BLANC - Blanc

WAve Function and DEnsity Functional MEthods COMbined: efficient treatment of challenging cases of electron correlation in molecules and materials – WADEMECOM

Submission summary

Density functional theory (DFT) is by now the most popular method for electronic structure calculations in quantum chemistry and condensed matter physics, because of its unique combination of low computational cost and reasonable accuracy. There are, however, a few but important cases where even the best approximations of its key ingredient, the exchange-correlation (xc) energy functional, fail. One of these important cases is the description of London dispersion (or van der Waals) forces, attributed to long-range dynamical electron correlation effects, which play a fundamental role in the cohesion of apolar species, in layered materials, in biological macromolecules, etc. Another example consists in a wide range of phenomena where the independent-particle approximation is even qualitatively wrong and which calls for a multireference treatment, as exemplified by bond-breaking and complicated open-shell systems of transition- metal, lanthanide and in particular actinide systems. The objective of this project is to explore systematically a new and unified strategy to overcome these problems of present-day DFT. Although practical solutions to these problems have already been proposed by several research groups, remaining either in the hierarchy of methods designated as 'Jacob's Ladder', or using some ad hoc extensions of the DFT, these approaches are often strongly empirical and not general, or have a prohibitive cost. In order to make a decisive step towards the solution of these specific shortcomings of DFT in describing delicate electron correlation problems, we join in this project the forces of research teams in Nancy, Paris, and Strasbourg, with different and complementary backgrounds in theory and modeling of the electronic structure of molecular systems and solids. Our strategy is based on the separation of the electron interactions by their range in the Hamiltonian. We expect that transferable short-range correlation effects can be handled efficiently via specific DFT functionals, while non-transferable long-range exchange and correlation will be treated by methodologies borrowed from wave function techniques. The short-range electron repulsion should be defined such that the system defined by the long-range interactions only can be described by a single determinant when there are no appreciable static correlation and/or dispersion. The long-range dynamic correlations, responsible for dispersion interactions will be handled by an adiabatic connection fluctuation-dissipation (ACFD) approach applied to the long-range interactions, via the calculation of the frequency-dependent charge density response. The strong, static correlation effects will be handled by a multiconfigurational treatment limited to the long range portion of the electron correlation, usually restricted to a subset of localized orbitals. We emphasize that the proposed approach has promising potential also in other cases, not addressed directly in the present proposal, where conventional DFT fails. Examples are the description of charge transfer complexes and charge transfer excitations, reaction barriers, dissociation energies and reaction heats. The range-separated DFT/WFT theory has been initiated and elaborated mainly the Paris group, active in the development of short-range functionals as well. The Nancy group has an experience with the implementation of range-separated hybrid functionals in both quantum chemical (MOLPRO) and solid state context (in collaboration with the VASP team in Vienna), especially for the treatment of the long-range part of the exchange and of the correlation for van der Waals forces. The Strasbourg group has a long-standing tradition in developing and implementing advanced methodologies (DIRAC), including DFT, in heavy-element quantum chemistry, including relativistic effects. The proposed methodologies will be validated on various challenging systems, like stacking dimers of aromatic rings, layered solids on the one side, actinide-actinide bonding, complexes of late actinides (Cm and Am), spin crossover systems, etc. Several specific aspects of this work will be done in collaboration with international partners from Austria, Denmark, Germany and Australia. As a result of our joint effort, we expect to produce computational tools, useful for both the quantum chemistry and the solid state physics communities, that are able to perform reliable total energy calculations for systems where dispersion forces, multireference situations, or both, play an important role. The availability of such tools would be a major breakthrough by considerably widening the range of applicability of DFT methods in chemistry, biology and physics.

Project coordination

Janos ANGYAN (Université)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE - DELEGATION REGIONALE MIDI-PYRENEES

Help of the ANR 463,748 euros
Beginning and duration of the scientific project: - 48 Months

Useful links

Explorez notre base de projets financés

 

 

ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter