Combining density functional theory (DFT) and pair distribution function (PDF) analysis to solve the structure of metastable materials: the case of metakaolin

TLDR

The atomic structure of complex metastable materials is critical yet difficult to resolve with standard techniques. The study introduces a novel thermodynamics‑based, constraint‑free method to determine the structure of metakaolin. The method iteratively refines neutron PDF data with least‑squares real‑space fitting and density‑functional theory geometry optimisation. The resulting structure is energetically feasible, matches experimental data, reveals tri‑coordinated aluminium, and enables atom‑level tailoring of metakaolin for improved macro‑scale performance.

Abstract

Understanding the atomic structure of complex metastable (including glassy) materials is of great importance in research and industry, however, such materials resist solution by most standard techniques. Here, a novel technique combining thermodynamics and local structure is presented to solve the structure of the metastable aluminosilicate material metakaolin (calcined kaolinite) without the use of chemical constraints. The structure is elucidated by iterating between least-squares real-space refinement using neutron pair distribution function data, and geometry optimisation using density functional modelling. The resulting structural representation is both energetically feasible and in excellent agreement with experimental data. This accurate structural representation of metakaolin provides new insight into the local environment of the aluminium atoms, with evidence of the existence of tri-coordinated aluminium. By the availability of this detailed chemically feasible atomic description, without the need to artificially impose constraints during the refinement process, there exists the opportunity to tailor chemical and mechanical processes involving metakaolin and other complex metastable materials at the atomic level to obtain optimal performance at the macro-scale.