Investigation and Optimization of Electrochemical Systems via Simulation and Theory

Abstract

Electrochemical systems are ubiquitous in our modern lives in many forms, including energy storage and conversion devices like batteries and fuel cells, which enable our modern consumer electronics and underpin our shift away from fossil fuels, and corrosion, which affects many metallic structures. However, the physical phenomena that underlie the working of such systems often involve several mechanisms spanning multiple length scales. Due to this complexity, the research and development (R&D) of electrochemical systems has been challenging. The goal of this Ph.D. dissertation is to accelerate the R&D of electrochemical systems by adopting a Materials Genome Initiative (MGI)-based approach, which integrates modern computational tools and data-driven methods with experiments. The approach is employed to achieve four goals. The first is to implement modeling and simulation techniques that incorporate the relevant physics and solve the resulting mathematical model in complex geometries. This aspect also includes a machine-learning-based automated parameterization algorithm developed to determine unknown model parameters. The second is to generate insights into the behavior of the system, especially into the characteristics like the reaction current density distribution that cannot be easily obtained via experiments. The third is to optimize the system design for superior performance by exploiting these insights. The last goal is to reduce the computational cost associated with the approach by developing analytical and semi-analytical frameworks. These goals are achieved to varying extents for Li-ion batteries, Mg alloys undergoing microgalvanic corrosion, and solid oxide fuel cells. For Li-ion batteries, the tradeoff between the energy density and fast-charging capability in conventional electrodes is overcome by employing a novel-electrode architecture and studying its effect on the fast-charging performance using electrode-level simulations parameterized by machine learning. The novel architecture is formed by ablating vertical channels along the thickness of the electrodes with laser, and it is referred to as the Highly Ordered Laser-Patterned Electrode architecture. Simulations and theoretical analyses resulted in scientific understanding and an approach to optimizing such an architecture. For Mg alloys, an open-source software application in PRISMS-PF has been developed to simulate the microgalvanic corrosion behavior. Using the application, the effect of the electrochemical properties and the spatial distribution of second phases on the corrosion behavior is elucidated, and design strategies are devised for the alloy microstructure to minimize the corrosion rate. Finally, for solid-oxide fuel cells, the impedance behavior of a mixed ion-electron conducting cathode with an experimentally determined microstructure is simulated, and the effect of material properties on the impedance behavior is studied. Furthermore, the Adler-Lane-Steele model, a widely used analytical model for determining the impedance response, is extended to account for the spatial variation of the vacancy-concentration amplitude due to the reaction at the pore/solid interface.