PNNL

Abstract

We study the vertical ionisation potential, electron affinity, fundamental gap and exciton binding energy values of small bare and hydroxylated TiO2 nanoclusters to understand how the excited state properties change as a function of size and hydroxylation. In addition, we have employed a range of many-body methods; including G0W0, qsGW, EA/IP-EOM-CCSD and DFT (B3LYP, PBE), to compare the performance and predictions of the different classes of methods. We demonstrate that for bare (i.e. non-hydroxylated) clusters all many-body methods predict the same trend with cluster size. The highest occupied and lowest unoccupied DFT orbitals follow the same trends as the electron affinity and ionisation potentials predicted by the many-body methods but are generally far too shallow and deep respectively in absolute terms. In contrast, the ?DFT method is found to yield values in the correct energy window. However, its predictions depend on the functional used and do not necessarily follow trends based on the many-body methods. The effect of hydroxylation of the clusters is to open up both the optical and fundamental gap. Finally, a simple microscopic explanation for the observed trends with cluster size and upon hydroxylation is proposed in terms of the Madelung onsite potential.