Intervalley Band Crossing and Transition of Fractional Chern Insulators
in Floquet Twisted Bilayer MoTe · 2026
Read Intervalley Band Crossing and Transition of Fractional Chern Insulators by in Floquet Twisted Bilayer MoTe (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-03104.txt.
Introduction
Yuhao Shi and Zhao Liu study twisted bilayer molybdenum ditelluride, a layered material whose stacked sheets form patterns with exotic electronic phases, while it is driven by circularly polarized light. Using Floquet theory in the high-frequency limit, they build an effective steady Hamiltonian from a Dirac model that keeps both the valence and conduction bands of a single layer.
Keeping both bands matters: the light-driven coupling between them produces time-reversal breaking terms that vanish if the conduction bands are removed too early. As the driving grows stronger, Floquet bands cross and holes redistribute between the two valleys. With interactions included, the system switches between two fractional Chern insulator phases with different valley polarization behavior, at a total hole filling of five thirds.
Catalog record
| CREATOR | in Floquet Twisted Bilayer MoTe |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-03104
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Intervalley Band Crossing and Transition of Fractional Chern Insulators in Floquet Twisted Bilayer MoTe Abstract \setCJKmainfont [Path=./,UprightFont=*-Regular.otf]FandolSong We study the twisted MoTe 2 homobilayer coupled to periodic driving of a circularly polarized light (CPL). Using Floquet theory in the high-frequency limit, we start from the Dirac model including both the valence and conduction bands of monolayer MoTe 2 to derive an effective time-independent Floquet Hamiltonian. This Floquet Hamiltonian contains explicit time-reversal symmetry breaking terms that are absent if conduction bands are neglected from the beginning of the derivation. Based on the Floquet Hamiltonian, we find the increasing of CPL driving intensity can cause the crossing of Floquet bands between the two valleys. When interactions are included, we identify the redistribution of holes during the intervalley Floquet band crossing. Accordingly, the ground state of the Floquet Hamiltonian at total hole filling 5/3 evolves from the Laughlin-type FCI in one valley to that in the other valley. Introduction. Floquet systems have emerged as a versatile platform for engineering novel states of quantum matter. By subjecting materials to periodic driving, one can dynamically modify their band structures and effective interactions, enabling topological phases and dynamical transitions that are inaccessible in equilibrium \!{}^{\text{\cite[cite]{[\@@bibref{}{castro2022floquet,khemani2016phase,potter2016classification,yao2017discrete,potirniche2017floquet,lerose2019prethermal}{}{}]}}} . These effects are typically described by an effective static Floquet Hamiltonian, obtained through high-frequency expansions \!{}^{\text{\cite[cite]{[\@@bibref{}{eckardt2015highfrequency,bukov2015universal,mikami2016brillouinwigner}{}{}]}}} or more general non-perturbative approaches \!{}^{\text{\cite[cite]{[\@@bibref{}{rodriguez-vega2021lowfrequency,vajna2018replica,vogl2020effectivea,vogl2019analog,verdeny2013accurate,vogl2019flow}{}{}]}}} . Over the past decade, this framework has been widely applied to both weakly and strongly correlated systems, highlighting Floquet engineering as a promising route to control quantum phases \!{}^{\text{\cite[cite]{[\@@bibref{}{oka2019floquet,tsuji2008correlated,tsuji2009nonequilibrium}{}{}]}}} . This framework has been naturally extended to moiré materials, which provide a fertile setting for exploring the interplay between strong correlations and topology. For example, circularly polarized light (CPL) shining vertically across the twisted bilayer graphene (TBG) has been shown to cause nontrivial modifications of the band topology \!{}^{\text{\cite[cite]{[\@@bibref{}{oka2019floquet,vogl2020effectivea,katz2020optically,topp2019topological,li2020floquetengineered,ikeda2020highorder,vogl2020floquet,rodriguez-vega2020floquet}{}{}]}}} , which can potentially support fractional Chern insulators (FCIs) \!