690
Density Functional Theory Study of the Alkali Metal Cation Adsorption on Pt(111), Pt(100), and Pt(110) Surfaces

Tuesday, May 13, 2014: 15:40
Indian River, Ground Level (Hilton Orlando Bonnet Creek)
I. Matanovic (University of New Mexico), P. Atanassov (University of New Mexico, Center for Emerging Energy Technologies), F. Garzon, and N. J. Henson (Los Alamos National Laboratory)
Vibrational spectroscopy provides the most definite means of identifying surface species generated upon molecular adsorption and the species generated by surface reactions. The ability to simulate observable vibrational spectra using density functional theory allows for a new level of interaction between the theoretical and experimental material science. The comparison between theoretically and experimentally observed spectra allows an indirect means for obtaining valuable information about the nature of the catalysts’ active sites or the mechanism of surface reactions. In previous work the local-density approximation (LDA) and different generalized-gradient approximations  (GGA) were assessed as to their ability to reproduce bond lengths and vibrational frequencies of small polyatomic molecules [1]; however, corresponding systematic studies involving surface bound species are lacking.  In this work we will discuss the capability of different approximations of the exchange-correlation functional to reproduce experimental vibrational spectra of a test suite of surface adsorbates [2] and then apply the identified optimal procedure to simulate experimentally observed vibrational spectra of more complex nitrogenous species on molybdenum-nitride.      

In the first part of the work we used a combination of different exchange-correlation functionals and pseudopotentials to reproduce adsorption energies, equilibrium bond lengths, and vibrational frequencies of N2, H2 and D2 on the Rh(111) surface. In all the cases, a softening of most vibrational modes was observed. For example, the frequency of the N-N stretching mode of N2 on Rh(111) was calculated as 2181, 2180 and 2198 cm-1 using GGA-PBE, metaGGA-MO6L, and metaGGA-TPSS, respectively while the experimental frequency was determined as 2256 cm-1[2]. All our further test calculations were motivated by identifying the most important parameters which lead to best agreement with experiment. For example, we tried to assess the importance of reproducing accurate bulk/surface structural parameters, coupling with phonons, and validity of the harmonic approximation.

The optimal level of theory obtained from the calculations on the model systems is used to calculate the vibrational spectra of nitrogenous species on molybdenum nitride and then compared to the spectra obtained with diffuse reflectance-FTIR and attenuated total reflectance-FTIR.  Namely, the transformation of nitrogen and hydrogen to ammonia on molybdenum nitride was monitored in-situ in order to study the catalytic activity of molybdenum nitride towards reduction of nitrogen to ammonia. To complement the experiment we studied the adsorption and vibrational spectra of different species that are considered to be intermediates in the ammonia synthesis, that is H, N, NHx, and NNHx, x=1-3. The comparison between the theoretical and experimentally observed IR frequencies provided valuable information on the reaction mechanism and the kinetics of molybdenum catalyzed ammonia synthesis.

[1] D. P. Patton, D. V. Porezag and M. R. Pederson, Phys. Rev. B, 55(12) 7454 (1998)

[2] J. P. Wey, H. D. Burkett, W. C. Neely, and S. D. Worley, J. Am. Chem. Soc., 1132919 (1991)

Figure1. DFT-optimized structures of N2 on, Rh(111) (left) and γ-Mo2N(111) (right)