the importance of these materials as catalysts.1,2 Syntheses of clusters of group 7 and group 8 metals have been especially well-investigated.1 However, reports of clusters of early transition metals on supports are still rare, although there is an extensive literature of the synthesis and chemistry of such clusters in solution.3 Our earlier work demonstrated that tantalum clusters could be prepared on SiO2 supports by treatment in H2 of the physisorbed precursor pentabenzyltantanlum,4 and the materials catalyze the conversion of alkanes, including methane, albeit only at low rates.4a To understand the surface chemistry better, we have now extended the work to a family of supports with systematically varied compositions, including SiO2, silica-aluminas of various compositions, and γ-Al2O3. The surface species formed from pentabenzyltantalum have been characterized by infrared (IR) spectroscopy, extended X-ray absorption (EXAFS) spectroscopy, and transmission electron microscopy (TEM), and the *To whom correspondence should be addressed. E-mail: bcgates@ucdavis. edu. (1) (a) Gates, B. C. Chem. Rev. 1995, 95, 511. (b) Xu, Z.; Xiao, F.-S.; Purnell, S. K.; Alexeev, O.; Kawi, S.; Deutsch, S. E.; Gates, B. C. Nature 1994, 372, 346. (2) Cariati, E.; Roberto, D.; Ugo, R. Chem. Rev. 2003, 103, 3707. (3) (a) M€uller, A.; Jostes, R.; Cotton, F. A. Angew. Chem., Int. Ed. 1980, 19, 875. (b) Babaian-Kibala, E.; Cotton, F. A.; Shang, M. Inorg. Chem. 1990, 29, 5148. (c) Cotton, F. A.; Diebold, M. P.; Feng, X.; Roth, W. J. Inorg. Chem. 1988, 27, 3413. (4) (a) Nemana, S.; Gates, B. C. Chem. Commun. 2006, 3998. (b) Nemana, S.; Gates, B. C. J. Phys. Chem. B 2006, 110, 17546. (c) Nemana, S.; Gates, B. C. Langmuir 2006, 22, 8214. (d) Nemana, S.; Okamoto, N. L.; Browning, N. D.; Gates, B. C. Langmuir 2007, 23, 8845. (e) Nemana, S.; Sun, J. M.; Gates, B. C. J. Phys. Chem. C 2008, 112, 7477.
10754 DOI: 10.1021/la901295d
products evolved during the formation of tantalum clusters were characterized by mass spectrometry. The data demonstrate a significant effect of the support surface composition on the conversion of physisorbed pentabenzyltantalum, leading to the formation of extremely small cationic clusters, with an average Ta-Ta coordination number of ∼3. The support surface composition has a slight effect on the Ta-Ta bonding distance and the Ta-O coordination number, but the Si:Al ratio of the support evidently affects the redox chemistry of the tantalum. The tantalum clusters are inferred to be anchored to the support via bridging O atoms of the support surface.
Experimental Methods Materials and Synthesis. Materials and Reagents. The supports included aerosil SiO2 (200 m2/g) and γ-Al2O3 (110 m2/g) (Degussa); the silica-alumina samples were prepared from Al2(SO4)3 3 18H2O (Sigma-Aldrich, 98%+) and hexamethyldisilazane (HMDS, Sigma-Aldrich, 99.9%); anhydrous hexane (Sigma-Aldrich, 95%) was used as a solvent. H2 was generated by electrolysis of water in a Balston generator (purity, 99.999%) or was supplied by Airgas (99.999%). Preparation of Supports. Each support was a powder. The SiO2 and γ-Al2O3 were used as received. Silica-alumina samples (SixAlOy) were synthesized by a reported procedure5 involving condensation reactions. Briefly, calculated amounts of SiO2 and Al2(SO4)3 3 18H2O were dissolved in deionized water; after 30 min of stirring, the pH was adjusted to 7.5-8 via the addition of 10% ammonium hydroxide in water. After stirring at room temperature for another 2 h, the resultant mixture was transferred to an (5) Wu, S.; Han, Y.; Zou, Y.-C.; Song, J.-W.; Zhao, L.; Di, Y.; Liu, S.-Z.; Xiao, F.-S. Chem. Mater. 2004, 16, 486.
Published on Web 07/16/2009
Langmuir 2009, 25(18), 10754–10763
Sun et al.
