Extension of local PNO-Møller-Plesset and PNO-coupled-cluster electronic structure methods to periodic systems

• Applicant: Arman Nejad
• Subject Area: Theoretical Chemistry: Electronic Structure, Dynamics, Simulation
• Term: since 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 517466522

The molecular treatment of electron correlation has seen rapid advances in the recent past, in particular with regard to the maturity of local pair natural orbital (PNO) and explicit correlation (F12) methods. In the solid state, the situation is very different. Several methods have established themselves to treat electron correlation in solids, e.g. density functional theory (DFT), quantum Monte Carlo (QMC), the density matrix renormalisation group (DMRG), full configuration interaction quantum Monte Carlo (FCIQMC), and Hartree-Fock (HF) and correlated post-HF methods. In contrast to the molecular case, HF and post-HF correlation methods, such as Møller-Plesset and coupled-cluster theory, are much less established for treating electron correlation in solids. In their canonical implementations, their high polynomial scaling limits their applicability to only small systems. Major breakthroughs were pioneered by Pisani et al. who published the first periodic HF code in the 1980s (CRYSTAL), paving the way for the later implementation of post-HF methods which was pioneered by Pisani and Schütz et al. in 2005 with a low-scaling local projected atomic orbital (PAO) LMP2 code (CRYSCOR), combining the periodic HF code from CRYSTAL with at that time state of the art methods to treat molecular electron correlation in MOLPRO. Currently, only a few of the existing codes extend beyond canonical implementations (high polynomial scaling) or MP2, i.e. coupled-cluster. In the same spirit that lead to the development of CRYSCOR, it is proposed here to combine the existing periodic HF code in TURBOMOLE with state of the art molecular local PNO-MP2 and PNO-coupled-cluster codes that are also available. On the one hand, this would significantly extend the existing toolkit to treat electron correlation in solids, as the local MP2 implementation in CRYSCOR is not based on pair natural orbitals (PNO) and is thus limited to smaller systems. Similarly, the periodic coupled-cluster implementation by Grüneis et al. is based on plane waves and with this project, an alternative formulation based on LCAO (linear combination of atomic orbitals) could be provided. (This project would further pave the way for future improvements to also interface with the explicit F12 molecular code in TURBOMOLE to further extend the applicability of post-HF methods in solids.)One possible application of interest would be the dehydrogenation of formic acid (HCOOH) in zeolites which is gaining interest as a renewable energy source in recent years. Conceptually, post-HF methods are very well-suited to treat the dynamic electron correlation in such sparsely packed insulating systems but efficient scaling is mandatory if one also wants to go further and also investigate nuclear vibrational effects. These are deemed important since the reaction most likely involves first the isomerisation from trans- (hydrogens on opposite sides of CO bond) to cis-HCOOH (hydrogens on the same side of CO bond).

• DFG Programme: WBP Fellowship
• International Connection: United Kingdom
• Host: Professor Dr. David P. Tew