Toward an Understanding of the Local Origin of Negative Thermal Expansion in ZrW2O8: Limits and Inconsistencies of the Tent and Rigid Unit Mode Models

Although zirconium tungstate (ZrW2O8) is the most popular negative thermal expansion (NTE) material, the exact mechanism responsible for its NTE still remains controversial. Specifically, the “Tent” model [Cao et al. Phys. Rev. Lett. 2002, 89, 215902; Bridges et al. Phys. Rev. Lett. 2014, 112, 045505] and the “rigid unit mode” (RUM) model [Tucker et al. Phys. Rev. Lett. 2005, 95, 255501] were subjects of debate during recent years. This work aims to shed light on this issue by means of molecular dynamics simulations which allow us to separate, for each bond distance, the “true” thermal expansion from the “apparent” thermal expansion, as well as to study the effective bond strength and the anisotropy of relative thermal motion. In spite of the good agreement with the experimental data of Cao, Bridges, and co-workers, a decrease of the “true” W−Zr distances has been observed accompanied by large transverse vibrations of the O atoms in the middle of the W−O−Zr linkage, in sharp contrast to the “tent” model. Moreover, in contrast to the RUM model, it has been found that the WO4 and ZrO6 polyhedra are strongly distorted by thermal motion, and, more importantly, that intra-polyhedra contributions to NTE are present. Accordingly, we can conclude that both the tent and RUM models are inadequate to explain NTE in ZrW2O8, and a more flexible model, simply based on rigid nearest W−O and Zr−O bonds and tension effect, should be adopted.