# Fileset

[ds5p-763x.pdf](https://mdr.nims.go.jp/filesets/74766421-7fc6-46fe-a0c7-5fdef5635ba9/download)

## Creator

Eric Anderson, Heonjoon Park, Kaijie Yang, Jiaqi Cai, [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), Liang Fu, Ting Cao, Di Xiao, Xiaodong Xu

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Magnetoelectric Control of Helical Light Emission in a Moiré Chern Magnet](https://mdr.nims.go.jp/datasets/60b8025e-773d-4079-8f6f-c5082d90c433)

## Fulltext

Magnetoelectric Control of Helical Light Emission in a Moiré Chern MagnetMagnetoelectric Control of Helical Light Emission in a Moiré Chern MagnetEric Anderson ,1,* Heonjoon Park,1 Kaijie Yang,2 Jiaqi Cai ,1 Takashi Taniguchi ,3Kenji Watanabe ,4 Liang Fu,5 Ting Cao,2 Di Xiao ,2,1,† and Xiaodong Xu 1,2,‡1Department of Physics, University of Washington, Seattle, Washington 98195, USA2Department of Materials Science and Engineering,University of Washington, Seattle, Washington 98195, USA3Research Center for Materials Nanoarchitectonics, National Institute for Materials Science,1-1 Namiki, Tsukuba 305-0044, Japan4Research Center for Electronic and Optical Materials, National Institute for Materials Science,1-1 Namiki, Tsukuba 305-0044, Japan5Department of Physics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA(Received 27 February 2025; revised 23 June 2025; accepted 22 July 2025; published 27 August 2025)Magnetoelectric effects and their coupling to light helicity are important for both fundamental scienceand applications in sensing, communication, and data storage. Traditional approaches require complexdevice architectures, involving separate spin-injection, ferromagnetic, and optically active layers. Recently,the emergence of 2D semiconductor moiré superlattices with flat Chern bands and strong light-matterinteractions has established a simple yet powerful platform for exploring the coupling between photon,electron, and spin degrees of freedom. Here, we report efficient current control of spontaneousferromagnetism and associated helicity of light emission in moiré MoTe2 bilayer—a system which hostsa rich variety of topological phases, including newly discovered zero-field fractional Chern insulators. Weshow that the current control is effective over a wide range of doping of the first moiré Chern band,implying the uniformity of the Berry curvature distribution over the flat band. By setting the system into theanomalous Hall metal phase, a current as small as 10 nA is sufficient to switch the magnetic order, asubstantial improvement over both conventional spin-torque architectures and other moiré systems. Therealized current control of ferromagnetism leads to continuous tuning of trion photoluminescence helicityfrom left to right circular via spin-valley Hall torque at zero magnetic field. Our results pave the way fortopological optospintronics based on semiconductors with synthetic flat Chern bands.DOI: 10.1103/ds5p-763x Subject Areas: Optoelectronics, Spintronics,Strongly Correlated MaterialsI. INTRODUCTIONMoiré superlattice engineering has been proven to be aneffective approach for creating new phases of matter withexotic physical properties [1–3]. A striking example is therecent discovery of the fractional quantum anomalousHall (FQAH) effect in twisted MoTe2 bilayer [4–10](tMoTe2) and rhombohedral-stacked multilayer graphene,[11] realizing long-sought zero-field fractional Cherninsulators. [12–18] Besides hosting the FQAH effect, thesemiconducting transition-metal-dichalcogenide (TMD)tMoTe2 is a direct band-gap semiconductor, with spin-valley locking and a strong excitonic response [19–21].These properties have enabled optical probing of the Cherninsulator [4] and putative composite Fermi liquid [10] statesvia measurement of the intensity and helicity of trionphotoluminescence. In addition, both electric field anddoping have been demonstrated to be effective controls ofthe spontaneous ferromagnetic order in tMoTe2 [21],enabling powerful magnetoelectric manipulation.The physical properties of tMoTe2 mentioned aboveoffer a new means to manipulate intrinsic magnetism—andthus light-emission helicity, which is tightly linked withspin-valley polarization—via electric current. Current con-trol of light-emission helicity is highly sought after, as itinvolves coupling the relevant degrees of freedom inelectronics, spintronics, and photonics. This functionalityhas recently been achieved in complex multilayerarchitectures [22]. For these devices, charge current in a*Present address: Université Paris-Saclay, CNRS, Centrede Nanosciences et de Nanotechnologies (C2N), Palaiseau91120, France.†Contact author: dixiao@uw.edu‡Contact author: xuxd@uw.eduPublished by the American Physical Society under the terms ofthe Creative Commons Attribution 4.0 International license.Further distribution of this work must maintain attribution tothe author(s) and the published article’s title, journal citation,and DOI.PHYSICAL REVIEW X 15, 031057 (2025)2160-3308=25=15(3)=031057(9) 031057-1 Published by the American Physical Societyhttps://orcid.org/0000-0002-1357-6645https://orcid.org/0000-0002-7829-9554https://orcid.org/0000-0002-1467-3105https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-0165-6848https://orcid.org/0000-0003-0348-2095https://ror.org/00cvxb145https://ror.org/00cvxb145https://ror.org/026v1ze26https://ror.org/026v1ze26https://ror.org/042nb2s44https://crossmark.crossref.org/dialog/?doi=10.1103/ds5p-763x&domain=pdf&date_stamp=2025-08-27https://doi.org/10.1103/ds5p-763xhttps://doi.org/10.1103/ds5p-763xhttps://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/layer