20762
2006, 110, 20762-20764 Published on Web 10/04/2006
Weak Hydrogen Bonding Can Initiate Alkane C-H Bond Activation in Acidic Zeolites Laura S. Sremaniak,‡ Jerry L. Whitten,‡ Matthew J. Truitt,‡ and Jeffery L. White*,† Department of Chemistry, Oklahoma State UniVersity, Stillwater, Oklahoma 74075, and Department of Chemistry, North Carolina State UniVersity, Raleigh, North Carolina 27695 ReceiVed: September 8, 2006
Ab initio calculations at the Hartree-Fock self-consistent field/single determinant (SCF) and configuration interaction multi-determinant (CI) expansion levels have been used to show that isobutane primary C-H bond activation occurs via direct protium exchange with the zeolite surface via a weakly hydrogen-bonded complex. The calculated 15 kcal/mol activation barrier agrees with the 13.7 kcal/mol value from a recently reported experimental study (J. Am. Chem. Soc. 2006, 128, 1847-1852). Overall, the mechanism described in this contribution demonstrates that weak C-H to O hydrogen bonding leads to complexes at the zeolite acid site that can facilitate C-H bond activation.
Mechanistic steps for the initiation and conversion of saturated hydrocarbons over solid acids are ambiguous relative to analogous reactions involving alkenes or oxygenates.1,2 Reactions involving alkanes, such as isomerizations and alkylations, are industrially relevant and fundamentally intriguing from a first principles chemistry perspective. How does an alkane react in a solid acid? Currently, carbenium ion chemistry catalyzed by trace olefinic impurities is considered the most probable route to alkane initiation via hydride transfer, since it is believed that zeolites, unlike superacid solutions, lack sufficient acid strength to directly protonate an alkane and form a pentavalent carbonium ion.3 However, a sufficient number of recent publications still propose this direct protonation step.4-7 Recently, we have published detailed experimental investigations of isobutane reactions on acidic zeolite HZSM-5, in which we observed direct proton exchange between the methyl groups of isobutane and the Brønsted acid site of the zeolite.8,9 The experimentally determined activation barrier for this process, as observed by in-situ solid-state NMR and isotopic labeling methods, was 13.6 kcal/mol. From this work, the following novel mechanism, involving a weakly hydrogen-bonded adsorption complex, was proposed for the experimentally detected H/D exchange pathway between isobutane and the acidic zeolite surface (Scheme 1).8 Here, we provide computational confirmation of Scheme 1 using ab initio calculations at the (i) Hartree-Fock self-consistent field/single determinant (SCF) and (ii) configuration interaction multi-determinant (CI) expansion levels.10,11 In contrast to some recently published density functional theory (DFT) studies for the isobutane on an acidic zeolite,12 the current results employ a realistic surface model (-OH termination instead of -H), using a more accurate CI level treatment for transition state optimization, and demonstrate through agreement with published experimental data that the direct proton exchange pathway is operable. One may argue that a cluster model is insufficiently * To whom correspondence should be addressed. E-mail: jeff.white@ okstate.edu † Oklahoma State University. ‡ North Carolina State University.
10.1021/jp0658703 CCC: $33.50
large to accurately model this process, as confinement effects cannot be accurately addressed without considering cage/channel topology. While cage effects are no doubt important, the most dominant interaction is that of the molecule with the acid site and its closest atoms. Indeed, Sauer has already shown that even isobutene, a more reactive molecule than isobutane, cannot approach the acid site at the methine position due to steric interference with the cluster of atoms immediately surrounding the acid site.13 The literature is currently filled with publications using a lower level of theory to treat a larger collection of atoms using commercial software packages. This fact alone should not exclude the consideration of more rigorous computations on the necessarily smaller cluster sizes, as long as the cluster most accurately represents the active site as allowed by the computational capabilities. As Figures 1 and 2 show, our cluster is terminated with hydroxyl groups instead of hydrides to more properly emulate connectivity to an extended system. We also emphasize that the mechanism addresses only the very first steps involved in alkane activation at the active site of the zeolite; in no way do the previously published experimental data or the new calculations reported here exclude the possibility of subsequent reactions involving the traditionally proposed loss of hydrogen by the surface-adsorbed pentavalent transition state (see Scheme 1), and further reaction of the carbenium ion.14 Indeed, the surface adsorbed transition state shown in Scheme 1 is a pentavalent carbonium ion stabilized by the zeolite surface, similar to that proposed by Kramer et al. for methane.7 The calculations we describe here address the experimentally proven (stoichiometric and regiospecific) H/D exchange between a single isobutane molecule and the acid site,8,15 the formation of which is facilitated by nonclassical hydrogen bonding, and which can act to actually start the catalytic cycle. Any subsequent steps involving hydrogen loss and carbenium ion chemistry via bimolecular pathways will require additional calculations. The SCF/CI methods are used to determine the energy barrier to proton exchange, and all reported charges are from an atomic population analysis. A flexible basis set is used for Si {13 s-type, © 2006 American Chemical Society
