Spontaneous fractional Chern insulators in transition metal dichalcogenides Moiré

Heqiu Li,1 Umesh Kumar,2 Kai Sun,1 and Shi-Zeng Lin3 1Department of Physics, University of Michigan, Ann Arbor, Michigan 48109, USA 2Theoretical Division, T-4, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA 3Theoretical Division, T-4 and CNLS, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA (Dated: September 14, 2021) Moiré realized in two-dimensional heterostructures offers an exciting platform to access strongly- correlated electronic states. In this work, we study transition metal dichalcogenides (TMD) Moiré superlattices with time-reversal symmetry and nontrivial spin/valley-Chern numbers. Utilizing realistic material parameters and the method of exact diagonalization, we find that at certain twisting angle and fractional filling, gapped fractional topological states, i.e., fractional Chern insulators, are naturally stabilized by simply introducing the Coulomb repulsion. In contrast to fractional quantum Hall systems, where the time-reversal symmetry has to be broken explicitly, these fractional states break the time-reversal symmetry spontaneously. We show that the Chern number contrasting in the opposite valleys imposes a strong constraint on the nature of fractional Chern insulator and the associated low energy excitations.

Introduction.— When two layers of two-dimensional mate- In this work, we show that by simply increasing the Coulomb rials are placed atop of each other with slight misalignment, interaction strength in such TMD Moiré superlattices, the sys- it creates a superlattice with periodicity much larger than the tem undergoes a quantum phase transition that spontaneously atomic lattice parameter. Because of the large lattice period- breaks the time-reversal symmetry by polarizing electrons into icity, one can fill or empty the entire band by electrode gating. one of the two valleys. Further increase of Coulomb interaction This Moiré superlattice provides a tunable platform to control will trigger a second quantum phase transition, and stabilize the electronic band structure [1,2], and therefore enables ac- a FCI at a fractional filling. For excitations, our numerical cess to a plethora of interesting quantum states. Because the studies observe both (intravalley) fractional excitations from band width in these systems can be tuned to be extremely nar- the fractional and (intervalley) valley-wave row [2], these Moiré superlattices open up a new pathway to excitations from the spontaneous symmetry breaking. We stabilize various strongly-correlated phases such as supercon- argue that the symmetry breaking state and low-energy excita- ductivity and correlated insulators [3–27]. Furthermore, such tions are constrained by the valley contrasting Chern number electronic band structure can also be topologically nontrivial, in TMD Moiré superlattice. e.g., with a nonzero integer Chern number [28–32]. Combined Model.— We consider twisted homobilayer TMD materials. with their strong coupling nature, such Moiré superlattices of- For each single layer, the low energy electronic states reside at fer a promising route to realize the long-sought fractionalized the valence band maxima at ±푲 valleys. Contrary to bilayer topological order [33–39]. systems where the valley and spin degrees of freedom Recently, gapped electronic states at various fractional fill- are both present, in TMD each valley in the top valence band ings (e.g., 1/3) were observed in transition metal dichalco- has fixed spin orientation due to strong spin-orbit coupling and genide (TMD) Moiré superlattices, e.g., WSe2/WS2 [18, 40– the broken inversion symmetry [55]. With a small twist angle 43]. In general, gapped electronic states at fractional filling 휃 between two layers, the +푲 valley for the top and bottom may have two origins: (a) charge order that spontaneously layers are shifted to 푲푡 and 푲푏 in the Moiré Brillouin zone breaks the translational symmetry and (b) fractional topolog- (MBZ) respectively, [Fig.1(b)]. For convenience we choose ical order, e.g., fractional Chern insulators (FCI) [44–51]. In the rhombus-shaped MBZ and set the point 푴 = (푲푡 + 푲푏)/2 these TMD Moiré superlattices, the observed gapped states as the origin. We employ the continuum model [2] in which were interpreted as Wigner crystals of electrons, because the the Moiré Hamiltonian for the +푲 valley is: underlying single-particle bands are topologically trivial [52]. 2 2 ℏ (풌−푲푏) ! Encouraged by such exciting experimental progress, here we − 2푚∗ + Δ픟 (풓) Δ푇 (풓) 퐻+ (풌, 풓) = 2 2 (1) explore the feasibility of the second category in TMD Moiré † ℏ (풌−푲푡 ) Δ푇 (풓) − 2푚∗ + Δt (풓) superlattices. In particular, we focus on systems like MoTe2, arXiv:2101.01258v2 [cond-mat.mes-hall] 10 Sep 2021 which may host topologically nontrivial bands with non-zero Here 푚∗ is the effective mass. The form of Moiré potential, spin/valley-Chern numbers [53]. In contrast to a partially Δ푏,푡,푇 , is dictated by the 퐷3 crystalline symmetry and a com- filled Chern band [44–51], because these systems preserve the bination of 퐶2푧 rotation followed by switching the two layers, time-reversal symmetry, two types of fractional states are in and can be parameterized by [53]: principle allowed (a) time-reversal invariant fractional topo- logical insulators [54] and (b) FCIs via spontaneously break-  − · − ·  Δ (풓) = 푤 1 + 푒 푖푮2 풓 + 푒 푖푮3 풓 ing the time-reversal symmetry. The key focus of this study 푇 ∑︁  is whether Coulomb repulsion could stabilize some of these Δ푙 (풓) = 2푤푧 cos 푮 푗 · 풓 + 푙휓 , (2) fractional states in TMD Moiré superlattices. 푗=1,3,5 2 where 푙 ∈ {푏, 푡} = {+1, −1} and 푮 푗 is the Moiré re- 4휋 ciprocal lattice vectors with length |푮 푗 | = √ and po- 3푎 푀 K 휋 ( 푗−1) t lar angle . Here 푎푀 = 푎0/휃 is the Moiré lattice 3 M Γ constant for a small twisted angle 휃 and 푎0 is the lat- K ' Kb b tice parameter of TMD. The Hamiltonian for the valley −푲 can be obtained by the time-reversal symmetry 퐻− (풌, 풓) = ∗ 퐻+ (−풌, 풓) . To be specific, we focus on twisted MoTe2 ho- C = +1 2 ∗ 2 20 12 mobilayer with typical parameters (ℏ /2푚 푎 , 푤푧, 푤, 휓) = 0 = − ◦ 18 C 1

