# Fileset

[s41467-025-59691-5.pdf](https://mdr.nims.go.jp/filesets/d622c6fe-6e88-4e04-98e5-24643c238bfe/download)

## Creator

Liheng An, Haiyang Pan, Wen-Xuan Qiu, Naizhou Wang, Shihao Ru, Qinghai Tan, Xuran Dai, Xiangbin Cai, Qiuyu Shang, Xiufang Lu, Hao Jiang, Xiaodan Lyu, Shunshun Yang, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Fengcheng Wu, Wei-bo Gao

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Observation of ferromagnetic phase in the second moiré band of twisted MoTe2](https://mdr.nims.go.jp/datasets/68b8cae5-4101-409e-b680-89e0093f7e81)

## Fulltext

Observation of ferromagnetic phase in the second moiré band of twisted MoTe2Article https://doi.org/10.1038/s41467-025-59691-5Observation of ferromagnetic phase in thesecond moiré band of twisted MoTe2Liheng An1,12, Haiyang Pan1,12, Wen-Xuan Qiu2,12, Naizhou Wang 3,4,Shihao Ru 1,5, Qinghai Tan 6, Xuran Dai1, Xiangbin Cai 1, Qiuyu Shang1,Xiufang Lu 1, Hao Jiang1, Xiaodan Lyu1, Shunshun Yang1, Kenji Watanabe 7,Takashi Taniguchi 8, Fengcheng Wu 2 & Wei-bo Gao 1,5,9,10,11Flat bands and electron correlation in moiré lattices give rise to many exoticphases, including Mott insulators, superconductivity, and topological states.Within the first moiré band, integer and fractional quantum anomalous Halleffects have been observed in twisted bilayer MoTe2 (tMoTe2) at one holedoping and fractional doping per moiré unit cell, respectively. When the sec-ondmoiré band is fully hole doped, quantum spin Hall insulator has also beenreported in tMoTe2 at a certain twist angle. Exotic topological states togetherwith ferromagnetic (FM) states in the high moiré band can potentially exist aswell. In this study,we report the observation of a FMphase in the secondmoiréband in tMoTe2. The FM phase can be tuned by both the doping level anddisplacement field. At filling around 2.58 holes per moiré unit cell, the FMphase reaches a Curie temperature of 3.5 K. A large displacement field cansuppress the FM phase, like the FM phase at the filling of −1. Our resultsdemonstrate the realization of time-reversal symmetry-breaking states in thehigher moiré bands in tMoTe2.Stacking layeredmaterialswith small twist angles provides a solid-stateplatform for exploring the interplay of correlation and topology1–3.Extremely flat bands quench electron kinetic energy and enhanceelectronmany-body correlation effects,which canbe controlledby thetwisted angle4, stacking order5,6, and displacement field. Robust Cherninsulators have been observed in many moiré systems, like twistedbilayer graphene/hBN superlattice7,8, AB stacked MoTe2/WSe25,tMoTe29–14 and rhombohedral graphenemultilayers aligned to hBN15–17.Among these, tMoTe2, whichhosts an effectivehoneycomb latticefor low-energy carriers, has emerged as an important platformto studythe interplay between band topology and electron correlation. Thesystem at integer filling factor ν = −1 (one hole permoiré unite cell) andfractional fillings −2/3, and −3/5 have been confirmed to exhibit,respectively, integer and fractional quantum anomalous Hall effectsboth through optical and electrical transport method10,18. Many theo-retical works have been devoted to exploring the Chern insulating andfractional Chern insulating states19–28. Till now,most of the experimentwork on tMoTe2 mainly focus on the Chern insulator at filling −19 andfractional fillings below −1, like −2/3, and −3/510,11,14. These fractionalquantum anomalous Hall states can be understood from compositeReceived: 8 August 2024Accepted: 30 April 2025Check for updates1Division of Physics andApplied Physics, School of Physical andMathematical Sciences, Nanyang Technological University, Singapore, Singapore. 2School ofPhysics and Technology, Wuhan University, Wuhan, China. 3Department of Physics, School of Science and Research Center for Industries of the Future,Westlake University, Hangzhou, China. 4Institute of Natural Sciences, Westlake Institute for Advanced Study, Hangzhou, China. 5Centre for QuantumTechnologies, Nanyang Technological University, Singapore, Singapore. 6School of microelectronics, University of Science and Technology of China,Hefei, China. 7Research Center for Electronic and Optical Materials, National Institute for Materials Science, Tsukuba, Japan. 8Research Center for MaterialsNanoarchitectonics, National Institute for Materials Science, Tsukuba, Japan. 9School of Electrical and Electronic Engineering, Nanyang TechnologicalUniversity, Singapore, Singapore. 10National Centre for Advanced Integrated Photonics (NCAIP) Singapore, Nanyang Technological University,Singapore, Singapore. 11Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore, Singapore. 12These authors con-tributed equally: Liheng An, Haiyang Pan, Wen-Xuan Qiu. e-mail: wufcheng@whu.edu.cn; wbgao@ntu.edu.sgNature Communications |         (2025) 16:5131 11234567890():,;1234567890():,;http://orcid.org/0009-0005-0816-5963http://orcid.org/0009-0005-0816-5963http://orcid.org/0009-0005-0816-5963http://orcid.org/0009-0005-0816-5963http://orcid.org/0009-0005-0816-5963http://orcid.org/0000-0002-5119-0452http://orcid.org/0000-0002-5119-0452http://orcid.org/0000-0002-5119-0452http://orcid.org/0000-0002-5119-0452http://orcid.org/0000-0002-5119-0452http://orcid.org/0000-0003-4808-4795http://orcid.org/0000-0003-4808-4795http://orcid.org/0000-0003-4808-4795http://orcid.org/0000-0003-4808-4795http://orcid.org/0000-0003-4808-4795http://orcid.org/0000-0002-8634-3834http://orcid.org/0000-0002-8634-3834http://orcid.org/0000-0002-8634-3834http://orcid.org/0000-0002-8634-3834http://orcid.org/0000-0002-8634-3834http://orcid.org/0000-0002-3432-2986http://orcid.org/0000-0002-3432-2986http://orcid.org/0000-0002-3432-2986http://orcid.org/0000-0002-3432-2986http://orcid.org/0000-0002-3432-2986http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0003-3701-8119http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1467-3105http://orcid.org/0000-0002-1185-0073http://orcid.org/0000-0002-1185-0073http://orcid.org/0000-0002-1185-0073http://orcid.org/0000-0002-1185-0073http://orcid.org/0000-0002-1185-0073http://orcid.org/0000-0003-3971-621Xhttp://orcid.org/0000-0003-3971-621Xhttp://orcid.org/0000-0003-3971-621Xhttp://orcid.org/0000-0003-3971-621Xhttp://orcid.org/0000-0003-3971-621Xhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-59691-5&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-59691-5&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-59691-5&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-59691-5&domain=pdfmailto:wufcheng@whu.edu.cnmailto:wbgao@ntu.edu.sgwww.nature.com/naturecommunicationsFermi Liquid picture based on themapping from the first moiré Chernband to the lowest Landau level19,20. Theories also predict the possi-bility of the second moiré band mimicking the behavior of the firstLandau level29–31. Possible ν = −3/2 non-Abelian states are proposed inthe secondmoiré band30. A recent experiment has shown the evidenceof the existence of quantum spin Hall states32,33 in fillings −2 and −4.Evidence of Chern insulating states is also reported in twist bilayerWSe2 by finely controlling the twist angle, at fillings −1 and −334.In this study, we report the observation of FM states in the secondmoiré band, from ν = −2.4 to −2.58. The twist angle of tMoTe2 understudy is at 3.2° with moiré density around nM ≈ 2.7 × 1012cm−2. Weperformed reflective magnetic circular dichroism (RMCD) to investi-gate themagnetic properties. The RMCD hysteresis loops revealed theexistence of ferromagnetism at fillings around both −1 and −2.5. Thisfinding indicates the time reversal symmetry breaking states in highermoiré bands. Furthermore, we investigated the doping dependent andelectric field dependent magnetism in the second moiré band.ResultsDevice structure and characterizationThe tMoTe2 device is fabricated into dual gate structures as shown inFig. 1a. We adopt similar tear and stack method9–11 to fabricate tMoTe2as used in previous work9,10. Figure 1b shows the schematic image ofthe AA-stacked tMoTe2. In each moiré unit cell, there are three high-symmetry stacking sites as labeled by MM (transition-metal atoms ontransition-metal atoms), MX (transition-metal atoms on top of chal-cogen atoms), XM (chalcogen atomson topof transition-metal atoms).For the full filling doping (ν = −2), the holes equally populate on MXand XM sublattices, which form a honeycomb lattice in real space.Few-layer graphite is used to form both the top gate and bottomgate with hexagonal boron nitride (hBN) as dielectric material, shownin Supplementary Fig. 3d. In addition, the whole twisted bilayer isgrounded with another graphite electrode. The voltage sources con-trol the top (VTG) and bottom (VBG) gate voltages, which allows inde-pendent control of both the carrier density and the displacement field.For our devices, small intrinsic doping and built-in displacement fieldare observed, possibly owing to the imperfection sample fabricationprocess, which has been calibrated by the dual gate photo-luminescence (PL) map measurement and RMCD measurement.A continuous-wave 780 nm laser is used to excite the tMoTe2 forPL emission35. The