Computational Material Science - The Grand Challenges and Present Status

The need to understand and control material properties arises in the context of basic research, several industries and a host of areas of immediate relevance to society such as energy security, health, national security. The advent of high-speed computers, sophisticated theoretical techniques and powerful algorithms during past few decades has naturally led to the question is it possible to design materials with the desired properties using only computers, given only the compositional information? In fact realizing this goal is the ultimate dream of Computational Material Science (CMS). The realism associated with this dream has to be assessed in the light of the provocative statement One of the continuing scandals in the physical sciences is that it remains impossible to predict structure of even the simplest crystalline solids from a knowledge of their composition by Maddox in Nature in 1988. If structure determination itself is so difficult, how is it possible calculate properties? The enterprise of CMS has progressed by leaps and bounds since then. There exists methods of determining crystal structures with a fair degree of reliability, given only the composition. Also, one can proceed further and calculate many physical and chemical properties of crystalline materials using only the compositional information.

There is no royal route to calculate materials properties from scratches, even for perfectly crystalline materials. Even though the basic constituents of condensed matter and the microscopic law that governs their interaction are known, it has not been possible to predict materials properties. This inability is not only due to the intractability of the many-body Schrodinger/Dirac equations, but also due to the fact there are many instances where a many-body system shows emergence of fundamentally new kinds of phenomena such as occurrence of phase transitions. There is no simple way of even anticipating such emergent phenomena from an a priori knowledge of the microscopic laws. Unraveling new principles that govern the macroscopic realm from microscopic laws, is indeed a theoretical challenge. Recent advances in theory and computations have played a significant role in partially addressing this challenge.

In contrast to the crystalline and other homogeneous archetypes taught in the regular curricula, real materials are heterogeneous and replete with a variety of defects such as point defects, grain boundaries, multiple phases, surfaces & interfaces, and so on. The behaviour of such materials is also governed by collective, coordinated action of the phenomena associated with each length and time scale of material heterogeneity. The present day computational resources are grossly inadequate for providing a first principles description of properties of materials influenced by processes which straddle huge dynamical range of length- and time-scales. The present strategy is to use tools appropriate for each of these scales and link them seamlessly. For example, carry out ab initio calculations to determine the effective interaction between atoms; carry out molecular dynamics simulation using this interaction potential; carry out Monte Carlo simulations using the information and insight obtained from MD about the processes occurring at a longer time and spatial scale, and so on. World-wide, development of such powerful multiscale models is work still under progress. The present talk is intended to provide a bird’s eye view of this fascinating subject.