Dynamical control of Coulomb interactions and Hubbard bands in monolayer 1T-TaS$_2$

TL;DR

This work demonstrates that lattice distortions in monolayer 1T-TaS$_2$ directly tune electronic correlations by modulating the screened Coulomb interaction $U$. By combining constrained RPA with DFT+DMFT in a frozen-phonon description of the CDW amplitude mode, the authors map $U(a)$ and the Hubbard-band positions for each amplitude $a$, revealing a coherent shift of LHB and UHB that tracks the Mott gap. A ~3\% increase in CDW amplitude drives a transition from a Mott insulator to a correlated metal with a pronounced quasiparticle peak at $E_F$, illustrating dynamic control of correlations with light-driven lattice motion. This provides a microscopic mechanism for optical control of correlated phases in two-dimensional quantum materials and a framework for interpreting time-resolved spectroscopies in CDW TMDs.

Abstract

Monolayer 1T-TaS$_2$ hosts a star-of-David charge-density wave (CDW) that stabilizes a low-temperature Mott-insulating state. Recent time-resolved spectroscopies indicate a coupling between the CDW amplitude mode and the electronic correlation strength, yet the role of the screened Coulomb interaction remains unclear. Using the constrained random-phase approximation, we show that the CDW amplitude modifies the bare and screened on-site interactions, leading to sizable variations in the effective Hubbard U. Our combined density functional and dynamical mean-field theory calculations reveal that the Hubbard bands shift in concert with the CDW amplitude, and that a reduced distortion drives a transition from a Mott insulator to a correlated metal. These results demonstrate a direct link between lattice distortions and Coulomb interactions in transition-metal dichalcogenides, providing a microscopic mechanism for light-induced control of correlated phases in two-dimensional quantum materials.

Figures (4)

• Figure 1: (a) Illustration of a time-resolved STM experiment. A near-infrared (NIR) pulse indirectly excites the CDW amplitude mode in monolayer 1T-TaS$_2$, while a time-delayed THz probe pulse induces electron tunneling between the sample and STM tip to probe the local density of states. (b) Top view of the SoD supercell with Ta (brown) and S (yellow) atoms. $d$ denotes the distance between the central (site A) and the outer (site C) Ta atoms. Violet arrows indicate the CDW amplitude mode oscillation.
• Figure 2: DFT band structure and Wannier projection for varying CDW distortion.(a–d) Electronic band structure of monolayer 1T-TaS$_2$ for CDW amplitudes $a=-1\%$ to $a=+2\%$ around equilibrium, computed with PBE Perdew1996 in VASP Kresse1996b. The half-filled band at the Fermi level (red) is projected onto a maximally localized Wannier function Pizzi2020. (e–h) Wannier isosurfaces centered on the central Ta atom with smaller weight on surrounding atoms. Red/blue denote positive/negative values. The orbital becomes increasingly delocalized with larger CDW amplitude.
• Figure 3: Tuning the screened Coulomb interaction and Mott gap via CDW amplitude modulation.(a) Bare and screened on-site Coulomb interaction as a function of CDW amplitude $a$. The shaded region highlights the parameter range for which the electronic spectral functions $A(\omega)$ are shown in (b). Increasing the CDW amplitude from -1% to +2% shifts the lower (LHB) and upper Hubbard bands (UHB) and reduces the Mott gap. At $a=3\%$, the system undergoes a transition to a correlated metallic state. In (b) the chemical potential was set to the center of the gap.
• Figure 4: Momentum-resolved spectral function as a function of CDW amplitude. The electronic spectral function of the Ta $d_{3z^2-r^2}$ orbital (red) obtained from DFT+DMFT is shown together with the DFT band structure (black). As the CDW amplitude $a$ is varied from (a) -1% to (d) +2%, the Mott gap---defined as the spectral gap between the occupied LHB and the unoccupied UHB---gradually decreases from 0.5 to 0.1 eV.