with strong spin-orbit coupling generates a spincurrent, and the resulting spin-orbit torque enables mag-netization switching in an adjacent ferromagnetic layer.This further allows for injection of spin-polarized carriersinto a separate optically active layer. However, the degreeof circular polarization of emitted light is limited, largelydue to the spatial separation between the spin currentgenerating and the optically active layers.Here, we realized spin-valley Hall-torque-driven mag-netization switching and current-controlled helical lightemission in tMoTe2. Remarkably, the degree of circularpolarization can be continuously tuned, and reaches nearlyperfect circular polarization at a small applied current. Thisis achieved via three unusual properties of the system. First,the necessary ingredients—spin-torque control of magneticorder and spin-polarization-dependent helical emission—occur within a single direct band-gap semiconductorsystem, overcoming the challenge of multilayer structures.Second, flat Chern bands with nearly uniform Berrycurvature and spin-valley locking effects enable efficientspin-valley polarization control over a large doping range.Third, tMoTe2 hosts rich and electrically tunable quantumphases. Tuning the system into an anomalous Hall metalphase [23], i.e., a metallic phase with anomalous Halleffect, increases the efficiency of current control. Below, wepresent results from two tMoTe2 devices, D(3.7°) and D(3.9°), with twist angles of 3.7° and 3.9°. Both devices hostinteger and fractional QAH states, as detailed in priorreports [6,8].II. RESULTS AND DISCUSSIONA. Anomalous Hall metalWe begin with an exploration of the anomalous Hallmetal phase (AHM). Figure 1(a) shows a schematic of thedual-gated device and an atomic-force-microscopy (AFM)image of D(3.9°). The clean surface implies a high devicequality. Figure 1(b) is a plot of longitudinal resistance (Rxx)as a function of the doping and out-of-plane electric field(or displacement field) D=ε0. Data are taken at a temper-ature of 100 mK in D(3.7°) and symmetrized at magneticfield μ0H ¼ �100 mT. As identified in a prior report [6],(a)3 μm030Z (nm)(b) (c)1 2 3 4051015Rxx (kΩ) T (K)= -0.780 30Rxx (kΩ)-1.2 -1 -0.8 -0.6 -0.4-0.2-0.100.10.2D/ε0 (V/nm)0 10RMCD (%)-1.2 -1 -0.8 -0.6 -0.4-0.2-0.100.10.2D/ε0 (V/nm)0 200Rxx (kΩ)D/ε0 = 0-1 -0.9 -0.8 -0.7 -0.61234 T (K)(f)(e)(d)= -0.78-200 0 200-10010Rxx , Rxy (kΩ)μ0H (mT)FIG. 1. Anomalous Hall metal in moiré tMoTe2. (a) Top: schematic of the device cross section, with graphite gates in gray, BNdielectric layers in blue, and Pt contacts in yellow. Top Pt gates are used to improve contact transparency, and current is injected via thesample contacts. Optical probes can be used to reveal spontaneous spin-valley polarization with current controls. Bottom: AFM image ofdevice D(3.9°) with contacts for current injection visible. (b) Longitudinal resistance Rxx vs moiré filling ν and displacement field D=ε0in D(3.7°), consistent with previous measurements. Resistance in black regions is too high to be reliably measured. Rxx is symmetrized atjμ0Hj ¼ 200 mT. (c) As in (b), but measuring RMCD. Region with finite RMCD signal is ferromagnetic. Sample was measured atμ0H ¼ 25 mT to prevent domain flipping as gates were swept. (d) Longitudinal (Rxx) and Hall (Rxy) resistance vs magnetic field in theAHM metal regime. Rxy hysteresis and nonvanishing signal at 0 T are signatures of the anomalous Hall effect. (e) Temperaturedependence of Rxx vs ν in D(3.9°). (f) Temperature dependence of Rxx at ν ¼ −0.78, extracted along the dotted line in (e). Increasingresistance with the temperature indicates metallic behavior.ERIC ANDERSON et al. PHYS. REV. X 15, 031057 (2025)031057-2there are QAH and FQAH phases near hole fillings v ¼ −1and v ¼ −2=3, respectively, with suppressedRxx. Here, v isthe filling factor denoting the number of carriers per moiréunit cell. Between v ¼ −1 and −2=3, Rxx is also sup-pressed, while Rxy is finite but smaller than Rxx.A measurement of reflective magnetic circular dichroism(RMCD) at 25 mT in the same region of phase space isshown in Fig. 1(c). All optical measurements are taken at1.6 K unless otherwise indicated. The small appliedmagnetic field is to suppress spin fluctuations due to thedoping sweep (see Fig. S1 in Supplemental Material [24]).As RMCD is sensitive to spin-valley polarization, themeasurement reveals the ferromagnetic phase space versusv and D=ε0. Figure 1(d) illustrates Rxx and Rxy versus μ0Hat hole filling ν ¼ −0.78 and D=ε0 ¼ 0. As shown by theRMCD measurement in Fig. 1(c), the system is in a spin-valley-polarized (ferromagnetic) regime under these con-ditions. Rxy shows clear hysteresis and is on the order of afew kΩ, while Rxx is nearly constant at approximately10 kΩ, with a slight increase near the critical field HC ≈�20 mT where the sign of Rxy flips. Thus, the anomalousHall effect is observed at this partial filling of the first moiréChern band.We further performed temperature-dependent measure-ments. Figure 1(e) shows a measurement of Rxx versusfilling and temperature in D(3.9°). Around ν ¼ −1 andν ¼ −2=3, Rxx remains nearly vanishing at low temperatureand rises as the temperature is increased, when the thermalactivation of the carriers over the Chern gap dominates.Between −1 and−2=3, Rxx continuously increases with thetemperature. Taking a linecut at ν ¼ −0.78 (dashed line),Rxx increases from about 4 to 15 kΩ as the temperatureincreases from 100 mK to 4 K [Fig. 1(f)], suggesting ametallic phase. In addition, previous measurements of trionphotoluminescence versus doping [4,10] do not showsuppression of the