# Fileset

[s41467-022-34812-6.pdf](https://mdr.nims.go.jp/filesets/10e1bcc6-4916-4487-9b62-cd22007e2a34/download)

## Creator

Guanghui Cheng, Mohammad Mushfiqur Rahman, Zhiping He, Andres Llacsahuanga Allcca, Avinash Rustagi, Kirstine Aggerbeck Stampe, Yanglin Zhu, Shaohua Yan, Shangjie Tian, Zhiqiang Mao, Hechang Lei, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Pramey Upadhyaya, Yong P. Chen

## Rights

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

## Other metadata

[Emergence of electric-field-tunable interfacial ferromagnetism in 2D antiferromagnet heterostructures](https://mdr.nims.go.jp/datasets/427e0a31-d48f-4702-a897-54bdd6da3091)

## Fulltext

Emergence of electric-field-tunable interfacial ferromagnetism in 2D antiferromagnet heterostructuresArticle https://doi.org/10.1038/s41467-022-34812-6Emergence of electric-field-tunableinterfacial ferromagnetism in 2Dantiferromagnet heterostructuresGuanghui Cheng1,2,3,12, Mohammad Mushfiqur Rahman4, Zhiping He5,Andres Llacsahuanga Allcca2,3,6, Avinash Rustagi4,13,Kirstine Aggerbeck Stampe7, Yanglin Zhu8, Shaohua Yan9, Shangjie Tian 9,Zhiqiang Mao 8, Hechang Lei 9, Kenji Watanabe 10, Takashi Taniguchi 11,Pramey Upadhyaya3,4,6 & Yong P. Chen1,2,3,4,6,7Van derWaals (vdW)magnet heterostructures have emerged as newplatformsto explore exotic magnetic orders and quantum phenomena. Here, we studyheterostructures of layered antiferromagnets, CrI3 and CrCl3, with perpendi-cular and in-plane magnetic anisotropy, respectively. Using magneto-opticalKerr effect microscopy, we demonstrate out-of-plane magnetic order in theCrCl3 layer proximal to CrI3, with ferromagnetic interfacial coupling betweenthe two. Such an interlayer exchange field leads to higher critical temperaturethan that of either CrI3 or CrCl3 alone. We further demonstrate significantelectric-field control of the coercivity, attributed to the naturally brokenstructural inversion symmetry of the heterostructure allowing unprecedenteddirect coupling between electric field and interfacial magnetism. These find-ings illustrate the opportunity to explore exoticmagnetic phases and engineerspintronic devices in vdW heterostructures.Heterostructures are promising to host emergent phenomena anddevice functions not present in constituent parts1–10. One well-knownexample is the integration of two insulating complex oxides leading toa conducting two-dimensional electron gas at the interface2, withsurprising coexistenceof superconductivity and ferromagnetism3. Therecently explored van der Waals (vdW) magnets have pushed theresearch frontier to 2D magnetism where exotic magnetic groundstates and quantum phases can emerge11–18. Magnetic vdWheterostructures provide anew toolbox to exploremagnetic proximityand related effects4–10. A largely unexplored arena is to combine twodifferent magnetic orders and investigate the magnetic proximity atthe interface, which could allow modulation of magnetic interactionsand establish exotic magnetic properties. It is also of fundamentalsignificance to effectively control the exchange interactions andmagnetic anisotropy, with the latter being crucial to stabilize the long-range magnetic orders.Received: 9 January 2022Accepted: 8 November 2022Check for updates1Advanced Institute for Materials Research (WPI-AIMR), Tohoku University, Sendai 980-8577, Japan. 2Department of Physics and Astronomy, and BirckNanotechnology Center, Purdue University, West Lafayette, IN 47907, USA. 3Purdue Quantum Science and Engineering Institute, Purdue University, WestLafayette, IN 47907, USA. 4Elmore family school of electrical and computer engineering, Purdue University, West Lafayette, IN 47907, USA. 5Department ofPhysics, University of Science and Technology of China, Hefei, Anhui 230026, China. 6Quantum Science Center, Oak Ridge, TN 37831, USA. 7Institute ofPhysics and Astronomy and Villum Centers for Dirac Materials and for Hybrid QuantumMaterials, Aarhus University, Aarhus-C 8000, Denmark. 8Departmentof Physics, Pennsylvania State University, University Park, PA 16802, USA. 9Laboratory for Neutron Scattering, and Beijing Key Laboratory of OptoelectronicFunctional Materials MicroNano Devices, Department of Physics, Renmin University of China, Beijing 100872, China. 