Distinguished Talk #4
CNRS researcher, Université Paul Sabatier, Toulouse
MULTI-SCALE MODELS AND MACHINE LEARNING FOR A BETTER UNDERSTANDING AND PERFORMANCE PREDICTION OF THE CARBON ELECTRODE/ELECTROLYTE INTERFACE IN SUPERCAPACITORS (Wednesday, June 30)
Progress in the development of novel energy storage systems is hampered by our lack of understanding of the microscopic mechanisms that determine their performance. The key issue is that phenomena on the atomistic scale have consequences on macroscopic length and timescales. In particular, the effects of ionic confinement and diffusion are crucial for device performance, yet experiments that probe properties related to local structure and diffusion are challenging and difficult to interpret without a parallel modeling approach. In this talk, I will focus on carbon-carbon supercapacitors in which the energy is stored by ion adsorption at the electrode surface. In order to understand fundamentally the macroscopic properties of such systems, it is essential to characterize finely the porous materials used and the structural and dynamic properties of the fluid adsorbed. But, in order to screen materials for energy storage applications, it is necessary to develop computationally efficient methods. Here, I will present insights from different approaches. I will first describe molecular dynamics simulations, conducted using HPC facilities, which provide a microscopic understanding of the charging mechanisms in supercapacitors. I will then show the promising results we obtain with a mesoscopic model we develop, 10,000 times faster than molecular dynamics simulations, for the prediction of electrochemical performance in these systems. Finally, I will discuss recent developments in machine learning approaches to speed-up the calculation of relevant properties while staying at the atomic scale.
Céline Merlet is a CNRS researcher at Université Paul Sabatier in Toulouse. She received her PhD degree in 2013 from Université Pierre et Marie Curie in Paris where she had worked on molecular simulations of carbon-carbon supercapacitors. She then joined the University of Cambridge as a postdoctoral researcher working on simulating NMR spectra and diffusion of ions in energy storage materials such as porous carbons and lithium manganese oxides. Since 2017, she has been working at the CIRIMAT laboratory on the development and application of multi-scale models for a better understanding and performance prediction of electrochemical energy storage systems. Her project “SuPERPORES – Structure-PErformance Relationships in PORous carbons for Energy Storage” was awarded an ERC Starting Grant in 2017. In 2018, she received the Prix Louis Armand in Chemistry, awarded by the French “Académie des Sciences” and, in 2021, she received the PRACE Ada Lovelace Award and the Bronze Medal of the CNRS.