Abstract
Monolayer transitionmetal dichalcogenides are novel materials which at low energies constitute a condensedmatter realization of massive relativistic fermions in two dimensions. Here, we show that this picture breaks for optical pumping—instead, the added complexity of a realistic materials description leads to a new mechanism to optically induce topologically protected chiral edge modes, facilitating optically switchable conduction channels that are insensitive to disorder. In contrast to graphene and previously discussed toy models, the underlying mechanism relies on the intrinsic threeband nature of transitionmetal dichalcogenide monolayers near the band edges. Photoinduced band inversions scale linearly in applied pump field and exhibit transitions from one to two chiral edge modes on sweeping from red to blue detuning. We develop an ab initio strategy to understand nonequilibrium Floquet–Bloch bands and topological transitions, and illustrate for WS_{2} that control of chiral edge modes can be dictated solely from symmetry principles and is not qualitatively sensitive to microscopic materials details.
Introduction
Manipulating materials properties far from equilibrium recently garnered significant attention, with experimental emphasis on transient melting, enhancement, or induction of electronic order^{1,2,3,4}. A more tantalizing aspect of the matter–light interaction regards the possibility to access dynamical steady states with distinct nonequilibrium phase transitions to affect electronic transport^{5,6,7,8,9}. Conceptually simple, irradiation with a sufficiently broad pump pulse dresses the original electronic bands by multiples of the photon frequency, with electric dipole coupling resulting in an effective steadystate band structure; Floquet–Bloch theory then corresponds precisely to the classical limit of strong pump fields that are indistinguishable before and after photon absorption or emission. One paradigmatic model of such Floquet–Bloch bands is graphene^{10,11,12}, where circularly polarized light can break timereversal symmetry to dynamically lift the Dirac point degeneracies. While Floquet–Bloch states were indeed observed recently via microwave pumping of Dirac cones on the surface of topological insulators^{6,7}, an extension to proper topological phase transitions is still well beyond experimental reach due to the tremendous required electric field: to open a sizeable gap (ref. 10) for abovebandwidth pump frequencies. Conversely, experimentally realizable gaps at the Dirac points at lower pump frequencies^{11,12} or in semiconductor quantum wells^{5} come at the price of resonant absorption, heating the sample, or worse, at the required pump strengths.
Viewed naively as semiconducting analogues of graphene, trigonalprismatic monolayers of MoS_{2}, MoSe_{2}, WS_{2} and WSe_{2} possess sizeable intrinsic band gaps due to broken inversion symmetry^{13} that can be expected to sustain intense subgap pump pulses while limiting absorption for sufficient detuning from the band edge. Prior theoretical studies established that the band edges at K and K′ are dominated by transitionmetal dorbitals, which split into three groups with irreducible representations (IRs) , and of the C_{3h} point group^{14,15}. Generalizing graphene, these valleys are wellcaptured in equilibrium by a degenerate Kramers’ pair of massive Dirac fermions, giving rise to valleyHall^{13} and spinHall^{16,17} effects.
Out of equilibrium, dynamical breaking of timereversal symmetry was demonstrated to lift the valley degeneracy for WS_{2} and WSe_{2} via offresonant optical pumping with circularly polarized light^{8,9}. In this case, the selection rules for a massive Dirac fermion entail that the handedness of pump polarization selectively addresses either the K or K′ valley, imparting an AC Stark shift on only one of the valleys. Analogously, the photoexcitation can selectively populate valleys, enabling spin and valley currents using circular or linear polarization^{18,19,20}.
Even more tantalizingly, it was predicted that effective TMDC toy models—graphene with a gap—admit, in theory, an optically induced quantumHall effect with a single chiral mode localized at the sample edge. A highfrequency pump Ω→∞ well above the bandwidth can in principle close and invert the equilibrium band gap at a single valley^{21}; however, this requires a tremendous pump intensity. Alternatively, it was proposed that a resonant pump beam can hybridize the massive Dirac fermion valence and conduction bands (CBs), and thereby generate a single chiral edge mode^{8} at lower pump strength.