Article Table 1. Sample Notation and the Corresponding Preparation Methods
sample number
sample name
composition
method of preparation
1 2 3 4 5
Ta/Si30AlOy Ta/Si10AlOy Ta/Si2AlOy Ta/Al2O3 Ta/SiO2-cap
Si:Al = 30 (atomic) Si:Al = 10 (atomic) Si:Al = 2 (atomic) γ-Al2O3 SiO2 with CH3 groups on surface
TaBn5/Si30AlOy, treated in H2, at 523 K, 24 h TaBn5/Si10AlOy, treated in H2, 523 K, 24 h TaBn5/Si2AlOy, treated in H2, 523 K, 24 h TaBn5/Al2O3, treated in H2, 523 K, 24 h TaBn5/SilO2-cap, treated in H2, 523 K, 24 h
autoclave for further reaction (condensation) at 373 K for 2 days. The solid products were collected by centrifugation, washed intensively with deionized water, dried, calcined at 773 K in air for 8 h, and then heated to 973 K and held for another 24 h to remove impurities. Samples were stored in a glovebox filled with N2 or one filled with argon, with the O2 and H2O contents being 94%). Furthermore, toluene gas was evolved over the temperature range of 353-523 K and was subjected to GC analysis. When the support was any of the silica-aluminas (samples 1-3), bibenzyl was observed at lower temperatures (e.g., 398 K) along with toluene. However, the amount of bibenzyl was much less than that observed at the same temperature when the support was SiO2 (see the Supporting Information). This result indicates a greater stability of the Ta-C bond in the silica-alumina-supported species than in the SiO2-supported species, which is consistent with the IR observations previously stated. As the temperature of treatment in H2 was increased further (e.g., to 463 K when the support was a silica-alumina; see samples 1-3), diphenylmethane and benzene were also observed via GC analysis (Figure 2B; also see the Supporting Information). These compounds presumably resulted from the cleavage of bonds between the C atoms of methyl groups and C atoms in the aromatic rings of benzyl ligands. Moreover, the amount of diphenylmethane and benzene that formed increased as the alumina content of the support increased and, similarly, the temperature of the onset of diphenylmethane and benzene formation decreased. After the cleavage of C-C bonds and the release of benzene, methyl groups bonded to the tantalum became observable by IR spectroscopy (see Figure S2 in the Supporting Information). Specifically, for example, after treatment of sample 3 at 463 K in H2 for 12 h, the C-H stretching band at a frequency slightly higher than 3000 cm-1 and C-C stretching bands at ∼1454 and ∼1495 cm-1 (which is indicative of alkenes or aromatics) almost DOI: 10.1021/la901295d
10757
Article
Figure 2. Mass spectra of products evolved from samples during temperature ramps (black line, benzene; gray line, toluene; red line, ethylene): (A) TaBn5/SiO2; (B) TaBn5/Si30Al; (C) TaBn5/Si10AlOy; (D) TaBn5/Si2AlOy; and (E) TaBn5/Al2O3. The loadings of pentabenzyltantalum and the sample masses were the same for each experiment.
disappeared (see Figure S2 in the Supporting Information). However, the C-H stretching bands characterizing alkyl species (those located at wavenumbers slightly less than 3000 cm-1) were still clearly evident.17 These observations further confirm the greater Ta-C bond strength in the organotantalum species on silica-alumina, relative to those on SiO2. The C-H stretching vibrations of the methyl group were observed at ∼2972 and ∼2986 cm-1, corresponding to a ca. 10-cm-1 blue shift, compared with that of methyl groups bonded to the support in SiO2-cap (2963 cm-1; see Figure S1A in the Supporting Information). Therefore, we infer that the methyl groups were bonded to Ta atoms rather than to the support. The stable metal alkyls were completely removed from the samples when they were treated in flowing H2 at temperatures of >523 K for 24 h, as indicated by the disappearance of the bands at 2972 and 2986 cm-1 (data not shown). The removal of the alkyl groups resulted mainly in the formation of ethylene in the gas phase as well as traces of methane, as shown by GC analysis (see the Supporting Information). When the support was γ-Al2O3, bibenzyl and toluene were still the main decomposition products at low temperatures. When the temperature was increased, however, some ethylene and C4 species were observed (see Figure 2e, and see Tables S2-S4 in the Supporting Information). Triphenylmethane and traces of (17) Kalsi, P. S. Spectroscopy of Organic Compounds, 6th Edition; New Age International Publishers: New Delhi, India, 2004; pp 177-183.
