Abstract
The interplay of electronic correlations, spin–orbit coupling and topology holds promise for the realization of exotic states of quantum matter. Models of strongly interacting electrons on honeycomb lattices have revealed rich phase diagrams featuring unconventional quantum states including chiral superconductivity and correlated quantum spin Hall insulators intertwining with complex magnetic order. Material realizations of these electronic states are, however, scarce or inexistent. In this work, we propose and show that stacking 1TTaSe_{2} into bilayers can deconfine electrons from a deep Mott insulating state in the monolayer to a system of correlated Dirac fermions subject to sizable spin–orbit coupling in the bilayer. 1TTaSe_{2} develops a StarofDavid charge density wave pattern in each layer. When the StarofDavid centers belonging to two adyacent layers are stacked in a honeycomb pattern, the system realizes a generalized Kane–Mele–Hubbard model in a regime where Dirac semimetallic states are subject to significant Mott–Hubbard interactions and spin–orbit coupling. At charge neutrality, the system is close to a quantum phase transition between a quantum spin Hall and an antiferromagnetic insulator. We identify a perpendicular electric field and the twisting angle as two knobs to control topology and spin–orbit coupling in the system. Their combination can drive it across hitherto unexplored grounds of correlated electron physics, including a quantum tricritical point and an exotic firstorder topological phase transition.
Similar content being viewed by others
Introduction
Prospects of quantum information technologies have motivated an intense search for systems, which intertwine topology and electronic correlations^{1,2,3,4}. Strongly interacting and spin–orbitcoupled electrons on the honeycomb lattice, as theoretically described by the Kane–Mele–Hubbard model^{5}, feature quantum spin Hall (QSH), Mott–Hubbard, collective magnetic and chiral superconducting states. While several weakly interacting QSH systems are now wellstudied experimentally^{2,6}, material realizations of honeycomb Kane–Mele–Hubbard fermions with strong correlations and spin–orbit coupling are rare^{7,8}. Here, we introduce a new “van der Waals engineering” platform to serve this purpose.
We show that stacking of 1TTaSe_{2} into bilayers can deconfine electrons from a deep Mott insulating state realized in the monolayer to a system of correlated Dirac fermions subject to sizable spin–orbit coupling. Central to this transition is the possibility of van der Waals materials to stack in different configurations. For a specific honeycomb arrangement (Fig. 1) the kinetic energy associated with the electronic hopping t turns out to be of the same order of magnitude as the effective local Coulomb repulsion U. The system features, therefore, electronic correlations, which turn out to put the system right on the verge between QSH and correlated antiferromagnetic insulating states at charge neutrality and support chiral superconductivity under doping. We finally demonstrate that tuning the system via electric fields and twisting of the layers relative to each other sensitively affects the lowenergy electronic structure in terms of emerging Dirac mass and spin–orbit coupling terms and leads to completely unexplored regimes of correlated electrons.
Results and discussion
From a correlated insulator to emergent Dirac fermions
Layered groupV transition metal dichalcogenides (TMDCs) such as 1TTaSe_{2} or 1TTaS_{2} feature a lowtemperature commensurate charge density wave (CCDW) where Taatoms are displaced into StarofDavid (SoD) patterns (Fig. 1a)^{9,10}. In this phase, the SoDs form a triangular \(\sqrt{13}\times \sqrt{13}\) superlattice in every layer and host correlated electrons: 1TTaS_{2} shows a metaltoinsulator transition when entering the CCDW phase^{10}. In 1TTaSe_{2}, the bulk remains conductive till lowest temperatures, while the surface exhibits a Mott transition around 250 K^{11,12}. Recently, 1TTaSe_{2} has been fabricated down to monolayer thickness^{13,14,15} and a pronounced thickness dependence of the electronic structure has been reported^{15}. As bonds between the layers are mainly of van der Waals type, different stacking configurations are observed in experiments^{16,17} and have a strong impact on the electronic structure^{18,19,20}.
We compare the CCDW state in the monolayer to a bilayer with honeycomb stacking, where the SoD centers form a buckled honeycomb lattice (Fig. 1b). The bottom layer SoD centers form sublattice A and the top layer sublattice B. This stacking is one of many possible configurations, which can generally differ, both, by the local atomic stacking and by the stacking of the SoD centers. The local stacking considered here, has the Ta atoms of the bottom layer approximately beneath the lower Se atoms of the top layer (Fig. 1c). While this kind of local stacking is not the most commonly observed one of the Tphase TMDCs^{15,17}, a corresponding stacking sequence has been found in transmission electron microscopy studies of 1TTaS_{2}^{17} and turns out to be metastable in density functional theory (DFT) simulations of 1TTaSe_{2}^{21}. Regarding the stacking of the SoD centers further van der Waals DFT total energy calculations (see Supplementary Information) show that the particular honeycomb arrangement studied here, is metastable and energetically on the order of 10 meV per formula unit above the lowest energy configuration. This is similar to the configuration observed experimentally in ref. ^{15} and it is thus plausible that also the honeycomb configuration considered here, is within reach of experiments. Experimental approaches including tearandstack^{22} and Scanning Tunneling Microscopy voltage pulsebased manipulation schemes^{23} can present possible avenues to control and switch metastable stacking configurations and to reach the honeycomb configurations discussed in our paper. At small twist angles all different kinds of configurations can be realized locally in the moiré, including likely those honeycomb cases discussed here.