10Research Center for FunctionalMaterials, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan. 11International Center for Materials Nanoarchitectonics, NationalInstitute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan. 12Present address: Department of Physics, University of Science and Technology ofChina, Hefei, Anhui 230026, China. 13Present address: Intel Corp., Hillsboro, OR 97124, USA. e-mail: yongchen@purdue.eduNature Communications |         (2022) 13:7348 11234567890():,;1234567890():,;http://orcid.org/0000-0002-3412-4656http://orcid.org/0000-0002-3412-4656http://orcid.org/0000-0002-3412-4656http://orcid.org/0000-0002-3412-4656http://orcid.org/0000-0002-3412-4656http://orcid.org/0000-0002-4920-3293http://orcid.org/0000-0002-4920-3293http://orcid.org/0000-0002-4920-3293http://orcid.org/0000-0002-4920-3293http://orcid.org/0000-0002-4920-3293http://orcid.org/0000-0003-0850-8514http://orcid.org/0000-0003-0850-8514http://orcid.org/0000-0003-0850-8514http://orcid.org/0000-0003-0850-8514http://orcid.org/0000-0003-0850-8514http://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://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-34812-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-34812-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-34812-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-022-34812-6&domain=pdfmailto:yongchen@purdue.eduThe studies of layered semiconducting chromium trihalides haveshown exotic magnetic behaviors and rich tunability by stimuli14–18.Typically, the few-layer (FL) CrI3 is an antiferromagnet with Ising-likeperpendicular magnetic anisotropy (PMA)19,20, as schematically depic-ted in Fig. 1a. The interlayer antiferromagnetic coupling is ascribed tothe exchange interactions between Cr mediated by ligand atoms17,20. Incontrast, FL CrCl3 is an easy-plane interlayer antiferromagnet, wherespinsprefer to lie in the layers20,21. In particular, the single-ion anisotropyfromspin-orbit coupling (SOC)ofCr and the anisotropic exchange fromSOC of Cl nearly cancel out each other20,21. Therefore, CrCl3 is locatedclose to the boundary between PMA and in-plane anisotropy, suggest-ing that its magnetic properties may be particularly susceptible toexternal perturbations. The combined heterostructure of CrI3 andCrCl3is possibly a fertile system to realize rich magnetic phases and manip-ulate them. Such a heterostructure has not yet been explored.Here, we fabricate CrI3/CrCl3 heterostructures and demonstrateinterfacial ferromagnetism between the two antiferromagnets.Figure 1a left panel schematically depicts the expected spin config-urations for themagnetic ground states in bilayer (2L)CrI3 andFLCrCl3with PMA and in-plane anisotropy19–21, respectively. Due to the strongintralayer ferromagnetic coupling in chromium trihalides17, we candenote all spins in a given layer by amacroscopic spin (out-of-plane:↑,↓; in-plane:←,→). Theopticalmicrographof a representative 2LCrI3/FLCrCl3 heterostructure is shown in Fig. 1b. The 2L CrI3 is partiallystacked on top of FL CrCl3, allowing the comparison between regionsof 2L CrI3, FL CrCl3 and heterostructure. Atomic force microscopy(AFM) confirms flake thickness of 1.6 nm for 2L CrI3 (Fig. 1c) and 9.5 nmfor FL CrCl3, respectively.Results and discussionWeemploymagneto-optical Kerr effectmicroscopy (MOKE) under thepolar configuration as the primary measurement due to its high sen-sitivity to the magnetic moments perpendicular to the samplesurface22. Figure 1d shows the MOKE signal (θK) of the 2L CrI3 and the2L CrI3/FL CrCl3 heterostructure as a function of the perpendicularmagnetic field. In the 2L CrI3 region, θK stays close to zero at low field,corresponding to the antiferromagnetic states↓↑ or↑↓with zero netmagnetization. Beyond critical field ±0.76 T, θK abruptly jumps toferromagnetic states with finite magnetization. This is consistent withthe reported spin-flip transitions in 2L CrI323.In the 2L CrI3/FL CrCl3 heterostructure, the antiferromagnetic-to-ferromagnetic spin-flip transition of 2L CrI3 is still present and its cri-tical field decreases from ±0.76 T in 2L CrI3 region to ±0.57T in theheterostructure region. Remarkably, a significant square hysteresisloop is observed with coercive field ~±0.1 T, indicating a magnetictransition between two different phases with non-zero net magneti-zation, in sharp contrast to the antiferromagnetic ground states in 2LCrI3 (↓↑ or ↑↓). Such a ferromagnetic-like loop is absent in either 2LCrI3 or FL CrCl3, suggesting its origin from interfacial magnetic inter-action. This phenomenon should not be due to the charge transfer/doping-induced antiferromagnetic-to-ferromagnetic transition repor-ted in 2L CrI314,15, which does not exhibit such a coexistence ofantiferromagnetic-type and ferromagnetic-type transitions. Recentworks have shown that a twist of two chromium trihalides layers maylead to noncollinear antiferromagnetic-ferromagnetic domains andthus finite MOKE signals9,10,18,24–27. However, this scenario is also lesslikely to be relevant, as such domains are predicted to emerge forsufficiently large moiré periodicity. Due to large lattice constantmismatch18, theCrI3/CrCl3 heterostructure canhardly form largemoiréperiodicity even at zero twist angle. Furthermore, the magnetizationbehaviors (including thefield and temperaturedependence)measuredin twisted CrI39,10,25–27 exhibit qualitative difference from what weobserve in our samples. We further rule out several other possibleorigins (seedetailed discussions in SupplementaryText 1).Weproposethat at least three spin layers (two layers of CrI3 plus one neighboringlayer of CrCl3) are responsible for the observed transitions. Theneighboring CrCl3 layer is acting as the third spin layer with out-of-plane magnetic order after being stacked in proximity with CrI3, asschematically shown in Fig. 1a and denoted in the red dashed rec-tangles in Fig. 1d, e. Note that a perpendicular magnetic field inducescanting of the planar CrCl3 spins, giving rise to a continuously varyingMOKE background28 (Supplementary Fig. 1), which is typically sub-tracted and eliminated from our MOKE signal. Therefore, only per-pendicular spin-flip transitions are discussed in this work.Similar to trilayer CrI3, in principle several potential anti-ferromagnetic configurations can be considered: ↑↓↓ (−1), ↓↑↑ (+1),↓↑↓ (−1),↑↓↑ (+1)with thefirst two and the third spins referring to the2L CrI3 and the neighboring CrCl3 layer respectively and the numbersin brackets denoting the net magnetic moments. To figure out thecoupling type for theneighboringCrI3 andCrCl3 layers,we study the 1LCrI3/FL CrCl3 heterostructure where the magnetic behavior candirectly verify the interlayer coupling type. Figure 1e shows θK of the 1LCrI3 and the 1L CrI3/FL CrCl3 heterostructure, fabricated from the sameCrI3 flake. Interestingly, both show a ferromagnetic behavior with asingle hysteresis loop. This observation indicates that the neighboringCrI3 and CrCl3 layer is ferromagnetically coupled, in contrast to theinterlayer antiferromagnetic coupling in FL CrI319. Careful inspectionon the hysteresis loop in Fig. 1d shows more transition steps, possiblydue to the switching of magnetic domains23,29. Note that due to thin-film optical interference effect23,30. it is not possible to associate themagnitudeor sign of theMOKE signal to themagnetization of differentsamples (e.g., the doubling of the MOKE signal in the 1L CrI3/FL CrCl3heterostructure relative to that in the 1L CrI3 in Fig. 1e does notimply that the probed magnetization doubles). In the following text,Fig. 1 | CrI3/CrCl3 heterostructures and MOKE measurements. a Schematics ofthemagnetic ground states in bilayer (2L) CrI3 and few-layer (FL) CrCl3 before (left)and after (right) forming heterostructure. Only four layers of CrCl3 are shown forsimplicity. b Optical micrograph of a 2L CrI3/FL CrCl3 heterostructure. c Atomicforce microscopy of the heterostructure in the same position as in b. The heightprofile (along the yellow dotted line in the image) at the edge of CrI3 indicates thethickness of a bilayer. d MOKE signal (after subtracting a polynomial background,asdone for allMOKE curves in themain text) of the 2LCrI3 region and the 2LCrI3/FLCrCl3 heterostructure region as a function of perpendicular magnetic field. Twocurves of each region represent forward and backward sweeps of the