46131-G5
Ab Initio Study of Support Effects in the Direct Oxidation of Propylene to Propylene Oxide
1.Introduction
Propylene oxide (PO) is one of the most important materials produced by chemical industry. The current industrial methods of producing propylene oxide are either not profitable or are environmental unfriendly due to the production of either chlorinated or peroxycarboxylic acid waste 1. Despite the considerable research effort, there is no current industrial scale direct partial oxidation process for propylene epoxidation using molecular oxygen over a heterogeneous catalyst. If novel active and selective catalysts could be developed, their potentially economic and environmental impact can be considerable.
Stefan Vajda at Argonne National Lab is currently conducting experiments with size selected Ag clusters supported on alumina substrates using a combination of temperature programmed reaction and in situ grazing-incidence small angle X-ray scattering. We have been in contact with Vajda’s group and also that of Larry Curtiss of Argonne National Lab for exchange of ideas regarding the chemistry of direct partial oxidation of propylene over these catalysts. Results from Vajda’s work (which will be part of a planned joint publication) reveal a selectivity of about 75% to propylene oxide. In the temperature regime from room temperature to 100°C, the activation energy for propylene oxide production (attributed to Ag3) is estimated to be 27.5 kJ/mol. Above 100 ºC, the clusters begin to sinter, although the TOF of the catalyst does not change (i.e. loss of activity correlates with loss of surface area). Based upon these results, we hope to provide an understanding of the reaction mechanism and potential insight into improvement of the catalyst through our calculations.
2. Results
Calculations were performed using the VASP code, a plane wave periodic supercell density functional theory software package2, to explore potential pathways for the formation of propylene oxide on a gas phase Ag3 cluster in the presence of oxygen. For the purposes of this study we have chosen to focus on the D3h structure which is isoenergetic with linear Ag3. We find that molecular oxygen bonded to a single Ag atom is exothermic by 1.19 eV while side-on adsorption for the O2 molecule is slightly more favorable (-1.46 eV). Oxygen dissociation over the Ag3 cluster is also exothermic with respect to gas phase O2 by 0.66 eV, but therefore is endothermic with respect to adsorbed O2. Calculations indicate a barrier of 2.74 eV to dissociate O2 on the Ag3 cluster, too large to be involved in our reaction mechanism.
We investigated three pathways for propylene oxide formation: (1) a Langmuir Hinshelwood mechanism involving the direct abstraction of O atom from Ag3O to the double bond of propylene, (2) a Langmuir Hinshelwood mechanism involving the abstraction of O from an adsorbed O2 molecule on Ag3, i.e. Ag3O2, and (3) and Eley Rideal mechanism involving the abstraction of O from Ag3O via adsorption of C3H6 into an oxymetallacyle.
Transition state searches for Langmuir Hinschelwood type mechanism involving coadsorbed C3H6 and atomic oxygen could not locate a direct pathway to propylene oxide but rather involve an oxymetallocycle intermediate which is isoenergetic with respect to adsorbed oxygen and propene. The oxygen atom must first insert itself into the C-Ag bond in order to create the oxymetallocycle which involves a barrier of some 1.15 eV. The oxymetallocycle can then rearrange rather easily to form propene oxide with a barrier of only 0.16 eV in a mildly exothermic (-0.28 eV) step. The second pathway whereby propylene reacts with an adsorbed O2 molecule and leaves an oxygen atom remaining on the cluster is unfavorable with a barrier of ~2.5 eV, similar to O2 dissociation.
The most favorable mechanism that we have found involves an Eley Rideal type reaction of propene with atomic oxygen. The barrier for the oxametallocycle (0.38 eV) is consistent with the observed experimental barrier of 0.28 eV. We have not explored the barrier for oxygen dissociation on the amorphous alumina surface, but this could be larger than that for the desorption of the oxygen from the Ag3 cluster and could be rate limiting. As indicated above, the cluster by itself cannot break the O2 bond easily and the abstraction of O from O2 by propene is also difficult. Therefore, the formation of O atoms which can subsequently react to form propyelene oxide likely involves interaction with the support.
In the second year of this work, we will focus upon supported Ag clusters, specifically examining TiO2 as a model support. Previous work from Haruta involving Au/TiO2 catalysts has revealed large differences between anatase (which is active) and rutile (which is not)3.References
[1] K. Weissermel, H.J. Arpe, Industrial Organic Chemistry, Weinheim, Germany, Wiley-VCH, 2003, 146.
[2] G. Kresse, J. Furthmuller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys Rev B 54 (1996) 11169.
[3] M. Haruta, B.S. Uphade, S. Tsubota, A. Miyamoto, Selective oxidation of propylene over gold deposited on titanium-based oxides, Res Chem Intermediat 24 (1998) 329
