Algorithm Development

Bayesian Optimization for Material Design

Nikita_PhD Schematic of the materials informatics method which combines atomistic MD and BO. The algorithm starts by randomly creating a set of input parameters x. Structures characterized by these inputs, x, are then send as an input to MD simulations. Output from these MD runs in the form of total energy (of the training set) is used to train the Bayesian model. The predictive distribution then suggest the next best candidate to test based on the criterion of expected improvement. F, E, U, K, m, and a denote Force, Total Energy, Potential Energy, Kinetic Energy, Mass, and Acceleration, respectively.
We are currently developing a method which combines atomistic Molecular Dynamics (MD) and Bayesian optimization (BO) to predict the polymorphic single crystal structures crystals of a p-type (hole donor) organic semiconductor (BTBT variants). Finding the optimal energy structures is nonintuitive, but with BO it is possible to accelerate the pace of materials discovery because it allows predicting the new design based on previously tested designs without having to run time consuming expensive molecular simulations. This novel approach aims to improve the effectiveness of material informatics in designing next-generation multi-functional semiconducting nano-crystals with pre-chosen properties.
Related Publications:
Y. Diao, K. M. Lenn, W. Lee, M. A. Blood-Forsythe, J. Xu, Y Mao, J. A. Reinspach, S. Park, A. Aspuru-Guzik, G. Xue, P. Clancy, Z. Bao, S. C.B. Mannsfeld, Understanding polymorphism in organic semiconductor thin films through nanoconfinement, JACS, 136, 17046-17057 (2014).

Diffusion Rates Prediction using Machine Learning

We explore the use of machine learning methods in molecular simulations. Specifically, we develop techniques and tools for spatially-resolved fingerprinting (process of transforming Cartesian coordinates into representations suitable for machine learning studies) and how to leverage them for property predictions. Main area of application is currently on the study of diffusion processes in III-V semiconductors.

Related Publications:
M. Reveil and P. Clancy, Classification of spatially resolved molecular fingerprints for machine learning applications and 10 development of a codebase for their implementation, Mol. Syst. Des. Eng., 00, 1–11 (2018). [invited paper]

Nudged Elastic Band (NEB) Application Development

Developed by Jónsson et al., the Nudged Elastic Band (NEB) method is a powerful computational tool for uncovering energy barriers and consequently, transition states for reaction mechanisms. By developing a deep understanding of NEB, our group aims to determine the most efficient optimization method and approach for applying NEB to various molecular systems of interest. Accurate results via minimal time and computational resources is the goal – we are building upon the current NEB method to define a transferrable “recipe” that allows researchers to properly uncover their specific reaction mechanisms using the least arduous approach possible.

Related Publications:
H. C. Herbol, J. M. Stevenson and P. Clancy, Computational Implementation of Nudged Elastic Band, Rigid Rotation, and Corresponding Force Optimization, J. Chem. Theory Comput., 13 (7), 3250–3259 (2017).

Physical Analytics pipeLine (PAL)

PAL" An outline of PAL as applied to HOIPs, in which computational objective functions and experimental results feed into a statistical model. Bayesian optimization is then run to find ideal HOIP-Solvent mixtures, which can then be fed back to experiments.
PAL is a collaborative project geared towards the study of large, expensive cost functions. With Bayesian optimization, alongside a cheap computational ersatz for some desired experimental property, PAL aids researchers in studying previously too large to tackle type problems. Currently, PAL is being applied to Hybrid Organic Inorganic Perovskites (HOIPs) to isolate the best perovskite salt – solvent combination for dissolving HOIP reagents in solution.

Related Publications:

Simple Molecular Reactive Force Field (SMRFF)

SMRFF, originally developed in 2016 (link), is a novel reactive force field that strives to strike a balance between accuracy and use. By simplifying the expressions of reactions, the number of parameters necessary decreases, allowing for a more readily parameterized force field. Further, by isolating a “Long Range” and “Short Range” component to the force field, with smooth transitioning between the two, the computationally heavy portion (the short range) is minimized as much as possible, allowing for better speeds. Currently, SMRFF is being extended from the original implementation of Morse, Lennard-Jones, and Coulomb to that of a short-ranged Tersoff with long-ranged Lennard-Jones and Coulomb.

Related Publications:
J. Andrejevic, J.M. Stevenson and P. Clancy, Simple Molecular Reactive Force Field for Metal-Organic Synthesis, J. Chem. Theor. Comp., 12(2), 825-838 (2016).