10758 DOI: 10.1021/la901295d
Sun et al.
unknown products were also detected by GC among the nonvolatile solid products. In contrast to what was observed with the silica-alumina-supported samples, almost no benzene was detected (see Figure 2e). This comparison indicates that the C-C bond activation occurred differently on the different supports. In summary, all the data indicate significant influence of the support composition on the surface chemistry of the supported tantalum species. The IR and MS data indicate that the tantalum precursor was mainly physisorbed on each of the supports at room temperature. The reactions of the adsorbed tantalum complexes are dependent on the support surface composition, as demonstrated by the various products of the decomposition in H2. The supported tantalum species formed by decomposition of the adsorbates formed from pentabenzyltantalum were anchored to the support by groups that are inferred to be bridging O atoms. Formation of Tantalum Clusters. Evidence from STEM Images. STEM images clearly demonstrate the presence of clusters on the supports, formed as a result of decomposition of the adsorbates formed from pentabenzyltantalum in H2 (Figure 3; see the Supporting Information for more images), a conclusion that is consistent with the results of Nemana et al.4 Figure 3 shows typical aberration-corrected STEM images of the samples on a silica-alumina support and on γ-Al2O3 after a 24-h treatment of the physisorbed pentabenzyltantalum precursor at 523 K. The clusters in each sample are extremely small and raftlike, evidently consisting of only a few Ta atoms each, on average. Although the atomic-scale images suffered from imaging instability as a result of the electrostatic charging of the electrically insulating supports, some STEM images give evidence of raftshaped clusters with a nuclearity of ∼5-6 atoms, corresponding satisfactorily to the coordination number of ca. 3 determined by EXAFS spectroscopy (see Figures 3a and 3b). Although the STEM images show that many clusters consisted of 5-6 Ta atoms, we emphasize that single Ta atoms and clusters with more or fewer than 56 Ta atoms are also evident in the images. Specifically, some samples (especially those incorporating silica-alumina supports) include smaller clusters of 1-3 Ta atoms (Figure 3; also see Figure S3 in the Supporting Information). These images suggest that the cluster formation involved, as intermediates, binuclear and then trinuclear and larger clusters as Ta atoms migrated on the support surface. Thus, at higher treatment temperatures, we expect larger clusters, consistent with the observations of Nemana et al.4d Here, we do not include estimates of Ta-Ta distances based on the STEM images, because any slight scanning distortions in the images and delicate beam interactions with the clusters could introduce relatively large errors into the values. Characterization of Supported Tantalum Clusters by EXAFS Spectroscopy. The clusters formed on the supports as a result of treatment in H2 were also characterized by EXAFS spectroscopy. Ta-Ta contributions in the EXAFS data confirm the presence of tantalum clusters, in agreement with the STEM images and consistent with Nemana’s observations for SiO2supported samples.4 The EXAFS data were fitted to provide estimates of the average cluster size. Details of the models considered in the fitting and the goodness-of-fit parameters obtained with each of the best candidate models are summarized in Table 2. We emphasize that the EXAFS data determine only average structural information, giving a less-accurate measure of the cluster sizes than STEM. The model fitting the data best for each support included the following contributions: Ta-Ta, Ta-O (or TadC), Ta-Ol, and Ta-Al (or Ta-Si) (see Table 3), where the subscript “l” refers to “long” (some Ta-O distances were short (bonding) distances; the Langmuir 2009, 25(18), 10754–10763
Sun et al.
Article
Figure 3. STEM images of (a) sample 3 and (b) sample 4. Individual Ta atoms are visible in the images; superimposed on the right-hand image next to three of the clusters are models of the cluster structures shown as monolayers; N represents the number of Ta atoms in a cluster. Table 3. EXAFS Parameters Characterizing Samples 1-5a absorberbackscatterer pair
coordination number, N
distance between absorber and backscatterer Debye-Waller factor, Δσ2 ( inner potential correction, ΔE0 (eV) atoms, R (A˚) 103 A˚2) Sample 1
Ta-Ta Ta-O (or TadC) Ta-Olong Ta-Al (Si) Ta-Ta
3.0 (0.1) 3.9 (0.1) 3.8 (0.1) 1.7 (0.1) 3.0 (0.1)
2.88 (0.01) 1.88 (0.00) 3.01 (0.00) 3.57 (0.01) 2.87 (0.01)
6.5 (0.5) 4.9 (0.3) 2.3 (0.6) 3.9 (1.2) 7.5 (0.6)
-4.8 (0.7) -3.3 (0.2) -9.3 (0.2) -3.8 (0.5) -4.6 (0.7)
4.0 (0.4) 2.1 (0.7) 3.9 (1.1) 7.6 (0.9)
-3.4 (0.2) -9.4 (0.2) -5.5 (0.7) -4.3 (0.6)
4.9 (0.4) 3.7 (0.8)) 4.8 (1.2) 9.6 (0.5)
-1.7 (0.2) -9.4 (0.2) -2.6 (0.8) -3.0 (0.6)