field,respectively. The data are taken at the spots marked by red in b. Insets depictmagnetic ground states of 2LCrI3 and the CrI3/CrCl3 heterostructure (showing onlythe interfacial CrCl3 layer, highlighted by red dashed rectangles). eMOKE signal ofanother monolayer (1L) CrI3/FL CrCl3 heterostructure, compared with that mea-sured in the 1L CrI3 (from the same CrI3 flake as in the heterostructure region).Article https://doi.org/10.1038/s41467-022-34812-6Nature Communications |         (2022) 13:7348 2we mainly focus on other features (e.g., emergent hysteresis loop,transition/coercive fields, critical temperatures). To exclude anyunique causes related to the stacking sequence, we also studiedreversely stacked heterostructures with FL CrCl3 on top of 2L CrI3 andobserved similar hysteresis loops (Supplementary Figs. 2 and 3). Thelarger coercive field of the hysteresis loop observed in the reversedstackmay be due to sample differences or twist angle dependence andis out of the scope of this work.We next study the temperature dependence of the magnetism inthe heterostructure. Figure 2a, b shows the temperature dependenceof θK in 2L CrI3 and 2L CrI3/FL CrCl3 heterostructure. The extractedH1,H1*, H2* and Δθ1, Δθ1*, Δθ2* as a function of temperature are shown inFig. 2c, d, respectively. The antiferromagnetic spin-flip transitions (atH1 and H1*) in both 2L CrI3 and the heterostructure disappear at tem-peratures larger than TC ~ 40K and is consistent with previous mea-surements in 2L CrI331. The ferromagnetic-like hysteresis loop (at H2*)observed only in the heterostructure region survives up to a highertemperature TC* ~ 48 K. Another experiment on 1L CrI3/FL CrCl3 het-erostructure shows critical temperatures of TC ~ 33K and TC* ~ 37 K for1L CrI3 region and heterostructure region, respectively (Supplemen-tary Fig. 4). In 2D magnets, the critical temperature is determined bythe spin-wave excitation gap, which is dictated by the anisotropiespresent in the system12,23,32,33. Our density functional theory resultssuggest an increase in the effective single-ion anisotropy of CrI3 whenbrought in proximity to CrCl3. On the other hand, thanks to theinduced ferromagnetic coupling, the CrCl3 layer now sees an effectiveanisotropy field that depends both on the interfacial ferromagneticcoupling as well as the anisotropy of CrI3, which is expected to enlargethe spin-wave gaps for both the materials. This is consistent with theobserved increase of TC for both 1L CrI3/FL CrCl3 and 2L CrI3/FL CrCl3systems.To better understand the observations, we explore the magneticground states of the CrI3/CrCl3 bilayer using first-principles calcula-tions (Supplementary Text 2). We find that the perpendicular ferro-magnetic state (↑↑) is more favorable than three other magneticconfigurations: perpendicular antiferromagnetic state (↑↓), the statesthat one layer is out-of-plane polarized while the other in-planepolarized (↑→or→↑). Further considerationonmagnetic dipole-dipoleinteraction and different commensurate twist angles (0° and 30°) doesnot undermine the favorable perpendicular ferromagnetic state. Theinterlayer exchange energy Jinter in CrI3/CrCl3 can be approximated bythe energy difference between perpendicular antiferromagnetic andferromagnetic configurations17. We estimate Jinter ≈ −77 (−64)μJ/m2 for0° (30°)-twisted CrI3/CrCl3, compared to the reported interlayerexchange ~80 μJ/m2 in 2L CrI332,34,35. Such interfacial exchange couplingin the heterostructure wins over the in-plane anisotropy of CrCl3 andresults in the out-of-plane magnetic order in the CrCl3 layer next toCrI3, in agreement with our observations.We next turn to explore the electrical tunability of the observedinterfacial magnetism. A unique aspect of the CrI3/CrCl3 hetero-structure, when comparedwith previously exploredmonolayer and/orhomobilayer systems14,15,36, is the absence of structural inversion sym-metry. The Neumann’s principle37 states that the spin-charge couplingis dictated by the symmetries of the system.We thus expect to observespin-charge coupling phenomena for the interlayer magnetic order. Inparticular, breaking of structural inversion allows for direct electric-field modification of the magnetic