Skip to content

In the News:

Multi-university effort will advance materials, define the future of mobility
- MIT News Office

events

May 2022
M T W T F S S
 1
2345678
9101112131415
16171819202122
23242526272829
3031  

About

Battery modeling and simulation has matured at two widely disparate scales, at the extremes of scales: (i) macroscopic modeling as exemplified by porous electrode theories, which are fitted to current-voltage data and used for systems engineering, but lack predictive power and (ii) atomistic modeling as exemplified by the quantum (density-functional theory, DFT) calculations which predict the voltage and bulk diffusivity from first principles, but are limited to small periodic unit cells of atoms near equilibrium. It is increasingly recognized that practical energy, power density, and lifetime of rechargeable batteries are controlled by processes at neither the macroscopic nor the atomistic scale, but in between at the mesoscale (nm-μm) over intermediate time scales (ns-s), where nonlinear dynamics lead to emergent material properties.

The grand challenge confronting the field is to develop predictive models validated by direct experiments that span these scales by capturing mesoscale phenomena, such as intercalation kinetics, charge-transfer and space charge at interfaces, phase transformations in primary particles and composite porous electrodes, mechanical deformation, parasitic reactions (solid-electrolyte interphase, SEI), and metal growth. The massive datasets generated at the mesoscale by spatially and time-resolved experiments and simulations necessitate a data analytic approach, which has not yet been applied to battery research. Multiscale materials modeling challenges have been overcome in other fields, such as composites and metal alloys, but batteries are inherently more complicated, due to the nonlinear, multiphysics coupling of electrochemical reactions, phase transformations, and transport in complex microstructures, dominated by charged, multiphase interfaces.

The D3BATT project will develop a novel multiscale-modeling framework for rechargeable batteries to accelerate materials discovery and design, guided by machine learning on large databases of experimental/simulated images/videos at matching length and time scales, electrochemical characterization, and atomistic calculations of material properties. This unique approach can unify all of the energy materials design efforts at Toyota Research Institute by connecting atomistic with macroscopic properties, while paving the way to address other grand challenges in materials for catalysis, separations, light weighting, etc.

Success will be measured by both fundamental and practical metrics. The scientific goals include comprehensive, multiscale simulation and experimentation on selected model materials in order to understand the coupling of electrochemical reactions, transport and phase transformations across the relevant length and time scales shown in the cover figure. The engineering goals include the optimization of bulk materials, interfaces, microstructures, cells, and cycling protocols in order to improve key metrics for Li-ion batteries. In addition to the traditional Ragone plot (energy density vs. power density), attention will be paid to other essential, but often overlooked, properties for electrified transportation: recharging rate (for safety and long lifetime), state of health (degradation), temperature range (for normal use and storage), and optimal cycling protocols (for vehicle performance as well as factory forming).