Abstract

At the moment we see a rapid development of computers and simulation techniques. In 10 years time many chemists will first model their experiments computationally before implementing them at the bench to use existing knowledge and to predict the unknown. Several interdisciplinary challenges we will have to overcome to bring computational chemical modelling closer to predictive molecular engineering and in silico design of new materials and drugs have been identified in a recent review.(1)

The presentation focuses on the following questions:

- What does predictive mean in the context of chemical simulations?

- How can the concept of chemical accuracy be extended to include reliable risk assessments and quality control?

- How can uncertainties be classified in the computational chemistry modelling cycle?

- Which role do validation and verification play?

- What can be learned from other fields?

The answers to these questions will be explored based on examples from electronic structure and coarse grained simulation methods. In future, the interplay of theoretical and computational chemistry with other scientific disciplines such as mathematics, engineering, and visualisation science will lead to predictive simulations which enhance developments cycles of molecules, reduce costs and risks, and generate new ideas for experiments to design new drugs or materials and maybe one day will allow for the virtual testing and design of a whole airplane.

(1) J. B. Rommel. ‘From Prescriptive to Predictive: an Interdisciplinary Perspective on the Future of Computational Chemistry’. submitted, 2016. (2) J.B. Rommel, ’Reliable and Predictive Multi-scale and Multi-physics Simulations: from Atoms and Electrons to Drugs and Airplanes’, to be published 2016 (3) J.B. Rommel, ’Density Embedding Methods: Quality Control and Efficiency for Large Metal Clusters’, to be published 2016