To study the electronic structure of such engineered stackings, we combine abinitio calculations in the frameworks of DFT and the random phase approximation (RPA) with effective lowenergy models, which we investigate with dynamical meanfield theory (DMFT) and twoparticle selfconsistent (TPSC) manybody approaches, see “Methods” section.
In the CCDW phase, the band structure of monolayer 1TTaSe_{2} obtained with nonspinpolarized DFT is characterized by a single (Ta) flat band at the Fermi level^{10,18,19,20}, which has a bandwidth of less than 20 meV (Fig. 2a, left). Hence, the CCDW formation largely quenches inplane hopping of the electrons. In the honeycombstacked bilayer (with no twist), two dispersive bands with a bandwidth of the order of 200 meV emerge from the lowenergy flat band of the monolayer (Fig. 2a, right). Comparison of the mono and bilayer bandwidths shows that interlayer hopping effects must dominate over intralayer hoppings by approximately an order of magnitude. In this sense, CCDW TaSe_{2} bilayers are the exact opposite of graphene bilayer systems, since in the latter outofplane coupling is an order of magnitude weaker than inplane hopping^{24,25}. For the 1TTaSe_{2} honeycomb bilayer, the upper and lower lowenergy bands touch as Dirac points at the Brillouin zone corners K and K'. In the undoped system, these Dirac points are exactly at the Fermi level.
We next construct a Wannier Hamiltonian to describe the Dirac bands with one Wannier function for each SoD center, i.e., two Wannier orbitals per bilayer CCDW superlattice unit cell (Supplemental Fig. 2). The resulting nearestneighbor hopping between a sublattice A site in the bottom and a neighboring sublattice B site in the top layer amounts to t = −34 meV and is the leading term of the Wannier Hamiltonian. There are further terms in the Wannier Hamiltonian, which are, however, at least an order of magnitude smaller than t.
The effective Hubbard interaction U for the SoD Wannier orbitals of CCDW TaSe_{2} calculated in RPA is U ≈ 130 meV, which is also in line with the experimental estimates in ref. ^{15} and calculations for TaS_{2}^{19}. The ratio of hopping to Coulomb interaction is decisive in determining the strength and kind of electronic correlation phenomena taking place. Our calculations yield U/∣t∣ ≈ 3.8.
To study the resulting electronic correlations, we performed simulations of the Hubbard model for the nontwisted CCDW 1TTaSe_{2} bilayer in the framework of DMFT and the TPSC approach^{26,27}. The quasiparticle weight Z shown in Fig. 2c is a measure of the electronic correlation strength. In the temperature range T = 60–230 K, both DMFT and TPSC consistently yield essentially constant Z ≈ 0.75. Our system is thus at intermediate local correlation strength and far away from the paramagnetic Mott–Hubbard transition taking place above U_{Mott}/t ≈ 8.2. The DMFT and TPSC studies of the full Hubbard model for the nontwisted CCDW 1TTaSe_{2} bilayer are, thus, in line with studies of the idealized Hubbard model on the honeycomb lattice^{28}.
The TPSC calculations give insight to spinfluctuations taking place in the system (Supplemental Fig. 4). The inverse intra and intersublattice terms of the static magnetic susceptibility at wave number q = 0, as well as the antiferromagnetic correlation length ξ_{AFM} are shown in Fig. 2d. We observe that antiferromagnetic fluctuations with alternating spin orientation between the two sublattices (A, B) are dominant and strongly enhanced at temperatures T ≲ 100 K. These fluctuations indicate that the nontwisted CCDW 1TTaSe_{2} bilayer is close to a quantum phase transition from a Dirac semimetal to an antiferromagnetic insulator, which occurs for ideal Hubbard honeycomb systems exactly in the range of U_{c}/t ≈ 3.6–3.8^{6,29,30}.
Spin–orbit coupling
The aforementioned magnetic correlation phenomena are sensitive to details of the lowenergy electronic structure and spin–orbit coupling (SOC), which we discuss in the following based on a symmetry analysis. The space group of nontwisted CCDW 1TTaSe_{2} bilayer in the honeycomb structure is \(P\bar{3}\) (#147), comprising inversion symmetry and threefold rotations C_{3} around an axis perpendicular to the bilayer. Imposing also timereversal symmetry, every band must be twofold degenerate. Therefore, SOCinduced qualitative changes of the band structures can occur near the Dirac points at K and K′. A corresponding k ⋅ p expansion (see Supplementary Information) reads
where the pseudospin S describes the sublattice degree of freedom, σ acts on the electron spin, and τ = ±1 labels the valley (K, K′) degree of freedom. This Hamiltonian comprises three contributions; the first contribution is a twodimensional massless Dirac term, with the sublatticepseudospin playing the role of the spin inherent to the Dirac equation. This term is analogous to the massless Dirac term in graphene^{6,24,25}. SOC is responsible for the second and third contributions to H_{0}: a valley–spin–sublattice coupling λ_{SOC} = 0.74 meV, which is often called Kane–Mele spinorbit term^{31,32}, and a sublatticestaggered Rashba term α_{R2}, which belongs to the R2 class according to the classification form ref. ^{33}. A finite Kane–Mele term λ_{SOC} opens a gap and turns the system described by H_{0} into a QSH insulator^{31,32}. Importantly, the Kane–Mele term here is enhanced in comparison to its counterpart in graphene by two orders of magnitude^{24,25,34}, and corresponds to a temperature T ≈ 10 K, which is well accessible in experiments. Given that U/∣t∣ ≈ 3.8, the nontwisted CCDW 1TTaSe_{2} bilayer implements a material realization of the Kane–Mele–Hubbard model in a regime very close to the topological quantum phase transition from a QSH insulator to an antiferromagnetic insulator (Fig. 2e).
Dirac fermions and their topology are affected by different kinds of mass fields. An energy difference between electrons localized in sublattice A and B leads to a socalled Semenoff mass term M, which would enter the Hamiltonian H_{0} of equation (1) in the form MS_{z}. This term breaks sublattice invariance and thereby inversion symmetry and leads to a transition from a QSH to a band insulator at ∣M∣ = ∣λ_{SOC}∣^{31,35}. In the nontwisted CCDW 1TTaSe_{2} bilayer, the intrinsic Semenoff mass M_{0} = 0 is required to vanish by symmetry. Our Wannier construction yields M_{0} = 0.14 meV, which is small as compared to all other relevant terms in the system. Possible origins of this small symmetry breaking can be the Wannier constructions and also faint asymmetries accumulated during selfconsistency iterations in the DFT calculations. Vertical electric fields E_{z}, as realizable in field effect transistor geometries (Fig. 1c), break inversion symmetry, translate into staggered sublattice potentials, and therefore corresponding extrinsic Semenoff mass contributions ΔM (Fig. 2b). DFT calculations (Supplementary Fig. 3) yield the approximate relation ΔM ≈ eE_{z}d/ϵ_{⊥}, where d = 6.4 Å is the interlayer distance, e is the elementary charge, and ϵ_{⊥} = 3.72 plays the role of an effective dielectric constant. The QSH to band insulator transition is reached for E_{z} ≈ 0.5 mV Å^{−1} = 50 kV cm^{−1} (Fig. 2b), which is well within reach of experiments^{36}.
Taken together, our calculations show that the honeycombstacked nontwisted CCDW 1TTaSe_{2} bilayer is located in a region of the phase diagram (Fig. 2e) with three different phases (QSH insulator, band insulator and antiferromagnetic insulator) coming together. Electron correlation is known to change the order of the QSHband insulator transition from second to firstorder^{37}. Contrary to the standard noninteracting QSH to band insulator transition, where the gap closes and reopens continuously with vanishing gap at the transition point, the QSH gap remains finite and the system changes discontinuously to a band insulating state at the transition point. Application of vertical electric fields in the system at hand represents hence a possibility to realize this exotic interactioninduced firstorder transition.
Twisted CCDW 1TTaSe_{2} bilayers
Layered van der Waals systems allow to realize different stacking configurations via twisting, i.e., relative rotations between the layers. General twist angles θ lead to incommensurate moiré patterns superimposed to the CCDW lattice. Arguably the simplest case of twisting is a rotation angle of θ = 180^{∘}, which leads to a system with identical Bravais lattice but different symmetry of the supercell basis (Fig. 1c): in case of honeycomb stacking of the CCDW, the space group of the 180^{∘} twisted structure of CCDW TaSe_{2} bilayer is reduced to P3, meaning that inversion symmetry is lost with respect to the nontwisted case. The resulting band structure (Fig. 3a) is qualitatively similar to the nontwisted honeycomb case regarding the overall shape and width of the lowenergy bands. Thus, also the 180^{∘} twisted case will be far away from the paramagnetic Mott transition. However, the lowenergy band structure is markedly different in the 180^{∘} twisted case. First, the conduction band is almost flat between K and M. Second, inversion symmetry breaking lifts band degeneracies: our DFT calculations reveal a staggered potential and an associated intrinsic Semenoff mass term of M_{0} = 8.55 meV, which opens a gap at the K and K' points already without SOC and without external electric fields. Furthermore, additional SOC terms are now allowed by symmetry (see Supplementary Information) and completely lift the remaining band degeneracies except for the timereversal symmetric points Γ and M.
Based on a symmetry analysis, we obtain the following lowenergy model in the vicinity of K and K′:
The additional terms with respect to the nontwisted honeycomb structure equation (1) are the intrinsic Semenoff mass (M_{0}), the spin–valley coupling term (B), the Kane–Mele–Rashba interaction (λ_{R}), and a Rashba interaction belonging to the R1 class^{33} giving rise to pure Rashba spinpolarization patterns (α_{R1}). The last term (with coupling constant λ_{D}) can also be seen as an effective kdependent magnetic field parallel to the zaxis. Our DFT calculations yield M_{0} = 8.55 meV, B = −1.85 meV, ∣λ_{SOC}∣ ≲ 0.05 meV and λ_{R} = 3.21 meV. These terms affect the dispersion and imprint an intricate sublattice and spin structure to the lowenergy bands, which can be manipulated by vertical electric fields as shown in Fig. 3b.
Except for E_{z} = −4.2 meV Å^{−1} (ΔM = −6.68 meV and M = M_{0} + ΔM = 1.87 meV), the system is always gapped. We calculated the Z_{2} topological invariant for the noninteracting 180^{∘} twisted bilayer in comparison to the nontwisted bilayer case,, as well as for two cases in between where the ratio B/λ_{SOC} is varied (Fig. 3c, and Supplemental Information). In the nontwisted bilayer, the system is in a QSH state unless an extrinsic sufficiently large Semenoff mass term or an additional Rashba SOC term λ_{R} are added. In the 180^{∘} twisted case, the situation is very different regardless whether or not the intrinsic Semenoff mass term is compensated by an external electric field and regardless of λ_{SOC}. Indeed, many changes in the SOC terms suppress the QSH state in the 180^{∘} twisted bilayer: The comparably large Rashba λ_{R} and the spin–valley coupling terms B and a strong reduction in λ_{SOC}. Each of these alone is sufficient to suppress the QSH state. At vertical electric field E_{z} = −4.2 mV Å^{−1} the gaps at K and K' close, and a band touching point emerges.
While it is clear that there will be tendencies towards interactioninduced (quasi)ordered phases as well, here, the kind of ordering is likely different from the nontwisted case but largely unexplored. The band touching at E_{z} = −4.2 mV Å^{−1} implements a situation similar to saddle points in a twodimensional dispersion, where already arbitrarily weak interactions would trigger different kinds of magnetic or excitonic instabilities^{38}. How these instabilities translate into the intermediately correlated and strongly spin–orbitcoupled case of 180^{∘} twisted TaSe_{2} is a completely open question.
Conclusions and outlook
The field of twistronics with materials like bilayer graphene is based on the idea that weak interlayer coupling can flatten highly dispersive bands and thereby boost electronic correlations^{39,40,41}. The system introduced here takes the opposite route of deconfining formerly Mott localized electrons. This approach should be generally applicable to interfaces of Mott localized electrons under two conditions: the interlayer coupling should substantially exceed the inplane one and at the same time define a connected graph linking all sites of the system. Possible example systems range from stacking faults in the bulk of CCDW layered Mott materials^{17} to molecular systems^{42}.
Especially the twisting degree of freedom opens new directions to experiments. Since interlayer hopping is the dominant kinetic term in deconfined Mott systems like bilayers of CCDW 1TTaSe_{2}, we expect incommensurability effects to be much more pronounced than in twisted graphene systems^{39}. θ = 30^{∘} twisted CCDW 1TTaSe_{2} bilayer should realize a quasicrystal with twelvefold rotation symmetry and provide an experimental route to correlated electrons and emerging collective states in a quasicrystalline environment.
The prototypical case of CCDW 1TTaSe_{2} bilayer demonstrates how deconfinement of Mott electrons leads to exotic states of quantum matter: the nontwisted bilayer approaches a quantum tricritical region of competing QSH, trivial band insulating, and antiferromagnetic insulating states. At 180^{∘} twist angle different kinds of electrically controllable band degeneracies with associated manybody instabilities, hypothetically of excitonic type, emerge. Clearly, the phase space for manipulating deconfined Mott electrons is high dimensional. We here identified the combination of twist angle and perpendicular electric field as decisive for TaSe_{2} bilayers. Further means to control emerging electronic states include dielectric engineering^{43} and charge doping. Our calculations showed that the nontwisted CCDW 1TTaSe_{2} bilayer in honeycomb stacking approximates the (Kane–Mele) Hubbard model on the honeycomb lattice with U/∣t∣ ≈ 3.8 very well. In this regime, doping the system away from the Dirac point towards the van Hove singularity is expected to lead to chiral superconductivity^{44,45,46,47}, most likely of d + idtype.