{}^{\text{\cite[cite]{[\@@bibref{}{sun2011nearly,tang2011hightemperature,neupert2011fractional,sheng2011fractional,regnault2011fractional,parameswaran2013fractional,bergholtz2013topological,liu2024recent}{}{}]}}} , the lattice analogs of the celebrated fractional quantum Hall effect, in Floquet systems \!{}^{\text{\cite[cite]{[\@@bibref{}{grushin2014floquet,anisimovas2015role,hu2023floquet}{}{}]}}} . Advances have been also reported in twisted MoTe 2 homobilayers (tMoTe 2 ) \!{}^{\text{\cite[cite]{[\@@bibref{}{dong2024floquet,qin2024lightenhanced}{}{}]}}} , which host flat bands over a wide range of twist angles \!{}^{\text{\cite[cite]{[\@@bibref{}{wu2018hubbard,zhan2020tunability,devakul2021magic}{}{}]}}} . Unlike in TBG, recent work focusing on the valence bands of tMoTe 2 found that the leading effect of the vertically applied high-frequency CPL was only a constant overall quasienergy shift of static bands \!{}^{\text{\cite[cite]{[\@@bibref{}{vogl2021floquet}{}{}]}}} . This shift is independent of valley and driving frequency \Omega . Nevertheless, topological transitions can be driven by a longitudinal light generated in a waveguide \!{}^{\text{\cite[cite]{[\@@bibref{}{vogl2021floquet}{}{}]}}} . The valence and conduction band edges near the K and K^{\prime} points of a monolayer MoTe 2 are well described by massive Dirac fermions \!{}^{\text{\cite[cite]{[\@@bibref{}{xiao2012coupled}{}{}]}}} . In previous works about tMoTe 2 , this massive Dirac structure is mostly neglected by perturbatively dropping the conduction band which is separated from the valence band by a large gap, resulting in a free-electron description (parabolic dispersion) of the valence band edge of a monolayer MoTe 2 . While this approximation works well for describing the band dispersion, it fails to capture the feature of time-reversal symmetry breaking in the original massive Dirac model \!{}^{\text{\cite[cite]{[\@@bibref{}{PhysRevResearch.4.L032024}{}{}]}}} , which may manifest itself under the driving of CPL. This observation motivates us to revisit the Floquet problem in tMoTe 2 using the full consideration of the massive Dirac structure in monolayers. It is also interesting to ask if there exist Floquet FCIs induced by CPL in tMoTe 2 when interactions are included. In this work, we investigate the off-resonant high-frequency Floquet engineering of tMoTe 2 by vertical CPL shining. Instead of adopting the free-electron model for each MoTe 2 monolayer from the beginning, we start from the massive Dirac model to derive the effective Floquet Hamiltonian, and integrate out the conduction band only at the end. In the obtained free-electron approximation of the Floquet Hamiltonian, we identify overall quasienergy shifts that depend not only on the driving frequency but also on the valley. This time-reversal symmetry breaking feature is absent if one derives the Floquet Hamiltonian completely within the free-electron framework. We then incorporate the screened Coulomb interaction in the system, and confirm that it does not cause additional interaction terms in the second-order ( 1/\Omega^{2} ) Floquet corrections that may affect the many-body physics of the driven system. By performing exact diagonalization of the many-body Floquet Hamiltonian at total hole filling \nu_{h}=5/3 , we track the evolution of the ground state when increasing CPL intensity drives the crossing of Floquet bands in opposite valleys, revealing the competition between distinct Floquet Laughlin-type FCI phases. Dirac model of static tMoTe 2 . We start our discussion from the single-particle Hamiltonian of the static tMoTe 2 . We assume that the top and bottom layers are rotated by \pm\theta/2 from the AA stacked configuration. The moiré Hamiltonian in the framework of massive Dirac model takes the form of \!{}^{\text{\cite[cite]{[\@@bibref{}{wu2019topological}{}{}]}}} \displaystyle\mathcal{H}_{\text{kin}}=\begin{pmatrix}h_{t,+}&T_{+}\\ T_{+}^{\dagger}&h_{b,+}\end{pmatrix}\bigoplus\begin{pmatrix}h_{t,-}&T_{-}\\ T_{-}^{\dagger}&h_{b,-}\end{pmatrix}, (1) where h_{\ell,\xi} is the massive Dirac Hamiltonian of layer \ell=t,b (top and bottom) at the valley \xi=\pm ( K and K^{\prime} ), and T_{\xi} is the interlayer hopping matrix. The monolayer term h_{\ell,\xi} includes both the conduction and valence bands, as given by \!