5.6 (0.4) 6.2 (0.9) 8.1 (1.1) 7.0 (0.6)
-1.5 (0.2) -9.4 (0.2) -2.7 (0.7) -6.2 (0.8)
Sample 2 Ta-O (or TadC) Ta-Olong Ta-Al (Si) Ta-Ta
3.7 (0.1) 3.5 (0.1) 1.7 (0.1) 3.0 (0.1)
1.88 (0.00) 3.01 (0.01) 3.59 (0.01) 2.87 (0.01) Sample 3
Ta-O (or TadC) Ta-Olong Ta-Al (Si) Ta-Ta
3.5 (0.1) 3.6 (0.1) 1.6 (0.1) 2.9 (0.1)
1.86 (0.00) 3.01 (0.00) 3.58 (0.02) 2.86 (0.01) Sample 4
Ta-O (or TadC) Ta-Olong Ta-Al Ta-Ta
2.8 (0.1) 2.7 (0.1) 1.6 (0.1) 2.6 (0.1)
1.86 (0.00) 3.01 (0.01) 3.57 (0.02) 2.88 (0.01) Sample 5
Ta-O 3.8 (0.1) 1.88 (0.00) 4.2 (0.4) -3.7 (0.2) 3.6 (0.1) 3.01 (0.01) 2.6 (0.4) -9.7 (0.2) Ta-Olong Ta-Si 2.7 (0.1) 3.61 (0.01) 9.0 (1.1) -9.7 (0.5) a Numbers in parentheses are the calculated errors and represent precisions, not accuracies. Estimated accuracies are as follows. For Ta-Ta: N, (20%; R, (0.02 A˚; Δσ2, (20%; and ΔE0, (20%. For Ta-O: N, (30%, R, (0.02 A˚, Δσ2, (25%, and ΔE0, (20%. For Ta-Si (or Ta-Al): N, (50%, R, (0.03 A˚, Δσ2, (30%, and ΔE0, (20%.
“long” Ta-O distances (Table 3) are too long to be bonding distances). All these results are broadly consistent with those reported by Nemana et al.4 for SiO2-supported samples made via a method that is similar to ours. Analysis of the EXAFS data that characterize the tantalum species on the various supports (see Figures S4-S7 in the Supporting Information) consistently indicated the presence of Ta-Ta contributions, with the coordination number being ∼3 Langmuir 2009, 25(18), 10754–10763
(see Table 3), except for sample 5 (see the Supporting Information), Ta/SiO2-cap, for which the coordination number was ca. 2.6. The contribution was clearly identified by the Ta-Ta difference file in each case (see Figures S4C, S5C, S6C, and S7C in the Supporting Information; in these figures, the Ta-Ta difference files are phase- and amplitude-corrected and Fourier transformed, resulting, in each case, in a symmetrical imaginary part with its positive part peaking near the top of the absolute part13). DOI: 10.1021/la901295d
10759
Article
Figure 4. Dependence of average Ta-Ta distance determined by EXAFS spectroscopy on the composition of the support. 1
The k -weighted Fourier-transformed Ta-Ta difference files also include obvious Ta-Ta contributions, but, as expected, they are characterized by reduced amplitudes relative to the k3-weighted Fourier-transformed data (see Figures S4D, S5D, S6D, and S7D in the Supporting Information). The significant contrast in amplitude between k1- and k3-weighted Ta-Ta contributions confirms that a high-Z backscatterer (identified as Ta) was indeed necessary to fit the EXAFS data satisfactorily. The fact that the Ta-Ta coordination number was approximately the same for each of the samples, and approximately equal to 3, indicates that the clusters were small (consistent with the STEM results) and that the cluster nuclearity was not dependent significantly on the support composition under the conditions of our experiments. In contrast, the Ta-Ta bonding distance was determined to be dependent on the Si:Al ratio in the support, as shown in Figure 4 and Table 3. When the support was SiO2, the Ta-Ta bonding distance was determined to be 2.93 A˚.4 As alumina was increasingly incorporated in the support, the Ta-Ta bond distance decreased monotonically, reaching the value of 2.86 A˚ when the support was γ-Al2O3 (see Table 3). The EXAFS data that characterize the tantalum clusters on the various supports also indicate backscatterers other than Ta atoms. The coordination shell located at ca. 1.86-1.88 A˚ indicates the bonding of Ta atoms to the O atoms of the support4 (see Table 3 for the parameters that characterize the corresponding samples, and see the Supporting Information for difference files of the spectra). Because EXAFS spectroscopy does not distinguish clearly between low-atomic-weight backscatterers such as O and C atoms,18 the distance could alternatively be interpreted as evidence of TadC contributions,4e consistent with the TadC distances that have been observed by X-ray diffraction (XRD) crystallography for Ta(CH-t-Bu)-[NCN](O-t-Bu)2 (TadC bond length=1.914 A˚),19 and TaCl2(CH-t-Bu)(CNN) (TadC bond length=1.86 A˚).20 The Ta-O (or TadC) coordination numbers decreased as the Si:Al ratios in the support decreased (see Figure 5). The coordination number that characterized silica-alumina sample 1 (with a Si:Al atomic ratio of 30) was determined to be 3.9; however, when the support was γ-Al2O3 (sample 4), the value was only 2.8. (18) Pandya, K. I.; Koningsberger, D. C. Physica B 1989, 158, 386. (19) Abbenhuis, H. C. L.; Rietveld, M. H. P.; Haarman, H. F.; Hogerheide, M. P.; Spek, A. L.; Van Koten, G. Organometallics 1994, 13, 3259. (20) Rietveld, M. H. P.; Teunissen, W.; Hagen, H.; van de Water, L.; Grove, D. M.; van der Schaaf, P. A.; Muhlebach, A.; Kooijman, H. W.; Smeets, J. J.; Spek, A. L.; van Koten, G. Organometallics 1997, 16, 1674.