(495 meV, 8 meV, −8.5 meV, −89.6 ) [53]. W 8 / 16 The top valence band of a TMD single layer splits into mul- 1 2 ∆ 4 tiple Moiré bands due to the Moiré potential. As shown in Fig. 14 ◦ 12 1(c) and (d), when the twist angle is close to 휃0 = 1.38 , the 0 K ' Γ K K Kb' 1.0 1.2 1.4 1.6 1.8 2.0 top Moiré band becomes nearly flat. The flatness of a band b t b can be characterized by a ratio of the gap between the nearest bands to its band width. For the top Moiré band, the ratio can FIG. 1. (a): Schematic view of the Moiré superlattice. (b): We be as large as 13. When 휃 < 3.1◦, The top Moiré band is topo- choose the Moiré Brillouin zone (MBZ) to be the rhombus and the logical characterized by a valley/spin Chern number 퐶 = ±1 origin in momentum space is chosen at 푀. (c): Moiré band structure at 휃 = 1.38◦. The top Moiré band is nearly flat with Chern number due to the skyrmion lattice pseudo spin textures of the Moiré ( ( ))− ( ( )) ±1. (d): The gap ratio Δ12 = min 퐸1 풌 max 퐸2 풌 as a function potential [53]. The Chern number for the opposite valley/spin 푊 max(퐸1 (풌))−min(퐸1 (풌)) of twisted angle 휃, where 퐸 (풌) (퐸 (풌)) is the energy of the first is opposite as required by time-reversal symmetry. Thus, at 1 2 (second) topmost Moiré band. the single-particle level, such TMD homobilayer realizes a quantum valley/spin Hall insulator. We then introduce screened Coulomb interaction and project the time-reversal symmetry is spontaneously broken or pre- it to the nearly flat top Moiré band [56]: served [59], and our exact diagonalization below show that the 1 ∑︁ FCI is favored and stabilized in our system. 퐻 = 휌(풒)푉 (풒)휌(−풒) int : : Valley polarized FCI.— We define the filling factor 휈 = 2퐴 풒 2휌푒/휌푠, where 휌푒 is the electron density occupying the top ∑︁ 푈 0 ∗ = 푣(풒)휆 (풌)휆 0 (풌 ) Moiré band and 휌푠 is the electron density for the full filling of 2푁 휏,풒 휏 ,풒 풌,풌0,풒,휏,휏0 cell the two-fold degenerate top Moiré band. The factor 2 accounts † † 0 0 for the valley degree of freedom. Using exact diagonalization, 퐶 (풌)퐶 0 (풌 + 풒)퐶 0 (풌 )퐶 (풌 + 풒), (3) 휏 휏 휏 휏 at 휈 = 1/3 we observe numerical evidence of spontaneous where 휏 = ± is the valley index, and 휆휏,풒 (풌) = h푢휏,풌 |푢휏,풌+풒i valley polarization and FCI in the strong interaction limit, as shown in Fig.2(a). For 8 electrons in 4 × 6 unit cells (4 × 6 × 2 is the form factor originated√ from the projection. Here 푣(풒) = 4휋 tanh(푞푑)/ 3푞푎푀 is the dimensionless screened single-particle states including both valleys), the ground states Coulomb potential with 푑 the separation between the elec- are fully valley polarized with three nearly-degenerate ground trode and Moiré superlattice, which is set to 푑 = 2푎푀 in the states for each valley polarization, separated from the excited calculations. 퐴 is system area and 푁cell is the number of the states by an energy gap of the order of 2 K. We calculated unit cells in the calculations. The coefficient of 푣(풒) is chosen the many-body Chern number of each ground state [48], and to make 푈 equal to the bare Coulomb potential between two the topological index is found to be 1/3, characterizing a 1/3 particles separated by 푎푀 . 퐶휏 (풌) is the annihilation operator FCI phase. This conclusion is further supported by the total for single particle state |푢휏,풌 i. We neglect the weak intervalley momentum for each ground state, which obeys the generalized impurity scattering process associated with a large momentum Pauli exclusion rule [60]. transfer, and therefore the Hamiltonian also has a valley 푈(1)푣 The occupation number 푛(푘1, 푘2) of single particle states symmetry. In this model, there are two competing symme- for each of the three many-body ground states are plotted in try breaking states: an intervalley coherent state that breaks Fig.2(b). 푛(푘1, 푘2) is uniformly distributed for different the valley 푈(1)푣 symmetry and an valley/spin polarized state single particle states, consistent with the fact that the ground that breaks the time reversal symmetry. At half filling of state is an incompressible liquid. The spectrum evolution the topmost band (account for valley degree of freedom), our under flux insertion along 푘2 direction is shown in Fig.2(c). Hartree-Fock analysis and exact-diagonalization results both The excitation gap is maintained throughout the flux insertion suggest that a valley-polarized state is energetically favored, process. which spontaneously breaks the time-reversal symmetry and The topological nature of the ground states are further con- leads to a interaction-induced Chern insulator [57]. At frac- firmed by our calculation of the particle entanglement spec- tional filling, in principle, two types of fractional topological trum (PES) [60]. To compute PES, we divide the 푁 particles states might emerge, a fractional Chern insulator or a frac- into two collections of 푁퐴 and 푁퐵 = 푁 − 푁퐴 particles and tional topological insulator [54, 58], depending on whether trace out 푁퐵 particles to get the reduced density matrix 휌퐴. 3

0.5 cle Moiré band has the largest gap to bandwidth ratio [see Fig. (a) 16.10 (b) 1(d)]. This is consistent with the quantum Hall systems with 0.4

16.05 )

2

k flat Landau levels, where the interaction stabilizes simultane-

,

16.00 1 0.3

k

( ously the fractional quantum Hall state with spin polarization.