Methods
DFT calculations
We perform DFT^{48,49} calculations by using the Vienna abinitio simulation package (VASP)^{50,51} with the generalized gradient approximation of Perdew, Burke, and Ernzerhof (GGAPBE) for the exchangecorrelation functional^{52,53}. We obtain the total energies, relaxed structures and electronic structure of mono and bilayer 1TTaSe_{2} systems. We calculate the total energies for various possible stackings in undistorted and CCDW bilayer 1TTaSe_{2} using Γcentered kmeshes of 15 × 15 × 1 and 9 × 9 × 1, respectively, and taking into account van der Waals (vdW) corrections within DFTD2 and crosschecking with DFTD3^{54,55}, see Supplemental Fig. 1. The ionic relaxation is done using the conjugate gradient algorithm until all force components are smaller than 0.02 eV Å^{−1}.
Since the DFTD2 corrections yield correct interlayer distances but do not correctly capture the CCDW distortions, we adopted the following relaxation procedure to calculate the commensurate \(\sqrt{13}\times \sqrt{13}\) CCDW bilayer structures:

1.
We perform relaxations for a \(\sqrt{13}\times \sqrt{13}\) supercell of the monolayer (without vdW corrections), using a superlattice constant of a = 12.63 Å according to ref. ^{15}. We fix the vertical positions of the Ta atoms, while allowing for Ta inplane displacements. The Se atoms are allowed to freely relax in all three directions.

2.
We include then a second layer and optimize the interlayer distance, d, while keeping all relative intralayer distances fixed. We find d = 6.4 Å for the ideal honeycomb stacking in CCDW bilayer 1TTaSe_{2}.

3.
We relax the CCDW bilayer following the same procedure described for the monolayer, i.e., without vdW corrections and vertical positions of the Ta atoms according to the optimized interlayer distance d fixed, while allowing for inplane displacements. The Se atoms are allowed to freely relax in all three directions.
We cross check the results obtained by this procedure against calculations with vdW corrections according to the DFTD3 method^{55}. In contrast to DFTD2, the CCDW distortions are well described in the DFTD3. Thus, full relaxations of all atomic positions have been performed in the DFTD3 framework for the mono and bilayer. Both our stepbystep procedure using DFTD2 method, and the full relaxation using DFTD3 give equivalent results for the total energies (see Supplemental Fig. 1b), for the crystal and band structures.
For the noncollinear magnetic calculation, i.e., when SOC is included, we set the net magnetic moment to zero in all atoms of the unit cell, and use a Γcentered kmesh of 6 × 6 × 1.
Estimation of the screened Hubbard interaction U via RPA
We estimate the local Hubbard interaction U for the flat bands around the Fermi level in the CCDW 1TTaSe_{2} bilayer from the abinitio calculation of the screened Coulomb interaction, using RPA for the undistorted bilayer. We follow a similar procedure as in ref. ^{56}, which we summarize below:

We initially calculate the WANNIER90 tightbinding model for the three lowenergy Ta bands \({\mathcal{C}}\), whose orbital character is mostly \(\{{d}_{{z}^{2}},{d}_{{x}^{2}{y}^{2}},{d}_{xy}\}\) (see Fig. 2(a) in ref. ^{56}) in the undistorted monolayer 1TTaSe_{2}.

The static RPAscreened Coulomb interaction tensor W_{αβγδ}(q, ω → 0) is calculated for undistorted monolayer 1TTaSe_{2}, where q is a reciprocal wave vector on a Γcentered mesh of 18 × 18 × 1, and \(\alpha ,\beta ,\gamma ,\delta \in {\mathcal{C}}\). We neglect q = 0 terms in our RPA analysis in order to avoid unphysical effects.

In the CCDW 1TTaSe_{2} bilayer, the \({d}_{{z}^{2}}\) orbitals from Ta atoms in the SoD centers have the largest contribution for the bands around the Fermi level. Thus, for each q, we consider only the tensor element W(q) ≡ W_{αααα}(q) with \(\alpha ={d}_{{z}^{2}}\).

Then, the local Hubbard interaction U in a single star of David is calculated by averaging over the \({d}_{{z}^{2}}\) orbital weight from each Ta atom (labeled by \({w}_{{d}_{{z}^{2}}}({\bf{R}})\)) in the star of David:
$$U=\mathop {\sum}\limits_{{\bf{R}},{\bf{R}}^{\prime} \in {\davidsstar} }{w}_{{d}_{{z}^{2}}}({\bf{R}})U({\bf{R}}{\bf{R}}^{\prime} ){w}_{{d}_{{z}^{2}}}({\bf{R}}^{\prime} )$$(3)where U(R) is the discrete Fourier transform of W(q).