10760 DOI: 10.1021/la901295d
Sun et al.
Figure 5. Dependence of Ta-O (or TadC) coordination number determined by EXAFS spectroscopy on the composition of the support.
The Ta-O coordination number that characterized the capped silica (sample 5) was 3.8. These results further confirm our inference that tantalum was bonded to bridging O atoms of the support, because most of the hydroxyl groups were already replaced with trimethylsiloxyl groups on the capped silica. Furthermore, a relative long (nonbonding) Ta-Ol contribution was found with a distance of ca. 3.0 A˚. This result shows that some O atoms near the Ta atoms on the surface were farther from the Ta atoms than the O atoms that were bonded to the Ta atoms. The long (nonbonding) Ta-O distance is typical of metal-oxygen distances found by EXAFS spectroscopy for clusters of various group 7 and group 8 metals on oxide and zeolite supports (ca. 2.50-3.45 A˚, depending on the support, the pretreatment conditions, and the atmosphere4,6). The long Ta-O distance suggests weak interactions between the metal in the clusters and surface O atoms; these interactions are not well-understood.6b Another coordination shell also was observed, characterized by relatively long (nonbonding) Ta-backscatterer distances of 3.56-3.57 A˚ and coordination numbers of ∼1.6-1.7; this was included in the overall fit as a Ta-Al or a Ta-Si contribution. These contributions were too small to be determined with confidence; the distances are in agreement with those reported4 and are consistent with the presence of Si and/or Al atoms near bridging O atoms of the support that interact with the Ta atom clusters. Notwithstanding the uncertainty of the fits regarding the higher-shell contributions, an important point is that the parameters that characterize the Ta-Ta contributions are barely dependent on the choice of model to represent the EXAFS data (they vary little from one model to another). This result is important because it confirms that these contributions have been determined accurately. Oxidation State of Tantalum in Clusters on Various Supports. To characterize the oxidation states of tantalum in the supported clusters, XANES data were collected at the Ta LIII edge. The derivatives of the corresponding spectra are shown in Figure 6. The Ta LIII edge data that characterize the supported tantalum clusters show that the edge energy decreased gradually as the support was changed from SiO2 to silica-aluminas to γ-Al2O3. The difference in the values that characterize the clusters on SiO2 and on γ-Al2O3 was ∼0.4 eV, calculated on the basis of XANES data that characterize the reference compounds;4d the reference data indicate values of the formal oxidation state of the tantalum in the range of 2.6 to 3.0. The data are evidence of Langmuir 2009, 25(18), 10754–10763
this work
this work
this work
inference from product distribution, but with substantial uncertainty associated with complexity of product distribution inference from product distribution, but with substantial uncertainty associated with complexity of product distribution inference from product distribution, but with substantial uncertainty associated with complexity of product distribution silica-alumina surface (Si:Al = 30 [atomic]) silica-alumina surface (Si:Al = 2 [atomic]) γ-Al2O3 surface
C6D6 solution SiO2 surface
toluene bibenzyl (with some toluene) bibenzyl, toluene, benzene, diphenylmethane, ethylene, methane diphenylmethane, benzene, bibenzyl, toluene, methane, ethylene bibenzyl, toluene, triphenylmethane, ethylene
intramolecular hydrogen transfer free-radical chain process (with some intramolecular hydrogen transfer) free-radical chain process, acid-catalyzed cracking, and possibly some intramolecular hydrogen transfer acid-catalyzed cracking, free-radical chain process, and possibly some intramolecular hydrogen transfer free-radical chain process and possibly some intramolecular hydrogen transfer
inference from product distribution inference from product distribution
21
Langmuir 2009, 25(18), 10754–10763
predominant reaction classes
comments (21) (a) Schrock, R. R. J. Org. Chem. 1976, 122, 209. (b) Malatesta, V.; Ingold, K. U.; Schrock, R. R. J. Org. Chem. 1978, 152, C53.
