Research Interests

Postdoc 2

Development of machine learned orbital-free density functional theory (ML-OFDFT)

Postdoc 1

Search for a protocol for the calculation of 57Fe Mossbauer parameters (isomer shifts and quadrupole splitting) by employing scalar-relativistic X2C Hamiltonian and low-cost DFT functionals.

Comprehensive conformational analysis of disaccharides through a combination of computational techniques, including molecular dynamics (MD), CREST (Conformer-Rotamer Ensemble Sampling Tool), and ORCA’s own global geometry optimization and ensemble generator, GOAT.

Using QM/MM models, qualitative and quantitative investigation of various noncovalent interactions–such as hydrogen bonds, CH-π, π-π, and ion-π interactions–that play critical roles in small molecule binding and recognition within protein environments.

Linking spectroscopic properties with structural variations in metalloenzymes like Styrene oxide isomerase (SOI) and Carbon monoxide dehydrogenase (CODH). We often use MD and multi-scale modeling to identify the possible intermediates and transition states involved in catalytic cycles.

Development of general-purpose segmented all-electron relativistically contracted (SARC) basis sets for X2C calculations.

Ph. D.

1. Development of state-of-the-art empirical double hybrid functionals.

The main project of my Ph.D. degree is the development of new and improved empirical double hybrid density functional methods (DHs). We have published a series of papers that have gradually improved the previously available DHs. Our latest functional, XYG9[f1]@B44LYP is by far the best double hybrid functional one that can be found in the literature.

By re-parametrizing the original DSD functionals and including the D4 empirical dispersion correction, we developed revised DSD (revDSD) functionals.
Proposed ωDSD functionals by introducing range separation for the exchange term.

Developed dRPA-based DSD double hybrids to improve the performance of DSD functionals for systems with strong type A static correlation.
Introduced DSD3 and ωDSD3 functionals by including scaled canonical MP3 correlation term into the final energy expression, which significantly improved the performance of revDSD and ωDSD double hybrids.

Proposed XYG9 functionals by including scaled third-order Görling-Levy perturbation theory (GLPT3) term, which offered significant improvement over our previous DH functionals. These new double hybrids are even better than a few of the expensive composite WFT methods.

Do Double-Hybrid Functionals Benefit from Regularization in the PT2 Term? Observations from an Extensive Benchmark
Can we improve it further by using modified opposite-spin-scaled MP2 (MOS-MP2) instead of standard MP2?

2. Application of new double hybrid functionals.

We have already checked the performance of new revDSD, ωDSD and dRPA based DSD functionals for two additional datasets of systems exhibiting moderately strong static correlation (a.k.a. near degeneracy correlation) effects — one organometallic (10.1063/1.5137915, 10.1002/ijch.201900114, 10.1021/acs.jpca.1c01294, 10.1021/acs.jpca.1c01295), the other of extended porphyrins (10.1021/acs.jpca.9b10880, 10.1021/acs.jpca.1c01295, 10.1021/acs.jpca.1c01295).
We are currently exploring the applicability of our functionals for electronic excitation spectra, electric field derivative properties, NMR chemical shift, and so on.

3. Exploring unconventional density functionals (mainly density-corrected DFT).

We investigated under what circumstances HF-DFT (a form of density-corrected DFT ) offers an advantage over “standard” self-consistent DFT.

4. Noncovalent interactions

Performance of composite localized coupled-cluster methods for noncovalent interactions. [AIP Conference Proceedings; arXiv preprint: http://arxiv.org/abs/2111.01882]
Development of an extended dataset ion-pi interaction and performance assessment of different state-of-the-art DHs and composite and localized WFT methods.