DMFT and TPSC manybody calculations
For nontwisted CCDW 1TTaSe_{2} bilayer, we construct tightbindingHubbard Hamiltonians of the type
from a Wannierization of the abinitio DFT band structure (Supplementary Information), and estimate the onsite repulsion U to be about 130 meV by means of RPA calculations. We study these effective lowenergy models from a manybody perspective to judge the type of correlations in the system. In DMFT the lattice Hamiltonian is mapped onto a selfconsistently determined single impurity problem, solved—in our case—within numerical exact quantum Monte Carlo in the hybridization expansion flavor (CTHYB) (for a review, see^{57}). The resulting sublatticeresolved selfenergy, Σ, is local (kindependent) but it contains frequencydependent nonperturbative corrections beyond HartreeFock to all orders and can account for Mott–Hubbard metaltoinsulator transitions. All DMFT calculations are performed using w2dynamics^{58}. The double counting is accounted for using the fully localized limit. For twodimensional systems it is important to estimate nonlocal effects at the level of the selfenergy, not included in DMFT. To this goal, we apply the TPSC method^{59}, which produces accurate results in the weaktointermediate coupling regime, if compared to lattice quantum Monte Carlo calculations in the single band Hubbard model. For nontwisted CCDW 1TTaSe_{2} bilayer, which is modeled by a multiband system we use the multisite formulation of TPSC^{26} while neglecting the Hartree term to avoid double counting of correlation effects already accounted for in DFT. Moreover, to be able to apply TPSC to this system we project out spin offdiagonal terms and take only the diagonal spinup contributions from DFT while still assuming a paramagnetic state. The combination of TPSC accounting for the kdependence of Σ and DMFT, in which we can include all offdiagonal terms between spinorbitals and access antiferromagnetic ordering at strong coupling consitutes a powerful tool to determine the manybody nature of 1TTaSe_{2}.
Data availability
The data that support the findings of this study is available from the corresponding author upon reasonable request.
References
Tsui, D. C., Stormer, H. L. & Gossard, A. C. Twodimensional magnetotransport in the extreme quantum limit. Phys. Rev. Lett. 48, 1559–1562 (1982).
Qi, X.L. & Zhang, S.C. Topological insulators and superconductors. Rev. Mod. Phys. 83, 1057–1110 (2011).
Sato, M. & Ando, Y. Topological superconductors: a review. Rep. Prog. Phys. 80, 076501 (2017).
Rachel, S. Interacting topological insulators: a review. Rep. Prog. Phys. 81, 116501 (2018).
Hohenadler, M. & Assaad, F. F. Correlation effects in twodimensional topological insulators. J. Phys.: Condens. Matter 25, 143201 (2013).
Wehling, T. O., BlackSchaffer, A. M. & Balatsky, A. V. Dirac materials. Adv. Phys. 63, 1–76 (2014).
Marrazzo, A., Gibertini, M., Campi, D., Mounet, N. & Marzari, N. Prediction of a largegap and switchable KaneMele Quantum Spin Hall insulator. Phys. Rev. Lett. 120, 117701 (2018).
Wu, X., Fink, M., Hanke, W., Thomale, R. & Di Sante, D. Unconventional superconductivity in a doped quantum spin Hall insulator. Phys. Rev. B 100, 041117 (2019).
Wilson, J. A., Salvo, F. J. D. & Mahajan, S. Chargedensity waves and superlattices in the metallic layered transition metal dichalcogenides. Adv. Phys. 24, 117–201 (1975).
Rossnagel, K. On the origin of chargedensity waves in select layered transitionmetal dichalcogenides. J. Phys.: Condens. Matter 23, 213001 (2011).
Perfetti, L. et al. Spectroscopic signatures of a bandwidthcontrolled Mott transition at the surface of 1TTaSe_{2}. Phys. Rev. Lett. 90, 166401 (2003).
Colonna, S. et al. Mott phase at the surface of 1TTaSe_{2} observed by Scanning Tunneling Microscopy. Phys. Rev. Lett. 94, 036405 (2005).
Nakata, Y. et al. Selective fabrication of Mottinsulating and metallic monolayer TaSe_{2}. ACS Appl. Nano Mater. 1, 1456–1460 (2018).
Börner, P. C. et al. Observation of charge density waves in freestanding 1TTaSe_{2} monolayers by transmission electron microscopy. Appl. Phys. Lett. 113, 173103 (2018).
Chen, Y. et al. Strong correlations and orbital texture in singlelayer 1TTaSe_{2}. Nat. Phys. 16, 218–224 (2020).
MüllerCaspary, K. et al. Atomicscale quantification of charge densities in twodimensional materials. Phys. Rev. B 98, 121408 (2018).