measurement is kept at 1.6 K unless otherwisespecified. The PL signal for emission wavelength above 1064nm iscollected by Superconducting nanowire single-photon detector(SNSPD). As shown in Supplementary Fig. 4d, the real space PL map ismeasured for the whole device. We select the homogeneous areas tocollect the PL signal. The tMoTe2 system with small twist angles like 3-4° is a direct band gap semiconductor9. Figure 1d illustrates the PLsignal versus thedoping level along the yellowdashed linecut in Fig. 1e.We use trion emission as a sensitive probe to characterize incom-pressible states in tMoTe2. At particular integer fillings like −1 and −2,the free carriers are reduced due to charge gap formation, suppressingthe emission of trion. These dips in PL counts correspond to incom-pressible states at integer fillings like ±2 and ±1.To verify whether the filling −1 state realizes a Chern insulator, thePL counts aremeasured as a function of the doping and an applied out-of-plane magnetic field, as presented in Supplementary Fig. 5. Theincompressible state at filling −1 disperses linearly with the magneticfield. The Chern number is extracted to be −1, which is consistent withprevious work10.In Fig. 1e, wemeasure the PL emission as a function of VTG and VBGand plot it as a function of doping and displacement field. Through thePLmap, we deduce the ratio of the top gate and back gate capacitance.With these calibrations and themeasurement of hBN thickness (seen inSupplementary Fig. 6), it allows us to identify the twist angle for thisFig. 1 | Gate-tunable correlated states in tMoTe2 devices. a Schematic image ofthe device structure. The tMoTe2 sample is grounded. The top gate and bottomgate are applied through hBN (10 to 30 nm). b Moiré superlattice of AA stackedtMoTe2. The red, orange, and green circlesmark high-symmetry stacking sitesMM,MX, and XM, respectively. c Side view of MM, MX, and XM sites. d PL counts as afunction of doping level, with displacement field (D/ε0) almost at zero, along thelinecut of the yellow dashed line shown in (e). e Steady state photoluminescence(PL) versus carrier density (filling factor) and displacement field.Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 2www.nature.com/naturecommunicationsdevice #O1 to be around 3.2 ± 0.1°. We use the carrier density differ-ence between integer filling +1 and +2 as moiré density11, shown inSupplementary Fig. 7.The ferromagnetic phase in the second moiré bandWe performed RMCD measurements36,37 to probe the magnetism.Figure 2a plots the RMCD intensity as a function of the doping anddisplacement field. A small out-of-plane magnetic field B = 30 mT isapplied to suppress the RMCD fluctuation. The signature of FM state isobserved in the vicinity of filling −1, ranging from −0.6 to −1.2, anddisplacement field from +0.12 V/nm to −0.12 V/nm. The RMCD signalon different sample places is shown in Supplementary Fig. 8, whichindicates that our device exhibits a homogeneous RMCD signal over afew micrometers. In addition, another filling range with non-zeroRMCD is found between filling −2.4 to −2.58, a direct evidence of FMphase formation in the second moiré band.At fixed carrier densities, we performed RMCDmeasurement at alarger magnetic field range from B = −0.4T to 0.4T. As shown inFig. 2b, the RMCD intensity is plotted as a function of filling factor andmagnetic field. The displacement field is kept at around zero. TheRMCD signal is most pronounced in the first moiré band, from fillingν = −0.75 to −1.2. For the state at ν = � 2, no RMCD signal is observed,indicating the absence of ferromagnetism. Therefore, the filling −2state is likely a quantum spin Hall insulator, of which the band struc-ture calculated with Hartree-Fock (HF) approximation is shown inSupplementary Fig. 1b. The RMCD signal at filling −2.5 saturates above50mT and reach ~0.3%. The RMCD signal in the secondmoiré band is afew times smaller than the signal in the first moiré band. This can bedue to different bandwidth and interaction strength in the first andsecond bands.To resolve the hysteresis loop, we perform wavelength-dependent RMCD measurement9. Based on the PL spectrum oftMoTe2 as shown in Supplementary Fig. 9, we selected the excitationlaser wavelength of 1100nm to 1125 nm for the RMCD measurement.Although the RMCD signal shows different slope based on the wave-length used (Supplementary Fig. 10), an obvious jump of the RMCDsignal is observed from −0.05 T to 0.05 T in all cases, indicating theformation of FM states at filling −2.58. Based on these measurements,the