In this study, we instead show that such a simple description fails to hold for optical pumping; here, the added complexity of a more realistic model of TMDC monolayers opens up a novel avenue to engineer a Floquet topological insulator in a realistic experimental setting. We argue that correctly addressing optical excitations necessitates a minimally threeband description near the band edges that leads to a frequenytunable mechanism to photoinduce one or two chiral edge modes. To understand the nature of concurrent Floquet band inversions at both K and K′, we develop an ab initio Floquet k.p formalism that directly connects equilibrium densityfunctional theory (DFT) calculations with nonequilibrium Floquet theory. We illustrate these predictions for the example of a WS_{2} ribbon, and present nonequilibrium ribbon spectra as well as an ab initio characterization of the photoinduced valley topological band inversions. We find that control of chiral edge modes is determined solely by crystal symmetry and is insensitive to materials microscopics such as multiphoton processes or local interorbital dipole transitions that cannot be captured straightforwardly in a tightbinding model.
Results
Threeband nature of the light–matter interaction
Central to this study, dipole transitions to higherlying bands, as determined directly from ab initio calculations, and underlying symmetry considerations are crucial for a determination of photoinduced topological band inversions in TMDCs. To understand the breakdown of a description as graphene with a gap, consider a monolayer ribbon uniformly irradiated by circularly polarized light, with collinear sample and polarization planes (Fig. 1a). In graphene, the lowenergy bands near the Fermi level are separated from higherenergy CBs by >10 eV at K,K′ (ref. 22), hence optical frequencies can be treated safely within the canonical lowenergy Dirac model of π orbitals. In contrast, the band structures of prototypical TMDC monolayers possess a E′ band only ∼2 eV above the CB (Fig. 1b), and E′′ bands in the same vicinity^{23,24,25,26}. Circularly polarized light at closetoband gap pump frequencies therefore couples the A′ CB to both the E′ valence band (VB) and the higherenergy CB (XB), while leaving the E′′ bands decoupled in the absence of multipole transitions.
Consider first the case of slightly reddetuned pumping below the band edge (Fig. 2a). Here, a ring of states from the higherenergy XB is brought into resonance with the CB, while simultaneously avoiding resonant coupling between the VB and CB. Central to experimental feasability, this regime can be expected to substantially limit absorption and heating. At the band edge, C_{3h} dipole selection rules (Fig. 2e) dictate that absorption of a photon couples transitions . At K′, the bare CB m=0; A′, CB〉 (Floquet index m) couples to the XB dressed by a single emitted photon as well as the VB dressed by a single absorbed photon . Both transitions, although offresonant, are energetically favourable, leading to a significant Stark shift at K′ (Fig. 2b). Conversely, at K the IRs of VB and XB are reversed. Here, the CB couples to the VB dressed by a single emitted photon m=−1; E′, VB〉 as well as the XB dressed by a single absorbed photon . Both transitions are energetically unfavourable, leading to a negligible shift of the band edge. Slightly away from K and K′, electric dipole coupling lifts the ring of degeneracy between the CB and XB and opens a photoinduced hybridization gap at both valleys (Fig. 2b), which scales linearly with weak pump fields. Crucially, the resulting Floquet–Bloch bands exhibit a topological ‘band inversion’ with the orbital character flipped close to the valley minimum, at both valleys (Fig. 2b).
Topology of photoinduced band inversions
To discern whether the band inversions can be nontrivial, we devise effective Floquet twoband models of the hybridization gaps. We start from the generic description of a semiconductor in the absence of spin–orbit coupling and excitonic effects, where V(r) is the crystal potential. In equilibrium, starting from an ab initio Bloch eigenbasis at a single highsymmetry point, the dispersion and orbital content follow from canonical k.p theory by perturbing in momentum deviation k under replacement (ref. 27). In the presence of a timeperiodic field A(t)=A[cos(Ωt), sin(ωt)]^{T}, a straightforward generalization uses a Floquet eigenbasis of the generic nonequilibrium problem . This basis can be obtained from DFT calculations via knowledge of the equilibrium band energies and dipole transition matrix elements. Note that this Floquet k.p theory is nonperturbative in the applied pump field and naturally accounts for multiphoton coupling to higherenergy CBs, XBs and deeper VBs, as well as local interorbital dipole transitions. In the following, denotes the dimensionless field strength with lattice constant a_{0}=3.2 Å and electric field ; A=0.1 corresponds to for optical pump fields with Ω=1.5 eV. An effective lowenergy description of the photoinduced gaps can now be devised in Floquet basis in analogy to the equilibrium problem, by considering a perturbation in crystal momentum and downfolding onto effective twoband Floquet models using canonical Löwdin perturbation theory.