n 15.95 0.2 Note that the FCI can be stabilized in a relatively broader 15.90 0.1 region of the twisted angle here compared to that in magic an- 0 5 10 15 20 0 5 10 15 20 + k N + k k1N2 k2 1 2 2 gle twisted bilayer graphene [33–39], and the region of angle (c) (d) 14 for FCI increases with interaction. For interaction above the 15.9198 12 dashed line in Fig.3(b), our single-band approximation used 15.9194 ξ 10 in the numerical calculations breaks down and it requires to 15.9190 8 take other nearby bands into account. 15.9186 6 Excitations.— Here we study the charge neutral excitations 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 5 10 15 20 Φ / 2π k1N2 + k2 above the FCI ground states. As a consequence of the sponta- neous valley polarization, we consider the valley waves exci- Í † FIG. 2. Numerical diagonalization results for 8 particles in 4×6 Moiré tation |Ψ푣 (풒)i = 풌 푧풌퐶+ (풌 + 풒)퐶− (풌)|Ψ−i, where |Ψ−i is ◦ lattice. We choose 휃 = 1.38 and 푈 = 1.38 meV. The bandwidth the FCI ground state with 휏 = − valley fully occupied and 푧풌 at this twist angle is 푊 = 0.083 meV. Here 푁 = 푁1 × 푁2/3 is is variational parameter. The presence of the form factor in the number of particles. (a): Energy spectrum with three nearly Eq. (3) breaks the valley pseudospin SU(2) rotation system degenerate ground states in each valley. (b): The occupation number down to the valley 푈(1)푣 symmetry. As a result, the valley of single particle states 푛(푘1, 푘2) for each of the three many-body wave excitation are gapped as shown in Fig.4, which can be ground states. The nearly uniform distribution of 푛(푘1, 푘2) suggests 2 the ground state is an incompressible liquid. (c): Under flux insertion fitted by 퐸푤 (풒) = 퐽푞 + 퐴. The valley wave disperses weakly along 푘2 direction, the ground states evolve into each other. (d): in momentum and thus is well localized in real space. Particle entanglement spectrum (PES) for the separation of 푁퐴 = 4 The lowest intravalley many-body excitation has lower en- particles. ergy than valley wave excitation for the parameters we used, i.e., the energy difference between the lowest fully-polarized excited state and the FCI state is 퐸 = 0.167 meV < 퐸 , The PES levels 휉 are obtained from the logarithm of eigen- 푚푏 푤 see Fig.4. Nevertheless, the valley wave excitation remains values of 휌 , and are labeled by the total momentum of the 퐴 a stable excitation because the decay of the valley wave to the remaining 푁 particles, as shown in Fig.2(d). There is a 퐴 intravalley many-body excitations are forbidden. Intravalley clear entanglement gap with 2730 levels below the gap for many-body excitations has valley quantum number 0, while 푁 = 4, consistent with the counting of quasihole excitation 퐴 the valley wave has valley quantum number 2. in the 휈 = 1/3 FCI [60]. To exam the finite-size effect, we study the scaling of the In quantum Hall ferromagnets, a pair of skyrmions has lower many-body gap Δ with various system sizes. For a genuine energy than the particle-hole bound state [61]. The system FCI, Δ remain finite in the thermodynamic limit when both size limitation in the exact diagonalization does not allow us to study the valley skyrmion excitation in our numerical cal- 푁1 and 푁2 approach infinity. However Δ should vanish if only culations. Here we use an effective Hamiltonian density for one of 푁1 or 푁2 approaches infinity, because this limit is a one-dimensional system which should not support FCI [47]. This is confirmed in Fig.3(a), which shows Δ decreases when 푁1 is fixed at 3 and 푁2 increases from 4 to 8, but Δ increases (a) (b) when the system size changes from 3 × 8 to 4 × 6. 0.25 0.20 We then map out the phase diagram at 휈 = 1/3 filling as a 4×6 0.15 function of the interaction strength 푈 which can be controlled 0.10 × 3× 4 by distance between the electrodes and the Moiré superlattice 3×8 3×6 3 5 0.05 in experiments and the dielectric constant 휖. The results are 0.00 shown in Fig.3(b). We find a valley non-polarized Fermi liq- 0.00 0.05 0.10 0.15 0.20 0.25 uid at a small 1/휖 corresponding to a small 푈, a Fermi liquid with valley polarization at an intermediate interaction and the FIG. 3. (a): The many-body gap Δ for various system sizes at 푣 = 1/3 FCI phase with valley polarization at strong interaction. De- filling. The interaction strength is fixed to be 푈 = 1.38 meV. The pending on the twisted angle 휃, which controls the bandwidth, increase of Δ in 4 × 6 system suggests the gap persists in the two- the valley non-polarized Fermi liquid can transit directly to dimensional thermodynamic limit. (b): The phase diagram for Fermi FCI with valley polarization or through an intermediate Fermi liquid (FL), FL with valley polarization (VP) and fractional Chern 2 insulator (FCI) at different interaction strength 푈(휖) = 푒 and liquid with valley polarization. The phase transition between 4휋 휖 휖0 푎푀 the valley-polarized Fermi liquid and FCI can be described by twisted angle 휃. The dashed line corresponds to 푈(휖) = Δ12, above Ginzburg-Landau theory with a Chern-Simons term [57]. The which the interaction starts to mix different bands and the single-band direct transition occurs near 휃 = 1.38◦ where the single parti- approximation breaks down. 4

(a) (b) 20.65 15.1 0.223 k1 = 0,k2 = 0 15.0 20.55 0.222 = = 14.9 k1 0,k2 2 20.45 14.8 0.2 EW 0.221 20.35 14.7 Emb 0.1 0 5 10 15 0 2 4 6 8 10 12 14 0.220 + k1N2 + k2 k1N2 k2 0 E (c) (d) g 14 0.219 20.45 12 0.0 0.5 1.0 1.5 2.0 2.5 3.0 2 2 20.40 ξ 10 q aM 8 20.35 6 FIG. 4. Dispersion of valley wave excitation 퐸푤 (풒) for 8 particles in 0.0 0.2 0.4 0.6 0.8 1.0 0 5 10 15 Φ / 2π + 4 × 6 lattice. Excitation above the ground state with total momentum k1N2 k2 푘1 = 0, 푘2 = 0 (푘1 = 0, 푘2 = 2) are labeled by the squares and circles respectively. The slight energy difference for these two ground states FIG. 5. FCI at 푣 = 2/5. (a): the energy spectrum of 8 particles in is caused by finite size effect. The inset compares the energy of the 4*5 system, and (b): 6 particles in 3*5 system. (c): the flux insertion lowest valley wave excitation 퐸푤 , the lowest intravalley many-body for the system in (a), where the five ground states are marked in red excitation 퐸푚푏 and the ground state energy 퐸푔. (some of them are on top of each other). A finite gap remains during flux insertion. (d): the particle entanglement spectrum for (a) with 푁퐴 = 3. There are 51 × 20 = 1020 states below the dashed line, valley pseudospin n(푟) [61] consistent with quasihole counting. ∫ 퐽 2 퐴 2 1 02 0 0 퐻푛 (푟) = (∇n) − 푛푧 + 푑푟 푉 (푟 − 푟 )휌푠 (푟)휌푠 (푟 ), 2 2 2 Discussions.— We show that TMD Moiré superlattices can (4) host fractional topological states via spontaneously breaking where 퐽 and 퐴 are given by the valley wave spectrum. The the time-reversal symmetry, using realistic parameters of TMD presence of the valley pseudospin anisotropy can be traced Moiré superlattices. Comparing with graphene, the spin- back to the opposite Chern number for the opposite valley. One valley locking in TMD materials breaks the SU(2) spin rota- cannot rotate n from one valley to opposite valley adiabatically tion symmetry and eliminates the spin wave Goldstone modes, without closing the energy gap, which implies the existence which could help stabilize the FCI states. The valley contrast- of anisotropy for n. The last term accounts for the Coulomb ing Chern number in TMD Moiré superlattices also dictates interaction푉 (푟−푟 0), because a skyrmion is dressed with charge the symmetry breaking states, hence the nature of fractional- 푖 푗 푎 푏 푐 푖 푗 ized topological states, and also the low energy excitations in distribution 휌푠 (푟) = 휖 휖푎푏푐푛 휕푖푛 휕푗 푛 /8휋, where 휖푎푏푐 (휖 ) is the Levi-Civita tensor with 푖, 푗 being the space index and the FCI. The gapped nature of these states can be detected by 푎, 푏, 푐 being the spin index. The skyrmion topological charge transport, optical measurements etc, and its topological nature ∫ 2 can be accessed by Hall conductivity measurement. Due to the 푄푠 = 푑푟 휌푠 (푟) is quantized to an integer number. The easy axis anisotropy favors skyrmions with a small radius while strong analogy between FCI and chiral spin liquids, it is plau- the Coulomb repulsion favors skyrmions with a large radius. sible that Moiré superlattices may also help realize/stabilize Their competition determines the skyrmion size [62]. exotic spin liquid phases [63–65] by unitizing the valley/layer FCI at 푣 = 2/5.— In twisted graphene Moiré superlattices, pseudospin or real spin degrees of freedom. the Halperin (332) state is stabilized at 푣 = 2/5 due to the re- Acknowledgements.— This work done at LANL was carried maining SU(2) spin rotation symmetry in the valley polarized out under the auspices of the U.S. DOE NNSA under contract state [36]. In our TMD Moiré superlattices, the spin rotation No. 89233218CNA000001 through the LDRD Program. S. Z. symmetry is absent because of the spin-valley locking. The L. was also supported by the U.S. Department of Energy, Office Chern number contrasting valley degree of freedom disfavors of Science, Basic Energy Sciences, Materials Sciences and the (332) state. To demonstrate this explicitly, we calculate Engineering Division, Condensed Matter Theory Program. the energy spectrum, spectrum flow under flux insertion and H.L. and K.S. acknowledge support through NSF Grant No. entanglement spectrum at 푣 = 2/5, and the results are dis- NSF-EFMA-1741618. played in Fig.5. The 5-fold degenerate ground states are valley polarized, and are consistent with the 푣 = 2/5 FCI state. In the fractional quantum Hall, the 푣 = 2/5 state belongs to the second hierarchical Jain state, and similarly one can assign the [1] J. M. B. Lopes dos Santos, N. M. R. Peres, and A. H. Cas- = / 푣 2 5 FCI as a second hierarchical FCI. Our results highlight tro Neto, “Graphene bilayer with a twist: Electronic structure,” the importance of symmetry in dictating the ground state and Phys. Rev. Lett. 99, 256802 (2007). contrast the difference between the graphene and TMD Moire [2] Rafi Bistritzer and Allan H. MacDonald, “Moiré bands in superlattices. twisted double-layer graphene,” PNAS 108, 12233–12237 5