wavelength 1121 nm is selected since it gives the largest RMCDcontrast. We note that the filling -2.58 represents the limit of dopinglevel achievable in device #O1, but not the boundary of FM states.Figure 2d–g shows the RMCD hysteresis loop at fillings 0, −1, −2,and −2.5. The orange (blue) color line is for up (down) sweeps. Asdoping level is set at υ=0, no hysteresis loop is resolved and theRMCDsignal is almost zero between magnetic field from −40mT to 40mT.The coercive field for υ= � 1 is about 10 mT, and the difference inRMCD amplitude for up anddown sweeps is about 4%. As the doping isincreased to around υ= � 2 no hysteresis loop or RMCD signal isdetected. When further doping more carriers into the second moiréband, another pronounced hysteresis loop of the RMCD signal isshown, which is a direct evidence of FM state. The coercive field offilling −2.5 is around 10mT, and the RMCD amplitude is around 0.4%,smaller than that of υ= � 1The above feature of RMCD signal is consistent with our numer-ical results. Based on HF approximation, we perform calculation onout-of-plane FM phase for filling −1 to −3. The energy of this phaserelative to the symmetric phase is shown in Supplementary Fig. 2,which indicates that the FMphase is favored around filling ν = −3 ~ −2.2and −1.6 ~ −1. As the filling approaches −2 from either side, the ferro-magnetism is continuously suppressed. The ferromagnetism at υ= �2:5 is revealed by the HF band structure, as shown in Fig. 2c, where thesecond band with spin (valley) polarization is partially filled by holesand contributes to the observed ferromagnetism. Here υ= � 2:5 is ametallic state and has a weaker ferromagnetism compared to the FMChern insulator at υ= � 1:In our calculation,wealso consider the in-plane antiferromagneticstate, denoted as AFMx state, as the competing state. The energycompetitions among the three types of states (i.e., symmetric, FM,AFMx) is shown in the supplementary Fig. 1. Thenumerical results showthat the AFMx state could be the ground state for ν from�2:2 to�1:6.This state can be understood from the Kane-Mele-Hubbard model25,which is a phenomenological model for understanding the interactingphysics in tMoTe2. Electron-electron interaction candrive an instabilityin the Kane-Mele-Hubbard model towards an antiferromagnetic Motta b cd e f gFig. 2 | Ferromagnetism in the first and second moiré band. a Reflective mag-netic circular dichroism (RMCD) signal as a function of the doping filling factor νandelectricfield (D/ɛ0). Thewavelength is at 1120.9 nm.bRMCD signal as a functionof out-of-plane magnetic field B and filling factor ν from 0 to −2.4. c CalculatedHartree-Fock (HF) band structure at υ= � 2:5 in the metallic FM phase. The solidand dotted lines, respectively, plot spin up and down bands. The Fermienergy marked by the horizontal black dashed line is set to 0. d–f RMCD signalat filling factor 0, −1, and −2, respectively, with the displacement field nearD/ɛ0 = 0mV/nm. g RMCD signal at filling factor −2.5, with the displacement fieldD/ɛ0 ≈ 10mV/nm.Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 3www.nature.com/naturecommunicationsinsulating state at ν = � 2. However, the mean-field calculation canoverestimate this tendency towards symmetry-breaking states. Giventhe small mean-field energy differences between the AFMx and sym-metric states around ν = � 2, it is fully possible that the symmetricstate is the actual ground state around ν = � 2: In experiment, theRMCD signal vanishes at ν = � 2 even in the presence of an out-of-plane magnetic field, as shown in Fig. 2f. This experimental resultstrongly indicates that the ν = � 2 state is non-magnetic, which sug-gests the overestimation of the symmetry-breaking states in themean-field theory. However, our theoretical calculations can qualitativelycapture the doping ranges where FM state occurs, since the energygain of the FM state compared to the symmetric state can besignificant.Experimentally, the FM hysteresis loops are sensitive to the exci-tation laser power we used. We kept the laser power at 20 nW to avoidlaser heating effects. As shown in Supplementary Fig. 11, the hysteresisloop will gradually disappear when further increasing the power ofresonant laser power to 200 nW. This is possibly due to the heatingeffect that can break down the FM states.Doping and electrical control of the ferromagnetic phaseThe FM states can be tuned effectively by doping and displacementfield. RMCD signal near υ= � 1 is scanned as a function of filling factorwith zero displacement field, as shown in Supplementary Fig. 12. Thefilling −1 state gives the strongest coercivefield. Similarly, Fig. 3a showsRMCD signal as a function of the filling factor υ and the out-of-planemagnetic field. The difference of RMCD for the forward and backwardmagnetic field scanning is shown in Fig. 3b. The coercive field forυ<� 2 showsamaximumat aroundfilling −2.5. RMCDhysteresis loopsare shown in Fig. 3c. The FM phase in second moiré band is repro-ducible in another device#O2 with a twist angle 2.8° (SupplementaryFig. 