Central to the robustness of this proposal, the form of these effective models is determined solely from symmetry and is universal to trigonalprismatic TMDC monolayers. To see this, first consider K. Here, the Floquet eigenbasis (Fig. 2b) Ψ_{1}〉 (Ψ_{2}〉) admixes m=0; A′, CB〉 with m=−1; E′, VB〉, ( with m=−2; A′, CB〉, m=0; E′, VB〉), linear in field A. Constrained by crystal symmetry, we find that the effective Floquet physics at K is generically determined by a pd Dirac model (see the ‘Methods’ section):
An additional purely dispersive term breaks particlehole symmetry. Crucially, the offdiagonal couplings v_{p},v_{d} are linear functions in field strength, suggesting a sizable photoinduced gap already at weak fields. While the parameters depend on the details of the Bloch states near the Dirac points, overall topological considerations can be gleaned simply from equation (1). In the absence of v_{d}, equation (1) describes a conventional massive Dirac fermion, with M (B) the Dirac (inverse band) mass, and v_{p} the Dirac velocity. The orbital character exhibits a pwave winding around K and mirrors the quantum anomalous Hall effect in Hg_{y}Mn_{1–}_{y}Te quantum wells^{28}. Switching on v_{d} imparts a trigonal distortion by reducing the continuous rotational symmetry around K to C_{3}, and introduces instead a ‘dwave’ winding in the limit .
At K′ interchanged IRs entail a strongly admixed Floquet eigenbasis as well as a significant Stark shift. However selection rules forbid a coupling between Ψ_{1}〉 ,Ψ_{2}〉 linear in k—instead, one finds that the two bands couple to quadratic order in , via intermediate states m=0; E′, XB〉 or m=−1; A′, CB〉. The effective Hamiltonian for K′ in this case generically reads
with v′ a rotationally symmetric band mixing term.
At K (K′), the band ordering is inverted when M/B>0 (M′/B′>0). If the orbital character of Floquet–Bloch bands in the remaining Brillouin zone is sufficiently benign, we can draw conclusions on the global topology by understanding separately the band inversions at K and K′. Rewriting equations (1) and (2) in terms of Pauli matrices , the local Berry curvature follows from the winding with . One can see by inspection that the absence of in equation (2) enforces ; therefore, the photoinduced band inversion around K′ is necessarily trivial. Conversely, the band inversion at K is topological and triggers a change in the Chern number , which captures the contribution to topology arising from the band inversion in the vicinity of K (see Supplementary Note 3, Supplementary Fig. 1).
Consider first the limit of a massive Dirac model with v_{d}=0. In this case, the Chern number changes from for B/M < 0 to for B/M>0, inducing a single chiral edge mode, spanning the photoinduced hybridization gap, and localized at the boundary of a uniformly illuminated sample. Switching on v_{d} introduces a trigonal distortion of the Floquet–Bloch bands around K up to a critical strength , at which the Floquet–Bloch bands close the gap at three points away from K, related by C_{3}. Correspondingly, this topological transition changes by 3, to , entailing not one but two chiral edge modes at the sample boundary.
Red versus blue detuning
To get an understanding of the mechanism that might trigger this transition, note that while the relevant Floquet basis is predominantly built from only the m=0 CB and m=−1 XB for a reddetuned pump, the pwave coupling v_{p} between the two is necessarily mediated via the VB (or other bands of equivalent IR), highlighting the necessity of a minimal threeband description (see Supplementary Note 2). We stress that such an effective Diraclike contribution is generically impossible to obtain starting from a twoband equilibrium description as a massive Dirac fermion. In contrast, the dwave term results from direct coupling between CB and XB. Strong optical absorption in TMDC monolayers indicates a large dipole transition matrix element between VB and CB, suggesting that a sufficiently reddetuned pump will generically reach the C=1 phase. Note that the asymmetry between K,K′ stems from the choice of polarization, and reverses for opposite handedness of the circularly polarized pump beam.
In contrast, consider the opposite regime of a sufficiently bluedetuned pump (Fig. 2c). Here, a ring of VB states is brought into resonance with the CB near K,K′ while pushing the photondressed XB into the equilibrium band gap. Electric dipole coupling again opens photoinduced hybridization gaps, both at the bottom of the CB and top of the VB (Fig. 2d), and the symmetry analysis mirrors the discussion of the reddetuned case above, leading to equivalent effective Hamiltonians at K,K′ (1), (2). However, linear in k coupling between the m=0 CB and the m=+1 VB is now necessarily mediated via the XB (or other bands separated further in energy), whereas the dwave term v_{d} follows directly from dipole coupling between VB and CB, dominating over v_{p}. One can thus generically expect a frequencytunable Floquet Chern insulator in monolayer TMDCs.