anisotropy and the interlayerexchange interactions via terms of the form (see detailed discussionsin Supplementary Text 3):Eelec mi,σi� �= σ1 � σ2� �β1m2z1 + β2m2z2 +β3m1 �m2� �, ð1Þwhere mi, σi are the magnetization and charges of the respectivelayers, ðσ1 � σ2Þ∼ electric field and β1,2,3 parameterizes the strength ofrespective interactions. Microscopically, the electric-field control ofinterfacialmagnetic interactions could arise fromelectric-field-inducedchanges in the orbital occupancy in conjunction with spin-orbitinteractions. Such a mechanism has attracted significant interest forconstructing low-dissipation spintronic memory and logic devices38,39.To check the electric-field tuning of the observed interfacialmagnetism, we fabricated a dual-gated 1L CrI3/FL CrCl3 device, asshown in Fig. 3a, b. This structure allows us to study themagnetizationof the CrI3/CrCl3 heterostructure (as well as that of the 1L CrI3 region inthe same device) under the top-gate voltage Vtg and back-gate voltageVbg. The two voltages are converted to electrostatic doping density nand displacement fieldD (Methods). Figure 3c shows the coercive field(Hc) in 1L CrI3/FL CrCl3 heterostructure increases from ~700Oe to~1000Oe when the D is tuned from −1.4 V nm−1 to 1 V nm−1, indicatingthe enhancement of the magnetic anisotropy of the interfacial ferro-magnetism in the heterostructure. The full mappings in Fig. 3d, epresent the extractedHc as a function of both n andD in 1L CrI3 and the1L CrI3/FL CrCl3 heterostructure, respectively. A quite weak modula-tion of Hc is observed in 1L CrI3, suggesting that the magnetism of 1LCrI3 can hardly be tuned under the range of gating voltages of thiswork. Separate experiments on FL CrCl3 demonstrate that the mag-netism of CrCl3 also can hardly be tuned by electrostatic gating(Supplementary Fig. 5d). However, significant tunability of the Hc isobserved by the D applied to the heterostructure. Such a dramatictunability in the CrI3/CrCl3 is in agreement with the electric field con-trol of interfacial magnetic interactions allowed by the structuralsymmetry breaking, predicted in the above theoretical analysis.The intriguing electrical tunability allowed by symmetry breaking isalso observed in a heterostructure containing a bilayer CrI3 (Supple-mentary Text 4).Fig. 2 | Temperature dependence of the magnetism of 2L CrI3 and 2L CrI3/FLCrCl3 heterostructure. MOKE signal in the 2L CrI3 region (a) and the 2L CrI3/FLCrCl3 heterostructure region (b) as a function of perpendicular magnetic field atdifferent temperatures. Critical fields H1, H1*, H2* and magnitudes in the change ofMOKE signal Δθ1, Δθ1*, Δθ2* of magnetic transitions are labeled. c Temperaturedependence of the critical fields of magnetic transitions. PM paramagnetic. Thecritical fields are extracted from the peak of derivative dθK/dH and the error barsare the peak widths. d Temperature dependence of the magnitudes in the changeof MOKE signal at magnetic transitions. Solid curves are fitted by a power-lawequation11. The critical temperatures TC and TC* are indicated. The error bars are theuncertainties in extracting the transition magnitudes.Article https://doi.org/10.1038/s41467-022-34812-6Nature Communications |         (2022) 13:7348 3In summary, we studied the interfacial magnetism in CrI3/CrCl3heterostructures and demonstrated the interfacial ferromagneticcoupling between neighboring CrI3 and CrCl3 layers. The demon-strated ability to engineer magnetoelectric phenomena by breakingsymmetries via vdW heterostructures provides opportunities for vdWspintronics. The compatible hybrids of 2D magnets with other quan-tummaterials, suchasunconventional superconductors, ferroelectricsor topological materials are predicted to demonstrate exotic topolo-gical phases and many-body interactions6,12,13, as well as to design newspintronic devices and therefore are highly desirable for further study.MethodsCrystal growthSingle crystal CrI3 was synthesized using the chemical vapor transport(CVT) method40. The Cr powder and iodine pieces were mixed with astoichiometric ratio and loaded into a quartz tube (inner diameter,10mm; length, 180mm). The quartz tube was sealed under vacuumand then transferred