
Thermal robustness of the quantum spin Hall phase in monolayer WTe2
Antimo Marrazzo1, ∗
1Dipartimento di Fisica, Universit`a di Trieste, Strada Costiera 11, I-34151 Trieste, Italy
(Dated: February 13, 2023)
Monolayer 1T’-WTe2has been the first two-dimensional crystal where a quantum spin Hall phase
was experimentally observed. In addition, recent experiments and theoretical modeling reported
the presence of a robust excitonic insulating phase. While first-principles calculations with hybrid
functionals and several measurements at low temperatures suggest the presence of a band gap of
the order of 50 meV, experiments could confirm the presence of the helical edge states only up
to 100 K. Here, we study with first-principle simulations the temperature effects on the electronic
structure of monolayer 1T’-WTe2and consider the contributions of both thermal expansion and
electron-phonon coupling. First, we show that thermal expansion is weak but tends to increase
the indirect band gap. Then, we calculate the effect of electron-phonon coupling on the band
structure with non-perturbative methods and observe a small reduction of the band inversion with
increasing temperature. Notably, the topological phase and the presence of a finite gap are found
to be particularly robust to thermal effects up to and above room temperature.
In 2014 [1], Qian et al. predicted through first-
principles simulations that two-dimensional (2D) tran-
sition metal dichalcogenides (TMDs) in the 1T’ phase
would exhibit a sharp quantum spin Hall effect (QSHE),
characterized by strong band inversions and relatively
large band gaps. These TMDs are defined by the chemi-
cal formula MX2, where M is the transition metal (W or
Mo) and X is the chalcogenide (Te, Se or S), and they
exhibit a variety of polytypic structures such as 1H, 1T,
and 1T’. In particular, the 1T structure in MX2TMDs is
typically unstable in freestanding conditions and under-
goes a spontaneous lattice distortion in the xdirection
to form a period-doubling 2 ×1 distorted superstructure,
the 1T’ structure, made of 1D zigzag chains along the
ydirection [2]. In all the TMDs, but WTe2[3], the 1T’
structure is dynamically stable but it does not correspond
to the lowest energy polymorph [1] and it is a metastable
phase, although it can be stabilized under appropriate
conditions [2] thanks to the large energy barrier between
the 1H and the 1T’ phases [1]. In WTe2instead, the 1T’
structure corresponds to the most stable phase [2,3].
Conductance experiments have confirmed the presence
of a quantum spin Hall insulating (QSHI) state in 1T’-
WTe2until 100 K [4,5].
As QSHI material, 1T’-WTe2excels in many aspects.
First, it exhibits a strong band inversion [1] of about
1 eV and a band gap above kBTR[6], where TRis
room temperature. In addition, WTe2is a simple binary
compound, which simplifies the experimental synthesis–
especially with bottom-up approaches–compared to the
numerous ternary QSHI compounds [7]. Finally, WTe2
has a layered crystal structure and the binding energy be-
tween the layers is relatively low [8] (around 30 meV/˚
A2),
such that the material is exfoliable into monolayers [9].
It is then compelling that monolayers of 1T’-WTe2have
become among the most promising 2D materials to re-
∗antimo.marrazzo@units.it
alize the QSHE and the target of several experimental
investigations [4,5,10–13].
Although predictions report an extremely robust band
inversion and indirect band gap above room temperature,
experiments could find signatures of the QSHI phase un-
til 100 K only, suggesting a possible role of temperature
on the band structure. Here we show that the topologi-
cal phase and the presence of a finite gap are particularly
robust to thermal effects up to high temperatures, sug-
gesting that extrinsic effects or thermally activated bulk
conductance [14] might be responsible for the measured
transition temperature.
The presence of a finite indirect band gap in mono-
layer WTe2has been debated. Density-functional theory
(DFT) simulations [1,10] with the semilocal PBE func-
tional predict a metallic state, with bands overlapping
at the Fermi level and a direct gap throughout the Bril-
louin zone (BZ). On the contrary, calculations [6,15,16]
with the HSE hybrid functional predict an indirect gap
of about 50 meV. The predicted band gap is also strongly
sensitive to the lattice constant [6,11,17], where even a
small amount of strain can open a gap as predicted at the
DFT level with different functionals such as LDA, PBE
and PBEsol [17] and confirmed experimentally [11]. In
this work, we calculate the PBE and HSE band structure
both with spin-orbit coupling (SOC) on the equilibrium
structure relaxed at the DFT-PBE level, as reported in
Fig. 1. The Wannier-interpolated HSE band structure
has a finite indirect gap of 62 meV, while at the PBE
level the system is metallic.
Angle-resolved photoemission spectroscopy [6]
(ARPES) and scanning-tunneling microscopy and
spectroscopy [10] (STM/STS) experiments indicate the
presence of a finite gap of around 50 meV, in agreement
with the hybrid-functional predictions. Some recent
STM/STS experiments instead suggested the presence of
a metallic band structure with a Coulomb gap induced
by disorder [18]. Finally, more recent transport [13] and
STM/STS [12] experiments, backed-up by many-body
calculations [13,19,20], provided strong evidence of the
arXiv:2210.11258v2 [cond-mat.mes-hall] 10 Feb 2023