Ab initio Floquet analysis of WS_{2} monolayers
To illustrate these predictions, consider WS_{2} as a prototypical TMDC monolayer. First, we perform ab initio DFT calculations to derive an effective minimal tightbinding model of three Wannier orbitals localized on the W transition metal (Fig. 1b,c). The resulting Floquet spectrum on a ribbon is depicted in Fig. 3, calculated from the periodaveraged singleparticle spectral function (see the ‘Methods’ section). In equilibrium (Fig. 3a), WS_{2} already hosts a pair of trivial edge states in analogy to zigzag edges of graphene, with right (left) propagating modes at the K (K′) point that span the band gap. A weak, reddetuned pump field opens a hybridization gap at the bottom of the CB, spanned by a single chiral mode at K (Fig. 3b), localized at the sample edge. The photoinduced gap scales linearly with weak A, but closes and reopens at a critical pump strength, transitioning again to a trivial phase without chiral modes. Conversely, for a bluedetuned pump (Fig. 3c) a second chiral edge mode appears, spanning the hybridization gaps both at the bottom of the CB and the top of the VB.
The appearances of edge modes in ribbon spectra are in excellent agreement with effective model parameters (equations (1) and (2)) derived from the Wannier tightbinding description. Figure 4b depicts M/B and the ratio of p−/d−wave couplings that determine the Chern number for the band inversion at K (Fig. 4c), in perfect correspondence with a rigorous calculation of the 2+1D Floquet winding number^{29} of the driven Wannier tightbinding model (Fig. 4d, see Supplementary Note 3), with the circularly polarized pump entering via Peierls substitution. For weak fields, deep within both the red and bluedetuned regimes, the sign of the Dirac M and band B mass are equal in the topologically nontrivial phase. Here, C≠0, and the Chern number follows from trigonal distortion and changes from C=1 for red detuning to C=2 for blue detuning. Increasing A instead closes and reopens the Floquet gap at K, flips the sign of M and uninverts the bands to reach a trivial phase with C=0. We note that this picture breaks down for intense pump fields with energy scales on the order of the equilibrium band gap; here, additional topological phase transitions can arise (Supplementary Fig. 2, Note 4).
Having checked the validity of Floquet k.p theory in the tightbinding model, we now turn to the full ab initio problem. To quantify the effects of multiphoton resonances, as well as local interorbital dipole transitions not captured in a tightbinding model, we consider an ab initio 185band description of band energies and dipole transition matrix elements at K and K′ and calculate the model parameters of equations (1) and (2), taking into account up to fourphoton processes. The bands closest to the equilibrium gap are depicted in Fig. 5a. The resulting k.p classification is depicted in Fig. 5b. Crucially, while the resonance lines distort due to effects not accounted for in the tightbinding model, the frequencydependent switch from C=1 to C=2 as well as the reclosing of the hybridization gap and transition back to a trivial regime with increasing pump strength remain qualitatively similar. This suggests that the mechanism of photoinduced chiral edge modes described in this work is largely robust at weak fields to the microscopic details of the material.
Discussion
A key challenge for a condensedmatter realization of Floquet topological insulators regards driving the system strongly to induce the required changes to the equilibrium band structure while simulteanously mitigating the inevitable absorption and heating in such a scheme. However, a common thread for pioneering works on graphene^{11,12}, semiconductor quantum wells^{5} and topological insulator surface states^{6,7} is that these require either highfrequency pumping at tremendous pump strengths or resonant pumping, thereby injecting substantial energy into the system and heating the sample. Recent work has tried to tackle these problems by discerning whether special couplings to bosonic or fermionic heat sinks^{30,31,32} can dissipate enough energy and nevertheless stabilize a Floquet steadystate ensemble. Analogously, the bluedetuned regime entails a ring of resonance between CB and VB, leading to enhanced absorption. Conversely, and central to the experimental feasability of this proposal, reddetuned pumping of monolayer TMDCs circumvents these issues by entirely avoiding resonant coupling between VB and CB while nevertheless inducing nontrivial band topology by virtue of the minimally threeband nature of electron–photon coupling. Naively, this can be understood by noting that, for weak recombination rates, the rate of carrier photoexcitation scales as g^{2}A^{2}/δ^{2} whereas the photoinduced hybridization gap scales linearly in pump strength A (see the ‘Methods’ section), where δ is the laser detuning from resonance and g an effective dipole matrix element. Nevertheless, this residual photoexcited population will lead to broadening of the Floquet–Bloch bands due to recombination and electron–phonon scattering and set a lower bound on observable bulk hybridization gaps, while at the same time serving as a necessary ingredient to reach a steadystate population. Extending the topological characterization to open quantum systems in the limit of strong dissipation remains an interesting topic for future study^{33}.
