| dc.description.abstract | Topological phases of matter such as topological insulator (TI), quantum anomalous Hall insulator (QAHI), nodal line semimetal (NLSM), and triple point metal (TPM) can be realized by manipulating the spin-orbit coupling (SOC) and symmetry present in the crystalline solids. To understand many of these phases in novel materials, we attempted to develop symmetry-based methods using the density functional theory in combination with the high-throughput (HT) and machine learning (ML) approaches. To take advantages of crystalline symmetry in preserving triply degenerated band crossing in the Brillouin zone, we first studied a new set of semimetals X2YZ (X = {Cu, Rh, Pd, Ag, Au, Hg}, Y = {Li, Na, Sc, Zn, Y, Zr, Hf, La, Pr, Pm, Sm, Tb, Dy, Ho, Tm} and Z = {Mg, Al, Zn, Ga, Y, Ag, Cd, In, Sn, Ta, Sm}), which show the existence of multiple topological triple point fermions along four independent C3 axes in this cubic lattice. Next, we report the topological phases of the hydrogenated group 13 monolayers (aluminane, gallenane, indinane, and thallinane), where time-reversal (TRS), inversion (IS), and mirror symmetry (MS) protect the topological NLSM state. Interestingly, under 2.6% tensile strain along the x-direction, gallenane evolves to TI, which could be promising for spintronics applications. On the other hand, TRS and IS breaking with strong SOC effect in the hexagonal lattices produce valley-polarized QAH effect, having potential applications in dissipation-less valleytronics devices. To explore large search space of VP-QAH insulators, a HT method has been developed, and applied to “aNANt” MXene database, which resulted in 14 MXenes exhibiting the VP-QAH effect. These screened MXenes have non-zero Berry curvature at the Brillouin-zone corner and a single chiral edge state connecting from valence to conduction band within the bulk bandgap, resulting in a valley-dependent dissipation-less current flow. The strong spin-orbit coupling effect in the 2D buckled inversion asymmetric crystal could provide another important striking phenomenon Rashba effect, where the orbital momentum is locked with the spin. By employing HT-based computational screening method, we extract 206 Rashba semiconductors, among which 20 have Rashba constant greater than 1eVÅ. These could have promising applications in spin-based field effect transistors. Predicting existence of all these phases via DFT is a time and resource extensive process, which hinders the accelerated search. Therefore, we have developed ML based models in imbalanced dataset that can classify and predict new magnetic nodal line semimetals and Rashba materials. The classification and regression models to predict MNLSMs, and corresponding nodal positions use only the basic elemental features. With an excellent classification accuracy of 94% to predict NLMSMs, the regression model predicts the nodal line positions (N1 and N2) having R2 of 0.96 and 0.92, respectively. Similarly, the classification model classifies the Rashba materials with 89% accuracy. The symmetry analysis-based HT and ML approaches developed here could be utilized to search for novel materials having these exotic properties in an accelerated manner. | en_US |
