Towards a physically sound description of zinc oxide (nano-)systems: SCC-DFTB – bridging a gap?

Stefan E. Huber (Technische Universität München, Garching, Germany), Matti Hellström, Michael Probst, Kersti Hermansson, and Peter Broqvist

Zinc oxide (ZnO) has a wide range of applications due to its unique catalytic, photo-catalytic, electronic and many other properties. For most of these properties, surface chemistry plays a crucial role, and thus, due to the enhanced surface-to-volume ratio, ZnO nano-systems are regarded as promising candidates for a variety of (future) technological applications such as for water-splitting, the synthesis of hydrogen peroxide or the reduction of graphene oxides to name just a few. For computational modeling, ZnO nanosystems represent a difficult task. Whereas density functional theory (DFT) methods are associated usually with severe limits in system size due to their computational demand, computationally cheaper force-field approaches lack in the description of electronic properties. In addition, the parameterization of force fields can be a tedious task. This is also related to the computational demand associated with the electronic structure methods usually used for the calculation of reference data against which force fields are then parameterized. Here, an attempt to bridge the gap between the quantum mechanical and atomistic scale, shall be presented. The goal is to use an approximate DFT method, namely density-functional tight-binding with self-consistent charges (SCC-DFTB) [1], as an intermediate step in the parameterization procedure. In particular, DFT is used for the parameterization of SCC-DFTB and the latter is then used for the major part of the parameterization of force-fields. This idea is based on the fact, that usually relatively little reference data is required to parameterize SCC-DFTB in contrast to the considerably larger training sets usually required for the parameterization of force fields. Thus, use of SCC-DFTB as an intermediate step can yield a considerable speed-up of the overall parameterization procedure. However, as an approximate DFT method, the reliability of SCC-DFTB relies per se heavily on the quality of the used parameters.

In this presentation, the focus is on the recent development, assessment and application of such an SCC-DFTB set of parameters for ZnO systems [2, 3]. It is found that SCC-DFTB in conjunction with the new set of parameters yields results in agreement with experimental and theoretical data concerning energetic, geometrical and electronic properties of ZnO bulk systems, low-index polar and non-polar surfaces and also surface-water interfaces. The method is thus capable of a reasonable description of ZnO nanosystems and can be set to use for the said purpose. In addition, its lower computational cost compared to DFT allows the treatment of considerably larger systems than is possible with DFT, while the method still retains the electrons in the model which allows the investigation of electronic properties. The latter feature makes the method also an interesting candidate in multiscale modeling frameworks in which the specific advantages of force fields and (approximate) electronic structure methods are combined to extract data at different spatial and temporal scales.

[1] Aradi, B.; Hourahine, B.; Frauenheim, Th; J. Phys. Chem. A 111 (2007) 5678.
[2] Hellström, M.; Jorner, K.; Bryngelsson, M.; Huber, S.E.; Kullgren, J.; Frauenheim, Th.; Broqvist, P.; J. Phys. Chem. C 117 (2013) 17004.
[3] Huber, S.E.; Hellström, M.; Probst, M.; Hermansson, K.; Broqvist, P.; Surf. Sci. 628 (2014) 50.