Computational Modelling Group

Seminar  22nd November 2011 10 a.m.  27/2003

Modelling the Solid State: From oxygen storage capacity to transparent semiconductors

Professor Graeme W. Watson
School of Chemistry and CRANN, Trinity College Dublin, Dublin 2, Ireland

Web page
http://www.tcd.ie/Chemistry/staff/people/gww/gw_new/group/members/watson/
Categories
AMBER, Biomechanics, Biomolecular simulations, Carbon nanotubes, CASTEP, Catalysis, Complex Systems, Computer Science, FFT, Finite differences, Finite elements, Fortran, Gaussian, Git, GPU, HECToR, HPC, Linux, Metals, Micromagnetics, Molecular Dynamics, Molecular Mechanics, Monte Carlo, NWCHEM, Onetep, Photonics, Pylab, Quantum Chemistry, Quantum Dynamics, Semiconductors, Sensors, Software Engineering, Structural biology, Structural dynamics, Superconductivity, SVN, Systems biology, Tribology, VTK, Xmgrace
Submitter
Chris-Kriton Skylaris

Professor Graeme Watson

Abstract

Density Functional Theory (DFT) has become a standard approach to modelling solid state materials and we will demonstrate its utility in two areas. The first will investigate the doping of ceria (CeO2) with divalent noble metal ions. This has been shown to improve the reducibility and enhance the oxygen storage capacity (OSC) of ceria, although the reasons for this are not well understood.[1] We have examined the interaction of a range of divalent dopants with CeO2 using density functional theory, and found that the dopant preferentially adopts the coordination of its own oxide within CeO2, instead of the cubic coordination of Ce(IV) in ceria.[2] Depending on the electronic structure of the dopants, the different coordinations can create weakly or undercoordinated oxygen ions that are more easily removed than in pure CeO2. We have used these insights to identify dopants which will increase the reducibility of CeO2, whilst being economically more viable than the presently used noble metals.[3] To achieve this we have had to carefully consider the self-interaction error within DFT and we will discuss briefly the approach taken. The second area is that of transparent conducting oxides. CdO has been studied for decades as a prototypical wide band gap transparent conducting oxide with excellent n-type ability. Despite this, uncertainty remains over the source of conductivity in CdO, and over the lack of p-type CdO, despite its VBM being high with respect to other wide band gap oxides. In this study, we use screened hybrid density functional theory to study intrinsic defects and hydrogen impurities in CdO, and identify for the first time the source of charge carriers in this system. We explain why the oxygen vacancy in CdO acts as a shallow donor, and does not display negative-U behaviour similar to all other wide band gap n-type oxides. We also demonstrate that p-type CdO is not achievable, as n-type defects dominate under all growth conditions. Lastly we estimate theoretical doping limits, and explain why CdO can be made transparent by a large Moss-Burnstein shift caused by suitable n-type doping. We argue that the defect chemistry of CdO serves as a guide for the design of next generation transparent conducting oxides.[4]

References

1 M. S. Hegde, G. Madras, K. C. Matil, Acc. Chem. Res. 42 (2009) 704.

2 D. O. Scanlon, B. J. Morgan, G. W. Watson, Phys. Chem. Chem. Phys. 13 (2011) 4279.

3 A. B. Kehoe, D. O. Scanlon, G. W. Watson, Chemistry of Materials 23 (2011) 4464.

4 M. Burbano, D. O. Scanlon and G. W. Watson, Journal of the American Chemical Society, 133 (2011) 15065.