Density functional theory in solution: Implementing an implicit solvent model for CASTEP and ONETEP

SpeakerJames C. Womack (University of Southampton)
Time11am on 24th November 2016

Density functional theory (DFT) has come to dominate the landscape of modern quantum chemistry and condensed matter physics, largely because of its favourable balance of accuracy and computational efficiency. The combination of efficient implementations of the theory with effective use of modern parallel computer hardware allows the practical application of DFT to relatively large assemblies of atoms. Following this approach, electronic structure software packages such as CASTEP [1] and ONETEP [2] have been successfully applied in the study of complicated chemical systems, such as biomolecules, nanoparticles and surfaces.

When studying systems of technological interest, with real-world applications, it is vital to capture the effect of the environment. In many cases this is some kind of liquid solvent. Even with the benefit of modern computer hardware and theoretical techniques, explicit modelling of solvent molecules is often impractical. An alternative approach is to implicitly represent the solvent as a dielectric continuum, placing the solute in a vacuum cavity surrounded by this unstructured medium. In a DFT calculation using an implicit solvent model, the effect of the solvent on the solute enters into the Hamiltonian via an interaction potential. This approach simplifies the simulation of solute-solvent interactions by vastly reducing the number of atoms requiring a quantum mechanical treatment and avoiding the need to statistically average solvent molecule positions.

In this talk, I will describe a powerful minimal-parameter implicit solvent model which is based on the direct solution of the nonhomogeneous Poisson equation (NPE) and which employs a cavity derived directly from the quantum mechanical electron density [3]. This model was originally implemented in ONETEP for application in large-scale DFT calculations, using an efficient multigrid solver, DL_MG, to solve the NPE [4]. A number of improvements to the model and its implementation are currently underway. In particular:

For each of these projects, I will outline the theoretical and computational aspects of the work and present our recent progress.


  1. S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. I. J. Probert, K. Refson, and M. C. Payne, Z. Kristallogr. 220, 567 (2005)
  2. C.-K. Skylaris, P. D. Haynes, A. A. Mostofi, and M. C. Payne, J. Chem. Phys. 122, 084119 (2005))
  3. J. Dziedzic, H. H. Helal, C.-K. Skylaris, A. A. Mostofi, and M. C. Payne, EPL 95, 43001 (2011)
  4. L. Anton, J. Dziedzic, C.-K. Skylaris, and M. Probert, Multigrid solver module for ONETEP, CASTEP and other codes, Technical report, dCSE, 2013