luminescence intensity in this regionof phase space, in contrast to the suppression observed atChern insulator states. This is consistent with the com-pressible nature of a metallic phase. Taken together withRxx being several times larger than Rxy, the data demon-strate the system is an AHM over this region of doping, i.e.,a spontaneous ferromagnetic metal phase exhibiting ananomalous Hall signal in a partially filled Chern band.B. Current control of spin-valley polarizationNext, we examine current control of ferromagnetism,focusing on the anomalous Hall metal phase. We employRMCD to probe the ferromagnetic order. For instance,Fig. 2(a) shows the RMCD signal as μ0H is cycled at ν ¼−0.77 and D=ε0 ¼ 0, i.e., in the AHM. The clear RMCDhysteresis with finite signal at zero magnetic field is(b)(d)(a)(f)(e)(c)(g)-100 -50 0 50 100-20-1001020RMCD (%)μ0H (mT)-μ0H +μ0H I = -35 nAMI = -35 nA MI = 35 nAMI = 35 nAMI = 0 nA-10 10RMCD (%)M3 μ mMIJSFIG. 2. Transverse spin and charge current in a ferromagneticmetal. (a) RMCD signal as a function of μ0H swept down and up.RMCD hysteresis demonstrates ferromagnetic order. (b) RMCDspatial map in the anomalous Hall metal regime (ν ¼ −0.77) withmagnetization pointing up. Gray dashed lines indicate Pt contactarea. (c) RMCD map under a current of I ¼ −35 nA (indicatedby the arrows) flowing between the upper-left contacts of thedevice. The sign change in RMCD implies the reversal of spin-valley polarization. (d) RMCD map with the current directionreversed. The sign-reversed domain appears on the opposite sideof the current channel. (e) RMCD map with magnetizationpointing down and I ¼ −35 nA, as in (c). Despite the reversedmagnetization, domain signs on either side of the current channelremain consistent with (c). (f) RMCD map with the reversedcurrent direction and magnetization pointing down. Signs ofdomains on either side of the current channel match those in (d).(g) Schematic of the spin-valley Hall effect.MAGNETOELECTRIC CONTROL OF HELICAL LIGHT … PHYS. REV. X 15, 031057 (2025)031057-3consistent with the observed anomalous Hall signal inFig. 1(d). To determine the effect of the injected chargecurrent I on the spin-valley polarization, we first took aspatial map of the RMCD signal over the entire sample atI ¼ 0. Figure 2(b) shows this map for the same ν and D=ε0as in Fig. 2(a), measured at μ0H ¼ 0 with the magnetiza-tion initialized to point up. A finite, positive RMCD signalexists across the entire sample area, with contacts visible asregions of decreased RMCD (outlined with gray dashedlines). This is the expected behavior for a homogeneoussample tuned to the ferromagnetic (FM) phase.In stark contrast, when an RMCD spatial map is takenunder the same conditions as Fig. 2(b), but with currentflowing between the two contacts on the top left,clear domains with negative RMCD signal are observed[Figs. 2(c) and 2(d)]. The current injection configuration isdiscussed further in Sec. IV. The opposite sign of theRMCD signal in these domains compared to the rest of thesample area demonstrates a current-driven spin-valleypolarization reversal. The spatial position of the domainwith reversed spin-valley polarization can be controlled bythe current flow direction. For I ¼ −35 nA [Fig. 2(c)], thisnegative domain with magnetization pointing down (blue)is to the lower right of the current channel. In contrast, forcurrent flowing in the opposite direction [Fig. 2(d)], thenegative domain is on the upper left side.We found that the spatial pattern of current-inducedpolarization-reversed domains is independent of the initialmagnetization orientation. The RMCD map in Fig. 2(e) istaken with the magnetization initialized to point down, asevidenced by the opposite sign of the RMCD signal on thebottom right of the map, far from the current channel.Comparing this map to Fig. 2(c) taken at I ¼ −35 nA,although the initial magnetization direction is reversed, thesigns of the domains on either side of the current channelare the same, i.e., positive on the upper left and negative onthe lower right. Similar behavior is observed when com-paring Figs. 2(d) and 2(f); for the same I ¼ 35 nA current,the signs of the domains on either side of the currentchannel match, even though the initial magnetizationdirections are reversed. Varying the source-and-drain pinconfiguration changes the current channel and thus thelocations of the domains (see Fig. S2 in SupplementalMaterial [24]), but the same qualitative behavior remains.Similar RMCD maps taken at selected current values (seeFig. S3 in Supplemental Material [24]) show that current-induced polarization-reversed domains are present forvalues of I as low as 10 nA [24].The above measurements demonstrate that the magneticdomain pattern is determined only by the current flowdirection. This results from the mechanism for current-induced magnetization flipping—spin-valley Hall torque—as shown in Fig. 2(g). The spin-valley Hall torque switch-ing of magnetic order has been observed in MoTe2=WSe2heterobilayer superlattices, but only near full filling of themoiré valence band (or ν ¼ −1) [25,26]. The observationwas attributed to the large Berry curvature nearv ¼ −1 [3,27,28], where the QAH effect was observed [29].This Berry curvature generates an intrinsic spin-valley Halleffect—the source of the spin-valley-polarized currentwhich exerts a torque on the magnetic ground state andcauses the observed magnetization flipping. Our resultsdemonstrate the same mechanism works in tMoTe2.A major distinction in tMoTe2, however, lies in twounique aspects of the system. The first is the low currentdensity required for