Hovden, R. et al. Atomic lattice disorder in chargedensitywave phases of exfoliated dichalcogenides (1TTaS_{2}). Proc. Natl Acad. Sci. USA 113, 11420–11424 (2016).
Freericks, J. K., Krishnamurthy, H. R., Ge, Y., Liu, A. Y. & Pruschke, T. Theoretical description of timeresolved pump/probe photoemission in TaS_{2} : a singleband DFT.DMFT(NRG) study within the quasiequilibrium approximation. Phys. Status Solidi (B) 246, 948–954 (2009).
Darancet, P., Millis, A. J. & Marianetti, C. A. Threedimensional metallic and twodimensional insulating behavior in octahedral tantalum dichalcogenides. Phys. Rev. B 90, 045134 (2014).
Ritschel, T., Berger, H. & Geck, J. Stackingdriven gap formation in layered 1TTaS_{2}. Phys. Rev. B 98, 195134 (2018).
Lazar, P., Martincová, J. & Otyepka, M. Structure, dynamical stability, and electronic properties of phases in TaS_{2} from a highlevel quantum mechanical calculation. Phys. Rev. B 92, 224104 (2015).
Weston, A. et al. Atomic reconstruction in twisted bilayers of transition metal dichalcogenides. Nat. Nanotechnol. 15, 592–597 (2020).
Ma, L. et al. A metallic mosaic phase and the origin of Mottinsulating state in 1TTaS 2. Nat. Commun. 7, 10956 (2016).
Castro Neto, A. H., Guinea, F., Peres, N. M. R., Novoselov, K. S. & Geim, A. K. The electronic properties of graphene. Rev. Mod. Phys. 81, 109–162 (2009).
Katsnelson, M. I. Graphene: Carbon in Two Dimensions (Cambridge University Press, 2012).
Zantout, K., Altmeyer, M., Backes, S. & Valentí, R. Superconductivity in correlated BEDTTTF molecular conductors: critical temperatures and gap symmetries. Phys. Rev. B 97, 014530 (2018).
Zantout, K., Backes, S. & Valentí, R. Effect of nonlocal correlations on the electronic structure of LiFeAs. Phys. Rev. Lett. 123, 256401 (2019).
Tran, M.T. & Kuroki, K. Finitetemperature semimetalinsulator transition on the honeycomb lattice. Phys. Rev. B 79, 125125 (2009).
Arya, S., Sriluckshmy, P. V., Hassan, S. R. & Tremblay, A.M. S. Antiferromagnetism in the Hubbard model on the honeycomb lattice: a twoparticle selfconsistent study. Phys. Rev. B 92, 045111 (2015).
Raczkowski, M. et al. The Hubbard model on the honeycomb lattice: from static and dynamical meanfield theories to lattice quantum Monte Carlo simulations. Phys. Rev. B 101, 125103 (2020).
Kane, C. L. & Mele, E. J. Z_{2} Topological order and the Quantum Spin Hall effect. Phys. Rev. Lett. 95, 146802 (2005).
Ezawa, M. Valleypolarized metals and quantum anomalous hall effect in silicene. Phys. Rev. Lett. 109, 055502 (2012).
Zhang, X., Liu, Q., Luo, J.W., Freeman, A. J. & Zunger, A. Hidden spin polarization in inversionsymmetric bulk crystals. Nat. Phys. 10, 387–393 (2014).
Kochan, D., Irmer, S. & Fabian, J. Model spinorbit Hamiltonians for graphene systems. Phys. Rev. B 95, 165415 (2017).
Di Sante, D. et al. Towards topological quasifreestanding stanene via substrate engineering. Phys. Rev. B 99, 035145 (2019).
Klein, J. et al. Electricfield switchable secondharmonic generation in bilayer MoS_{2} by inversion symmetry breaking. Nano Lett. 17, 392–398 (2017).
Amaricci, A., Budich, J., Capone, M., Trauzettel, B. & Sangiovanni, G. Firstorder character and observable signatures of topological quantum phase transitions. Phys. Rev. Lett. 114, 185701 (2015).
Kotov, V. N., Uchoa, B., Pereira, V. M., Guinea, F. & Castro Neto, A. H. Electronelectron interactions in graphene: current status and perspectives. Rev. Mod. Phys. 84, 1067–1125 (2012).
Bistritzer, R. & MacDonald, A. H. Moiré bands in twisted doublelayer graphene. Proc. Natl Acad. Sci. USA 108, 12233 (2011).
Cao, Y. et al. Unconventional superconductivity in magicangle graphene superlattices. Nature 556, 43–50 (2018).
Cao, Y. et al. Correlated insulator behaviour at halffilling in magicangle graphene superlattices. Nature 556, 80–84 (2018).
