About this Abstract |
Meeting |
MS&T23: Materials Science & Technology
|
Symposium
|
Ceramics and Glasses Modeling by Simulations and Machine Learning
|
Presentation Title |
A B-C Story, Investigated by A.I. and CALPHAD |
Author(s) |
Olivier Hardouin Duparc, Romuald Béjaud, Antoine Jay, Olivier Rapaud, Nathalie Vast |
On-Site Speaker (Planned) |
Olivier Hardouin Duparc |
Abstract Scope |
The highly polymorphous boron and carbide elements give birth to a phase diagram that has been hotly debated on the boron rich side where the extremely hard and useful ‘B4C’ phase is actually composed of several structures made of slightly distorted (B12-nC) icosahedra connected and chained by various types of carbon-boron chains. The exact determination of these many structures still implies well trained natural intelligence based on experimental studies and ab initio calculations with pressure and temperature with unit cells containing up to 414 atoms. The incorporation of newly proposed atomic structures into an adequately developed CALPHAD model has enabled us to generate a consistent B-C phase diagram that explains the large domain of the ‘B4C’ phase experimentally observed at high temperature (viz. from 8.7 to 20 C at.%). Our model also reveals the importance of configuration entropy for the stabilization of this large domain. |