Cationic substitution, dynamical stability, thermal stability, electronic and thermoelectric properties in 2D dialkali metal monoxides via DFT and ML approach

S S. Chellaiya T Thomas Rueshwin R R. D. Eithiraj

Abstract

Abstract Using the WIEN2k software, Density Functional Theory (DFT) was applied to analyse the impact of cationic substitution on the physical features of the 1T-K2O monolayer. Phonon dispersion analysis confirmed dynamical stability, Ab-initio Molecular Dynamics (AIMD) simulations validated thermal stability, and cohesive energy calculations ensured thermodynamic stability of 1T-K2O. Based on the phonon studies, both 1T-KNaO and 1T-KRbO are dynamically unstable with a slightly visible imaginary frequency. Specifically, for electronic property assessment, generalized gradient approximation (GGA) and hybrid exchange-correlation functionals were utilized. This study unveiled the 1T-KXO (X = Na, K, Rb) monolayers as an indirect band gap semiconductor, for 1T-K2O, 1T-KNaO and 1T-KRbO were 0.94 eV (1.84 eV), 1.03 eV (1.94 eV) and 0.84 eV (1.77 eV) obtained implementing GGA and hybrid functionals, respectively. Using a machine learning (ML) approach, the band gap was predicted as 1.45 eV (0.85 eV) for 1T-K2O, 1.79 eV (0.97 eV) for 1T-KNaO, and 1.17 eV (0.72 eV) for 1T-KRbO, with random forest regression (linear regression) method. The physical properties were tailored by the impact of cationic substitution on the 1T-K2O were studied. The variation in the physical properties were investigated. Optical analysis indicated a strong absorption coefficient, underscoring the 1T-KXO monolayers potential for photovoltaic applications in the UV region. The ZT value obtained at room temperature are 0.58, 0.86 and 0.69 for 1T-K2O, 1T-KNaO and 1TKRbO, respectively. Additionally, the 1T-KNaO demonstrated promising thermoelectric properties, at 400 K achieving a figure of merit (ZT) of 0.93, indicating its suitability for waste heat recovery. A ML model was trained to predict the ZT of 1T-KXO using random forest regression and linear regression.

Article Details

Volume / Issue Vol. 15, Issue 1
Published July 28, 2025
ISSN 2045-2322
Publisher Nature Portfolio

Journal Info

Scientific Reports

Nature Portfolio

ISSN: 2045-2322 Open Access Life Sciences

Authors (3)

S

S. Chellaiya

T

Thomas Rueshwin

R

R. D. Eithiraj