13). For device#O2, the FM hysteresis loop is most pronounced atfilling −2.78 (Supplementary Fig. 14) within the achievable dop-ing range.We also measure the RMCD signal as a function of the displace-ment field (Fig. 4). The hysteresis loop has the trend of suppression bythe increasing of displacement field. The coercive field graduallydecreases from 1.5mT at electric field 13mV/nm to 0.5 mT at 47mV/nm. An unsymmetric coercive field is observed for all the electric fieldas shown in Fig. 4c, of which the microscopic origin still needs moreinvestigation. The electric field dependence is also observed in Devi-ce#O2 (Supplementary Fig. 15).Cuire temperature of the FM phaseTo investigate the magnetic interaction strength, we measure RMCDsignal as a function of temperature. At a low temperature T = 1.6 K, aclear RMCD signal jump can be detected, which indicates FM interac-tion between spin moments at filling −2.58. For the high temperaturelike 7 K, the RMCD curve shows a linear relation with the magneticfield, which is the behavior of the paramagnetic phase. We extract theslope of RMCD to calibrate themagnetic susceptibility χ = ∂R∂H jH =09. Wefit the inverse of the slope in the paramagnetic phase as a function ofthe temperature following the Cuire-Weiss law described by χ = CT�θC,as shown in Fig. 5b. By fitting the curve, a positive θC =3:5K is extrac-ted, which is the FM Curie temperature.The temperature dependence of RMCD signal is also measuredwith optimal excitation laser wavelength 1120.9 nm for ν below−2.4, asshown in Fig. 5c. We measure the RMCD signal by sweeping the mag-netic forward and backward from T = 1.6K to 4.5 K. At this wavelength,the RMCD signal shows hysteresis loop that gradually disappears asthe temperature increases. The temperature at which the hysteresisloop vanishes leads to a Curie temperature of Tc ~ 3.5 K at ν = −2.58,consistent with the above estimate. TheCurie temperature obtained inthis way shows doping dependence when the filling is tuned from −2.4to −2.58, which gradually decreases when ν deviates from −2.58.Previous work in small twist angle 2.1° has shed light to realizefractional quantum spin Hall insulating states in secondmoiré band intMoTe2 at ν = � 3. The investigation of high order moiré band opensFig. 3 |DopingdependentRMCD fromfilling−2.3 to −2.58. aRMCDsignals versusfilling andmagneticfield swept up and sweptdown.bHysteresis loopof theRMCDasafunction of magnetic field and filling factor. c RMCD signal as a function of magnetic field from filling −2.38 to −2.58.Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 4www.nature.com/naturecommunicationsup the possibility of realizing more topological states, including non-Abelian states29,30,33,38,39. Our study demonstrated that FM phase canalso appear in the second moiré band and is tunable by both dopinglevel28 and the displacement field. As can be expected in the nearfuture, different twist angles and multi-layer structures40 can lead toother possible topological phases, making the second moiré band asan interesting place to study both many-body correlation and bandtopology.Fig. 4 | Electric filed dependent RMCD at filling −2.58. a RMCD signals versus displacement field and magnetic field swept up and swept down. bHysteresis loop of theRMCD versus displacement field. c RMCD signal at filling −2.58 with displacement field from 13mV/nm to 47mV/nm.Fig. 5 |TCof ferromagnetic phase atfilling−2.58. aTemperature dependent RMCDat filling factor υ= � 2:58. The measurement wavelength is at 1104.7 nm. b Curie-Weiss fit (dashed line) of the inverse RMCD slope at vanishing magnetic field versustemperature, which leads toTCof 3.5K. cTemperature dependent RMCDhysteresis atfilling factor υ= � 2:58. The measurement wavelength is at 1120.9nm. d Curie tem-perature, at which the RMCD hysteresis loop vanishes, as a function of filling factor.Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 5www.nature.com/naturecommunicationsMethodsDevice fabricationThe dual-gate tMoTe2 was fabricated via a tear-and-stack technique10,11,as shown in Supplementary Fig. 3. Bulk crystal MoTe2 was purchasedfromHQGraphene. First, few-layer graphite and hBN (10 to 30 nm), asshown in Supplementary Fig. 6, were mechanically exfoliated onto asilicon wafer. The cleanness of the top surface was checked using bothan optical microscope with dark field and Atomic Force Microscopy(AFM). Polydimethylsiloxane (PDMS) and Polycarbonate (PC) wereused to pick up the bottom hBN (the bottom substrate) and bottomgraphite (the bottom gate). The whole stack was annealed at 200 °C ina forming gas atmosphere to eliminate residues at the interface. Theexfoliation and transfer process was performed in a nitrogen-filledglovebox, where the water and oxygen concentrations were keptbelow 0.1 ppm. Monolayer MoTe2 was mechanically exfoliated onto aSi/SiO2 wafer using Scotch tape. The monolayer MoTe2 was cut in halfwith an AFM tip. Then, another layer of top graphite (5-10 layers) /tophBN (10 to 30 nm) stack was used to pick up the two parts of MoTe2sequentially. The twist angle was set by rotating the bottom transferstage by 3°. Subsequently, another 5 to 10 layers of graphite werepicked up and contacted with the bottom MoTe2 to form the ground.The whole stack was dropped onto the pre-patterned bottom elec-trode (Cr/Au, 5/30 nm) to form the complete electric contact. The PCfilm on top was washed with a chloroform rinse followed by an iso-propyl alcohol (IPA) rinse.Optical measurements set-upAll optical measurements were performed in a home-built confocaloptical microscope. The sample was mounted in the cryostat Auto-dry 2100 equipped with a 9 T superconducting magnet at basetemperature 1.6 K. A × 50 non-magnetic objective (numerical aper-ture, 0.63) with a spot size of 1 um was used to collect PL and RMCDsignal. For the PL measurement, a M-squared laser with tunningrange 700nm–1000 nm was used as the excitation source. A1064 nm long pass filter is placed before the entrance to the spec-trometer to remove the reflected laser excitation. The PL excitationpower ranges from 200nW to 2 uW, and the gate dependent PLcounts signal is obtained by superconducting single-photon detector(SNSPD). Princeton spectrometer equipped with a liquid-nitrogen-cooled InGaAs charge-coupled device (CCD) detector is used toanalyse the PL spectrum.The reflective magnetic circular dichroism (RMCD) measure-ment was performed with a continuous wave laser (HUBNER Pho-tonics C-WAVE VIS), which produced a narrow MHz bandwidthlaser, ranging 1080 nm to 1130 nm. The set up is shown in Supple-mentary Fig. 16. The excitation laser was first passed through anoptical chopper at frequency f1 912 Hz and then polarized 45° to thephotoelastic modulator (PEM) fast axis with a maximum retardanceof λ/4. The departing circularly polarized light oscillates betweenright circular polarized light and left circular polarized light in asinusoidal time-dependence with a frequency f2 50.5 kHz. Thereflected signal from the sample is reflected by a beamsplitter andsent into an InGaAs avalanche photodiode (Thorlabs APD430C/M).The photodiode current was amplified and turned into voltagesignal by SR570 and sent into two lock-in amplifiers set at 50.5 kHzand 912 Hz, giving the RMCD signal and the laser excitation inten-sity, respectively.Carrier density calculationThe carrier density is n = (VTGCTG +VBGCBG)/e-n0 and displacementfield D/ɛ0 = (VTGCTG − VBGCBG)/2ɛ0-D0/ɛ0, where CTG and CBG are thetop and bottom gate capacitance obtained from the devicegeometry, e is the electron charge, and ɛ0 is the vacuum permittivity.For device#O1, the top BN thickness is 19 nm, andCTG/e = ɛɛ0/d = 0.87 × 1012cm−2. The bottom BN thickness is 24 nm, andCBG/e = ɛɛ0 /d =0.69 × 1012cm−2. For device#O2, the top BN thickness is44 nm, and CTG/e = ɛɛ0/d =0.38 × 1012cm−2. The bottom BN thickness is40 nm, and CBG/e = ɛɛ0 /d = 0.41 × 1012cm−2. The offset carrier densityn0 is derived from fitting the integer filling in dual gate PL spectramap.The offset displacement field D0 is determined from the symmetricaxis of the dual gate RMCD map.Note added—Recently, two complementary and independentstudy41,42 appeared on Nature Physics.Data availabilityAll data within the article and the Supplementary Information thatsupport the findings of this study are available, or from the corre-sponding author F.W. orW.B.Guponrequest. Sourcedata areprovidedwith this paper.Code availabilityThe code that supports the findings of this study is available from thecorresponding author upon request.References1. Wu, F., Lovorn, T., Tutuc, E., Martin, I. & MacDonald, A. Topologicalinsulators in twisted transitionmetal dichalcogenide homobilayers.Phys. Rev. Lett. 122, 086402 (2019).2. Yu, H., Chen, M. & Yao,W. Giantmagnetic field frommoiré inducedBerry phase in homobilayer semiconductors. Natl Sci. Rev. 7,12–20 (2020).3. Xie, Y. et al. Fractional Chern insulators in magic-angle twistedbilayer graphene. Nature 600, 439–443 (2021).4. Devakul, T., Crépel, V., Zhang, Y. & Fu, L. Magic in twisted transitionmetal dichalcogenide bilayers. Nat. Commun. 