(2011). angle twisted bilayer graphene,” Nature 588, 610–615 (2020), [3] Yuan Cao, Valla Fatemi, Ahmet Demir, Shiang Fang, Spencer L. number: 7839 Publisher: Nature Publishing Group. Tomarken, Jason Y. Luo, Javier D. Sanchez-Yamagishi, Kenji [16] Minhao He, Yuhao Li, Jiaqi Cai, Yang Liu, K. Watanabe, Watanabe, Takashi Taniguchi, Efthimios Kaxiras, Ray C. T. Taniguchi, Xiaodong Xu, and Matthew Yankowitz, “Symme- Ashoori, and Pablo Jarillo-Herrero, “Correlated insulator be- try breaking in twisted double bilayer graphene,” Nature Physics haviour at half-filling in magic-angle graphene superlattices,” , 1–5 (2020), publisher: Nature Publishing Group. Nature 556, 80–84 (2018). [17] Xiaomeng Liu, Zeyu Hao, Eslam Khalaf, Jong Yeon Lee, Yuval [4] Yuan Cao, Valla Fatemi, Shiang Fang, Kenji Watanabe, Takashi Ronen, Hyobin Yoo, Danial Haei Najafabadi, Kenji Watanabe, Taniguchi, Efthimios Kaxiras, and Pablo Jarillo-Herrero, “Un- Takashi Taniguchi, Ashvin Vishwanath, and Philip Kim, “Tun- conventional superconductivity in magic-angle graphene super- able spin-polarized correlated states in twisted double bilayer lattices,” Nature 556, 43–50 (2018). graphene,” Nature 583, 221–225 (2020), number: 7815 Pub- [5] Xiaobo Lu, Petr Stepanov, Wei Yang, Ming Xie, Mohammed Ali lisher: Nature Publishing Group. Aamir, Ipsita Das, Carles Urgell, Kenji Watanabe, Takashi [18] Emma C. Regan, Danqing Wang, Chenhao Jin, M. Iqbal Taniguchi, Guangyu Zhang, Adrian Bachtold, Allan H. Mac- Bakti Utama, Beini Gao, Xin Wei, Sihan Zhao, Wenyu Zhao, Donald, and Dmitri K. Efetov, “Superconductors, orbital mag- Zuocheng Zhang, Kentaro Yumigeta, Mark Blei, Johan D. Carl- nets and correlated states in magic-angle bilayer graphene,” Na- ström, Kenji Watanabe, Takashi Taniguchi, Sefaattin Tongay, ture 574, 653–657 (2019). Michael Crommie, Alex Zettl, and Feng Wang, “Mott and [6] Matthew Yankowitz, Shaowen Chen, Hryhoriy Polshyn, Yux- generalized Wigner crystal states in WSe 2 /WS 2 moiré super- uan Zhang, K. Watanabe, T. Taniguchi, David Graf, Andrea F. lattices,” Nature 579, 359–363 (2020), number: 7799 Publisher: Young, and Cory R. Dean, “Tuning superconductivity in twisted Nature Publishing Group. bilayer graphene,” Science 363, 1059–1064 (2019). [19] Lei Wang, En-Min Shih, Augusto Ghiotto, Lede Xian, Daniel A. [7] Alexander Kerelsky, Leo J McGilly, Dante M Kennes, Lede Rhodes, Cheng Tan, Martin Claassen, Dante M. Kennes, Yu- Xian, Matthew Yankowitz, Shaowen Chen, K Watanabe, song Bai, Bumho Kim, Kenji Watanabe, Takashi Taniguchi, T Taniguchi, James Hone, Cory Dean, et al., “Maximized elec- Xiaoyang Zhu, James Hone, Angel Rubio, Abhay N. Pasu- tron interactions at the magic angle in twisted bilayer graphene,” pathy, and Cory R. Dean, “Correlated electronic phases in Nature 572, 95–100 (2019). twisted bilayer transition metal dichalcogenides,” Nature Mate- [8] Yuan Cao, Debanjan Chowdhury, Daniel Rodan-Legrain, Oriol rials 19, 861–866 (2020), number: 8 Publisher: Nature Publish- Rubies-Bigordà, Kenji Watanabe, Takashi Taniguchi, T Senthil, ing Group. and Pablo Jarillo-Herrero, “Strange metal in magic-angle [20] Ming Xie and A. H. MacDonald, “Nature of the correlated graphene with near planckian dissipation,” arXiv preprint insulator states in twisted bilayer graphene,” Phys. Rev. Lett. arXiv:1901.03710 (2019). 124, 097601 (2020). [9] Hryhoriy Polshyn, Matthew Yankowitz, Shaowen Chen, Yuxuan [21] Fengcheng Wu and Sankar Das Sarma, “Collective excitations Zhang, K Watanabe, T Taniguchi, Cory R Dean, and Andrea F of quantum anomalous hall ferromagnets in twisted bilayer Young, “Large linear-in-temperature resistivity in twisted bi- graphene,” Phys. Rev. Lett. 124, 046403 (2020). layer graphene,” Nature Physics 15, 1011–1016 (2019). [22] Ying Su and Shi-Zeng Lin, “Current-induced reversal of anoma- [10] Yonglong Xie, Biao Lian, Berthold Jäck, Xiaomeng Liu, lous hall conductance in twisted bilayer graphene,” Phys. Rev. Cheng-Li Chiu, Kenji Watanabe, Takashi Taniguchi, B Andrei Lett. 125, 226401 (2020). Bernevig, and Ali Yazdani, “Spectroscopic signatures of many- [23] Bikash Padhi, Chandan Setty, and Philip W. Phillips, body correlations in magic-angle twisted bilayer graphene,” Na- “Doped twisted bilayer graphene near magic angles: Prox- ture 572, 101–105 (2019). imity to wigner crystallization, not mott insulation,” [11] Yuhang Jiang, Xinyuan Lai, Kenji Watanabe, Takashi Nano Letters 18, 6175–6180 (2018), pMID: 30185049, Taniguchi, Kristjan Haule, Jinhai Mao, and Eva Y Andrei, https://doi.org/10.1021/acs.nanolett.8b02033. “Charge order and broken rotational symmetry in magic-angle [24] Bikash Padhi, R. Chitra, and Philip W. Phillips, “Gener- twisted bilayer graphene,” Nature 573, 91–95 (2019). alized wigner crystallization in moiré materials,” (2020), [12] Youngjoon Choi, Jeannette Kemmer, Yang Peng, Alex Thom- arXiv:2009.13536 [cond-mat.str-el]. son, Harpreet Arora, Robert Polski, Yiran Zhang, Hechen Ren, [25] Bikash Padhi and Philip W. Phillips, “Pressure-induced metal- Jason Alicea, Gil Refael, et al., “Electronic correlations in insulator transition in twisted bilayer graphene,” Phys. Rev. B twisted bilayer graphene near the magic angle,” Nature Physics 99, 205141 (2019). 15, 1174–1180 (2019). [26] Nikolaos Stefanidis and Inti Sodemann, “Excitonic laughlin [13] U. Zondiner, A. Rozen, D. Rodan-Legrain, Y. Cao, R. Queiroz, states in ideal topological insulator flat bands and their pos- T. Taniguchi, K. Watanabe, Y. Oreg, F. von Oppen, Ady Stern, sible presence in moiré superlattice materials,” Phys. Rev. B E. Berg, P. Jarillo-Herrero, and S. Ilani, “Cascade of phase tran- 102, 035158 (2020). sitions and Dirac revivals in magic-angle graphene,” Nature 582, [27] Nick Bultinck, Shubhayu Chatterjee, and Michael P. Zaletel, 203–208 (2020), number: 7811 Publisher: Nature Publishing “Mechanism for anomalous hall ferromagnetism in twisted bi- Group. layer graphene,” Phys. Rev. Lett. 124, 166601 (2020). [14] Dillon Wong, Kevin P. Nuckolls, Myungchul Oh, Biao [28] Ya-Hui Zhang, Dan Mao, Yuan Cao, Pablo Jarillo-Herrero, and Lian, Yonglong Xie, Sangjun Jeon, Kenji Watanabe, Takashi T. Senthil, “Nearly flat chern bands in moiré superlattices,” Phys. Taniguchi, B. Andrei Bernevig, and Ali Yazdani, “Cascade of Rev. B 99, 075127 (2019). electronic transitions in magic-angle twisted bilayer graphene,” [29] Cécile Repellin, Zhihuan Dong, Ya-Hui Zhang, and T. Senthil, Nature 582, 198–202 (2020), number: 7811 Publisher: Nature “Ferromagnetism in narrow bands of moiré superlattices,” Phys. Publishing Group. Rev. Lett. 124, 187601 (2020). [15] Kevin P. Nuckolls, Myungchul Oh, Dillon Wong, Biao Lian, [30] Aaron L. Sharpe, Eli J. Fox, Arthur W. Barnard, Joe Finney, Kenji Watanabe, Takashi Taniguchi, B. Andrei Bernevig, and Kenji Watanabe, Takashi Taniguchi, M. A. Kastner, and Ali Yazdani, “Strongly correlated Chern insulators in magic- David Goldhaber-Gordon, “Emergent ferromagnetism near 6

