Hi all,
Tomorrow's group meeting will be given by Thomas. The title and abstract
are included inline after this message.
Best,
Ian
-------------------------------------
Speaker: Thomas Markovich
Title: Enabling large-scale simulation of many-body dispersion forces in
condensed phase systems
Abstract: Dispersion interactions are ubiquitous in nature, and extremely
important for explaining the structure and function of many systems, from
soft matter to surfaces and solids. Due to their long-range and scaling
with system size, dispersion interactions can prove particularly important
in modeling nanostructured systems where reduced dimensionality creates
large polarizable surfaces. Standard pairwise approximations are
insufficient for such systems, and the true non-additive and many-body
character of dispersion plays a crucial role. The many-body dispersion
(MBD) method of Tkatchenko and co-workers [A. Tkatchenko et al., Phys. Rev.
Lett. 108, 236402 (2012); A. Ambrossetti et al., J. Chem. Phys. 2014, 140,
18A508] seeks to address this behavior by computing the full many-body
correlation energy for a fictitious set of coupled quantum harmonic
oscillators that mimic the fluctuations of the real polarizable
valence-electron density. Much of the work on MBD, to date, has focused on
the energetics of various molecules and materials, with all necessary
gradient information being obtained through numeric differentiation. We
recently presented an implementation of the relevant analytic gradients
with respect to nuclear displacements, which permitted fast and accurate
geometry optimizations of many gas-phase systems and showed that PBE+MBD
geometries matched those of highly accurate wavefunction theories at a
fraction of the cost. In this talk I will describe an efficient
implementation of the MBD energy and analytic gradients, which has enabled
their application to larger simulations of condensed-phase systems. I will
show examples of geometry and unit-cell optimizations with MBD corrections
in the condensed phase, as well as the first ever MBD corrected band
structure calculations.
Show replies by date