12, 6730 (2021).5. Li, T. et al. Quantum anomalous Hall effect from intertwined moirébands. Nature 600, 641–646 (2021).6. Li, T. et al. Continuous Mott transition in semiconductor moirésuperlattices. Nature 597, 350–354 (2021).7. Serlin, M. et al. Intrinsic quantized anomalous Hall effect in a moiréheterostructure. Science 367, 900–903 (2020).8. Sharpe, A. L. et al. Emergent ferromagnetism near three-quarters filling in twisted bilayer graphene. Science 365,605–608 (2019).9. Anderson, E. et al. Programming correlated magnetic states withgate-controlled moiré geometry. Science 381, 325–330 (2023).10. Cai, J. et al. Signatures of fractional quantum anomalous Hall statesin twisted MoTe2. Nature 622, 63–68 (2023).11. Zeng, Y. et al. Thermodynamic evidence of fractional Chern insu-lator in moiré MoTe2. Nature 622, 69–73 (2023).12. Xu, F. et al. Observation of integer and fractional quantumanomalous Hall effects in twisted bilayer MoTe 2. Phys. Rev. X13, 031037 (2023).13. Ji, Z. et al. Local probe of bulk and edge states in a fractional Cherninsulator. Nature 635, 578–583 (2024).14. Redekop, E. et al. Direct magnetic imaging of fractional Cherninsulators in twisted MoTe2. Nature 635, 584–589 (2024).15. Lu, Z. et al. Fractional quantum anomalous Hall effect in multilayergraphene. Nature 626, 759–764 (2024).16. Han, T. et al. Large quantum anomalous Hall effect in spin-orbitproximitized rhombohedral graphene. Science 384, 647–651(2024).17. Sha, Y. et al. Observation of a Chern insulator in crystallineABCA-tetralayer graphene with spin-orbit coupling. Science 384,414–419 (2024).18. Park, H. et al. Observation of fractionally quantized anomalous Halleffect. Nature 622, 74–79 (2023).19. Li, B. &Wu, F. Variationalmapping of Chern bands to Landau levels:Application to fractional Chern insulators in twisted MoTe 2. Phys.Rev. B 111, 125122 (2025).Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 6www.nature.com/naturecommunications20. Dong, J., Wang, J., Ledwith, P. J., Vishwanath, A. & Parker, D. E.Composite Fermi liquid at zero magnetic field in twisted MoTe 2.Phys. Rev. Lett. 131, 136502 (2023).21. Goldman, H., Reddy, A. P., Paul, N. & Fu, L. Zero-field compositeFermi liquid in twisted semiconductor bilayers. Phys. Rev. Lett. 131,136501 (2023).22. Zhang, X.-W. et al. Polarization-driven band topology evolution intwisted MoTe2 and WSe2. Nat. Commun. 15, 4223 (2024).23. Jia, Y. et al. Moiré fractional Chern insulators. I. First-principlescalculations andcontinuummodels of twistedbilayerMoTe 2. Phys.Rev. B 109, 205121 (2024).24. Reddy, A. P. & Fu, L. Toward aglobal phasediagramof the fractionalquantum anomalous Hall effect. Phys. Rev. B 108, 245159 (2023).25. Qiu, W.-X., Li, B., Luo, X.-J. & Wu, F. Interaction-driven topologicalphase diagram of twisted bilayer MoTe 2. Phys. Rev. X 13, 041026(2023).26. Yu, J. et al. FractionalChern insulators versus nonmagnetic states intwisted bilayer MoTe 2. Phys. Rev. B 109, 045147 (2024).27. Xu, C., Li, J., Xu, Y., Bi, Z. & Zhang, Y. Maximally localized Wannierfunctions, interaction models, and fractional quantum anomalousHall effect in twisted bilayer MoTe2. Proc. Natl Acad. Sci. USA 121,e2316749121 (2024).28. Gonçalves, M. & Lin, S.-Z. Doping-induced Quantum AnomalousHall Crystals and Topological Domain Walls. arXiv preprintarXiv:2407.12198 (2024).29. Ahn, C.-E., Lee, W., Yananose, K., Kim, Y. & Cho, G. Y. Non-Abelianfractional quantum anomalous Hall states and first Landau levelphysics of the second moiré band of twisted bilayer MoTe 2. Phys.Rev. B 110, L161109 (2024).30. Xu, C.,Mao, N., Zeng, T. & Zhang, Y.MultipleChern bands in twistedMoTe 2 and possible non-Abelian states. Phys. Rev. Lett. 134,066601 (2025).31. Wang, T. et al. Topology, magnetism and charge order in twistedMoTe2 at higher integer hole fillings. arXiv preprint arXiv:2312.12531(2023).32. Kang, K. et al. Evidence of the fractional quantum spin Hall effect inmoiré MoTe2. Nature 628, 522–526 (2024).33. Zhang, Y.