A Single Microscopic State to Characterize Ordering Tendency in Discrete Multicomponent System

TL;DR

This paper introduces a novel approach to characterize ordering tendencies in multicomponent systems using a single microscopic state, called the projection state, which simplifies analysis without requiring detailed energy or temperature information.

Contribution

The authors derive a modified formulation to determine pair probabilities in multicomponent systems using only one projection state, extending explicit relationships up to five components.

Findings

Validated the formulation through thermodynamic simulations.

Reduced computational complexity for multicomponent systems.

Provided explicit formulas up to R=5 components.

Abstract

Our recent study reveals that macroscopic structure in thermodynamically equilibrium state and its temperature dependence for classical discrete system can be well-characterized by a single specially-selected microscopic state (which we call projection state: PS), whose structure can be known a priori without any information about energy or temperature. Although PS can be universally constructed for any number of components R, practical application of PS to systems with R >= 3 is non-trivial compared with R = 2 (i.e., binary system). This is because (i) essentially, multicomponent system should inevitably requires linear transformation from conventional basis functions to intuitively-interpreted cluster probability basis, i.e., multiple PS energies are required to predict one chosen pair probability, leading to practically accumulating numerical errors, and (ii) additionally, explicit formula for the transformation from basis functions to pair probabilities should be required, which has been explicitly provided up to ternary (R = 3) system so far. We here derive modified formulation to directly determine probability for likeand unlike-atom pair consisting of any chosen elements by using a single PS energy, with providing explicit relationship between basis functions and pair probabilities up to quinary (R = 5) systems. We demonstrate the validity of the formulation by comparing temperature dependence of pair probabilities for multicomponent systems from thermodynamic simulations.