Role of on-site Coulomb energy and negative-charge transfer in a Dirac semi-metal

TL;DR

NiTe$_2$, a Type-II Dirac semimetal, has had conflicting reports on the role of electron correlations. The authors combine SXPES, HAXPES, XAS, and Ni $2p$-$3d$ resonant-PES with charge-transfer cluster-model analysis to quantify $U_{dd}$ and $\Delta$, and to compare NiTe$_2$ with NiO. They find $U_{dd}=3.7$ eV and $\Delta=-2.8$ eV for NiTe$_2$, indicating a negative-charge-transfer metal with a ground state dominated by $d^9\underline{L}^1$ weight; the hybridization $T_{eg}$ is smaller than in NiO. The analysis shows $U_{dd} > |\Delta|$, placing NiTe$_2$ in a moderately correlated, $p$-type Dirac semimetal regime and highlighting the essential role of finite on-site repulsion in stabilizing Dirac physics. Overall, the work demonstrates that correlation effects, though modest, are crucial for the Dirac semimetal state in NiTe$_2$ and helps reconcile previous conflicting views.

Abstract

Angle-resolved photoemission spectroscopy in combination with band structure calculations have shown that the layered transition metal dichalcogenide NiTe$_2$ is a type-II Dirac semimetal. However, there are conflicting conclusions in the literature regarding the role of electron correlations in NiTe$_2$. We study the core-level and valence band electronic structure of single crystal NiTe$_2$ using soft and hard X-ray photoemission spectroscopy (SXPES, HAXPES), X-ray absorption spectroscopy (XAS) and Ni $2p-3d$ resonant photoemission spectroscopy(resonant-PES) to quantify electronic parameters in NiTe$_2$. The on-site Coulomb energy ($U_{dd}$) in the Ni $3d$ states is quantified from measurements of the Ni $3d$ single particle density of states and the two-hole correlation satellite. The Ni $2p$ core level and $L$-edge XAS spectra are analyzed by charge transfer cluster model calculations using the experimentally estimated $U_{dd}$ (= 3.7 eV), and the results show that NiTe$_2$ exhibits a negative charge-transfer energy ($Δ$ = -2.8 eV). The same type of cluster model analysis of NiO $L$-edge XAS confirms its well-known strongly correlated charge-transfer insulator character, with $U_{dd}$ = 7.0 eV and $Δ$ = 6.0 eV. The $d$-$p$ hybridization strength $T_{eg}$ for NiTe$_2$$<$NiO, and indicates that the reduced $U_{dd}$ in NiTe\textsubscript{2} compared to NiO is not due to an increase in $T_{eg}$. The increase in $d^n$ count on the Ni site in NiTe$_{2}$ by nearly one electron is attributed to negative-$Δ$ and a reduced $U_{dd}$. However, since $U_{dd}$$>$$|Δ|$, the results indicate the important role of a finite repulsive $U_{dd}$ in making NiTe$_{2}$ a moderately correlated $p$-type Dirac semi-metal.

Figures (10)

• Figure 1: Ni 2p and Te 3$p$ core level spectra of NiTe$_2$ single-crystal measured at $T$ = 80 K with $h\nu$ = 1.5 keV (SXPES) and at $T$ = 20 K with $h\nu$ = 6.5 keV (HAXPES)
• Figure 2: Least square fitting of Ni 2$p$ and Te 3$p$ core levels of NiTe$_2$ single-crystal measured at $T$ = 80 K with $h\nu$ = 1.5 keV (SXPES) and at $T$ = 20 K with $h\nu$ = 6.5 keV (HAXPES)
• Figure 3: Least-squares fitting of Te 3$d$ core levels of NiTe$_2$ single-crystal measured at $T$ = 80 K with $h\nu$ = 1.5 keV (SXPES) and at $T$ = 20 K with $h\nu$ = 6.5 keV (HAXPES)
• Figure 4: (a). The Ni 2$p$-3$d$ resonant-PES valence band intensity map plotted as a function of incident photon energies ($h\nu$ = 849-879 eV) versus BE ( = -1.2 to 45.8 eV). (b) The Ni $L_3$- and $L_2$-edge XAS plotted as a function of $h\nu$ (top X-axis). (c) Valence band spectra (BE = -1.2 to 35.0 eV) measured at select $h\nu$ values (labelled $a - v$) across the $L_3$- and $L_2$-edges of Fig. 1(b). (d) The kinetic energy of the Resonant Raman peak which becomes the $L_3VV$ Auger peak, plotted as a function of $h\nu$ relative to the XAS $L_3$ peak energy(bottom X-axis).
• Figure 5: NiTe$_2$ valence band measurements at different off-resonant $h\nu$ values of $h\nu$ = 849.25 eV, 1.5 keV, 6.5 keV, and one in the Resonant Raman region with $h\nu$ = 851.75 eV.