magnetization switching in the anoma-lous Hall metal phase. At approximately 102 A=cm2, theswitching current is nearly an order of magnitude lowercompared to other moiré systems [25,26]. This is mostlikely due to the fact that in other systems the FM phaseexists only close to the insulating state, which has arelatively large spin-valley energy splitting. In contrast,in the AHM phase, the spin splitting is decreased due tocarrier screening and/or partial valley polarization. Thesecond aspect is that tMoTe2 exhibits this current controlbehavior over a much larger region of phase space.Although in the above discussion we focus on theAHM, efficient current switching is observed over a broadrange of fillings of the first moiré Chern band. As shown inFig. S4 in Supplemental Material [24], the sign of RMCDchanges with applied current not only near ν ¼ −0.77 butover the entire FM phase space for hole fillings belowν ¼ −1, as well as in the wings of the FM phase abovev ¼ −1 with large electric field. Because the transversespin current arises from the Berry curvature of the moirébands, this broad range demonstrates that the Berrycurvature remains substantial far from the ν ¼ −1 bandedge. Our observations are consistent with theoretical workpredicting uniform Berry curvature [23,27] as well as withthe nearly uniform Berry curvature condition needed for theobserved fractional Chern insulator (FCI) states in thesystem [30].C. Current-controlled helical light emissionHaving established efficient current control of spin-valley polarization, we now turn to a second key behaviorin moiré MoTe2: spin-valley polarization-dependent emis-sion helicity. Although magnetism-related phenomena havebeen observed in several moiré TMD systems, the sponta-neous helical light emission is unique to tMoTe2 because itsbilayer form remains a direct band-gap semiconductor,while other TMD moiré systems become an indirect bandgap. Details of the emission mechanism have been dis-cussed in depth in previous works [4,10]. In brief, spin-valley-polarized carriers bind to electron-hole pairs in theopposite valley to form singlet trions, which emit inthe σþðσ−Þ channel for the electron-hole recombinationin the þKð−KÞ valley. In the FM phase, due to sponta-neous spin-valley polarization of the doped holes, lightemission occurs only in a single valley, leading to circularlyERIC ANDERSON et al. PHYS. REV. X 15, 031057 (2025)031057-4polarized emission at μ0H ¼ 0. This circularly polarizedemission is observed across the entire sample and dis-appears above the Curie temperature TC (SupplementalMaterial Fig. S5 [24]). As shown below, current control ofspin-valley polarization would also allow us to control trionemission helicity.Figures 3(a) and 3(b) present the spatial maps ofcircular-polarization-resolved trion photoluminescence.The degree of circular polarization is defined asρ ¼ f½PLðσ−Þ − PLðσþÞ�=½PLðσ−Þ þ PLðσþÞ�g. The dataare taken with linearly polarized excitation at 632.8 nm,with v ¼ −0.77 and I ¼ �35 nA, matching the experi-mental conditions of Figs. 2(c) and 2(d). In these spatialmaps, domains with opposite signs of ρ on either side of thecurrent channel appear, which switch with the reversal of I.This behavior demonstrates that the switching of the lighthelicity is caused by the reversal of spin-valley polarizationby current, as demonstrated in our RMCDmeasurements inFig. 2. Selecting a spot on one of these domains [black dotin Fig. 3(b)], we plot polarization-resolved PL spectra[Figs. 3(c) and 3(d)] for I ¼ �35 nA at μ0H ¼ 0. The dataclearly show that as we reverse the current flow direction,the trion luminescence switches helicity with near-unitypolarization. This establishes current control of light-emission helicity. In addition, comparing the polariza-tion-resolved PL spatial maps between magnetizationinitialized up [Figs. 3(a) and 3(b)] and down [Figs. 3(e)and 3(f)], we observe the same domain behavior as for theRMCD maps in Fig. 2, as expected from the spin-valleyHall torque mechanism.We next consider the emission helicity as a function ofboth the applied magnetic field and current. As seen inFig. 4(a), ρ vs μ0H swept down and up at I ¼ 0 and v ¼−0.77 shows hysteretic behavior and finite ρ at zeromagnetic field. This is expected for a ferromagnetic systemwhere near-unity ρ arises from spin-valley polarization.The same measurement with injected current [Figs. 4(b)and 4(c)] shows that while hysteresis vs μ0H is still visible,the center of hysteresis moves away from μ0H ¼ 0,depending on the current direction. A full dependence ofρ vs both I and μ0H is shown in Fig. 4(d). We observe thatthe ρ can be continuously tuned as a function of bothparameters. ρ vs I and μ0H for both magnetic field sweepdirections is shown in Supplemental Material Fig. S6 [24].The tuning of the trion PL emission helicity as a function ofthe current is highlighted in Fig. 4(e); it can be continuouslyvaried between σþ and σ− at zero magnetic field. While wedo not observe hysteresis in the ρ vs I measurements in theAHM regime, hysteresis in spin-valley polarization vs I isobserved at v ¼ −1 and finite D=ε0 (SupplementalMaterial Fig. S7 [24]). This discrepancy is likely due tostronger domain-wall pinning in gapped phases, leading tononvolatile domains.Our experimental results can be compared to quantita-tive calculations using the spin-diffusion equations in aslab geometry [Fig. 4(f)] [26,31]. We define δMz ascurrent-induced magnetization and M0z ¼ gμBn as samplemagnetization. Here, g is the g factor of charge carriers, μBis the Bohr magneton, and n is the carrier density. Using(b)(a)(d)(c)1.11 1.115Energy (eV)0200PL (a.u.)