products
Discussion Nemana et al. reported the formation of SiO2-supported tantalum clusters prepared from physisorbed pentabenzyltantlum by treatment in flowing H2 at 523 K. Our results extend Nemana’s observations to a family of supports including silica-aluminas and γ-Al2O3; the data reported here are the first to characterize early-transition-metal clusters on supports other than SiO2. The data show that the support surface composition exerts significant influence on the chemistry of the cluster synthesis but only modest influence on the structure of the tantalum clusters that are formed. The cluster formation evidently proceeds via decomposition of the pentabenzyltantalum that is initially physisorbed on the various supports (SiO2, silica-aluminas, and γ-Al2O3). The data presented here show that both the rate of decomposition and the products formed in the decomposition are dependent on the support composition, as summarized in Table 4. When the support was SiO2, the evolved products were mainly biphenyl and a small amount of toluene, indicating that the Ta-C bond in pentabenzyltantalum was activated. As the alumina content of the support increased, diphenylmethane, benzene, and ethylene were increasingly formed, and these products show that the activation of the C-C bonds linking the C atoms of methyl groups to the aromatic rings of the benzyl ligands was facilitated. These decomposition products formed from the physisorbed pentabenzyltantalum are contrasted to those formed in the decomposition of pentabenzyltantalum in C6D6 solvent, whereby only toluene was produced.21 It is clear that the support exerts a significant effect on the reactivity of the physisorbed tantalum species, and changes in the support surface composition lead to marked changes in the decomposition chemistry. Next, we summarize some results to begin to unravel this chemistry. The fact that the reaction on SiO2 differs substantially from that in C6D6 suggests that bonding of the tantalum to electron-donating 4a
reaction medium
a systematic change in the tantalum oxidation state as a result of changes in the support composition,4b indicating that the support affected the redox chemistry of the tantalum clusters.
Table 4. Summary of Data Characterizing Decomposition of Pentabenzyltantalum in Various Media
Figure 6. XANES edge position represented by the peak of the derivative of the X-ray absorption spectra characterizing (a) sample 4 (Ta/Al2O3), (b) sample 3 (Ta/Si2AlOy), (c) sample 1 (Ta/ Si30AlOy), and (d) Ta/SiO2. The inset shows a magnified view of the figure in the energy range of 9881-9885 eV, which clearly indicates the difference of edge position characterizing the tantalum clusters supported on the various supports.
this work
Article ref
Sun et al.
DOI: 10.1021/la901295d
10761
Article
bridging O atoms of the support may influence the chemistry, as had been inferred previously4d and considered further above. The observation that the incorporation of alumina in the support changes the chemistry of the decomposition raises the question of whether the support acid-base properties are important. SiO2, silica-alumina, and γ-Al2O3 differ from each other in their acidbase properties, and, specifically, in the electron-donor tendencies of the surface O atoms. SiO2 is a very weak acid and a weak base (having weak electron-donor tendencies). According to Sanderson’s principle,22 the incorporation of Al atoms in the SiO2 framework results in negatively charged framework O atoms, the electron-donor tendency of which is increased accordingly. Consistent with this picture are reports of the basicity of framework O atoms of zeolites, controlled by AlO(SiO)AlO sequences.23 The number of SiO units incorporated between AlO pairs can be used to tune the basicities of the framework O atoms;23a for example, when this number is 1, the electron-donor strength of the oxygen is strong, but when it is 2, the electron-donor strength becomes medium or weak. On the other hand, the incorporation of alumina into SiO2 also creates Lewis and/or Broensted acid sites, depending on the synthesis method, composition, and activation temperature. Usually, the number of Broensted acid sites per unit area on amorphous silica-aluminas decreases as the alumina content increases and the number of Lewis acid sites increases.5,24 The nature of the acidity also is dependent on the activation temperature: the higher this temperature, the greater the number of Lewis acid sites.24 On γ-Al2O3, dehydroxylation to remove various types of OH groups attached to multiply coordinated Al atoms gives rise to Lewis acid-base pairs. Lewis basic O atoms are expected to bridge tetrahedral and octahedral cations.25 Evidence that acidic sites on the silica-alumina supports participated in the activation of C-C bonds is provided by the observation of the cracking product benzene, the formation of which is expected to have been catalyzed by Broensted acid sites, whereas on the support with just Lewis acid sites (γ-Al2O3), almost no benzene was observed (see Figure 1, as well as the Supporting Information). This comparison supports the inference that the surface acid sites were involved in the decomposition of tantalum precursors;when the supports had substantial acidic character. However, the data that characterize the reactions on SiO2 and on γ-Al2O3, which lack the strongly Broensted acidic groups that catalyze cracking reactions and do not give benzene as a decomposition product, indicate that there is more to the pentabenzyltantalum decomposition chemistry than just cracking. When the support was SiO2, the decomposition products were toluene and bibenzyl, but not benzene. Toluene was also formed as the main product in the reaction of pentabenzyltantalum in C6D6, as observed by Schrock et al.,21 who inferred that the reaction proceeded by an intramolecular hydrogen transfer process. We suggest that, in this respect, the chemistry on SiO2 was comparable to that of Schrock et al.21 However, in the reaction in C6D6, the yield of bibenzyl formed from pentabenzyltantalum-which was considered to be the product of a free-radical chain process-was almost too low to (22) Sanderson, R. T. Chemical Bonds and Bond Energy; Academic Press: New York, 1976. (23) (a) Barthomeuf, D. Microporous Mesoporous Mater. 2003, 66, 1. (b) Barthomeuf, D. J. Phys. Chem. B 2005, 109, 2047. (24) Crepeau, G.; Montouillout, V.; Vimont, A.; Mariey, L.; Cseri, T.; Mauge, F. J. Phys. Chem. B 2006, 110, 15172. (25) Kn€ozinger, H.; Ratnasamy, P. Catal. Rev.-Sci. Eng. 1978, 17, 31.
