Understanding ceria/noble metal interactions for water -gas -shift catalysts using density functional theory

Nathaniel Aaron Deskins, Purdue University

Abstract

The water-gas-shift (WGS) reaction (CO + H2O → H 2 + CO2) is important for CO removal. Ceria (CeO2) is an active potential support for WGS when combined with various metals (Pt, Au, Pd, Rh, Cu), and an understanding of the mechanism for WGS on ceria-supported catalysts is missing. Using density functional theory I have attempted to gain better understanding of ceria surfaces under WGS conditions. Ceria-supported metals that are active for WGS are known to exist in a cationic form as small particles. This work includes modeling substitution of lattice Ce4+ cations with Pt, Rh, Pd, Au, and Cu as well as adsorption of single atoms. Charge analysis suggests that substitution is feasible, while adsorption does not necessarily produce cationic metals. My results also show that substitution occurs with an increase in O defects, which could lead to increased catalyst reactivity, while increased O defect formation did not occur with adsorbed metals. Adsorption of CO on doped and non-doped surfaces generates carbonate (CO3) structures, with very strongly adsorbed carbonates (adsorption energy as high as 117 kcal/mol) forming on the doped surfaces. The presence of O defects near the doped metal gives weaker carbonates, but these are still fairly strongly bound, to the surface (< 40 kcal/mol). Energetics of a CO oxidation cycle show that doping leads to a more 'downhill' reaction pathway. Further work involved modeling hydroxylation of ceria surfaces via H 2 and H2O dissociation. My results show that hydroxyls from H2 are more stable at lower temperatures than from H2O. Co-adsorption was shown to decrease the stability of surface hydroxyls. My work also shows that O defects are likely H2O dissociation sites. I also examined formation of carbonates and formates, which are possible WGS intermediates. Comparison of my calculated vibrational frequencies and experimental frequencies shows good correlation. Filling O defect sites is proposed to lead to intermediate in-surface formates. Finally, the energetics of the proposed WGS mechanisms, redox and formate, are compared. The redox mechanism appears to be more energetically 'downhill', suggesting that it may be more feasible than the formate mechanism, though further work is needed.

Degree

Ph.D.

Advisors

Thomson, Purdue University.

Subject Area

Chemical engineering|Chemistry

COinS