three-quarters filling in twisted bilayer graphene,” Science 365, [46] Titus Neupert, Luiz Santos, Claudio Chamon, and Christopher 605–608 (2019). Mudry, “Fractional quantum hall states at zero magnetic field,” [31] M. Serlin, C. L. Tschirhart, H. Polshyn, Y. Zhang, J. Zhu, Phys. Rev. Lett. 106, 236804 (2011). K. Watanabe, T. Taniguchi, L. Balents, and A. F. Young, “Intrin- [47] N. Regnault and B. Andrei Bernevig, “Fractional chern insula- sic quantized anomalous hall effect in a moiré heterostructure,” tor,” Phys. Rev. X 1, 021014 (2011). Science (2019), 10.1126/science.aay5533. [48] D. N. Sheng, Zheng-Cheng Gu, Kai Sun, and L. Sheng, “Frac- [32] Guorui Chen, Aaron L Sharpe, Eli J Fox, Ya-HuiZhang, Shaoxin tional quantum hall effect in the absence of landau levels,” Na- Wang, Lili Jiang, Bosai Lyu, Hongyuan Li, Kenji Watanabe, ture Communications 2, 389 (2011). Takashi Taniguchi, et al., “Tunable correlated chern insulator [49] Siddharth A. Parameswaran, Rahul Roy, and Shivaji L. Sondhi, and ferromagnetism in trilayer graphene/boron nitride moir\’e “Fractional quantum Hall physics in topological flat bands,” superlattice,” arXiv preprint arXiv:1905.06535 (2019). Comptes Rendus Physique Topological insulators / Isolants [33] Patrick J. Ledwith, Grigory Tarnopolsky, Eslam Khalaf, and topologiques, 14, 816–839 (2013). Ashvin Vishwanath, “Fractional chern insulator states in twisted [50] Emil J. Bergholtz and Zhao Liu, “Topological flat band models bilayer graphene: An analytical approach,” Phys. Rev. Research and fractional chern insulators,” International Journal of Mod- 2, 023237 (2020). ern Physics B 27, 1330017 (2013), publisher: World Scientific [34] Cécile Repellin and T. Senthil, “Chern bands of twisted bilayer Publishing Co. graphene: Fractional chern insulators and spin phase transition,” [51] Yang-Le Wu, B. Andrei Bernevig, and N. Regnault, “Zoology Phys. Rev. Research 2, 023238 (2020). of fractional chern insulators,” Phys. Rev. B 85, 075116 (2012). [35] Ahmed Abouelkomsan, Zhao Liu, and Emil J. Bergholtz, [52] Fengcheng Wu, Timothy Lovorn, Emanuel Tutuc, and A. H. “Particle-hole duality, emergent fermi liquids, and fractional MacDonald, “Hubbard model physics in transition metal chern insulators in moiré flatbands,” Phys. Rev. Lett. 124, dichalcogenide moiré bands,” Phys. Rev. Lett. 121, 026402 106803 (2020). (2018). [36] Zhao Liu, Ahmed Abouelkomsan, and Emil J. Bergholtz, “Gate- [53] Fengcheng Wu, Timothy Lovorn, Emanuel Tutuc, Ivar Martin, tunable fractional chern insulators in twisted double bilayer and A. H. MacDonald, “Topological insulators in twisted transi- graphene,” Phys. Rev. Lett. 126, 026801 (2021). tion metal dichalcogenide homobilayers,” Phys. Rev. Lett. 122, [37] Patrick Wilhelm, Thomas C. Lang, and Andreas M. Läuchli, 086402 (2019). “Interplay of Fractional Chern Insulator and Charge-Density- [54] Michael Levin and Ady Stern, “Fractional topological insula- Wave Phases in Twisted Bilayer Graphene,” arXiv:2012.09829 tors,” Phys. Rev. Lett. 103, 196803 (2009). [cond-mat] (2020), arXiv: 2012.09829. [55] Di Xiao, Gui-Bin Liu, Wanxiang Feng, Xiaodong Xu, and [38] Ramanjit Sohal, Luiz H. Santos, and Eduardo Fradkin, “Chern- Wang Yao, “Coupled spin and valley physics in monolayers of simons composite fermion theory of fractional chern insulators,” mos2 and other group-vi dichalcogenides,” Phys. Rev. Lett. 108, Phys. Rev. B 97, 125131 (2018). 196802 (2012). [39] Ramanjit Sohal and Eduardo Fradkin, “Intertwined order in [56] Jong Yeon Lee, Eslam Khalaf, Shang Liu, Xiaomeng Liu, Zeyu fractional chern insulators from finite-momentum pairing of Hao, Philip Kim, and Ashvin Vishwanath, “Theory of corre- composite fermions,” Phys. Rev. B 101, 245154 (2020). lated insulating behaviour and spin-triplet superconductivity in [40] Yang Xu, Song Liu, Daniel A. Rhodes, Kenji Watanabe, Takashi twisted double bilayer graphene,” Nature Communications 10, Taniguchi, James Hone, Veit Elser, Kin Fai Mak, and Jie Shan, 5333 (2019). “Correlated insulating states at fractional fillings of moiré super- [57] See Supplemental Materials for (1) variational calculations of lattices,” Nature 587, 214–218 (2020), number: 7833 Publisher: the energy for the valley polarized and intervalley coherent state, Nature Publishing Group. (2) Hartree-Fock calculations, (3) discussion on the possibility [41] Chenhao Jin, Zui Tao, Tingxin Li, Yang Xu, Yanhao Tang, of the Halperin state and (4) effective theory for the transition Jiacheng Zhu, Song Liu, Kenji Watanabe, Takashi Taniguchi, between the valley polarized Fermi liquid and fractional Chern James C. Hone, Liang Fu, Jie Shan, and Kin Fai Mak, “Stripe insulator. phases in WSe2/WS2 moir\’e superlattices,” arXiv:2007.12068 [58] Ady Stern, “Fractional Topological Insulators: A Pedagogical [cond-mat] (2020), arXiv: 2007.12068. Review,” Annual Review of 7, 349– [42] You Zhou, Jiho Sung, Elise Brutschea, Ilya Esterlis, Yao Wang, 368 (2016), publisher: Annual Reviews. Giovanni Scuri, Ryan J. Gelly, Hoseok Heo, Takashi Taniguchi, [59] One may think of another possibility, analogous to the Halperin Kenji Watanabe, Gergely Zaránd, Mikhail D. Lukin, Philip (푚1, 푚2, 푛1) states. However, this state is not favored because Kim, Eugene Demler, and Hongkun Park, “Signatures of of the opposite Chern number in the opposite valley. [57]. bilayer Wigner crystals in a transition metal dichalcogenide [60] B. Andrei Bernevig and N. Regnault, “Emergent many-body heterostructure,” arXiv:2010.03037 [cond-mat] (2020), arXiv: translational symmetries of abelian and non-abelian fractionally 2010.03037. filled topological insulators,” Phys. Rev. B 85, 075128 (2012). [43] Xiong Huang, Tianmeng Wang, Shengnan Miao, Chong Wang, [61] S. L. Sondhi, A. Karlhede, S. A. Kivelson, and E. H. Rezayi, Zhipeng Li, Zhen Lian, Takashi Taniguchi, Kenji Watanabe, “Skyrmions and the crossover from the integer to fractional Satoshi Okamoto, Di Xiao, Su-Fei Shi, and Yong-Tao Cui, “Cor- quantum hall effect at small zeeman energies,” Phys. Rev. B 47, related Insulating States at Fractional Fillings of the WS2/WSe2 16419–16426 (1993). Moir\’e Lattice,” arXiv:2007.11155 [cond-mat] (2020), arXiv: [62] Shubhayu Chatterjee, Matteo Ippoliti, and Michael P. Zaletel, 2007.11155. “Skyrmion Superconductivity: DMRG evidence for a topolog- [44] Evelyn Tang, Jia-Wei Mei, and Xiao-Gang Wen, “High- ical route to superconductivity,” arXiv:2010.01144 [cond-mat] temperature fractional quantum hall states,” Phys. Rev. Lett. (2020), arXiv: 2010.01144. 106, 236802 (2011). [63] Krishna Kumar, Kai Sun, and Eduardo Fradkin, “Chern-simons 1 [45] Kai Sun, Zhengcheng Gu, Hosho Katsura, and S. Das Sarma, theory of magnetization plateaus of the spin- 2 quantum xxz “Nearly flatbands with nontrivial topology,” Phys. Rev. Lett. heisenberg model on the kagome lattice,” Phys. Rev. B 90, 106, 236803 (2011). 174409 (2014). 7