Tsukahara, N. et al. Evolution of Kondo resonance from a single impurity molecule to the twodimensional lattice. Phys. Rev. Lett. 106, 187201 (2011).
Pizarro, J. M., Rösner, M., Thomale, R., Valentí, R. & Wehling, T. O. Internal screening and dielectric engineering in magicangle twisted bilayer graphene. Phys. Rev. B 100, 161102 (2019).
BlackSchaffer, A. M. & Doniach, S. Resonating valence bonds and meanfield d wave superconductivity in graphite. Phys. Rev. B 75, 134512 (2007).
Nandkishore, R., Levitov, L. S. & Chubukov, A. V. Chiral superconductivity from repulsive interactions in doped graphene. Nat. Phys. 8, 158–163 (2012).
Kiesel, M. L., Platt, C., Hanke, W., Abanin, D. A. & Thomale, R. Competing manybody instabilities and unconventional superconductivity in graphene. Phys. Rev. B 86, 020507 (2012).
BlackSchaffer, A. M. & Honerkamp, C. Chiral d wave superconductivity in doped graphene. J. Phys.: Condens. Matter 26, 423201 (2014).
Hohenberg, P. & Kohn, W. Inhomogeneous electron gas. Phys. Rev. 136, B864–B871 (1964).
Kohn, W. & Sham, L. J. Selfconsistent equations including exchange and correlation effects. Phys. Rev. 140, A1133–A1138 (1965).
Kresse, G. & Hafner, J. Normconserving and ultrasoft pseudopotentials for firstrow and transition elements. J. Phys.: Condens. Matter 6, 8245–8257 (1994).
Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmentedwave method. Phys. Rev. B 59, 1758–1775 (1999).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple [Phys. Rev. Lett. 77, 3865 (1996)]. Phys. Rev. Lett. 78, 1396–1396 (1997).
Grimme, S. Semiempirical GGAtype density functional constructed with a longrange dispersion correction. J. Comput. Chem. 27, 1787–1799 (2006).
Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFTD) for the 94 elements HPu. J. Chem. Phys. 132, 154104 (2010).
Kamil, E. et al. Electronic structure of single layer 1TNbSe_{2} : interplay of lattice distortions, nonlocal exchange, and Mott–Hubbard correlations. J. Phys.: Condens. Matter 30, 325601 (2018).
Gull, E. et al. Continuoustime Monte Carlo methods for quantum impurity models. Rev. Mod. Phys. 83, 349–404 (2011).
Wallerberger, M. et al. w2dynamics: Local one and twoparticle quantities from dynamical mean field theory. Comput. Phys. Commun. 235, 388–399 (2019).
Vilk, Y. M. & Tremblay, A.M. Nonperturbative manybody approach to the Hubbard model and singleparticle pseudogap. J. Phys. 7, 1309–1368 (1997).
Acknowledgements
We thank D. Di Sante, P. Eck, E. van Loon, M. Schüler, and C. Steinke for useful conversations. J.M.P. and T.W. acknowledge funding from DFGRTG 2247 (QM^{3}) and the European Graphene Flagship. S.A. and G.S. are supported by DFGSFB 1170 Tocotronics, and further acknowledges financial support from the DFG through the WürzburgDresden Cluster of Excellence on Complexity and Topology in Quantum Matter–ct.qmat (EXC 2147, projectid 390858490). T.M., K.Z., and R.V. acknowledge funding from the DFG through grant VA117/151. P.B. acknowledges financial support from the Italian Ministry for Research and Education through PRIN2017 project “Tuning and understanding Quantum phases in 2D materialsQuantum 2D” (ITMIUR Grant No. 2017Z8TS5B). This research was supported in part by the National Science Foundation under Grant No. NSF PHY1748958. We gratefully acknowledge the Gauss Center for Supercomputing e.V. (www.gausscenter.eu) for funding this project by providing computing time on the GCS Supercomputer SuperMUC at Leibniz Supercomputing Center (www.lrz.de). Computing time at HLRN (Berlin and Göttingen) is acknowledged.
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Contributions
J.M.P. performed the DFT and RPA calculations. S.A. calculated the topological invariant diagrams. S.A., K.Z., and T.M. performed the DMFT and TPSC calculations. P.B. derived the k ⋅ p model. R.V., G.S., and T.W. supervised the project. J.M.P., S.A., G.S., and T.W. analyzed the results. All authors contributed to write the paper.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Pizarro, J.M., Adler, S., Zantout, K. et al. Deconfinement of Mott localized electrons into topological and spin–orbitcoupled Dirac fermions. npj Quantum Mater. 5, 79 (2020). https://doi.org/10.1038/s41535020002773
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41535020002773
This article is cited by

Heavy fermions vs doped Mott physics in heterogeneous Tadichalcogenide bilayers
Nature Communications (2024)

Hall anomalies of the doped Mott insulator
npj Quantum Materials (2023)