The holy grail for researchers in mixed quantum-classical dynamics is to develop a method fast enough to allow millions of single point calculations, but providing wave functions with uniform quality over different diabatic characters and over large sections of the configurational space. Unfortunately, we could not be farther from this goal. There are in fact many inexpensive methods available for electronic structure simulations, but all of them fail in these wave-function quality criteria.<br/>This disappoint scenario, however, should be balanced back by the fact that it is still possible to retrieve relevant information out of nonadiabatic dynamics based on low-level methods, even if in principle they are not completely suited for the problem in question. I have often observed this when using single-reference methods, either based on Hartree-Fock or density-functional theory, to deal with ultrafast processes in photoexcited organic systems. Naturally, the main question is to know how much can we stretch such methods before we are left with nothing but qualitatively wrong results.<br/>In this talk, I will illustrate this point through a series of case studies featuring surface hopping nonadiabatic dynamics of large molecular systems, based on several linear-response single-reference methods, including TD-DFT, TD-DFTB, TDA, and ADC(2).

Single-reference dynamics of strongly correlated systems: How far can we go?

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.