{}^{\text{\cite[cite]{[\@@bibref{}{xiao2012coupled}{}{}]}}} \begin{aligned} h_{\ell,\xi}(\bm{k},\bm{r})&=\begin{pmatrix}\Delta_{g}+\Delta_{\ell,c}(\bm{r})&0\\ 0&\Delta_{\ell,v}(\bm{r})\end{pmatrix}\\ &+e^{+i\ell\xi\frac{\theta}{4}\sigma_{z}}[\hbar v_{F}\Delta\bm{k}_{\xi}^{\ell}\cdot(\xi\sigma_{x},\sigma_{y})]e^{-i\ell\xi\frac{\theta}{4}\sigma_{z}}\end{aligned}, (2) where c and v denote the conduction and valence bands, respectively, \Delta_{g}\approx$1.1\text{\,}\mathrm{e}\mathrm{V}$ is the average band gap, and v_{F} is the Fermi velocity. The relative momentum \Delta\bm{k}_{\xi}^{\ell}=(\bm{k}-\bm{\kappa}_{\ell,\xi}) , with \bm{\kappa}_{\ell,\xi} is the valley \xi of layer \ell . \Delta_{\ell,p} measures the variation of the extrema energy in band p=c,v as a function of position \bm{r} , and takes the form of \Delta_{\ell,p}(\bm{r})=2V_{p}\sum_{j=1}^{3}\cos(\bm{b}_{2j-1}\cdot\bm{r}-\ell\psi_{p}). (3) Here \bm{b}_{n}=\frac{4\pi}{\sqrt{3}a_{0}/\theta}R_{z}(\frac{n-1}{3}\pi)\hat{x} is the moire reciprocal lattice vectors, R_{z} generates a counterclockwise rotation around the z -axis, and a_{0}\approx$3.52\text{\,}\mathrm{\SIUnitSymbolAngstrom}$ is the lattice constant of monolayer MoTe 2 . The interlayer tunneling T_{\xi} in valley \xi is parametrized as \displaystyle T_{\xi}(\bm{r}) \displaystyle=\begin{pmatrix}w_{c}&w_{cv}\\ w_{vc}&w_{v}\end{pmatrix} (4) \displaystyle+\begin{pmatrix}w_{c}&w_{cv}e^{-i2\xi\pi/3}\\ w_{vc}e^{i2\xi\pi/3}&w_{v}\end{pmatrix}e^{i\xi\bm{b}_{2}\cdot\bm{r}} \displaystyle+\begin{pmatrix}w_{c}&w_{cv}e^{i2\xi\pi/3}\\ w_{vc}e^{-i2\xi\pi/3}&w_{v}\end{pmatrix}e^{i\xi\bm{b}_{3}\cdot\bm{r}}. Floquet engineering. In the following, we include the effect of the vertically applied CPL driving in the single-particle physics by applying the Peierls substitution \bm{k}\to\bm{k}+e\bm{A}/\hbar with the vector potential \bm{A}=A_{0}(\cos{\Omega t},-\sin{\Omega t}) in Eq. ( 2 ). The light field is represented by an in-plane electric field {\bm{E}}=-\frac{\partial{\bm{A}}}{\partial t} . Here A_{0} measures the driving strength and \Omega is the driving frequency. Then Eq. ( 1 ) is modified to a time-dependent single-particle Hamiltonian \mathcal{H}_{\text{kin}}(t) . Note that we keep the interlayer tunneling as in the static case, because it is dominated by hopping between atoms that are exactly on top of each other, thus mostly contributed by the z -component of the vector potential which is absent in our setup. According to the Floquet theory, the stroboscopic evolution of the system can be captured by a time-independent effective Floquet Hamiltonian H_{\text{eff}} (upon a unitary transformation from micromotion) \!{}^{\text{\cite[cite]{[\@@bibref{}{rahav2003effective,goldman2014periodically,bukov2015universal,eckardt2015highfrequency}{}{}]}}} . At high frequencies, H_{\text{eff}} can be represented by a series expansion of 1/\Omega : H_{\text{eff}}\approx H_{\text{eff}}^{(0)}+H_{\text{eff}}^{(1)}+H_{\text{eff}}^{(2)}, (5) where we keep to the second order. We use H_{m}=\frac{1}{T}\int_{0}^{T}\mathcal{H}_{\text{kin}}(t)e^{-im\Omega t}dt to denote the Fourier component of \mathcal{H}_{\text{kin}}(t) . In our setup, H_{m} is nonzero only if m=0,\pm 1 . Using the Magnus expansion \!{}^{\text{\cite[cite]{[\@@bibref{}{rahav2003effective,goldman2014periodically,bukov2015universal,eckardt2015highfrequency}{}{}]}}} , we obtain \displaystyle H_{\text{eff}}^{(0)} \displaystyle=H_{0}=\mathcal{H}_{\text{kin}}, (6a) \displaystyle H_{\text{eff}}^{(1)} \displaystyle=\frac{1}{\hbar\Omega}[H_{1},H_{-1}]=\frac{(\hbar v_{F}A_{0})^{2}}{\hbar\Omega}\sigma_{z}\otimes\mathds{1}_{2}^{\ell}\otimes\xi_{z}, (6b) \displaystyle H_{\text{eff}}^{(2)} \displaystyle=\frac{1}{2(\hbar\Omega)^{2}}[H_{1},[H_{0},H_{-1}]]+h.c. \displaystyle=\frac{(\hbar v_{F}A_{0})^{2}}{(\hbar\Omega)^{2}}\bigg[-\mathcal{H}_{\text{kin}}+\frac{\Delta_{\ell,v}}{2}(\mathds{1}_{2}^{p}+\sigma_{z})\otimes\mathds{1}_{2}^{\ell}\otimes\mathds{1}_{2}^{\xi} \displaystyle+\frac{\Delta_{g}+\Delta_{\ell,c}}{2}(\mathds{1}_{2}^{p}-\sigma_{z})\otimes\mathds{1}_{2}^{\ell}\otimes\mathds{1}_{2}^{\xi} \displaystyle+T^{\prime}\otimes\frac{\gamma_{x}+i\gamma_{y}}{2}\otimes\mathds{1}_{2}^{\xi}+(T^{\prime})^{\dagger}\otimes\frac{\gamma_{x}-i\gamma_{y}}{2}\otimes\mathds{1}_{2}^{\xi}\bigg], (6c) where \mathds{1}_{2}^{p} , \mathds{1}_{2}^{\ell}