(6gd) Machine Learning and Data-Enabled Design and Discovery of Nano and Soft Materials

Research Interests: High-throughput materials design, Energy storage materials, Polymer nanocomposites, Superlattice, Self-healing materials, Self-assembly, Phase transition and free energy calculation, and Machine learning

Teaching Interests: Thermodynamics, Transport phenomena, Polymer Physics, Numerical Methods and Molecular modeling

Soft materials are highly correlated many body systems with complex structures and dynamics spanning a wide range of length and time scales. Their relaxation processes involve non-equilibrium phenomena such as vitrification, jamming and semi-crystallization, which are highly process-dependent and for which no comprehensive theoretical framework and understanding exist. A significant advancement in the understanding of these processes in soft materials are vital for their successful implementation in future electronics, medicine and energy devices. The current understanding of soft materials is limited by many factors. First, atomic level interactions in many soft materials, for a wide range of physicochemical conditions, is not well-established. Second, unlike biomolecules where informatics-based design has been accelerated by the availability of large, open, and homogenous structure/property relation databases such as the protein data bank, similar data in soft materials tend to be sparse, heterogenous, and often outright unavailable. Third, due to a general lack of rapid, parallelizable techniques for measuring materials properties, building such databases de novo for a specific problem is often cost-prohibitive. Furthermore, many critical design problems, such as the search for highly flexible barrier materials, target materials properties well outside the range of what have been discovered so far.

The overarching goal of the proposed research program is to address these limitations by combining molecular modeling, machine leaning and evolutionary computing. By combing these tools, I plan to develop more accurate and transferable force fields for a wide range of soft materials, including polymers, small organic molecules, colloids; discover their hitherto unknown structure-property relationship; and accelerate materials design. By integrating data science and first principle calculations, my goal is to address some of the open questions of soft matters – such as the non-universal growth of relaxation time upon cooling a material, the interrelationship between ion conductivity, modulus and glass formation behavior of a polymeric ionic liquids, and the universal relationship between the governing forces and structural order parameter of nanoparticles assembly within a polymer matrix. Addressing these questions are crucial for many future technologies, such as safer battery with high energy density and longer durability; capsule that prevents drugs from unwanted release and degradation caused by temperature variation; membrane that pulls salts out of sea water rapidly; flexible smart phone that can be wrapped around human’s wrist; and a coating that prevents roads and bridges from corrosion. These limitless possibilities can only be realized by efficient screening of the vast configurational space of soft materials, and investigate the above mentioned open questions. The vision of the proposed program is, therefore, to address these challenging problems and establish new materials design rules by building large amount of accurate homogeneous structure property data; use artificial intelligence (AI) and machine learning to understand their interrelationship; and employ evolutionary algorithms for discovering new structures and properties. This way, the proposed research program aims to push the frontier of materials research towards more predictive designing and phenomenological modeling.

With successful implementation of this program, I hope to study complex physical science problems and attract external funding. I am confident that this research program will lead to the development of useful design rules for flexible electronics; polyelectrolytes for energy storage application; and nanostructure materials for gas storage, water purification and many other applications. In this program, students will be trained in statistical mechanics, wide range of machine learning approaches and artificial intelligence and their applications in chemical engineering and materials science.

Selected Publications (Total: 14, Under review: 3, Under preparation: 4)

1. Patra T K, Zhang F, Schulman D; Chan H, Cherukara M, Terrones M; Das S, Narayanan B, Sankaranarayanan S KRS, Defect dynamics in 2D materials probed by combining machine learning, molecular simulation and high- resolution microscopy, Accepted for publication, ACS Nano (2018)
2. Patra T K, Meenakshisundaram V, Hung J, and Simmons D S, Neural network biased genetic algorithm for materials design: Evolutionary algorithms that learn, ACS Combinatorial Science 19, 96 (2017)
3. Patra T K and Singh J K, Polymer directed aggregation and dispersion of anisotropic nanoparticles, Soft Matter 10, 1823 (2014)
4. Patra T K and Singh J K, Coarse-grained molecular dynamics simulations of nanoparticle-polymer melts: Dispersion vs. Agglomeration, Journal of Chemical Physics 138, 144901 (2013)
5. Patra T K, Hens A, Singh JK, Vapor-liquid phase coexistence and transport properties of two dimensional oligomers, Journal of Chemical Physics 137, 084701 (2012)