Enhancing the reliability of Reverse Monte Carlo simulations of metallic glass structure by imposing strict constraints from partial pair correlation functions

XR Wei and ZC Lu and YB Zhang and JE Dong and Y Huang and HB Ke and FQ Meng and JK Zhao and BS Shang and D Ma, COMPUTATIONAL MATERIALS SCIENCE, 244, 113169 (2024).

DOI: 10.1016/j.commatsci.2024.113169

Reverse Monte Carlo (RMC) simulations are widely used to reconstruct 3-dimensional (3D) metallic glass (MG) structure from neutron/X-ray diffraction data, but their reliability is debated. Here, we employ RMC to simulate a Zr50Cu50 MG's structure using two types of "experimental" inputs as constraints, one with a total pair correlation function (TPCF) and the other with 3 partial PCFs (PPCFs), all of which were generated from a 3D glass structure obtained from molecular dynamics (MD) simulations of the same alloy. This MD-generated structure is then used as a benchmark to evaluate RMC's accuracy. Notably, both methods can replicate the MG's TPCFs, and total bond-angle/bond-length distributions. However, only the PPCFs-constrained RMC is able to re- produce the local structure, as manifested by nearly replicated partial bond-length/bond-angle distributions, partial coordination numbers, atomic volume, etc. This work demonstrates that PPCFs can impose stricter constraints than a TPCF does, significantly improving the reliability of RMC simulations.

Return to Publications page