Povel Methods for Predicting Properties of Complex Reactive Solid Interfaces

Brandon Wood | 18-FS-019


Disordered interfaces are critical to the functionality of a wide variety of modern energy storage and conversion devices. However, the intrinsic structural and chemical inhomogeneity introduces significant challenges for computational modeling that many current techniques are ill-suited to handle. To address this critical need, this feasibility study focused on developing and integrating multiscale methods for simulating structural and dynamic properties of disordered and interfacial regions.

Background and Research Objectives

The structure, dynamics, and reactivity of disordered interfaces determine key macroscale properties of many functional materials of interest for energy storage and conversion. These regions tend to be chemically and structurally complex, making physicochemical behavior extraordinarily difficult to study and predict. Although advances in experimentation have aided our understanding of interfaces and amorphous materials, interpreting these data relies on equivalent advances in the ability to predict properties of complex disordered systems.

Disorder presents several key challenges for predictive modeling. Methods exist for modeling simple interfaces such as well-ordered grain boundaries within metals or phase boundaries between very similar crystallographic structures. However, the far more complex interface compositions and structures exhibited by multicomponent functional materials are difficult to model explicitly at the atomic scale, since the configurational complexity requires sampling many different configurations. This complexity can be overcome in part by abandoning computationally demanding first-principles methods in favor of descriptions based on simpler classical pair-potential interactions, but many functionalities and processes rely on chemical reactions or solute interactions in these disordered regions, necessitating a high degree of predictive accuracy. Accordingly, robust simulation methodologies for treating complex reactive solid interfaces are rare, even though they are key to modeling a wide variety of chemical and electrochemical solid-state reactions. Our feasibility study addressed this need by integrating accurate first-principles simulations with flexible mesoscale methods that can span broader ranges of length and time and incorporate configurational complexity. The goal was to lay the groundwork for a comprehensive approach for simulating chemical and dynamic kinetics in disordered materials and complex interfaces. Although the original proposed scope also encompassed chemical reactions at interfaces, the decision was made to focus the project on simulating dynamic properties.

We developed and successfully demonstrated three different multiscale simulation techniques.

  1. We demonstrated an approach to analyzing ion diffusion through amorphous materials based on combining first-principles simulations and kinetic Monte Carlo approaches. Applying these to lithium diffusion in titanium-oxide (TiO 2 ) films for battery applications, we showed that structural disorder fundamentally changes ion-diffusion profiles.

  2. We demonstrated full integration of our first-principles models with our new mesoscale framework for simulating ion diffusion through complex microstructures and interfaces. This technique was applied to ceramic solid-state battery materials to show the effect of microstructure on ion transport, including disordered grain boundaries and interfaces.

  3. As a proof of concept, we applied a novel mixed-quantum classical interface simulation that accounts for dynamic disorder to the practical interfacial energy-storage problem of supercapacitors.

Demonstration of these three multiscale techniques is a first step toward implementing comprehensive modeling capabilities in disordered-interface simulations of hydrogen production and storage, electrical energy storage, corrosion and materials degradation, and geochemical phenomena.

Impact on Mission

This study supports DOE energy missions and advances the science, technology, and engineering competencies that are the foundation of the NNSA mission. Our work also supported Lawrence Livermore National Laboratory's mission focus area and research and development challenges in energy security, while supporting its core competencies in high-performance computing, simulation, and data science. By demonstrating the feasibility of our three multiscale integration techniques for studying disordered materials and interfaces, the project also helped promote implementation of computational modeling capabilities that can be applied to a variety of mission spaces.


This project demonstrated the integration of first-principles simulations based on density functional theory with three different mesoscale techniques for handling disorder. Each was tested on a model system related to energy storage. The new modeling techniques developed and demonstrated during this project will find widespread use in existing and future projects. For example, the mesoscale effective diffusion scheme (technique 2) was highlighted as a new capability in a proposal that was recently funded under the DOE’s Vehicle Technologies Office. In addition, the demonstrated capabilities from techniques 1 and 3 will be directly applied to improve and expand the Laboratory’s applied computational modeling activities within the Hydrogen Storage Materials-Advanced Research Consortium and HydroGEN Advanced Water Splitting Materials Consortium, funded under the DOE’s Fuel Cell Technologies Office.

Publications and Presentations

Heo, T. W., et al. 2018. "Mesoscale Modeling Study of Microstructural Impacts on the Effective Ionic Diffusivity of Solid Electrolytes for Li Batteries." Materials Research Society Fall Meeting, Boston, MA, October 2018. LLNL-ABS-733792.

Shea, P., et al. 2018. "Ion Transport in Disordered Materials." Bay Area Battery Summit, Berkeley, CA, October 2018. LLNL-POST-759545.

Wang, B., et al. 2018. "Microstructure Effects on Ionic Conductivity of Solid Electrolytes for Li Batteries: A Multiscale Approach." Lawrence Livermore National Laboratory Annual Student Poster Symposium, Livermore, CA, August 2018. LLNL-POST-755513.

Wood, B. C., et al. 2018. "Understanding and Optimizing Ionic Conductivity in All-Solid-State Batteries Through Multiscale Simulations." Oak Ridge National Laboratory Electrochemical Energy Storage Lunch Seminar, Oak Ridge, TN, October 2018. LLNL-PRES-759804.

Zhan, C., et al. 2018. "Origins and Implications of Interfacial Capacitance Enhancements in C 60 -Modified Graphene Supercapacitors." ACS Applied Materials & Interfaces , 10(43). doi: 10.1021/acsami.8b10349. LLNL-JRNL-754189.