10762 DOI: 10.1021/la901295d
Sun et al.
observe, whereas there was a substantial yield of bibenzyl resulting from the reaction on SiO2. Thus, we regard the formation of mainly bibenzyl from the tantalum species on SiO2 as evidence suggesting that there was more to the chemistry on SiO2 than to that in C6D6 and that the free-radical process was significant. We attribute the difference to the bridging oxygen ligands on SiO2, which we infer become coordinated to tantalum. Thus, we suggest that the tantalum complexes on SiO2 undergo predominantly a free-radical chain reaction process with possibly an intramolecular hydrogen transfer process, as occurs in C6D6.21 We consider the data to be consistent with a significant role of the support oxygen ligand to which tantalum is bonded in the chemistry of decomposition of the pentabenzyltantalum. The increased basicity (electron-donor tendency) of the bridging O atoms resulting from increases in the aluminum content of the support-to which tantalum was evidently bonded-is inferred to have modified the electronic properties of tantalum and, thus, increased the strength of the metal-carbon bond.26 However, the increased basicity of these surface O atoms induced by the aluminum was evidently not sufficient to activate the C-C bonds in the supported organotantalum species. We suggest that the increased Ta-C bond strength stabilized the benzyl ligands on tantalum, thus providing the basis for the activation of the C-C bonds with the increased temperature and the influence of surface acidity. Therefore, we suggest that both the electron-donor properties of the support O atoms and the acidity of the support are important. The Broensted acidity decreased as the aluminum content of the support increased, but our results indicate that the yield of benzene (inferred to be formed by cracking on Broensted acid sites) increased and the temperature of formation of benzene decreased as the aluminum content increased, and we have not resolved the issues that affect this pattern of the reactivity. In summary, the results indicate a significant effect of the incorporation of aluminum in the SiO2 support on the chemistry of decomposition of physisorbed pentabenzyltantalum. The various electron-donor tendencies of bridging O atoms induced by the alumina played the key role in tuning the strength of the Ta-C bond in the tantalum precursors. Support acidity is inferred to be involved in the activation of C-C bonds of supported organotantalum species at high temperatures. The combined influence of the electron-donor tendencies of the bridging O atoms and the support acidity is inferred to determine the decomposition chemistry of the supported organotantalum species. As a result of the decomposition of the physisorbed pentabenzyltantalum on the supports, tantalum clusters formed, as shown by EXAFS data and STEM images. The clusters formed on these supports were all small and raft-shaped, and the cluster structure was determined to be dependent on the support surface composition. The EXAFS data show that the average Ta-Ta bonding distance in the clusters on SiO2 was 2.93 A˚, but, as the support aluminum content increased, this distance decreased, to a value of 2.88 A˚ when the support Si:Al atomic ratio was 30 and to a value of 2.86 A˚ when the support was γ-Al2O3. The decrease in the Ta-Ta bonding distance with the decrease in Si:Al ratio corresponds to the change in the oxidation state of Ta. The XANES results show that the Ta LIII edge positions that characterize the clusters on the silica-aluminas and on γ-Al2O3 are lower in energy than those that characterize the clusters on SiO2. All of these values are greater than that which characterizes (26) (a) Davidson, P. J.; Lappert, M. F.; Pearce, R. Chem. Rev. 1976, 76, 219. (b) Clot, E.; Megret, C.; Eisenstein, O.; Perutz, R. N. J. Am. Chem. Soc. 2006, 128, 8350. (c) Qi, X.-J.; Li, Z.; Fu, Y.; Guo, Q.-X.; Liu, L. Organometallics 2008, 27, 2688.