Smokinggun evidence for the presence of chiral edge modes necessitates either measurement of the sample edge via nano angleresolved photoemission (nanoARPES) and local admittance probes such as microwave impedance spectroscopy, or direct transport measurements, scaling of conductance with sample length, where care must be taken regarding coupling between leads and the chiral Floquet edge mode^{34,35}. Conversely, ARPES stands as an immediate tool to verify the predicted photoinduced hybridization gaps in the 2D bulk, as a function of pump strength and when sweeping from red to blue detuning. For red detuning, the CB can be enhanced in twophoton photoemission, in analogy to probing the unoccupied higherenergy topological surface states of Bi_{2}Se_{3} (ref. 36). An interesting followup question concerns potential matrix element dependencies of photoemission from the W dorbitals, to directly observe and characterize the band inversions at K,K′ via bulk measurement.
The guiding theme of this work has been to build a bridge between the rapidly developing field of monolayer transitionmetal dichalcogenides and topological phase transitions out of equilibrium, to provide a route towards achieving the latter in an experimentally attainable setting. We have shown that the threeband nature of the valleys in prototypical WS_{2} leads to a new mechanism to ‘switch’ on or off, one or two chiral edges with near band gap optical irradiation. The resulting photoinduced gap in the singleparticle spectrum scales linearly with pump strength, suggesting substantial energy scales already at low fields, while simultaneously ensuring minimal heating with sufficient detuning from the band edge. Our theoretical analysis of the outofequilibrium valley band inversions connects directly with equilibrium ab initio calculations, whereas the ensuing topology of Floquet–Bloch bands relies purely on generic symmetry arguments, suggesting that the predictions are robust to microscopic detail and should be observable in a range of monolayer TMDC materials. Finally, our firstprinciples and theoretical analysis provides a promising strategy to predict and design topological states out of equilibrium in other semiconductor materials.
Methods
Ab initio calculations
Ab initio calculations were performed in the framework of the Perdew–Burke–Ernzerhof type generalized gradient approximation of DFT using the fullpotential linearized augmented plane wave method implemented in Wien2k (ref. 37). We consider a single monolayer of WS_{2} with a 30Å vacuum space perpendicular to the layer along the zdirection. The inplane lattice constant and the S position have been relaxed by optimization of the total energy and total force, respectively. For electronic structure calculations, we utilized a 15 × 15 × 1 kspace grid. Momentum matrix element calculations were performed using the OPTIC package implemented in Wien2k, with a 60 × 60 × 1 kspace grid. Maximally localized Wannier functions for the five W 5d orbitals were obtained using wien2wannier (ref. 38) and Wannier90 (ref. 39) with initial projections set to the spherical harmonics Y_{2m} (m=−2, −1, 0, 1, 2). Due to the symmetry of the hexagonal lattice, the calculated Hamiltonian in the new Wannier basis naturally decouples into the two standard subspaces {, } and {}.
Floquet theory of the singleparticle spectrum on a ribbon
Floquet theory captures the effective steady states that arise from a timedependent (quasi)periodic modulation. Consider a Hamiltonian with a periodic time dependence with frequency Ω. Then, solutions of the timedependent Schrödinger equation for can be written as , where is the Floquet quasienergy, and u_{m} are Fourier coefficients of the timeperiodic part of the wave function. Substitution of Φ(t) into Schrödinger’s equation recasts the timedependent problem as an effective timeindependent Floquet problem: the Floquet states can be found by finding eigenstates of the Floquet Hamiltonian
where, are the Fourier expansion coefficients of . If the original Hamiltonian has a static eigenbasis α〉, then the eigenstates of can be written as , with the original timedependent eigenstates of becoming . The next step is to connect back to observables of the original fermion operators. In the main text, we consider the spectral function
where, is the retarded Green’s function. Rewriting the fermion operators Ψ_{α}(x,t) in Floquet basis, one finally arrives at the Floquet spectral function
where Γ is a phenomenological broadening of the spectrum.
