Santosh Adhikari
Santosh Adhikari
PhD , Physics
PhD , Physics
Embedded Files
Research Activities
Machine Learning for understanding materials' electro and thermochemistry
Accurate Prediction of HSE06 Band Structures for a Diverse Set of Materials Using Δ-Learning
Accurate Prediction of HSE06 Band Structures for a Diverse Set of Materials Using Δ-Learning
Santosh Adhikari, Jacob Clary, Ravishankar Sundararaman, Charles Musgrave, Derek Vigil-Fowler, Christopher Sutton
Chemistry of Materials, 35, 20, 8397–8405 (2023)
Violin plot of k-resolved orbital eigenvalues predicted by various methods relative to HSE06. With an MAE of 0.13 eV, Linear+KRR (indicated by simply ML in next two subfigures in the right), is the best perfoming model. DOI: https://doi.org/10.1021/acs.chemmater.3c01131
Band gaps (eV) predicted by various methods as a function of HSE06 band gaps.
Band structure of one of the representative systems (AlP) predicted by the ML model.
Parity and histogram plot of the machine learning (ML, RFR) model trained to correct DFT predicted materials formation enthalpy employing features derived using charge-transfer and bonding analysis from DFT calculations (PBE and SCAN) itself. DOI: https://doi.org/10.48550/arXiv.2307.07609
Plot shows PBE formation enthalpy errors are drastic for materials where the PBE predicted charge-transfer between atoms therein is significant, indicating PBE predicted formation enthalpy is not reliable for such similar cases.
Self-interaction correction to semilocal DFT functionals.
The Fermi–Löwdin self-interaction correction for ionization energies of organic molecules
The Fermi–Löwdin self-interaction correction for ionization energies of organic molecules
Santosh Adhikari, Biswajit Santra, Shiqi Ruan, Puskar Bhattarai, Niraj K Nepal, Koblar A Jackson, Adrienn Ruzsinszky
The Journal of Chemical Physics, 153, 184303 (2020)
Perdew-Zunger self-interaction correction (PZSIC, 1981) introduces further errors in the cases where the electron density is not one-electron like (zσ=1). This work modifies existing schemes of scaling PZSIC (LSIC+).
The scheme, LDA-rLSIC+ predicts highly accurate ionization energies of 14 diverse organic molecules,. on par of theoretical benchmark, GW calculations.
LDA-rLSIC+ generalizes well and performs on par of LDA-LSIC+ and LDA-LSIC for predicting ionization energies of atoms and molecules in G21- dataset.
Dispersion correction to semilocal DFT functionals.
Molecule-surface interaction from van der Waals-corrected semilocal density functionals: The example of thiophene on transition-metal surfaces
Molecule-surface interaction from van der Waals-corrected semilocal density functionals: The example of thiophene on transition-metal surfaces
Santosh Adhikari, Hong Tang, Bimal Neupane, Gábor I Csonka, Adrienn Ruzsinszky
Physical Review Materials, 4, 025005 (2020)
Santosh Adhikari, Niraj K Nepal, Hong Tang, Adrienn Ruzsinszky
The Journal of Chemical Physics, 154, 124705 (2021)
DOI1: https://doi.org/10.1063/5.0042719
https://doi.org/10.1103/PhysRevMaterials.4.025005
Figure (center) shows the schematic of the van der Waals energy computed using damped Zaremba-Kohn (dZK) scheme that when added to the semilocal DFT energy (PBE/SCAN) predicts accurate adsorption energy, distances, site and angles for thiophene (left), benzene (right) in addition to adsorption of CO and rare gas (Xe) over the metallic surfaces.
Contact Info:
Contact Info:
Email: santos.adh114[at]gmail.com
Phone: (+1) (267) 909-1660
Page updated
Google Sites
Report abuse