Letters
J. Phys. Chem. B, Vol. 110, No. 42, 2006 20763
SCHEME 1
9 p-type, 1 d-type Gaussians} and Al {13 s-type, 7 p-type, 1 d-type Gaussians}, where the d functions have an exponent of 0.8. Oxygen atoms are also described by a basis of comparable quality {10 s-type, 5 p-type, 1 d-type Gaussians}. The dfunctions on Si and O are primarily for electron correlation. The C atoms of the isobutane molecule are described with 10 s-type, 5 p-type and 1 d-type Gaussians, and all hydrogen atoms in the zeolite and the isobutane are described with 5 s-type functions; additional p-type functions are added to the basis of the two hydrogen atoms that are transferred. The zeolite cluster is as depicted in Scheme 1 and the Si-O bonds are terminated with hydrogen-like atoms that model bonding to the extended network.16 The energy minimization was carried out via a Nelder Mead17 simplex algorithm that allowed the simultaneous optimization of 27 geometrical parameters. Most of the optimizations were carried out at the SCF level, and then followed by CI calculation of the energy at the determined geometry. However, the optimization was done directly at the CI level for the transition state. The ability to include correlation within the optimization is the main advantage of this method and offsets the disadvantage of numerous energy calculations. Steps along the proton exchange pathway are shown in Figure 1 and the transition state with atomic labels is shown in Figure
Figure 1. Calculated energy landscape for the surface catalyzed H/D exchange for isobutane/HZSM-5 depicted by Scheme 1. Atoms are denoted by the following color codes: Al ) light blue; H ) dark green; O ) blue; Si ) light green. Note that Si and Al atoms are bound to oxygen. Energies in kcal/mol are relative to the desorbed isobutane.
2. Together, Figures 1 and 2 confirm the qualitative proposed mechanism in Scheme 1. After weakly interacting with the zeolite cluster, isobutane rotates such that the unprotonated oxygen (O4) of the zeolite comes within bonding distance of H28 of the methyl group (Figure 1a). Structures (a) and (e) depict specific geometries calculated for isobutane; structure (a) is equivalent to full desorption and (b) nearly completely desorbed and only weakly interacting with the surface. The acidic hydrogen of the zeolite cluster (H36) then moves such that the C-H36-O5 angle increases to 164°, allowing for a bridging structure to form. The C-H28 and C-H36 distances are both 1.5 Å, and the O4-H28 distance is 1.3 Å while the O5-H36 is 1.1 Å (Figure 2). The computed activation energy is 14.5 kcal/mol at the CI level, which is in good agreement with the experimental value. The final position of the H36 on the desorbing isobutane is the same as the initial position of H28. The Mulliken charges of the oxygen and hydrogen atoms participating in the proton exchange are shown in Figure 3, where the labeled points (a-e) correspond to the configurations identified in the reaction coordinate of Figure 1. The charge on the Brønsted acid site cluster, in absence of the isobutane, is +0.74|e|. This dissociated limit is not shown in the figure. Once the isobutane has bonded, the charge on this proton has already decreased to +0.29|e| (point a for H36). As the complex proceeds along the reaction coordinate, this charge continues to decrease as the charge on H28, initially of the isobutane, increases from its most negative value of 0.11 on free isobutane to its final value of +0.29|e|. The oxygen atoms likewise show
Figure 2. Optimized transition state showing “tilted” adsorption complex. Atoms are denoted by the following color codes: Al ) light blue; H ) dark green; O ) blue; Si ) light green. Note that Si and Al atoms are bound to oxygen.
20764 J. Phys. Chem. B, Vol. 110, No. 42, 2006
Letters chemistry via the computationally and experimentally verified surface-stabilized complex shown here, isobutane is unique among them in its low activation barrier pathway. The relationship between the structure of the isobutane and the zeolite channel most likely provide an ideal “solvent” for facilitating H/D exchange; the details of specific zeolite structure solvent effects relative to alkane transition states will be emphasized in future work. Acknowledgment. The PI gratefully acknowledges the ACS Petroleum Research Fund (#38293-G5), North Carolina State University, and Oklahoma State University for support of this work. Support of the NC State portion of the project by the U.S. Department of Energy is gratefully acknowledged. References and Notes
Figure 3. Fractional charges on the two lattice oxygens O4 and O5, and initial/final Brønsted acid site protons (H36/H28) involved in the transition state.
