The scope of computational modeling

In view of the most serious environmental and energy problems, the need for appropriate materials with specific physical, mechanical and chemical behaviors is of uttermost importance. Despite the huge variety of technological materials, the search for new solutions adapted to special needs continues to be a major focus area of material development. Traditional material design activities are based on the vast amounts of empirical knowledge accumulated over the centuries. This pragmatic approach leads to increased performance but often fails to identify optimal solutions because of the lack of information about the actual micro-, nano- and atomic-level mechanisms.

Intelligent materials design

A more efficient and intelligent design of new materials should be based on accurate composition-processing-property relationships, which can be used as compass to look for solutions with preassigned properties. Such correlation maps can in principle be generated from a number of measured parameters describing the properties and processes. It is often assumed that a well-designed experiment can always be used to find the true value of any physical parameter and its connection to chemistry, structure and processing. But what happens if the quantities we are interested in are difficult or perhaps even impossible to measure? Even accessing the measurable physical parameters could sometimes be tremendously time-consuming, resource-demanding and expensive due to the large number of physical laboratory tests needed to establish robust design maps.

The state-of-the-art solution to the above problem is given by the modern computational materials science based on well-established models, theories and databases. During the last few decades, this computational materials modeling approach has reached a level of accuracy where an easily accessible input information, such as the atomic numbers and composition, is sufficient to yield values of physical quantities with an accuracy equal to or better than experiments.

Providing further understanding

In addition to data-generation, another important aspect of computational modeling is its potential to provide understanding. The traditional trial-and-error material development routes as well as the modern large-scale modeling efforts can answer the question, ‘what happens if we add a certain element or alter the processing’, but provides less information on ‘why that happens’.

This understanding becomes especially important today when we have access to a great calculation machinery in terms of accurate methods, computer programs and powerful resources. We can easily produce an astronomical amount of data which besides the contribution to the correlation maps, also open towards deep and reliable understanding. Efficient and sustainable material development assumes that we can look inside the materials, to the networks of atoms, molecules, phases, grains, and components, and reveal various interactions and mechanisms controlling their behavior. Only advanced first-principles and empirical modeling techniques formulated in terms of mathematical equations and scientific computer programs allow for such in-depth analysis.