σ+σ- I = 35 nAμ0H = 0 mT1.11 1.115Energy (eV)0200PL (a.u.)I = -35 nAμ0H = 0 mTI = -35 nAρ-1 1MI = 35 nA M(f)(e)I = -35 nAMI = 35 nAMFIG. 3. Current control of helical light emission. (a) Spatial mapof PL degree of circular polarization ρ with magnetizationinitialized to point up. Measurement taken with I ¼ −35 nAcurrent and in the anomalous Hall metal regime (ν ¼ −0.77) atμ0H ¼ 0 T. Domains with opposite signs of ρ are visible on eitherside of the current channel. Finite, but smaller, ρ is visible far fromthe current channel, on the lower-right side of the device. (b) Sameas (a), but with I ¼ 35 nA. The domain signs are reversedcompared to (a). (c) Circular-polarization-resolved PL spectrafrom the position marked by the black dot in (b). The PL showsclear helicity dependence of the light emission at μ0H ¼ 0 T, withthe signal predominantly in the σ− channel. (d) Same as (c), butwith I ¼ −35 nA, extracted from the same position in (a). Thehelicity dependence of the PL is switched, with signal now mainlyin the σþ channel. (e)As in (a), butwithmagnetization initialized topoint down. While the opposite sign of ρ is visible far from thecurrent channel on the lower right of the device, the sign of thedomains on either side of the current channel remains the same as in(a), with the same current flow direction but opposite magnetiza-tion. (f) As in (b), but with magnetization pointing down. Behaviorof domainswith switching of current andmagnetization direction isconsistent with RMCD behavior in Fig. 2.MAGNETOELECTRIC CONTROL OF HELICAL LIGHT … PHYS. REV. X 15, 031057 (2025)031057-5realistic sample parameters (see Sec. IV), Fig. 4(g) plots thecalculated spatial distribution of δMz=M0z for selectedcharge currents. The calculations show that δMz hasopposite sign at opposite sides of the current channel,and becomes comparable to M0z for a charge current ofapproximately 30 nA [Fig. 4(g)]. This estimate is on thesame order of magnitude as our experimental observationof current switching. A comparison between RMCD dataand the modeled spin-diffusion behavior (SupplementalMaterial Fig. S9 [24]) shows remarkably similar profile,lending additional experimental support for the spin-dif-fusion picture.III. CONCLUSIONSUsing RMCD and polarization-resolved PL, we haveestablished current control of spin-valley polarization andhelical light emission in moiré MoTe2. Although the devicerequires cryogenic temperatures, the strong couplingbetween electronic, spintronic, and photonic degrees offreedom demonstrated by these results could be appealingin a range of device applications. A particular advantage ofmoiré MoTe2 over other topological or optospintronicsystems such as magnetically doped topological insulatorsor spin light-emitting diodes is that tMoTe2 is intrinsicallytopological, ferromagnetic, semiconducting, and opticallyactive, with substantial potential for applications. Onepossibility is to use these properties to couple magneticinformation storage to optical communication within asingle device. In addition, the topological phases whichoccur in tMoTe2 could open new frontiers in topologicaloptospintronics. As the integer and fractional Chern insula-tors recently observed in this system have a topologicalindex which depends on the magnetization orientation, theobserved spin-valley Hall torque implies that currentinjection can be used to manipulate these topologicalstates. One interesting direction could be to use currentinjection to establish two magnetic domains with oppositeorientations and investigate the physics arising at theinterface between two zero-field FCIs with opposite topo-logical indices. We foresee that establishing current controlof magnetic order in moiré MoTe2 will add an importanttuning knob to our toolbox as investigation of these zero-field FCIs and their anyonic excitations continues.δMZ/MZ0010.5-1-0.50 5 10-5-10x (μm)I = 10 nAI = 20 nAI = 30 nA(b)(a)(c)(d)-μ0H -11ρ-50-2502550I (nA)-50 -25 0 25 50μ0H (mT)(e)(f)(g)IJS(0,+W/2)(0,-W/2)xyρ-1-0.500.51-50 -25 0 25 50I (nA)1.11 1.115Energy (eV)0150PL (a.u.)I = 70 nA1.11 1.115Energy (eV)0150PL (a.u.)σ+σ-I = -70 nA-200 -100 0 100 200-11-μ0H +μ0H I = 0 nAμ0H (mT)ρ-200 -100 0 100 200-11μ0H (mT)ρI = 35 nA-200 -100 0 100 200-11μ0H (mT)ρI = -35 nAFIG. 4. Continuous tuning of emission helicity via current. (a) ρ vs μ0H swept down and up, with I ¼ 0 nA. Hysteresis is visible,indicating that light-emission helicity is controlled by the spin-valley polarization. (b),(c) Same as (a), but with I ¼ 35 nA (b) andI ¼ −35 nA (c). While hysteresis of ρ vs μ0H is still visible, applied current favors opposite spin-valley polarizations for opposite flowdirections. (d) ρ vs I swept down, as a function of μ0H, starting from positive μ0H. Emission helicity can be continuously tuned betweenthe σþ and σ− channels by both current and magnetic field. (e) ρ as a continuous function of I, at μ0H ¼ 0 T, initialized withμ0H > 0 T. Insets: helicity-resolved PL emission for large positive and negative current. Helicity can be continuously tuned betweenthese two limits by changing I. All data are taken in the anomalous Hall metal regime, with ν ¼ −0.77, and with the beam spot on theopposite side of the current channel as for Figs. 3(c) and 3(d) (see Supplemental Material Fig. S8 [24]). (f) Schematic of sample currentflow configuration, with I and JS injected charge and transverse spin currents, respectively. Current source and drain are at the blackdots, with the current channel lengthW. (g) Induced magnetization δMZ versus x, derived from the spin-diffusion equations and realisticsample parameters at y ¼ 0. The