ENVIRON: A Library for Complex Electrostatic Environments in Electronic-structure Simulations

PI: Stefan Goedecker (University of Basel)

Co-PIs: Nicola Marzari, Juerg Hutter

January 1, 2014 - December 31, 2016

Project Summary

The computational study of chemical reactions in complex, wet environments is critical for applications in many fields, and of cross-disciplinary interest to physics, chemistry, materials science, chemical engineering, and biology. Solar-energy harvesting in a dyesensitized cell or electro-catalytic water splitting are just two examples of relevance for applications in energy and environment; examples from wet chemistry and biology are way too numerous to list, but could e.g. start with the key role that solvation, hydrophobicity, and water-assisted reactions have for proteins and for the entire cell environment. While the study of isolated molecules in solution (typically water) has been pioneered by the quantum chemistry community for over 30 years, very little has been done in the computational community of condensed matter physicists and materials scientists, typically running codes able to study periodic or extended systems. This is at variance with the manifold applications of these codes to dryenvironments, with thousands of papers published yearly. In addition, and very much driven by solid-state and materials applications, it is often essential to study chemical reactions at applied electrochemical potentials, establishing the correct relation between charge and potential, and taking into account the complex electrostatic screening coming from the solvent and the electrolytes - an almost completely unexplored area in the electronic-structure community, but of overarching importance to many of the applications alluded to above. Driven by very recent advances and developments in this field, we propose to develop and distribute an open-source library of verified and validated electrochemical and solvation modules, able to run efficiently on advanced computing architectures, and interface it with some of the core electronic-structure codes developed or co-developed in Switzerland - namely ABINIT, BigDFT, CP2K, and Quantum ESPRESSO. Given its documented and open-source nature, it is also expected that such tool will become of wide use outside these communities, and be adopted by other public or distributed codes. The core objective of this library will be to describe complex electrostatic environments where an explicit solvent becomes implicit, with a position-dependent dielectric constant, or where mobile ions can shield the charge or multipoles of the system of interest; it will embed the quantum simulation engines into a robust and efficient Poisson-Boltzmann solver that has been extensively verified and validated. This library will satisfy all the requirements for use in combination with the codes mentioned above, that represent the electronic charge density distribution and the corresponding potential on regular Bravais grids, but with a clear eye on future developments extending the electrostatic embedding described here towards more varied physical models, that could e.g. include the applied potential of operation for a micro-electronic device, or the electrostatic environment of the quaternary structure of a protein or of an entire cell membrane.

In addition, thanks to some of the established advantages of these codes, such as the possibility to do Born-Oppenheimer or Car-Parrinello molecular dynamics with tight energy conservation, or efficient linear-response calculations both in density-functional perturbation theory (DFPT) and in time-dependent density functional theory (TDDFT), this library will significantly extend the range of applications that can be addressed, to study optical effects such solvatochromism, the potential energy surfaces of electrocatalytic reactions, nucleation and growth in solution, phase stability, dissolution, and Pourbaix phase diagrams of nanoparticles, and more generally drive the detailed study of the liquid-solid interface; as mentioned, several of the great technological challenges in energy production and storage belong to these categories. The libray will be highly efficient on all common present and emerging hardware platforms, robust in convergence, user friendly and well documented. The user will be able to choose among several modules to describe the electrostatic environments, and the different physical models will be thoroughly tested and validated. All the three groups involved in the project have an excellent track record in the development and implementation of novel methods for electronic-structure calculations, in distributing them to the community, and in education and training for these tools. In addition, the project participants are in close contact with computer scientist such as Torsten Hoefler to get support in reducing the amount of effort for porting the library to diverse computer architectures. All the groups have also well established connections to the Swiss National Supercomputing center CSCS, for close assistance and interactions on different technical and hardware problems.