[64] Krishna Kumar, Kai Sun, and Eduardo Fradkin, “Chiral spin filling,” Phys. Rev. X 10, 031034 (2020). liquids on the kagome lattice,” Phys. Rev. B 92, 094433 (2015). [67] Eduardo Fradkin, Field Theories of Condensed Matter Physics, [65] W. Zhu, Shou-Shu Gong, Tian-Sheng Zeng, Liang Fu, and D. N. 2nd ed. (Cambridge University Press, Cambridge, 2013). Sheng, “Interaction-driven spontaneous quantum hall effect on [68] Hua Chen and Kun Yang, “Interaction-driven quantum phase a kagome lattice,” Phys. Rev. Lett. 117, 096402 (2016). transitions in fractional topological insulators,” Phys. Rev. B 85, [66] Nick Bultinck, Eslam Khalaf, Shang Liu, Shubhayu Chatterjee, 195113 (2012). Ashvin Vishwanath, and Michael P. Zaletel, “Ground state and hidden symmetry of magic-angle graphene at even integer 8

Supplemental Material: Spontaneous fractional Chern insulators in transition-metal-dichalcogenides Moiré superlattices

I. Energy for valley polarized (VP) and intervalley coherent (IVC) state

Here we compare the energy for the valley polarized and intervalley coherent state. We consider half filling of the topmost band including the valley degree of freedom. The wave function for the VP state can be written as

Ö † |ΨVPi = 퐶+ (푘)|0i. 푘

Here we choose 휏 = + valley to be fully occupied. Its energy hΨVP|퐻|ΨVPi can be decomposed into single particle contribution Í Í 2 Í 2 퐸0 = 푘 휖+ (푘); Hartree contribution: 퐸Ha = 푉 (0)[ 푘 휆+,0 (푘)] and the Fock contribution: 퐸Fo = − 푘,푞 푉 (푞)|휆+,푞 (푘)| . Here the summation of 푘 is over the whole Moiré Brillouin zone. We note that the system Hamiltonian is invariant under the following gauge transformation

퐶휏 (푘) → exp[푖휃휏 (푘)]퐶휏 (푘),

휆휏,푞 (푘) → exp[푖휃휏 (푘) − 휃휏 (푘 + 푞)]휆휏,푞 (푘).

