Error loading player: No playable sources found

3553553

New variants of the multistate CASPT2 method

Date
April 15, 2021

The accuracy of multistate CASPT2 (MS-CASPT2) for vertical transition energies and the robustness of extended MS-CASPT2 (XMS-CASPT2) in case of near-degenerate CASSCF states are combined in a new complete active space second-order perturbation theory approach.
The new method, building on top of the well-established machinery of the original CASPT2 approach, uses a transformation akin to XMS-CASPT2 to ensure approximate invariance under unitary rotations of the zeroth-order states and smoothly interpolates the Fock operator used in the Hamiltonian partition between state-specific and state-average regimes according to a dynamic weighting scheme.
The resulting methodology possesses the most desirable features of both MS-CASPT2 and XMS-CASPT2, that is, the ability to provide accurate transition energies and correctly describe avoided crossings and conical intersections.
In this contribution we present the results obtained with different weighting schemes and parameters, assessing the reliability on typical benchmark systems of photochemical relevance. The current limitations of the new method are discussed, especially from the point of view of excited states dynamics and recent developments addressing them are reported.

Speaker

Speaker Image for Roland Lindh
Professor, Uppsala University

Related Products

Thumbnail for COMP Poster Session:
COMP Poster Session:
: [COMP] Division of Computers in Chemistry
Thumbnail for COMP Poster Session:
COMP Poster Session:
: [COMP] Division of Computers in Chemistry
Thumbnail for Trans histone pathway of Ubiquitination of nucleosomal H2A by BRCA1/BARD1-UbcH5c Complex | Poster Board #444
Trans histone pathway of Ubiquitination of nucleosomal H2A by BRCA1/BARD1-UbcH5c Complex | Poster Board #444
The Breast Cancer Associated Protein 1 and its binding partner BRCA1 Associated Ring Domain Protein (BARD1) are a heterodimeric protein that, upon mutation, possess a high risk of breast and ovarian cancer…
Thumbnail for Machine learning supported molecular geometry optimizations: The restricted variance optimization procedure
Machine learning supported molecular geometry optimizations: The restricted variance optimization procedure
We will here present the the Restricted Variance optimization (RVO) procedure -- a method based on the gradient enhanced Kriging (a Gaussian Process Regression) in combination with a step restriction based on a cap on the allowed predicted variance[1,2]…