L-to-R: Pablo, Jaren, Sudheer, Rupali, Alessandro, Michele, Shou-Feng
Our research is divided into two main topics: dynamics of charge and excitation energy transfer, and subsystem DFT method development. We pursue state-of-the-art theory development as well as applications related to photosynthesis, design of novel materials for molecular electronics, electron transfer in biosystems and molecule-surface interactions.
Non-Adiabatic Dynamics with Subsystem DFT
This project involves development of electronic structure methods to carry out non-adiabatic molecular dynamics simulations in the framework of subsystem DFT. The code development for this project will take place within ADF and Quantum ESPRESSO suites of softwares.
The new methods will be applied to a range of problems:
- Electron and exciton transfer in biosystems
- Dye sensitized solar cells
- Organic-Metal and Organic-Semiconductor interfaces
Non-Local Potentials for Subsystem DFT
Interactions between subsystems, covalent bonds as well as weak van der
Waals, are tough to get with Susbsystem DFT.
This project involves development of non-local orbital-free and
orbital-dependent embedding potentials for the
correct treatment of these interactions between subsystems.
Non-additive correlation energy = van der Waals interactions
We have formulated a novel van der Waals theory inspired by density embedding ideas. In a JCP article, we devise a van der Waals theory for fragments with overlapping electron densities - an outstanding problem for the full correlation energy of interaction. The method is theoretically exact and amenable to approximations. The RPA approximation to this theory yielded excellent agreement against CCSD(T) benchmark interaction energies for the S22 set.
How does the environment affect superexchange couplings of hole transfer in DNA?
Graduate student Pablo Ramos has shed light on the role of the environment when holes (positive charges) move along a DNA oligomer. His all-electron calculations published in JCTC involved systems with up to 1300 electrons, and show that the DNA counter-strand has an overpowering effect over other environmental effects (such as state polarization and energy shift from the ribose groups). In addition, he confirmed previous semiempirical predictions on the variability of the nucleobase's ionization potentials going from the 3' to the 5' end of DNA.
Organic crystals: QM/MM is not enough, while QM/QM/MM offers a more complete description. The price? 850 QM atoms
Dr. Ruslan Kevorkyants has demonstrated how FDE can get two birds with one stone: correclty describe the spin density of an embedded radical, and achieve convergence of the QM/MM simulation with respect to the size of the QM region. The result is a quantitative agreement with EPR experiments of a X-ray irradiated Guanine hydrochloride crystal. His paper, defines FDE as a suitable method for simulating organic molecular crystals.
Decomposition of response functions (polarization propagator) into subsytem contributions
The response function of a system can be split into subsystem contributions. This is only apparently easy because the response should be subsystem-additive. As always, the devil is in the details: in a JCP article, we report that each subsystem response function can be decomposed into two terms: one is subsystem-specific, and the other involves couplings with all other subsystem response functions. This generalizes previous formulations of subsystem TD-DFT and has implications for understanding dynamical screening effects. We are still working on this project and more exciting findings will follow soon.
Charge transfer couplings and excitations with subsytem DFT
Implementation in ADF of electronic couplings in the context of Subsystem Density-Functional Theory (specifically the FDE formulation). Results published in JCP.