Seminars

Home Centers & Programs AI and Natural Sciences Seminars

FIELD: AI and Natural Sciences

DATE: Nov 04 (Tue), 2025

TIME: 15:00 ~ 17:00

PLACE: 7323

SPEAKER: Ha, Min Young

HOST: Yu, Ji Woong

INSTITUTE: 경희대학교

TITLE: Scale-bridging in supercritical fluids by machine learning: Critical phenomena and non-equilibrium phase separation

ABSTRACT: Supercritical fluids have a wide variety of applications in chemical industries as media for separation, extraction, and reaction, due to the anomalous blend of liquid-like and gas-like traits. Supercritical fluids simultaneously manifest microscopic density inhomogeneities and macroscopic crossover phenomena, yet the concrete relation between the anomalies at different length scales remains vague. Here, a framework is proposed to capture the physics of supercritical fluids to understand and predict the behavior of supercritical fluids at different length scales. Motivated by the mixture model approach on the local density distribution of supercritical fluid, particles in a model supercritical fluid are labeled as liquid-like or gas-like. Supercritical anomalies are strongly correlated to the transitions between the microstates, maximized when the number fractions are even. The thermodynamic Widom line, the traditional loci of macroscopic crossover phenomena, is redefined as the line of equal microstate fraction. Since the Widom line is enclosed in the deltoid region of microstate coexistence, the domain of supercritical coexistence is suggested to be called the Widom delta. Furthermore, the scaling function is defined as the gradient of the microstate fraction, which shows a power-law divergence behavior in the vicinity of the critical point. The liquid-gas critical point can be accurately located from the exponent of the scaling function, indicating that the macroscopic physics is effectively encoded in the microscopic partitioning of microstates. By reframing the thermodynamic Widom line within the Widom delta and linking it to microstate coexistence, we offer a perspective for bridging local density inhomogeneities and large-scale behavior in supercritical fluids. Finally, we apply the Widom delta framework to the local structure analysis of non-equilibrium phase separation in supercritical fluids to advance the multi-scale understanding of experimentally relevant processes.

FILE

List