Simulating Li-ion Batteries and Beyond: Developing Electrode Materials for Next-Generation Batteries

Abstract

In order to curve the environmental impact of fossil fuels, many countries have turned to "cleaner” energy sources and storage devices. One hope is that next-generation batteries may completely revolutionize energy storage. For in- stance, much research has been done concerning replacing the internal-combustion engine of automobiles with electric powertrains. However, it is essential that these batteries not just be energy and power dense, but also sustainable. This means that next-generation battery materials must be acquired through (relatively) en- vironmentally friendly and humane means. Towards this goal, there are two main options: select battery materials that can be responsibly obtained, or design bat- teries to not contain those unsustainable materials. With careful research, com- putational researchers can test which materials and designs theoretically work to save manufactures time and money, and to help explain their experimental results.To guide experimentalists, theorists run quantum mechanical simulations to test the stability, electronic structures, and diffusion on such candidate materials. One of the most popular methods is DFT, an approximate method that is usually accurate and computationally efficient. For more exact answers, one can use more advanced methods such as QMC. In this work, we explain our work using DFT to study anode materials for various ion batteries. In particular, we studied Li absorption and diffusion on 2D regular and Janus TMDs, and found that the sides with the lower electronegative chalcogenides had lower diffusion energy barriers than regular TMDs. We performed similar studies of adsorption and diffusion of various ions on three types of transition-metal carbide monolayers with Sulfur terminations (referred to as a type of ”MXene”). In one study, we closely examined Na and Mg ion interaction with these structures, and found that Na diffuses more efficiently on them than Mg. In another study, we build a physics-based machine learning model to use on our adsorption energy data. For the dilute case, ionization potential of the adatoms and workfunction of the materials are the most significant features to the model. For the full-coverage case, repulsion energy of the adatoms is the most significant feature. Finally, we also studied the discharge products of a novel battery design, the Li-air battery. We performed QMC calculations on the discharge products at the cathode, and found that Fo ?ppl Li2O2 is the most stable product.