Missions

Recruited from the CNRS, the researcher will apply and develop a CSP (crystal structure prediction) methodology for the in silico prediction and design of materials for space propulsion. This could be based on generative prediction using diffusion models, genetic/evolutionary algorithms, or even randomization (etc.). Binary and ternary phase diagrams under pressure will be explored. From the generated crystal structure, he/she will determine various properties (energy gap, thermodynamic, dynamic and thermal stabilities (AIMD), chemical bond analysis, etc.) using quantum chemical methods. One of the tasks will be to write regular activity reports and scientific publications in English, including activity reports for CNES. He/she will have to interact with, among others, CNES scientific managers within the framework of the defined Study. He/she will participate in the scientific coordination of the research group, as well as in the supervision of trainees.

Activités

- Exploit AI databases of crystalline structures. Use generative AI to predict crystalline structures, taking crystalline properties into account. Predict new materials according to several selection criteria (energy, mechanical properties, detonation, etc.). Use multi-objective approaches for materials design.
- Calculate electronic structures of crystalline structures generated in silico, thermodynamic/dynamic (phonon)/mechanical properties, DFT/MLIP molecular dynamics simulations and exploit results.
- Present results via regular reports, conferences and scientific articles.
- Installation and monitoring of calculation codes on the GENCI computing centers and the IC2MP local cluster
- Participate in the IC2MP applied quantum chemistry group and the consortium associated with ANR and other research group projects. Interaction in English.
- Supervision of students on research internships (max. 3/year, L3 to M2) and co-supervision of PhD students.

Compétences

-- PhD in theoretical chemistry applied to materials, materials physics, computer science/applied mathematics
- strong experience in generative prediction algorithms and/or DFT applied to periodic systems (e.g. VASP code, xtb, QE etc.) and/or MLIP (Machine Learned Potentials) approaches, and/or multi-objective approaches.
- In-depth knowledge of Python programming languages (or even C++, Fortran) and the Unix system; demonstration of scripting and open-source code
- Co-author of several internationally-renowned scientific articles.
- Certified level in written and spoken English.

Contexte de travail

The work will be carried out at the CNRS Institut de Chimie des Milieux et Matériaux de Poitiers (IC2MP - UMR CNRS 7285) in the Catalyse et Milieux Non-conventionnels team (https://ic2mp.labo.univ-poitiers.fr/). He/she will work within the "Chimie Quantique Appliquée" research group at IC2MP, under the supervision of the project's scientific manager, specific agreement CNES n°250131/00 and CNRS n°LSP307751.

Contraintes et risques

Short mission possible