We choose a gauge which fixes the valley pseudospin associated with |ΨIVCi in the 푥 direction, and the wave function of the intervalley coherent state can be written as

1 Ö † † |ΨIVCi = √ [퐶+ (푘) + 퐶− (푘)]|0i. 2 푘 0 Í 0 The energy for the IVC is the sum of single particle contribution 퐸0 = 푘 [휖+ (푘) + 휖− (푘)]/2; Hartree contribution: 퐸Ha = 푉 (0) Í 2 4 [ 푘,휏=± 휆휏,0 (푘)] and the Fock contribution

0 1 ∑︁  2 2  퐸 = − 푉 (푞) |휆+ (푘)| + |휆− (푘)| + 휆+ (푘)휆− − (푘 + 푞) + 휆− (푘)휆+ − (푘 + 푞) . Fo 4 ,푞 ,푞 ,푞 , 푞 ,푞 , 푞 푘,푞

It easy to check that the single particle and Hartree part of the energy for the VP and IVC are the same. We compare the Fock ∗ part of the energy. Noticing that 휆휏,−푞 (푘 + 푞) = 휆휏,푞 (푘), we have

휆+,푞 (푘)휆−,−푞 (푘 + 푞) + 휆−,푞 (푘)휆+,−푞 (푘 + 푞) ≤ 2|휆+,푞 (푘)휆−,−푞 (푘 + 푞)|, where the bound is saturated when 휆+,푞 (푘)휆−,−푞 (푘 + 푞) is positive real. The energy difference between the IVC and VP is

0 1 ∑︁  2 퐸 − 퐸 ≥ 푉 (푞) |휆+ (푘)| − |휆− (푘)| . Fo Fo 4 ,푞 ,푞 푘,푞

By considering time reversal oepration and 퐶2 rotation combined with layer flipping symmetry, we can show |휆+,푞 (푘)| = |휆−,푞 (푘)|. After we choose the form of |ΨIVCi by fixing a gauge, it is not guaranteed that 휆+,푞 (푘)휆−,−푞 (푘 + 푞) is positive real for all 푘 and 푞. Therefore the IVC always has higher energy than that of the VP state. This conclusion is further supported by more detailed Hartree-Fock calculations below.

II. Hartree-Fock calculations

We treat the interaction part of the Hamiltonian using Hatree-Fock mean-field theory [S66] in which the Hamiltonian can be written as

∑︁ † 휏,휏0 1 푇 H = 퐶 (풌)[(ℎ (풌) − 휇)훿 0 + ℎ (Δ , 풌)]퐶 0 (풌) − tr ℎ (Δ , 풌)Δ 푀 퐹 휏 0 휏,휏 퐻 퐹 푘 휏 2 푀 퐹 푘 푘 (S1) 푘,휏,휏0 9

(a) (b) � = 1 � = 1/3

Í FIG. S1. 푃푣 (= 풌 [Δ++ (풌) − Δ−− (풌)]) dependence on interaction (푈) for filling (휈 = 1) in panel (a) and fractional filling (휈 = 1/3) in panel (b). We observe unpolarized metal for smaller 푈, polarized metal for intermediate 푈 and a valley polarized state for large 푈.

1 Here, ℎ0 (풌) = ℎ퐵푀 (풌) − 2 ℎ퐻 퐹 (Δ0, 풌) where Δ0 is the reference density matrix such that H푀 퐹 = ℎ퐵푀 of symmetry unbroken 휈/2 0  state when Δ = Δ [S66]. We therefore, choose Δ = ∀ 풌. 푘 0 0 0 휈/2 Also, the ℎ퐻 퐹 is given by,

휏,휏0 푈 ∑︁ ∑︁ † 0 푇 ℎ (Δ , 푘) = 푣 휆 (풌)훿 0 tr[휆 (풌 )Δ 00 (풌)] 퐻 퐹 푘 2푁 푮 휏,푮 휏,휏 휏00,푮 휏,휏 cell 푮 휏00,풌0 (S2) 푈 ∑︁ † 푇 0 − 푣 + 0휆 + 0 (풌)휆 0 0 (풌)Δ 0 (풌 + 풌 ) 2푁 푮 풌 휏,푮 풌 휏 ,푮+풌 휏,휏 cell 푮,풌0

In the above equation, the first and second terms are the Hartree and Fock contributions, respectively. Here, 푁cell is the total area of the MBZ and we have used 풒 = 푮 + 풌0. 풌, 풌0 are momentum vectors in the first Brillouin zone (BZ) and 푮 is the reciprocal vector connecting different BZs. The matrix, 휆휏,풒 (풌) = 휆휏,푮+풌0 (풌) contains the form factors for the single particle given by, 4휋 tanh(푞푑) 휆휏,푮+풌0 (풌) = h푢휏,풌 |푢휏,풌+푮+풌0 i. We also have 푣풒 = √ for dual-gate screened Coulomb interaction. 푞 3푎푀 ∗ 푡ℎ For writing the mean-field equation, we use the following condition; a) 휆+,풒 (풌) = 휆−,−풒 (−풌), and b) the 푮 푗 component of 푡ℎ momentum, 풒 = 푮푙 + 풌 in the 푙 Moiré Lattice is generated using central BZs as, |푢휏,풌+푮푙 (푮 푗 )i = |푢휏,풌 (푮 푗 + 푮푙)i, so as to have a consistent gauge. † † † One can write the new by solving the above equation; 푉 퐻푀 퐹푉푉 |휓i = 퐸푛푉 |휓i. One can evaluate the † † Hamiltonian as h휓|푉푉 퐻MF푉푉 |휓i = h휙|퐷|휙i where 퐶 (풌) 푢 (풌) 푢 (풌) 훾 (풌) |휓(풌)i = + = 푉 (풌)|휙(풌)i = 1 2 1 (S3) 퐶− (풌) 푣1 (풌) 푣2 (풌) 훾2 (풌)

† The gap Δ휏,휏0 (풌) = h퐶휏 (풌)퐶휏0 (풌)i has to be written in terms of these new We now write the gap equation in the new basis,

 2 2 ∗ ∗  † |푢1 (풌)| h푛훾 (풌)i + |푢2 (풌)| h푛훾 (풌)i 푢 (풌)푣1 (풌)h푛훾 (풌)i + 푢 (풌)푣2 (풌)h푛훾 (풌)i Δ 0 (풌) = h퐶 (풌)퐶 0 (풌)i = 1 2 1 1 2 2 . 휏,휏 휏 휏 ( ) ∗ ( )h ( )i + ( ) ∗ ( )h ( )i | ( )|2h ( )i + | ( )|2h ( )i 푢1 풌 푣1 풌 푛훾1 풌 푢2 풌 푣2 풌 푛훾2 풌 푣1 풌 푛훾1 풌 푣2 풌 푛훾2 풌 (S4) † Here 푛훾푚 (풌) = 훾푚 (풌)훾푚 (풌) Also, filling in the new basis is given by