-H. Non-Abelian and Abelian descendants of a vortex spinliquid: fractional quantum spin Hall effect in twisted MoTe 2. Phys.Rev. B 110, 155102 (2024).34. Foutty, B. A. et al. Mapping twist-tuned multiband topology inbilayer WSe2. Science 384, 343–347 (2024).35. Tan, Q. et al. Layer-dependent correlated phases in WSe2/MoS2moiré superlattice. Nat. Mater. 22, 605–611 (2023).36. Zhang, Z. et al. Controlled large non-reciprocal charge transport inan intrinsic magnetic topological insulator MnBi2Te4. Nat. Com-mun. 13, 6191 (2022).37. Qiu, W.-X. & Wu, F. Quantum geometry probed by chiral excitonicoptical response of Chern insulators. Phys. Rev. B 111, L121104(2025).38. Wang,C. et al. Higher Landau-Level Analogs andSignatures ofNon-Abelian States in Twisted Bilayer MoTe 2. Phys. Rev. Lett. 134,076503 (2025).39. Reddy, A. P., Paul, N., Abouelkomsan, A. & Fu, L. Non-Abelianfractionalization in topological minibands. Phys. Rev. Lett. 133,166503 (2024).40. Song, Z., Qi, J., Liebman, O. & Narang, P. Collective spin in twistedbilayer materials. Phys. Rev. B 110, 024401 (2024).41. Park, H. et al. Ferromagnetism and topology of the higher flat bandin a fractional Chern insulator. Nat. Phys. 21, 1–7 (2025).42. Xu, F. et al. Interplay between topology and correlations in thesecond moiré band of twisted bilayer MoTe2. Nat. Phys. 21,1–7 (2025).AcknowledgementsWe thank Jinxin Hu, Justin Song, Bo Yang, Jiaqi Cai, Xiaodong Xu forhelpful discussions. This work is supported by ASTAR (M21K2c0116,M24M8b0004), Singapore National Research foundation (NRF-CRP22-2019-0004), Temasek Laboratories (TLSP24-04) and CQT bridge funds.W.-X. Q. and F.W. are supported by National Key Research and Devel-opment Program of China (Grants No. 2021YFA1401300 and No.2022YFA1402401), National Natural Science Foundation of China (GrantNo. 12274333 and No. 12404084). W.-X. Q. is also supported by theChina Postdoctoral Science Foundation (Grants No. 2024T170675 andNo. 2023M742716). The numerical calculations in this paper have beendone on the supercomputing system in the Supercomputing Center ofWuhan University.Author contributionsH.P. fabricated the devices. L.A. performed the optical measurements.X.D. and S.R. developed the software and analysis code. L.A., H.P., F.W.and W.G. analyzed the data. F.W. and W.-X. Q. performed theoreticalstudies. K.W. and T.T. grew the bulk hBN crystals. L.A. andW.G. togetherwith all coauthors, wrote the manuscript. All authors, including N.W.,Q.T., X.C., Q.S., S.Y., X.F.L., H.J. and X.D.L., discussed the results andcommented on the manuscript.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-025-59691-5.Correspondence and requests for materials should be addressed toFengcheng Wu or Wei-bo Gao.Peer review information Nature Communications thanks U. Chandniwho co-reviewed with Saisab Bhowmik, and the other, anonymous,reviewer(s) for their contribution to the peer review of this work. A peerreview file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard tojurisdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution-NonCommercial-NoDerivatives 4.0 International License,which permits any non-commercial use, sharing, distribution andreproduction in any medium or format, as long as you give appropriatecredit to the original author(s) and the source, provide a link to theCreative Commons licence, and indicate if you modified the licensedmaterial. Youdonot havepermissionunder this licence toshare adaptedmaterial derived from this article or parts of it. The images or other thirdparty material in this article are included in the article’s CreativeCommons licence, unless indicated otherwise in a credit line to thematerial. If material is not included in the article’s Creative Commonslicence and your intended use is not permitted by statutory regulation orexceeds the permitted use, you will need to obtain permission directlyfrom the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.© The Author(s) 2025Article https://doi.org/10.1038/s41467-025-59691-5Nature Communications |         (2025) 16:5131 7https://doi.org/10.1038/s41467-025-59691-5http://www.nature.com/reprintshttp://creativecommons.org/licenses/by-nc-nd/4.0/http://creativecommons.org/licenses/by-nc-nd/4.0/www.nature.com/naturecommunications Observation of ferromagnetic phase in the second moiré band of twisted MoTe2 Results Device structure and characterization The ferromagnetic phase in the second moiré band Doping and electrical control of the ferromagnetic phase Cuire temperature of the FM phase Methods Device fabrication Optical measurements set-up Carrier density calculation Data availability Code availability References Acknowledgements Author contributions Competing interests Additional information