Floquet k.p theory and effective Hamiltonians at K,K′
Consider a generic timedependent starting point for K (and equivalently for K′)
where, k is the deviation in crystal momentum from K or K′, with a respective shift to the K or K′ point absorbed in . We consider circularly polarized light with A(t)=A[cos(Ωt), sin(Ωt)]^{T}. Now decompose into equilibrium [] and nonequilibrium [] constituents that determine the eigenbasis at K,K′, as well as a perturbation in k []:
where, . In equilibrium, the (timeindependent) eigenbasis α,n〉 of can be determined from ab initio calculations and transforms according to C_{3h}, with n,α indexing the nth band with IR α. In the absence of the pump field A(t)=0, conventional k.p theory proceeds by considering the deviation in Bloch momentum as a perturbation, described by .
In Floquet k.p theory, one instead starts from the exact Floquet eigenbasis of at K and K′. Consider a single spin manifold, and for simplicity denote bands by the C_{3h} single group IRs (see Supplementary Table 2, Supplementary Note 1 for equivalent double group identifications and selection rules in the full SOC problem). The selection rules (Fig. 2e, Supplementary Table 1) then entail that involve transitions
Here, m (n) are Floquet (band) indices, and are the momentum matrix elements for allowed dipole transitions (Fig. 2e), obtained from ab initio calculations. Using a sufficiently large number of ab initiodetermined Bloch states at K,K′, their dipole matrix elements and Floquet side bands, the Floquet eigenbasis at K,K′ can formally be determined exactly as functions of A,Ω.
The effective twoband Hamiltonians (equations (1) and (2)) described in the main text now follow via choosing two Floquet eigenstates for K and K′ each, that are adiabatically connected to the A=0 CB with m=0 as well as the m=−1XB (m=+1VB) for red (blue) detuning, and using Löwdin perturbation theory to downfold onto this twostate Floquet eigenbasis (a detailed derivation can be found in Supplementary Note 2).
Crucially, to distinguish K and K′, note that their irreducible representations for VB, XB interchange. A simple way to arrive at the effective Hamiltonians (1) and (2) follows from observing for (10) that the Floquet eigenbasis at K,K′ necessarily decomposes again into three IRs of the joint electron–photon problem, which are subsequently coupled by the Bloch momentum perturbation . In this picture, at K, the IRs of the two Floquet basis states of (1) differ; hence offdiagonal coupling enters already at linear order . Note that this coupling can necessarily only arise in the present minimally threeband description (see Supplementary Material). Conversely, at K′ the IRs of the basis of (2) are the same; offdiagonal coupling therefore necessarily enters only to quadratic order and higher, leading to a trivial band inversion.
The inclusion of full spin–orbit coupling does not qualitatively alter these conclusions. First, spinflip terms weakly admix E′′ bands of opposite spin^{40,41}; however, the full crystal double group again decomposes into two spinorbital manifold with equivalent selection rules and effective physics (Supplementary Note 1). Second, the valley Zeeman shift simply leads to a shift of the relevant resonance energies. Similarly, while monolayer TMDCs have been shown to give rise to large excitonic binding energies^{42,43}, in the context of our work their role is confined to shifting the relevant resonance energies, given appropriate tuning of the pump frequency.
Data availability
The data that support the findings of this study are available from the authors on request.
Additional information
How to cite this article: Claassen, M. et al. Alloptical materials design of chiral edge modes in transitionmetal dichalcogenides. Nat. Commun. 7, 13074 doi: 10.1038/ncomms13074 (2016).
References
 1
Schmitt, F. et al. Transient electronic structure and melting of a charge density wave in TbTe3 . Science 321, 1649–1652 (2008).
 2
Fausti, D. et al. Lightinduced superconductivity in a stripeordered cuprate. Science 331, 189–191 (2011).
 3
Kim, K. W. et al. Ultrafast transient generation of spindensitywave order in the normal state of BaFe2As2 driven by coherent lattice vibrations. Nat. Mater. 11, 497–501 (2012).
 4
Mankowsky, R. et al. Nonlinear lattice dynamics as a basis for enhanced superconductivity in YBa2Cu3O6.5 . Nature 516, 71–73 (2014).
 5
Lindner, N. H., Refael, G. & Galitski, V. Floquet topological insulator in semiconductor quantum wells. Nat. Phys. 7, 490–495 (2011).
 6
Wang, Y. H., Steinberg, H., JarilloHerrero, P. & Gedik, N. Observation of FloquetBloch states on the surface of a topological insulator. Science 342, 453–457 (2013).
 7
Mahmood, F. et al. Selective scattering between FloquetBloch and Volkov states in a topological insulator. Nat. Phys. 12, 306–310 (2016).
