(7b) Towards Accurate Atomistic Description of Reactive Interfaces for in silico design of Novel Functional Materials

The complex interplay of chemical reactions, defect-chemistry, transport phenomena, and structural evolution at reactive interfaces underpin much of energy capture, conversion and storage by materials. A fundamental understanding of these dynamical processes at angstrom-to-mesoscopic length scales over several nanoseconds is crucial, and urgently needed to enable development of novel functional materials for various energy applications. However, development of accurate atomistic models of reactive interfaces remains a grand challenge, mainly due to the non-systematic and intuition-driven methods employed. Here, I will first present a robust, efficient, and automatic computational framework to develop reactive interatomic potentials (force-fields; FFs) that significantly advances the state-of-the-art by completely obviating the need for human intuition. Using supervised machine learning empowered by genetic algorithms, and large datasets obtained from high-throughput ab initio calculations, we sample the parameter landscape efficiently to obtain a highly transferable set of FF parameters that accurately capture structural, dynamical, energetic, and other physical/chemical properties. We have successfully employed this framework for a wide range of material systems, including gold, ZrN_x, CoC_x, IrO₂, and 2D tin.^1,2 Subsequently, I will demonstrate how a synergistic integration of large-scale reactive molecular dynamics (RMD), electronic structure calculations, and ab initio MD (AIMD) can pave path towards an accurate description of reactive interfaces, and consequently enable computational discovery of novel materials. Specifically, I will highlight two representative examples: (1) using AIMD, and million-atom RMD simulations, we discovered a novel catalytic phenomena possible only under the extreme local conditions afforded by friction at sliding interfaces; this enabled design of novel catalytically active coatings (e.g., MoN_x-5%Cu) with exceptional wear resistance (~25% lower friction than best blended oils). We found that such coatings transform base lubricating oils (long-chain alkenes) into low-friction diamond-like carbon (DLC) film via dehydrogenation, backbone scission of long chain alkenes to form short fragments, and their subsequent polymerization. Our atomistic simulations also indicate that the formation of DLC-like tribofilm is crucially linked to the propensity of the transition metal catalyst to form carbides; carbide-forming metals, e.g., V preclude the formation of DLC tribofilm.³ (2) RMD simulations revealed a new transfer-free route to synthesize wafer-scale high quality single/multi-layer graphene directly on diamond.⁴We found that a Ni thin film placed on ultra-nanocrystalline diamond (UNCD) causes amorphization of diamond lattice via diffusion of Ni into the UNCD grain boundaries (GBs); concurrently, C from UNCD GBs diffuse through the Ni film, and form a graphene layer on the surface. This opens up avenues to explore novel synthesis pathways, and atomistic mechanisms underlying exotic phenomena in other 2D materials, e.g., stanene, phosphorene etc.

Research Experience

My research experiences lie in the broad area of theoretical/computational materials science, which have made me well-versed with developing accurate models at multiple length scales (angstrom to meso) and employing them to gain novel physical insights into exotic materials behavior. My expertise includes density functional theory calculations, ab initiomolecular dynamics, large-scale classical molecular dynamics simulations, multi-parameter optimization for empirical force-field development, machine learning, advanced sampling and mesoscopic length-scale models. Specifically, I have employed these tools to investigate (i) structural evolution, ionic diffusion, and solvation dynamics near electrochemical interfaces during aqueous corrosion, (ii) atomic-scale diffusion mechanisms in ionic liquid based electrolytes for Al-batteries, (iii) mechanisms underlying friction and chemical reactions at sliding interfaces, (iv) growth mechanisms for 2D materials on transition metal surfaces, and (v) self-assembly of biomolecules (like collagen, peptoids) on inorganic surfaces. Furthermore, as a part of the user program at Center for Nanoscale Materials (DOE nanoscience user facility at Argonne National laboratory), I have fostered successful collaborations with several experimental research groups at US universities; Through these user-collaborations, I have worked on materials science problems at various length/timescales including investigations on large area synthesis of 2D materials via electrochemical exfoliation, chemical reactions under tribological conditions, and self-assembly of ligand decorated nanoparticles.

Postdoctoral Project

â??Bridging the electronic and atomistic scales: force field development for reactive interfaces from first principlesâ?Under supervision of Dr. Subramanian Sankaranarayanan, Center for Nanoscale Materials, Argonne National Laboratory

PhD dissertation

â??Understanding structure-property relationships in b-eucryptite through atomistic simulationsâ?Under supervision of Prof. Cristian V. Ciobanu (Mechanical Engineering) and Prof. Ivar E. Reimanis, (Metallurgical and Materials Engineering), Colorado School of Mines.