Langmuir 2009, 25(18), 10754–10763
Sun et al.
Article
metallic tantalum (9881 eV), and this comparison indicates that the tantalum in the clusters was cationic. Nemana et al.4d correlated the edge positions recorded for a family of tantalum compounds with the tantalum oxidation state, suggesting that a change in edge position by ∼1 eV corresponds to an oxidation state change of ∼0.7 electronic charge per Ta atom. Accordingly, we suggest that the oxidation sate of tantalum in our samples was in the range of 2.6-3.0, which corresponds to the 0.4-eV shift in edge position with Si:Al ratios in our samples. In the synthesis of early-transition-metal clusters, the type of ligands present affects the cluster structure (e.g., the nuclearity and M-M bonding distance).3 For example, Cotton et al.3b determined that the symmetric trinuclear tantalum cluster core structure became unsymmetrical upon replacement of one anionic electron-donor ligand with neutral electron-donor oxygen ligands at one of the Ta atoms in the trinuclear tantalum clusters. ([Ta3Cl10(THF)3]- was converted to Ta3Cl9(THF)4.) Concomitantly, the Ta-Ta bonding distance decreased. Another example is provided by the comparison of [Nb3Cl10(PEt)3]- with [Nb3Cl7(PMe2Ph)6]-; Cotton et al. determined that, as a result of replacement of the anionic chloride ligands with the stronger electron-donor phosphine ligands, the clusters were reduced, with the number of Nb-Nb bonding electrons increasing from six to eight.3c Our results indicate the formation of tantalum clusters that are attached to the supports via bridging O atoms. The electron-donor abilities of the bridging O atoms increased as the Si:Al ratios of the supports decreased, which led to a significant effect on the decomposition of physisorbed pentabenzyltantalum. Correspondingly, we infer that the increased electron-donor abilities contributed to the decreasing oxidation state of the tantalum clusters on the supports. The EXAFS data demonstrate a change in the average Ta-O (or possibly TadC) coordination number with decreasing support Si:Al ratios. The number increased from 3.5 for SiO2 to 3.9 for silica-alumina with a Si:Al atomic ratio of 30 and then decreased monotonically to 2.83 for γ-Al2O3. The reason for the change is still not clear (and we emphasize that the data fall short of showing the details of how the clusters bond with the supports), but it does not affect our conclusion that the support composition affects the redox chemistry of the cationic tantalum clusters.
Conclusions Extended X-ray absorption fine structure (EXAFS), X-ray absorption near-edge structure (XANES), and scanning
Langmuir 2009, 25(18), 10754–10763
transmission electron microscopy (STEM) characterizations demonstrate that raft-shaped cationic tantalum clusters with a Ta-Ta coordination number of ∼3 were obtained on a series of supports including SiO2, silica-aluminas, and γ-Al2O3. The clusters were anchored to the supports via bridging O atoms. The different electron-donor abilities of the bridging O atoms, corresponding to the different support compositions, led to different chemistries of the decomposition of the physisorbed pentabenzyltantalum from which the supported clusters were formed; however, the effect of the support composition of the oxidation states and structures of the resultant tantalum clusters was small. Acknowledgment. We thank Jun Yang for helpful comments. This research was supported by the National Science Foundation (GOALI Grant No. CTS-05-00511) and by the U.S. Department of Energy, Office of Energy Research, Basic Energy Sciences (under Contract No. FG02-87ER13790, to R.J.L.-L.). X-ray absorption spectra were collected at the Stanford Synchrotron Radiation Laboratory (SSRL), a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Science, Basic Energy Sciences. The SSRL Structural Molecular Biology Program is supported by the Department of Energy, Office of Biological and Environmental Research, and by the National Institutes of Health, National Center for Research Resources, Biomedical Technology Program. We also acknowledge the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Beamline X-18B, for access to beam time; the NSLS is supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences (under Contract No. DEAC02-98CH10886). Beamline X-18B is supported by the NSLS, through the Divisions of Materials and Chemical Sciences of DOE and the Synchrotron Catalysis Consortium (U.S. DOE Grant No. DE-FG02-05ER15688). We thank the beamline staffs for their assistance. The electron microscopy experiments were performed at the Oak Ridge National Laboratory SHaRE User Facility, which is supported by the Division of Scientific User Facilities, DOE Office of Science, Basic Energy Sciences. Supporting Information Available: Reactor scheme, IR spectra, STEM images, EXAFS spectra, mass spectra of effluents, and effluent product distributions quantified by GC in the gas phase, and solid phase. This information is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la901295d
10763