 8
Sie, E. J. et al. Valleyselective optical stark effect in monolayer WS2 . Nat. Mater. 14, 290–294 (2014).
 9
Kim, J. et al. Ultrafast generation of pseudomagnetic field for valley excitons in WSE2 monolayers. Science 346, 1205–1208 (2014).
 10
Oka, T. & Aoki, H. Photovoltaic Hall effect in graphene. Phys. Rev. B 79, 081406 (2009).
 11
Kitagawa, T., Oka, T., Brataas, A., Fu, L. & Demler, E. Transport properties of nonequilibrium systems under the application of light: photoinduced quantum Hall insulators without landau levels. Phys. Rev. B 84, 235108 (2011).
 12
Sentef, M. A. et al. Theory of Floquet band formation and local pseudospin textures in pumpprobe photoemission of graphene. Nat. Commun. 6, 7047 (2015).
 13
Xiao, D., Liu, G.B., Feng, W., Xu, X. & Yao, W. Coupled spin and valley physics in monolayers of MoS2 and other groupvi dichalcogenides. Phys. Rev. Lett. 108, 196802 (2012).
 14
Bromley, R. A., Murray, R. B. & Yoffe, A. D. The band structures of some transition metal dichalcogenides. III. group VIA: trigonal prism materials. J. Phys. C: Solid State Phys. 5, 759 (1972).
 15
Mattheiss, L. F. Band structures of transitionmetaldichalcogenide layer compounds. Phys. Rev. B 8, 3719–3740 (1973).
 16
Feng, W. et al. Intrinsic spin Hall effect in monolayers of groupVI dichalcogenides: a firstprinciples study. Phys. Rev. B 86, 165108 (2012).
 17
Shan, W.Y., Lu, H.Z. & Xiao, D. Spin Hall effect in spinvalley coupled monolayers of transition metal dichalcogenides. Phys. Rev. B 88, 125301 (2013).
 18
Yu, H., Wu, Y., Liu, G.B., Xu, X. & Yao, W. Nonlinear valley and spin currents from fermi pocket anisotropy in 2d crystals. Phys. Rev. Lett. 113, 156603 (2014).
 19
Shan, W.Y., Zhou, J. & Xiao, D. Optical generation and detection of pure valley current in monolayer transitionmetal dichalcogenides. Phys. Rev. B 91, 035402 (2015).
 20
Muniz, R. A. & Sipe, J. E. Alloptical injection of charge, spin, and valley currents in monolayer transitionmetal dichalcogenides. Phys. Rev. B 91, 085404 (2015).
 21
Tahir, M., Manchon, A. & Schwingenschlögl, U. Photoinduced quantum spin and valley Hall effects, and orbital magnetization in monolayer MoS2 . Phys. Rev. B 90, 125438 (2014).
 22
Kogan, E., Nazarov, V. U., Silkin, V. M. & Kaveh, M. Energy bands in graphene: comparison between the tightbinding model and ab initio calculations. Phys. Rev. B 89, 165430 (2014).
 23
Zahid, F., Liu, L., Zhu, Y., Wang, J. & Guo, H. A generic tightbinding model for monolayer, bilayer and bulk MoS2 . AIP Adv. 3, 052111 (2013).
 24
Cappelluti, E., Roldán, R., SilvaGuillén, J. A., Ordejón, P. & Guinea, F. Tightbinding model and directgap/indirectgap transition in singlelayer and multilayer MoS2 . Phys. Rev. B 88, 075409 (2013).
 25
Rostami, H., Moghaddam, A. G. & Asgari, R. Effective lattice hamiltonian for monolayer MoS2: tailoring electronic structure with perpendicular electric and magnetic fields. Phys. Rev. B 88, 085440 (2013).
 26
Liu, G.B., Shan, W.Y., Yao, Y., Yao, W. & Xiao, D. Threeband tightbinding model for monolayers of groupvib transition metal dichalcogenides. Phys. Rev. B 88, 085433 (2013).
 27
Winkler, R., Papadakis, S. J., De Poortere, E. P. & Shayegan, M. SpinOrbit Coupling in TwoDimensional Electron and Hole Systems 41, (Springer (2003).
 28
Liu, C.X., Qi, X.L., Dai, X., Fang, Z. & Zhang, S.C. Quantum anomalous Hall effect in Hg1− yMnyTe quantum wells. Phys. Rev. Lett. 101, 146802 (2008).