critical charge current IC for magnetization flipping is about 30 nA, the same order of magnitude asexperimental observations (see Sec. IV).ERIC ANDERSON et al. PHYS. REV. X 15, 031057 (2025)031057-6IV. METHODSA. Device fabricationThe moiré MoTe2 samples equipped with electrical con-tacts used in this study were fabricated using the procedurediscussed in more depth in Ref. [6]. In brief, van der Waalsflakes used in the heterostructure devices—graphite, h-BN,and MoTe2—were mechanically exfoliated on oxygen-plasma-cleaned Si=SiO2 substrates and identified using anoptical microscope. Atomic force microscopy was used tocheckh-BNthickness.Agraphiteandh-BNbottomgatewithPt contacts was prepared using conventional dry-transfer,electron-beamlithography,e-beamevaporation,andcontact-mode atomic force microscope (AFM) cleaning techniques.In a glovebox with < 0.1 ppmH2O and O2 concentrations,monolayer MoTe2 was exfoliated, and an h-BN–encapsu-lated moiré bilayer was created with the cut-and-stack dry-transfer technique, before putting downon the prepared backgate.AfterwashoffandAFMcleaningof thedevice,asetofPtcontactgates andAuwirebondingpadsweredepositedusinge-beam lithography and evaporation. Finally, the devicewasAFM cleaned again, before transferring on a top graph-ite gate.B. Transport measurementsTransportmeasurementswere taken in a Bluefors dilutionrefrigeratorwith a 9-Tmagnet and base electron temperatureof about 80 mK. An ac current bias of 0.2–0.5 nA wasgenerated using a 100-MΩ resistor in series with an acvoltage source (SR830), with the current monitored usinga DL1211 current amplifier. Four-terminal Rxx and Rxysignals were amplified using the differential mode of anSR560 voltage preamplifier with an input resistance (about100 MΩ) much larger than the contact resistance of thedevice. The amplified voltage signalswere demodulated andmeasured using SR830 and SR860 lock-ins.C. Optical measurementsOptical measurements were taken in a closed-loopmagneto-optical cryostat (attoDRY 2100) with an attocubexyz piezostage and xy scanners, 9-T z-axis superconductingmagnet, and with a base temperature of 1.6 K. Current wasapplied to the sample using a 100-MΩ resistor in serieswith a dc voltage source (Keithley 2450) connected to thesource pin, with the drain pin grounded. Contact gates wereset to −3 V to ensure transparent electrical contact to themoiré MoTe2, as discussed in Ref. [6]. Polarization-resolved photoluminescence measurements were takenwith linearly polarized 632.8-nm HeNe laser excitationfocused on the sample by a high-NA nonmagneticcryogenic objective to an approximately 1-μm beam spot.Sample emission was collected by the same objective andpassed through a quarter wave plate and linear polarizer toselect the right- and left-circularly-polarized channels.Signal was then passed through a 75-μm pinhole anddispersed with a diffraction grating (Princeton Instruments,600 grooves/mm at 1-μm blaze) and detected by a liquid-nitrogen-cooled infrared CCD (Princeton InstrumentsPyLoN-IR 1.7).RMCD measurements were taken with excitation nearthe trion resonance by filtering a broadband supercontin-uum source (NKT SuperK Fianium FIR-20) by dualpassing through a monochromator to achieve a narrowexcitation bandwidth. The out-of-plane magnetization ofthe sample induces an MCD signal ΔR, the differencebetween the reflected right- and left-circularly-polarizedlight. To obtain the normalized RMCD ΔR=R, the laserintensity was chopped at p ¼ 850 Hz, and the phase wasmodulated by λ=4 via a photoelastic modulator atf ¼ 50 kHz. An InGaAs avalanche photodiode detectorwas used to collect the reflected signal, and the output wasread by two lock-in amplifiers (SR830). The ratio betweenthe p-component signal Ip and f-component signal Ifgives the RMCD signal: ΔR=R ¼ If=½J1ðπ=2Þ × Ip� whereJ1 is the first-order Bessel function.D. Determination of doping density and electric fieldA parallel plate capacitor model was used to determinethe carrier density n and electric field D from the appliedtop and bottom gate voltages VTG and VBG. Gate capac-itances CTG and CBG are calculated using the h-BNthickness determined by atomic force microscopy, takingthe h-BN dielectric constant to be 3.0. Thus, n andD can becomputed as n ¼ ðVTGCTG þ VBGCBGÞ=e − noffset andD=ϵ0 ¼ ðVTGCTG − VBGCBGÞ=2ϵ0 −Doffset, with e theelectron charge and ε0 the vacuum permittivity. Carrierdensity offset noffset is derived from fitting to the insulatingstates in the PL spectra and transport measurements. Doffsetis determined from the symmetry axis of the RMCD phasediagram.E. Induced magnetization and spin Hall currentThe magnetization induced by the spin Hall currents canbe derived from the spin-diffusion equations [26,31]. Theconfiguration considered is shown in Fig. 4(f). A chargecurrent I is injected in the y direction into a slab of widthW, generating a spin current JS in the x direction to flip themagnetization. Solving the diffusion equation in the slab,the induced magnetization [26,31] isδMzðx; yÞ ¼2gμBI tan θSHel2S=τSZdk2πeikxcoshðωkyÞωk cothðkW=2Þ sinhðωkW=2Þ þ 4k tan2ðθSHÞ coshðωkW=2Þ ;MAGNETOELECTRIC CONTROL OF HELICAL LIGHT … PHYS. REV. X 15, 031057 (2025)031057-7where g is the g factor of charge carriers, μB is the Bohrmagneton, θSH is the spin Hall angle, e is the electroncharge. lS is the spin-diffusion length, and τS is the spinrelaxation time ωk ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffik2 þ l−2Sp.The calculation parameters are estimated from exper-imental values. W ¼ 3 μm is the distance between elec-trical contacts. To estimate the value of lS, we note thatexperimentally the