∑︁ † ∑︁ ∑︁ 푛¯ = h퐶 (풌)퐶 (풌)i = tr[Δ 0 ] = h푛 (풌)i + h푛 (풌)i = 휈 휏 휏 휏,휏 훾1 훾2 (S5) 휏,풌 풌 풌 Eqs. S4 and S5 are then solved self-consistently for a fixed filling, until 휇 and mean field order parameter or all 풌 converges. But, we use a relatively relaxed condition for convergence, as Δ휏,휏0 (풌) can have multiple degenerate configurations, therefore, 푛 푛+1 Í we use max(Δ − Δ ) < tolerance limit, where Δ = 풌 Δ휏,휏0 (풌). In the numerical simulation, we observe that only diagonal element of Δ휏,휏0 (풌) are populated whereas the off-diagonal elements are zero, meaning that valley polarized state is the ground Í state. We define a valley polarization order parameter 푃푣 = 풌 [Δ++ (풌) − Δ−− (풌)]. 10

In Fig. S1, we plot 푃푣 dependence on the interaction (푈) for (휈 = 1) and fractional (휈 = 1/3) fillings using the parameters discussed in the main text and at 휃 = 1.38◦. In the case of filling (휈 = 1) shown in Fig. S1 (a), we observe that till around 푈 = 0.3푊 we have unpolarized metal, 푖.푒. equal number of electron in both the valleys. For an intermediate interaction, 0.3푊 ≤ 푈 ≤ 0.8푊, we observe partially polarized metal and from 푈 = 0.8푊 onward the system is fully polarized. Here, 푊 is the bandwidth of the non-interacting bands. The band in one valley is fully occupied and the system is a valley polarized insulator. This behavior is consistent with the results reported for twisted bilayer graphene in the Ref. [S27]. On the other hand, in the case of the fractional filling (휈 = 1/3) shown in Fig. S1 (b), we observe the system to be unpolarized metal till 푈 = 0.85푊. In the 0.85푊 ≤ 푈 ≤ 1.4푊 regime, partially polarized metal is observed and finally above, 푈 = 1.4푊, the system saturates into a completely polarized metal. Note that in this filling fraction, one can only partially fill the lower mean-field band. Hence, the system remains a metal in contrast to the case with 푣 = 1.

III. Possibility of the Halperin state

Here we ague that it is unfavorable to host the Halperin (푚1, 푚2, 푛1) states in TMD Moiré superlattice, because of the opposite Chern number in the opposite valleys. For a flat Chern band, a good starting point to understand the physics is the Landau level by neglecting the variation of the Berry curvature and band dispersion. The Chern bands in the ± valleys in TMD Moiré superlattice can be treated as Landau levels stabilized by opposite effective magnetic field, ±퐵. The wave function for the Halperin (푚1, 푚2, 푛1) state is " # Ö ∑︁ 1 |Ψ i = 푧 − 푧 푚1 푧 − 푧 푚2 푧 − 푧 푛1 exp − |푧 |2 , 퐻 +푖 + 푗 −푖 − 푗 +푖 − 푗 2 휏푖 푖< 푗 푖,휏=± 4푙퐵 √︁ where 푧±푖 = 푥 + 푖푦 and 푙퐵 = ℏ푐/푒퐵 is the magnetic length. Introducing composite particle through the flux attachment [S67]

휙휏 (푟) = exp (푖Θ휏) 휓휏 (푟),

∫ ∫ Θ+ = 푚1 푑푟2휃(푟1 − 푟2)휌+ (푟2) + 푛1 푑푟2휃(푟1 − 푟2)휌− (푟2),

∫ ∫ Θ− = 푚2 푑푟2휃(푟1 − 푟2)휌− (푟2) + 푛1 푑푟2휃(푟1 − 푟2)휌+ (푟2), where 휃(푟1 − 푟2) is the angle between the vector 푟2 − 푟1 and the 푥 axis. 휙휏 (푟) is the composite particle wave function and 휓휏 (푟) is the electron wave function. 휌휏 is the charge density. The effective magnetic experienced by the composite particles are

퐵eff,+ = 퐵 − Φ0 (푚1 휌+ + 푛1 휌−) , 퐵eff,− = −퐵 − Φ0 (푚2 휌− + 푛1 휌+) , with Φ0 = 2휋ℏ푐/푒. It is not possible to make 퐵eff,± vanishing for any positive integers (푚1, 푚2, 푛1). Therefore the Halperin state is generally unflavored in energetics.

IV. Effective theory for the transition between the valley polarized Fermi liquid and fractional Chern insulator

If we take the Landau level point of view by assuming completely flat bands with uniform Berry curvature, we can write down Ginzburg-Landau free energy to describe the phase transition between the Fermi liquid with valley polarization and fractional Chern insulator (FCI), analogous to the case for the transition between the fractional topological insulator and superfluid [S68]. In this consideration, we can think that two valleys experience an opposite magnetic field 퐵휏 = ∇ × 푎휏. The Lagrangian can be written as

 1 2 1 휇휈휆 L휏 = 휓¯ 푖휕푡 − 퐴0 − 푎휏,0 휓 − |(−푖∇ − 퐴 − 푎휏)휓| − 푉 (휓) + 휖 푎휇휕휈푎휆, 2푚휓 4휋푚

1 2 2 훽 4 L휙 = 휙¯휏 (푖휕푡 − 퐴0) 휙휏 − |(−푖∇ − 퐴)휙휏 | − 훼|휙휏 | − |휙휏 | , 2푚 휙 2 11

2 2 L휙 = −푔|휙| |휓| .

Here 푎휏 is the dynamical gauge field with Chern-Simons term. 퐴 is the external electromagnetic gauge fields. 휓 describes the † 2 2 FCI condensate and 휙휏 = h퐶휏퐶휏i is valley polarization order parameter, and their spatial average value obey h|휓| + |휙휏 | i = 휌0 with 휌0 the electron density. Here 푚 is an integer and we consider electron filling at 1/푚. The term with 푔 > 0 describes the repulsion between the FCI condensate and valley polarization condensate. 푉 (휓) is the potential for the FCI condensate which depends on the electron interaction. From this construction, it is clear why the Fermi liquid with valley polarization (VP) is favored over the the Fermi liquid with inter-valley coherent state (IVC). In VP, the effective magnetic field due to the valley contrasting Chern band cancels. In L휙, there is no 푎휏 gauge field, which is energetically favorable because of energy cost associated with the Meissner screening of 푎휏 † if it is present. Whereas for the IVC, we need to condense 휙IVC ∼ h퐶+퐶−i. In this case, we will have coupling to the 푎휏 field: − 1 |(−푖∇ − 퐴 − 2푎 )휙 |2, which costs energy due to the Meissner effect. 2푚IVC 휏 IVC The transition from the Fermi liquid with VP to FCI can be understood as follows. For an intermediate interaction, the Fermi liquid with VP is favored. Due to the repulsion between the FCI and VP condensates, electrons condense into the 휙휏 channel and the Fermi liquid with VP is stabilized. When the interaction becomes strong, the FCI is favored by the 푉 (휓) term, and more electrons condense into the FCI state.