Overview

Our team investigates the assembly of biological and nanostructured materials at length scales from atoms to 1000 nm using diverse computational and AI tools. We develop, validate, and apply the most accurate and interpretable atomistic potentials for compounds across the periodic table to solve challenges in soft matter, bioscience, and nanoscience.

This slideshow requires JavaScript.

Why Modeling and Simulation?
Monitoring intermolecular dynamics, chemical reactions, and other processes from the atomic scale to the micrometer scale is often impossible or expensive in experiments. Visualization, fundamental understanding of interactions, and precise numerical predictions help elucidate physiological processes, the function of advanced materials and devices, chemical and mechanical transformations. For example, modeling and simulation can help identify molecular signatures of diseases, aid in the design of drugs and delivery systems, accelerate the development of catalysts and electrode materials, and provide insights for the development of corrosion resistant alloys. The same techniques can also be used towards developing stronger carbon fiber, increasing the selectivity of biosensors, produce more efficient photovoltaic devices, and guide in formulations for building materials to customize performance. Novel feature representations and large data sets for machine learning of properties can be identified.

picture2

Our Innovation
We developed the first compatible simulation platform for compounds across the periodic table including inorganic and organic compounds (INTERFACE molecular dynamics). Simulation parameters (force fields) and surface models are derived such that the classical Hamiltonian reproduces structures and energies of individual compounds in highest accuracy, typically better than DFT, using quantum mechanically justified polynomial energy expressions that are working well for biomacromolecules, such as CHARMM, PCFF, AMBER, GROMACS, and OPLS-AA. This new approach, called INTERFACE MD (learn more), builds on the interpretation of atomic charges in the context of measured electron densities (learn more) and advances quantitative understanding of chemical bonding (learn more). Simulations allow the mobility of all atoms and predictions of interfacial energies in impressive agreement with experimental observations up to the large nanometer scale, reducing deviations up to several 100% in earlier models to the order of 5%. By use of modeling and laboratory tests, we can explain the selective adsorption of biomolecules to nanostructured surfaces, nanocrystal growth and shape development, preferences in morphology development and phase separation, catalytic activity, as well as nanomechanical properties up to failure. Accessible time and length scales reach many orders of magnitude beyond quantum methods. We integrate quantum-mechanical and coarse-grain models to understand electronic effects, access multiple length scales, and develop models of even better predictive ability.

Achievements To-Date
The focus to-date has been on force field parameters for inorganic compounds, such as metals, oxides, and minerals, including reliable surface models for (h k l) facets, specific surface chemistry, and pH sensitive mineral sufaces which have not been available before or misleading by orders of magnitude. Starting in 2003 (learn more), we began to develop  parameters and surface models that are now collected in the INTERFACE force field and enable compatibility between materials-oriented simulation and biomolecular simulations by using a uniform energy expression for all compounds across the periodic table, e.g., INTERFACE-CHARMM and INTERFACE-PCFF. For the first time, reliable simulations of an unlimited number of new inorganic-biological and inorganic-organic materials (trillions+) are feasible in one single platform. Multiphase systems that may include biomolecules, metals, minerals, polymers, and solvents are becoming amenable to realistic simulations to predict properties in high accuracy at the 1 to 100 nm scale (up to 1000 nm scale in 1 D).

Examples of Applications 
Our research team as well as others apply these concepts and new developments to understand, for example, biomineralization processes, catalyst efficiency, structural materials, and the function of photovoltaic cells. We integrate measurements (e.g., XRD, TEM, AFM, NMR, IR, zeta potential, binding constants, conductivity, impedance, DSC, mechanical, QCM, interface tensions) and molecular-level simulations to analyze and design nanoscale building blocks for various applications. We discovered a soft epitaxial recognition mechanism of organic ligands on metallic surfaces (learn more), explained the mechanisms of biomolecule and surfactant binding to oxidic surfaces as a function of pH (learn more), helped predict the growth and shape control of platinum nanocrystals (learn more), predicted reaction rates of metal nanoparticle catalysts in aqueous solution (learn more), explained the action of organic modifiers in major cement phases (learn more), the role of surfactants on clay minerals and cleavage energies for nanocomposites (learn more). Ongoing efforts aim at understanding corrosion, polymer-graphitic nanocomposites, and catalysts. Cutting-edge models for polyelectrolytes, metals, alloys, layered materials, and organic semiconductors with refined electronic structure (n electron “lone” pairs, d electrons, π electrons) as well as coarse-grain models are being developed.

As new materials are made in laboratories, the impact of our computational work also depends on experimental, clinical, and industrial partners.

Current and past supporters of our work include NSF-DMR, NSF-CBET, AFOSR, ONR, NASA, Procter & Gamble, Corning, UES Inc., the Air Force Research Laboratory (Dayton, Ohio), the Ohio Department of Development, Sika AG, the ETH Zurich Foundation, the University of Akron, as well as the Ohio Supercomputing Center, CU Research Computing, large-scale computing resources at Argonne National Laboratory, and the University of Colorado at Boulder.