size of the flipped domain is on theorder of μm, indicating lS should be on the same order. Wechoose ls ¼ 1 μm. The spin Hall angle is defined bytan θSH ¼ eσSH=ℏσxx, where σSH is the spin Hall conduc-tivity, σxx is the longitudinal conductivity, and ℏ is thereduced Planck constant. σSH ¼ ðσH;K − σH;K0 Þℏ=2e(Ref. [32]) due to spin-valley coupling in TMDs, withσH;K and σH;K0 the Hall conductivities of valleys K and K’.If we assume the system is fully polarized to K, σSH issimply σH;Kℏ=2e ¼ ℏσAH=2e, and σAH is the anomalousHall conductivity. Thus, tan θSH ¼ σAH=2σxx. The exper-imentally measured longitudinal resistance ρxx and anoma-lous Hall resistance ρxy are both on the same order of a fewkΩ. We thus take tan θSH ¼ σAH=2σxx ¼ ρxy=2ρxx ¼ 1=2.Finally, while τs cannot be determined directly from ourmeasurement, τs in TMD heterobilayers is estimated to beon the order of μs in Refs. [26,33]. We take τs to be on thesame order here, τs ¼ 5 μs.Figure 4(g) shows the induced magnetization δMz vsposition x at the center of the slab (y ¼ 0) for differentinjection currents I. The magnetization flipping occurswhen the peak of δMz is on the order of M0z , where M0z isthe magnetization of the ground state. M0z ¼ gμBn, andcarrier density n ¼ −3.8 × 1012 cm−2 for the filling ν ¼−0.78 in our experimental sample. With our choice of τs,the critical current for magnetization flipping is on the orderof 30 nA, consistent with the experimental values.ACKNOWLEDGMENTSThis project is mainly supported by the U.S. Departmentof Energy, Office of Science, Basic Energy Sciences, underAward No. DE-SC0018171. The electrical control offerromagnetism is partially supported by Vannevar BushFaculty Fellowship (Award No. N000142512047), andtheory is supported by Grant No. DE-SC0012509. E. A.acknowledges support by the National Science FoundationGraduate Research Fellowship Program under GrantNo. DGE-2140004. The authors also acknowledge theuse of the facilities and instrumentation supported byNSF Grant No. MRSEC DMR-1719797. K.W. and T. T.acknowledge support from the JSPS KAKENHI (GrantsNo. 21H05233 and No. 23H02052), the CREST (GrantNo. JPMJCR24A5), JST and World Premier InternationalResearch Center Initiative, MEXT, Japan. X. X. acknowl-edges support from the State of Washington funded CleanEnergy Institute and from the Boeing DistinguishedProfessorship in Physics.X. X. conceived and supervised the experiment. H. P.fabricated and performed the transport measurements,assisted by J. C. E. A. performed the magneto-opticalmeasurements. E. A., X. X., D. X., T. C., and L. F. analyzedand interpreted the results. K. Y. and D. X. performed thecalculations. T. T. and K.W. synthesized the h-BN crystals.E. A., X. X., and D. X. wrote the paper with input from allauthors. All authors discussed the results.The authors declare no competing interests.DATA AVAILABILITYThe data that support the findings of this article areopenly available [34].[1] Y. Cao, V. Fatemi, S. Fang, K. Watanabe, T. Taniguchi, E.Kaxiras, and P. Jarillo-Herrero, Unconventional supercon-ductivity in magic-angle graphene superlattices, Nature(London) 556, 7699 (2018).[2] R. Bistritzer and A. H. MacDonald, Moiré bands in twisteddouble-layer graphene, Proc. Natl. Acad. Sci. U.S.A. 108,12233 (2011).[3] F. Wu, T. Lovorn, E. Tutuc, I. Martin, and A. H.MacDonald, Topological insulators in twisted transitionmetal dichalcogenide homobilayers, Phys. Rev. Lett. 122,086402 (2019).[4] J. Cai et al., Signatures of fractional quantum anomalousHall states in twisted MoTe2, Nature (London) 622, 7981(2023).[5] Y. Zeng, Z. Xia, K. Kang, J. Zhu, P. Knüppel, C. Vaswani,K. Watanabe, T. Taniguchi, K. F. Mak, and J. Shan,Thermodynamic evidence of fractional Chern insulator inmoiré MoTe2, Nature (London) 622, 7981 (2023).[6] H. Park et al., Observation of fractionally quantizedanomalous Hall effect, Nature (London) 622, 7981 (2023).[7] F. Xu et al., Observation of integer and fractional quantumanomalous Hall effects in twisted bilayer MoTe2, Phys.Rev. X 13, 031037 (2023).[8] E. Redekop et al., Direct magnetic imaging of fractionalChern insulators in twisted MoTe2, Nature (London) 635,584 (2024).[9] Z. Ji, H. Park, M. E. Barber, C. Hu, K. Watanabe, T.Taniguchi, J.-H. Chu, X. Xu, and Z.-X. Shen, Local probeof bulk and edge states in a fractional Chern insulator,Nature (London) 635, 578 (2024).[10] E. Anderson et al., Trion sensing of a zero-field compositeFermi liquid, Nature (London) 635, 590 (2024).[11] Z. Lu, T. Han, Y. Yao, A. P. Reddy, J. Yang, J. Seo, K.Watanabe, T. Taniguchi, L. Fu, and L. Ju, Fractionalquantum anomalous Hall effect in multilayer graphene,Nature (London) 626, 759 (2024).[12] T. Neupert, L. Santos, C. Chamon, and C. Mudry, Frac-tional quantum Hall states at zero magnetic field, Phys. Rev.Lett. 106, 236804 (2011).[13] N. Regnault and B. A. Bernevig, Fractional Chern insu-lator, Phys. Rev. X 1, 021014 (2011).ERIC ANDERSON et al. PHYS. REV. X 15, 031057 (2025)031057-8https://doi.org/10.1038/nature26160https://doi.org/10.1038/nature26160https://doi.org/10.1073/pnas.1108174108https://doi.org/10.1073/pnas.1108174108https://doi.org/10.1103/PhysRevLett.122.086402https://doi.org/10.1103/PhysRevLett.122.086402https://doi.org/10.1038/s41586-023-06289-whttps://doi.org/10.1038/s41586-023-06289-whttps://doi.org/10.1038/s41586-023-06452-3https://doi.org/10.1038/s41586-023-06536-0https://doi.org/10.1103/PhysRevX.13.031037https://doi.org/10.1103/PhysRevX.13.031037https://doi.org/10.1038/s41586-024-08153-xhttps://doi.org/10.1038/s41586-024-08153-xhttps://doi.org/10.1038/s41586-024-08092-7https://doi.org/10.1038/s41586-024-08134-0https://doi.org/10.1038/s41586-023-07010-7https://doi.org/10.1103/PhysRevLett.106.236804https://doi.org/10.1103/PhysRevLett.106.236804https://doi.org/10.1103/PhysRevX.1.021014[14] D. N. Sheng, Z.