a crossing of net charge values centered about the transition state (point d). As an alternative mechanism, we also calculated energy barriers for a direct protonation of the methyl group via a symmetric transition state (methyl C3 axis directly above Brønsted proton on surface) to form a classical pentacoordinated carbonium ion. The carbonium ion would then diffuse to the adjacent oxygen and reprotonate the surface. While this mechanism is geometrically less complicated, the resulting 89 kcal barrier is almost an order of magnitude too high to agree with the experimental data.8 Overall, the mechanism described here demonstrates that even a small zeolite cluster (properly terminated) has the ability to solvate alkanes and can facilitate their low-barrier reactions via weak C-H to O hydrogen bonded complexes. In addition, the combined implications of the computational and experimental evidence on isobutane8,9 indicate that weak, nonclassical hydrogen bonds involving aliphatic C-H groups and O atoms are active in heterogeneous catalysis. Hydrogen bonds of this type are distinctly different than previously described cases in zeolites involving adsorbates with sp2 carbons.18 Indeed, such weak, nonclassical hydrogen bonds have been recently reviewed in the small-molecule literature,19 and this growing area of C-H to O hydrogen bonds is currently receiving widespread attention in the broader literature.20,21 We note that relative to the published literature dealing with H/D exchange between small alkanes (methane,7 ethane, propane,6 isobutane,15 and neopentane22) and HZSM-5, only isobutane exhibits a detectable reaction rate at room temperature and below.8,9 While all of these alkanes may be expected to initiate
(1) Weisz, P. B. Microporous Mesoporous Mater. 2000, 35-36, 1. (2) Milas, I.; Nascimento, M. A. C. Chem. Phys. Lett. 2001, 338, 67. (3) Haw, J. F.; Nicholas, J. B.; Xu, T.; Beck, L. W.; Ferguson, D. Acc. Chem. Res. 1996, 29, 259. (4) Esteves, P. M.; Nascimento, M. A. C.; Mota, C. J. A. J. Phys. Chem. B 1999, 103, 10417. (5) Mota, C. J. A.; Esteves, P. M.; Hernandez-Lamoneda, R. J. Am. Chem. Soc. 1997, 119, 5193. (6) Arzumanov, S. S.; Stepanov, A. G.; Parmon, V. N.; Reshtikov, S. I.; Freude, D. J. Phys. Chem. B 2005, 109, 19748. (7) Kramer, G. J.; van Santen, R. A.; Emeis, C. A.; Nowak, A. K. Nature (London) 1993, 363, 529. (8) Truitt, M. J.; Rovira, R.; White, J. L. J. Am. Chem. Soc. 2006, 128, 1847-1852. (9) Truitt, M.; Toporek, S.; Rivera, R.; White, J. L. J. Am. Chem. Soc. 2004, 126, 11144. (10) Whitten, J. L.; Yang, H. Surf. Sci. Rep. 1996, 24, 55. (11) Whitten, J. L.; Yang, H. Int. J. Quantum Chem. Quantum Chem. Symp. 1995, 29, 41. (12) Zheng, X.; Blowers, P. J. Phys. Chem. A 2006, 110, 2455. (13) Tuma, C.; Sauer, J. Angew. Chem., Int. Ed. 2005, 44, 4769. (14) Stocker, M. Microporous Mesoporous Mater. 2005, 82, 257. (15) Sommer, J.; Habermacher, D.; Jost, R.; Sassi, A.; Stepanov, A. G.; Luzgin, M. V.; Freude, D.; Ernst, H.; Martens, J. J. Catal. 1998, 181, 265. (16) Whitten, J. L.; Zhang, Y.; Menon, M.; Lucovsky, G. J. Vac. Soc. Technol. 2002, B20, 1710-1719. (17) Nelder, J. A.; Mead, R. Comput. J. 1965, 7, 308. (18) White, J. L.; Beck, L. W.; Haw, J. F. J. Am. Chem. Soc. 1992, 114, 6182. (19) Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond in Structural Chemistry and Biology; Oxford Science Publications: Oxford, 1999. (20) Vargas, R.; Garza, J.; Dixon, D.; Hay, B. P. J. Am. Chem. Soc. 2000, 122, 4750. (21) Sola`, J.; Riera, A.; Verdaguer, X.; Maestro, M. A. J. Am. Chem. Soc. 2005, 127(39) 13629-13633. (b) Ratajczyk, T.; Czerski, I.; KamienskaTrela, K.; Szymanski, S.; Wojcik, J. Angew. Chem. Int. Ed. 2005, 44, 1230. (c) Wang, X. B.; Woo, H. K.; Kiran, B.; Wang, L. S. Angew. Chem. Int. Ed. 2005, 44, 4968. (d) Caminati, W.; Lopez, J. C.; Alonso, J. L.; Grabow, J. U. Angew. Chem. Int. Ed. 2005, 44, 3840. (e) Shi, Z.; Olson, C. A.; Bell, A. J., Jr.; Kallenbach, N. R. Biophys. Chem. 2002, 101, 267. (22) Walspurger, S.; Sun, Y.; Sido, A. S. S.; Sommer, J. J. Phys. Chem. B, in press.