 29
Rudner, M. S., Lindner, N. H., Berg, E. & Levin, M. Anomalous edge states and the bulkedge correspondence for periodically driven twodimensional systems. Phys. Rev. X 3, 031005 (2013).
 30
Dehghani, H., Oka, T. & Mitra, A. Dissipative Floquet topological systems. Phys. Rev. B 90, 195429 (2014).
 31
Iadecola, T., Neupert, T. & Chamon, C. Occupation of topological Floquet bands in open systems. Phys. Rev. B 91, 235133 (2015).
 32
Seetharam, K. I., Bardyn, C.E., Lindner, N.H., Rudner, M.S. & Refael, G. Controlled population of FloquetBloch states via coupling to Bose and Fermi baths. Phys. Rev. X 5, 041050 (2015).
 33
Budich, J. C. & Diehl, S. Topology of density matrices. Phys. Rev. B 91, 165140 (2015).
 34
Kundu, A. & Seradjeh, B. Transport signatures of floquet majorana fermions in driven topological superconductors. Phys. Rev. Lett. 111, 136402 (2013).
 35
Farrell, A. & PeregBarnea, T. Photoninhibited topological transport in quantum well heterostructures. Phys. Rev. Lett. 115, 106403 (2015).
 36
Sobota, J. A. et al. Direct optical coupling to an unoccupied dirac surface state in the topological insulator Bi2Se3 . Phys. Rev. Lett. 111, 136802 (2013).
 37
Blaha, P., Schwarz, K., Madsen, G. K. H., Kvasnicka, D. & Luitz, J. Wien2K: an augmented plane wave and local orbitals program for calculating crystal properties Technische Universität Wien (2001).
 38
Kuneš, J. et al. Wien2wannier: from linearized augmented plane waves to maximally localized Wannier functions. Comp. Phys. Commun. 181, 1888 (2010).
 39
Mostofi, A. A. et al. wannier90: a Tool for obtaining maximallylocalised Wannier functions. Comp. Phys. Commun. 178, 685 (2008).
 40
Kormányos, A. et al. Monolayer MoS2: trigonal warping, the Γ valley, and spinorbit coupling effects. Phys. Rev. B 88, 045416 (2013).
 41
Gibertini, M., Pellegrino, F. M. D., Marzari, N. & Polini, M. Spinresolved optical conductivity of twodimensional groupVIB transitionmetal dichalcogenides. Phys. Rev. B 90, 245411 (2014).
 42
Chernikov, A. et al. Exciton binding energy and nonhydrogenic Rydberg series in monolayer WS2 . Phys. Rev. Lett. 113, 076802 (2014).
 43
Zhu, B., Chen, X. & Cui, X. Exciton binding energy of monolayer WS2 . Sci. Rep. 5, 9218 (2015).
Acknowledgements
We gladly acknowledge helpful discussions with Alexander Kemper, Ruixing Zhang, Patrick Kirchmann, Tony Heinz, and ZhiXun Shen. This work was supported by the U.S. Department of Energy, Office of Basic Energy Science, Division of Materials Science and Engineering under Contract No. DEAC0276SF00515. Computational resources were provided by the National Energy Research Scientific Computing Center supported by the Department of Energy, Office of Science, under Contract No. DE AC0205CH11231.
Author information
Affiliations
Contributions
M.C. developed the theoretical formalism and Floquet–Bloch simulations. C.J. and B.M. performed the densityfunctional theory simulations. The manuscript was written by M.C. with input from all authors. T.P.D. supervised the project.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Supplementary information
Supplementary Information
Supplementary Figures 12, Supplementary Tables 12, Supplementary Notes 14 and Supplementary References (PDF 1311 kb)
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Claassen, M., Jia, C., Moritz, B. et al. Alloptical materials design of chiral edge modes in transitionmetal dichalcogenides. Nat Commun 7, 13074 (2016). https://doi.org/10.1038/ncomms13074
Received:
Accepted:
Published:
Further reading

Twistangle engineering of excitonic quantum interference and optical nonlinearities in stacked 2D semiconductors
Nature Communications (2021)

Probing lightdriven quantum materials with ultrafast resonant inelastic Xray scattering
Communications Physics (2020)

Observing photoinduced chiral edge states of graphene nanoribbons in pumpprobe spectroscopies
npj Quantum Materials (2020)

Femtosecond exciton dynamics in WSe2 optical waveguides
Nature Communications (2020)

Electric Field and Light Field Modulated Josephson Effect in SiliceneBased SNS Josephson Junction: Andreev Reflection and Free Energy
Silicon (2020)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.