Title: Computer simulations of co-crystalline electrolytes for all-solid alkali metal batteries

Speaker: Arun Venkatnathan (IISER, Pune, India)
Time: May 28, 2020, 10:00–11:00
Location: Online (MS Teams)

Dr. Arun Venkatnathan is an Associate Professor in Chemistry at the Department of Chemistry and the Centre for Energy Science at IISER, the Indian Institute of Science Education and Research in Pune, India. Dr. A. Venkatnathan group's webpage may be found at: http://www.iiserpune.ac.in/~arun/

Abstract | Soft-solid co-crystalline electrolytes for lithium/sodium ion batteries exhibit excellent thermal stability and high ionic conductivity. These electrolytes are matrices of organic solvents (e.g. DMF, adiponitrile) coordinated with cations (Li+/Na+), where counter anions occupies the remaining space in crystals. In this talk, I will present a molecular understanding of structure, thermal stability and ion transport in some electrolytes from Molecular Dynamics simulations. The simulations mimic the processes of decomposition/melting of electrolytes and show a liquid-like surface layer on the nano-sized grains of these electrolytes as seen from several experiments performed by colloborators. The simulations also examine the dynamics of ions on the surface and bulk. The insights from simulations combined with experiments is expected to lead to the development of better and safer electrolytes for batteries.

P. Prakash, J. Aguirre, M. Van Vliet, P. Chinnam, D. Dikin, M. Zdilla, S. Wunder, A. Venkatnathan, Unravelling the structural and dynamical complexity of the equilibrium liquid grain-binding layer in highly conductive organic crystalline electrolytes, J. Mat. Chem. A, 6, 4394 (2018).

B. Fall, P. Prakash, M. R. Gau, S. L. Wunder, A. Venkatnathan and M. J. Zdilla, Experimental and theoretical investigation of the ion conduction mechanism of tris(adiponitrile)perchloratosodium, a self-binding, melt-castable crystalline sodium electrolyte, Chemistry of Materials, 31, 8850 (2019).

Seminar | The CCER seminars are aimed at researchers interested in computational approaches to (energy) research. The seminar is small-scale, typically 15 participants, and interactive, offering lots of room for discussion. If you would like to attend, just This email address is being protected from spambots. You need JavaScript enabled to view it. so as to receive the MS Teams meeting link.

Title: Deep Learning – beyond training with labels

Speaker: Vlado Menkovski (TU/e WSK&I)
Time: May 14, 2020, 10:00–11:00
Location: Online (MS Teams)

Abstract | Machine learning based solutions are being developed for a growing number of tasks in a broad range of fields. Deep Learning techniques have extended the scope of these solutions to high dimensional input spaces of images, natural text, and speech. It is still most common that these solutions take a supervised learning formulation, where for either observed or computed input data an expert provides a target value. In this talk I discuss the limitations of these approaches, particularly regarding understanding generalization capabilities of the model, interpreting the models’ behavior as well as utilizing efficiently the available expert knowledge.
I reflect on this through the experience of developing a model for classification of tokamak plasma confinement states. Going forward I evaluate alternative formulations for machine learning, focusing specifically on developments in deep metric learning and learning disentangled representations as a way to address the shortcomings of supervised learning.

Seminar | The CCER seminars are aimed at researchers interested in computational approaches to (energy) research. The seminar is small-scale, typically 15 participants, and interactive, offering lots of room for discussion. If you would like to attend, just This email address is being protected from spambots. You need JavaScript enabled to view it. so as to receive the MS Teams meeting link.

Title: Brain-like computing using disordered dopant networks in silicon

Speaker: Peter Bobbert (CCER)
Time: April 30, 2020, 10:00–11:00
Location: Online (MS Teams)

Abstract
A disordered network that is capable of detecting ordered patterns: it sounds contradictory, but it comes close to the way our brain works. I will discuss in this seminar a system of interacting electrons hopping on a disordered dopant networks in silicon that does precisely this. The complex physics and reconfigurability of this system can be exploited to perform various computing tasks like the recognition of basic patterns. The use of a surrogate model of the system based on a deep neural network allows rapid reconfiguration of the network for a desired computing task without the need of any further experimentation. Coupling of these networks in superstructures can lead to a computing technology that has the potential to outcompete conventional CMOS technology regarding energy consumption and footprint.

The CCER seminars are aimed at researchers interested in computational approaches to (energy) research. The seminar is small-scale, typically 15 participants, and interactive, offering lots of room for discussion. If you would like to attend, just This email address is being protected from spambots. You need JavaScript enabled to view it. so as to receive the MS Teams meeting link.