Future Direction

Despite recent strides, the field of computation-driven materials design remains plagued by the dearth of accurate models (at nano-to-meso length scales), as well as lack of efficient strategies to design new realistic functional materials (i.e, with defects, grain boundaries etc.) and identify routes to synthesize them. Motivated by this challenge, as a faculty, I seek to establish a successful research program titled â??Multi-scale in silico materials designâ?, which focuses on devising new strategies/tools to address these challenges, and applying these tools to investigate various technologically relevant materials systems. Specifically, I intend to develop robust computational frameworks to (1) use advanced sampling methods (e.g., evolutionary algorithms, metadynamics) to search for new metastable/stable structures for low-dimension materials (i.e., clusters, 2D sheets), (2) couple high throughput-DFT calculations with advanced global (e.g., genetic algorithms) and local (e.g., Simplex) optimization routines to develop new reactive interatomic force field (FF) models, which capture bond formation/dissociation events and transition states accurately; such FFs are crucial to advance fundamental understanding of interfacial reactions, mechanistic origins of materials phenomena, and structure-property relationships at length scales spanning tens of nanometers, (3) develop machine learning strategies to rapidly design new materials for superlubricity, battery electrolytes and anti-corrosion coatings, as well as (3) develop accurate coarse-grained models to understand self-assembly of hierarchical architectures at mm sizes, including superlattices of ligated nanoparticles.

Teaching Interests:

I possess a strong interest in teaching and mentoring students. During my postdoctoral appointment at Argonne National Laboratory (ANL), I pursed and obtained the opportunity to mentor several graduate students who actively collaborate with my group on various computational materials science research problems. While working towards my PhD dissertation, I delivered several guest lectures for the course â??Ceramic Engineeringâ? hosted by the Department of Materials Science and Engineering in Colorado School of Mines (CSM); additionally, I have also served as a teaching assistant for various materials science courses at CSM [e.g., Electronic Properties of Materials, Mechanical Properties of Materials, Nanomechanical Engineering etc.]. These opportunities have helped me develop the qualities necessary for effective teaching, and for being a good mentor.

Selected Publications

1. B. Narayanan, A. Kinaci, F. Sen, M. J. Davis, S. Gray, M. K Y. Chan, and S. Sankaranarayanan, â??Describing the diverse geometries of gold from nanoclusters to bulk â?? a first-principles based hybrid bond order potentialâ?, Journal of Physical Chemistry C 120, 13787 (2016).
2. B. Narayanan, K. Sasikumar, A. Kinaci, F. Sen, M. J. Davis, S. Gray, M. K Y. Chan, and S. Sankaranarayanan, â??Development of a modified embedded atom force field for zirconium nitride using multi-objective evolutionary optimizationâ?, Accepted for publication, Journal of Physical Chemistry C (2016).
3. A. Eridemir, G. Ramirez, O. Eryilmaz, B. Narayanan, Y. Liao, G. Kamath, and S. Sankaranarayanan, â??Cabon-based protective tribofilms from lubricating oilsâ?, In Press, Nature (2016).
4. D. Berman, S. Deshmukh, B. Narayanan, S. Sankaranarayanan, Z. Yan, A. Baladin, A. Zinovev, D. Rosenmann, and A. Sumant, â??Metal induced ultra-fast transformation of diamond into single domain graphene on wafer scaleâ?, Nature Communications 7, 12099 (2016).
5. B. Narayanan, S. A. Deshmukh, L. K. Shreshta, K. Ariga, V. Pol, and S. K. R. S. Sankaranarayanan, â??Cavitation and radicals drive the sonochemical synthesis of functional polymer spheresâ?, Accepted for publication, Applied Physics Letters (2016)
6. (Invited) B. Narayanan, S. Deshmukh, S. Sankaranarayanan, and S. Ramanathan, â??Strong correlations between structural order and passive state at waterâ??copper oxide interfacesâ?, Electrochimica Acta 179, 386 (2015)
7. B. Narayanan, G. Gilmer, J. Tao, J. DeYoreo, and C. Ciobanu, â??Self-assembly of collagen on surfaces: the interplay of collagen-collagen and collagen-substrate interactionsâ?, Langmuir 30, 1343 (2014)
8. B. Narayanan, A. van Duin, B. Kappes, I. Reimanis, and C. Ciobanu, â??A reactive force field for lithiumâ??aluminum silicates with applications to eucryptite phasesâ?, Modelling and Simulation in Materials Science and Engineering 20, 015002 (2012)