-C. Gu, K. Sun, and L. Sheng, Fractionalquantum Hall effect in the absence of Landau levels, Nat.Commun. 2, 1 (2011).[15] K. Sun, Z. Gu, H. Katsura, and S. Das Sarma, Nearlyflatbands with nontrivial topology, Phys. Rev. Lett. 106,236803 (2011).[16] E. Tang, J.-W. Mei, and X.-G. Wen, High-temperaturefractional quantum Hall states, Phys. Rev. Lett. 106,236802 (2011).[17] D. Xiao, W. Zhu, Y. Ran, N. Nagaosa, and S. Okamoto,Interface engineering of quantum Hall effects in digitaltransition metal oxide heterostructures, Nat. Commun. 2, 1(2011).[18] B. A. Bernevig and N. Regnault, Emergent many-bodytranslational symmetries of Abelian and non-Abelian frac-tionally filled topological insulators, Phys. Rev. B 85,075128 (2012).[19] C. Ruppert, B. Aslan, and T. F. Heinz, Optical propertiesand band gap of single- and few-layer MoTe2 crystals,Nano Lett. 14, 6231 (2014).[20] I. G. Lezama, A. Arora, A. Ubaldini, C. Barreteau, E.Giannini, M. Potemski, and A. F. Morpurgo, Indirect-to-direct band gap crossover in few-layer MoTe2, Nano Lett.15, 2336 (2015).[21] E. Anderson, F.-R. Fan, J. Cai, W. Holtzmann, T. Taniguchi,K. Watanabe, D. Xiao, W. Yao, and X. Xu, Programmingcorrelated magnetic states with gate-controlled moirégeometry, Science 381, 325 (2023).[22] P. A. Dainone et al., Controlling the helicity of light byelectrical magnetization switching, Nature (London) 627,783 (2024).[23] V. Crépel and L. Fu, Anomalous Hall metal and fractionalChern insulator in twisted transition metal dichalcogenides,Phys. Rev. B 107, L201109 (2023).[24] See Supplemental Material at http://link.aps.org/supplemental/10.1103/ds5p-763x for additional RMCDmeasurements versus v and D=ε0, RMCD spatial mapswith alternate pin configurations and selected currents,emission helicity above TC, current sweep hysteresis mea-surements, and a comparison between modeled spin-dif-fusion and experimental data.[25] C. L. Tschirhart et al., Intrinsic spin Hall torque in a moiréChern magnet, Nat. Phys. 19, 807 (2023).[26] Z. Tao et al., Giant spin Hall effect in AB-stackedMoTe2=WSe2 bilayers, Nat. Nanotechnol. 19, 28 (2024).[27] T. Devakul, V. Crépel, Y. Zhang, and L. Fu,Magic in twistedtransition metal dichalcogenide bilayers, Nat. Commun. 12,6730 (2021).[28] T. Devakul and L. Fu, Quantum anomalous Hall effect frominverted charge transfer gap, Phys. Rev. X 12, 021031(2022).[29] T. Li et al., Quantum anomalous Hall effect fromintertwined moiré bands, Nature (London) 600, 7890(2021).[30] C. Wang, X.-W. Zhang, X. Liu, Y. He, X. Xu, Y. Ran, T.Cao, and D. Xiao, Fractional Chern insulator in twistedbilayer MoTe2, Phys. Rev. Lett. 132, 036501(R) (2024).[31] M. Beconcini, F. Taddei, and M. Polini, Nonlocal topo-logical valley transport at large valley Hall angles, Phys.Rev. B 94, 121408 (2016).[32] J. Sinova, S. O. Valenzuela, J. Wunderlich, C. H. Back, andT. Jungwirth, Spin Hall effects, Rev. Mod. Phys. 87, 1213(2015).[33] E. C. Regan et al., Mott and generalized Wigner crystalstates in WSe2=WS2 moiré superlattices, Nature (London)579, 359 (2020).[34] E. Anderson, H. Park, K. Yang, J. Cai, T. Taniguchi, K.Watanabe, L. Fu, T. Cao, and X. Xu, Dataset for: Mag-netoelectric Control of Helical Light Emission in a MoiréChern Magnet (2025) [Data set]. Zenodo, 10.5281/zenodo.16883546.MAGNETOELECTRIC CONTROL OF HELICAL LIGHT … PHYS. REV. X 15, 031057 (2025)031057-9https://doi.org/10.1038/ncomms1380https://doi.org/10.1038/ncomms1380https://doi.org/10.1103/PhysRevLett.106.236803https://doi.org/10.1103/PhysRevLett.106.236803https://doi.org/10.1103/PhysRevLett.106.236802https://doi.org/10.1103/PhysRevLett.106.236802https://doi.org/10.1038/ncomms1602https://doi.org/10.1038/ncomms1602https://doi.org/10.1103/PhysRevB.85.075128https://doi.org/10.1103/PhysRevB.85.075128https://doi.org/10.1021/nl502557ghttps://doi.org/10.1021/nl5045007https://doi.org/10.1021/nl5045007https://doi.org/10.1126/science.adg4268https://doi.org/10.1038/s41586-024-07125-5https://doi.org/10.1038/s41586-024-07125-5https://doi.org/10.1103/PhysRevB.107.L201109http://link.aps.org/supplemental/10.1103/ds5p-763xhttp://link.aps.org/supplemental/10.1103/ds5p-763xhttp://link.aps.org/supplemental/10.1103/ds5p-763xhttp://link.aps.org/supplemental/10.1103/ds5p-763xhttp://link.aps.org/supplemental/10.1103/ds5p-763xhttps://doi.org/10.1038/s41567-023-01979-8https://doi.org/10.1038/s41565-023-01492-2https://doi.org/10.1038/s41467-021-27042-9https://doi.org/10.1038/s41467-021-27042-9https://doi.org/10.1103/PhysRevX.12.021031https://doi.org/10.1103/PhysRevX.12.021031https://doi.org/10.1038/s41586-021-04171-1https://doi.org/10.1038/s41586-021-04171-1https://doi.org/10.1103/PhysRevLett.132.036501https://doi.org/10.1103/PhysRevB.94.121408https://doi.org/10.1103/PhysRevB.94.121408https://doi.org/10.1103/RevModPhys.87.1213https://doi.org/10.1103/RevModPhys.87.1213https://doi.org/10.1038/s41586-020-2092-4https://doi.org/10.1038/s41586-020-2092-4https://doi.org/10.5281/zenodo.16883546https://doi.org/10.5281/zenodo.16883546 Magnetoelectric Control of Helical Light Emission in a Moiré Chern Magnet I. INTRODUCTION II. RESULTS AND DISCUSSION A. Anomalous Hall metal B. Current control of spin-valley polarization C. Current-controlled helical light emission III. CONCLUSIONS IV. METHODS A. Device fabrication B. Transport measurements C. Optical measurements D. Determination of doping density and electric field E. Induced magnetization and spin Hall current ACKNOWLEDGMENTS DATA AVAILABILITY References