to a double temperature zones furnace. Thetemperatures of the hot and cold ends of the furnace were set at650 °C and 550 °C, respectively. The growth with such a temperaturegradient lasted for 7 days. Finally, the furnace was shut down, and thequartz tube naturally cooled down to room temperature. The blackplate-like CrI3 crystals can be found at the cold end of the quartz tube.Single crystal CrCl3 was grown by the CVT method. The com-mercial CrCl3 polycrystal powder (99.9%) was sealed in a silica tubewith a length of 200mmand an inner diameter of 14mm. The tubewaspumped down to 0.01 Pa and sealed under vacuum, and then placed ina two-zone horizontal tube furnace. The two growth zones were raisedslowly to 973 K and 823 K for 2 days, and then held there for another7 days. After that, the furnace was shut down and cooled down natu-rally. Shiny, plate-like crystals with lateral dimensions of up to severalmillimeters can be obtained from the growth.Device fabricationFL CrI3, CrCl3 and hexagonal boron nitride (hBN) flakes are exfo-liated onto the silicon wafer covered by 285-nm thermal oxide layer.Flakes with proper thickness are selected by optical contrast23 andlater confirmed by AFM and MOKE measurements. CrI3 flakes usedin this work have 1~2 layers and the FL CrCl3 flakes are around5~10 nm (0.6 nm for each layer) thick. Heterostructures of CrI3and CrCl3 are fabricated by the dry-transfer method and encapsu-lated between two hBN flakes with a typical thickness of ~10 nm.Specifically, a stamp made of a thin polycarbonate and poly-dimethylsiloxane is then employed to pick up the flakes in sequenceunder an optical microscope. In the end, the finished stack isdeposited onto the target substrate with polycarbonate on topwhich is removed by chloroform afterwards. The whole process isperformed inside a glovebox to avoid material degradation.The exposure time to air is kept below ten minutes before trans-ferring the fabricated sample into the measurement chamber andpumping down.For thedual-gatedheterostructuredevice andmagnetic tunnelingjunction device, FL graphene flakes are exfoliated and integrated intothe stack following the above processes. The target substrate is pre-patterned with electrodes fabricated by standard e-beam lithography,Au/Ti deposition and lift-off processes. The stack is carefully alignedand transferred onto the target pattern to make contact betweengraphene flakes and electrodes.MOKE microscopyThe polarization of a linearly polarized light reflected from amagneticmaterial will be rotated by a Kerr angle θK, which is proportional to themagnetization of thematerial. In this work, the incident light is normalto the sample plane and MOKE is in the polar geometry, meaning thatthe magnetic vector being probed is perpendicular to the samplesurface and parallel to the incident light. A balanced photodetectorand lock-in method are used to obtain theMOKE signal. A laser is usedhere with wavelength of 633 nm and power of 5 µW. The sample isplaced in a helium-flow optical cryostat with the temperature down to6K andmagnetic field (perpendicular to sample surface) up to 5 T. Thelaser is focused onto the sample surface by an objective with the spotdiameter of 0.5 µm.Electrical control of the dual-gated deviceTop-gate and back-gate voltages can be applied to the FL graphenegates in the heterostructure device, while the graphene contact to theheterostructure is grounded. The dual-gate structure allows indepen-dent control of the doping density and displacement field applied onthe heterostructure. The doping density n and displacement field D areextracted by the simple parallel plate capacitor model. For simplicity,the CrI3/CrCl3 heterostructure is regarded as one channel, on which thedoping density and electric field are applied. The quantum capacitanceof CrI3 and CrCl3 is much larger than that of graphene due to the nearlyflat bands of these two magnetic semiconductors15. Therefore, onlygeometric capacitances Cbg and Ctg are considered. The doping den-sity and displacement field can be written as n =Cbg � Vbg +Ctg � Vtgand D= ðDbg +DtgÞ=2 = ðεbg � Vbg=dbg � εtg � V tg=dtgÞ=2, respec-tively. The relative dielectric constant of hBN14 is εbg = εtg = 3. For thedevice in Fig. 3, the thicknesses of bottom hBN and top hBN areobtained by AFM measurement to be dbg = 19.6 nm and dtg = 14.9 nm,respectively.Data availabilityThedata supporting thefindings of this study are included in the paperand its Supplementary Information file. Further data sets are availablefrom the corresponding author on reasonable request.Fig. 3 | Electrical control of the magnetism in 1L CrI3/FL CrCl3 heterostructure.a, b Optical micrograph and schematic structure of a dual-gated 1L CrI3/FL CrCl3device. Three few-layer graphene (FLG) flakes are used as back/top gates and thecontact to the stack. c Normalized MOKE signal as a function of perpendicularmagnetic field in 1L CrI3/FL CrCl3 heterostructure under different displacementfields D = −1.4, −0.2, 1.0 V nm−1. Coercive field Hc as a function of electrostaticdoping density n and displacement field D of 1L CrI3 (d) and 1L CrI3/FL CrCl3 het-erostructure (e), respectively. Theblackdots ine correspond to theMOKEcurves inc. The data are taken at the spots marked by red in a.Article https://doi.org/10.1038/s41467-022-34812-6Nature Communications |         (2022) 13:7348 4Code availabilityThe code supporting the findings of this study is included in the paperand its Supplementary Information file. Further code sets are availablefrom the corresponding author on reasonable request.References1. Geim, A. K. & Grigorieva, I. V. Van der Waals heterostructures.Nature 499, 419–425 (2013).2. Ohtomo, A. & Hwang, H. Y. A high-mobility electron gas at theLaAlO3/SrTiO3 heterointerface. Nature 427, 423–426 (2004).3. Bert, J. A. et al. Direct imagingof the coexistenceof ferromagnetismand superconductivity at the LaAlO3/SrTiO3 interface. Nat. Phys. 7,767–771 (2011).4. Zhong, D. et al. Van der Waals engineering of ferromagnetic semi-conductor heterostructures for spin and valleytronics. Sci. Adv. 3,e1603113 (2017).5. Fu, H. X., Liu, C. X. & Yan, B. H. Exchange bias and quantumanomalous nomalous Hall effect in the MnBi2Te4/CrI3 hetero-structure. Sci. Adv. 6, eaaz0948 (2020).6. Kezilebieke, S. et al. Topological superconductivity in a van derWaals heterostructure. Nature 588, 424–428 (2020).7. Tong, Q. J., Liu, F., Xiao, J. & Yao, W. Skyrmions in the moiré of vander Waals 2D magnets. Nano Lett. 18, 7194–7199 (2018).8. Hejazi, K., Luo, Z.-X. & Balents, L. Heterobilayer moiré magnets:moiré skyrmions and commensurate-incommensurate transitions.Phys. Rev. B 104, L100406 (2021).9. Song, T. et al. Direct visualization of magnetic domains and moirémagnetism in twisted 2D magnets. Science 374, 1140–1144 (2021).10. Xu, Y. et al. Coexisting ferromagnetic–antiferromagnetic state intwisted bilayer CrI3. Nat. Nanotechnol. 17, 143–147 (2021).11. Fei, Z. Y. et al. Two-dimensional itinerant ferromagnetism in atom-ically thin Fe3GeTe2. Nat. Mater. 17, 778–782 (2018).12. Gibertini, M., Koperski, M., Morpurgo, A. F. & Novoselov, K. S.Magnetic 2D materials and heterostructures. Nat. Nanotechnol. 14,408–419 (2019).13. Burch, K. S.,Mandrus, D. & Park, J. G.Magnetism in two-dimensionalvan der Waals materials. Nature 563, 47–52 (2018).14. Huang, B. et al. Electrical control of 2D magnetism in bilayer CrI3.Nat. Nanotechnol. 13, 544–548 (2018).15. Jiang, S. W., Li, L. Z., Wang, Z. F., Mak, K. F. & Shan, J. Controllingmagnetism in 2DCrI3 by electrostatic doping.Nat. Nanotechnol. 13,549–553 (2018).16. Li, T. X. et al. Pressure-controlled interlayermagnetism in atomicallythin CrI3. Nat. Mater. 18, 1303–1308 (2019).17. Sivadas, N., Okamoto, S., Xu, X. D., Fennie, C. J. & Xiao, D. Stacking-dependent magnetism in bilayer CrI3. Nano Lett. 18,7658–7664 (2018).18. Akram, M. et al. Moiré skyrmions and chiral magnetic phases intwisted CrX3 (X = I, Br, and Cl) bilayers. Nano Lett. 21,6633–6639 (2021).19. Song, T. C. et al. Giant tunneling magnetoresistance in spin-filtervan der Waals heterostructures. Science 360, 1214–1218 (2018).20. Tartaglia, T. A. et al. Accessing newmagnetic regimes by tuning theligand spin-orbit coupling in van der Waals magnets. Sci. Adv. 6,eabb9379 (2020).21. Wang, Z. et al. Determining the phase diagram of atomically thinlayeredantiferromagnetCrCl3.Nat.Nanotechnol. 14, 1116–1122 (2019).22. Wu, M., Li, Z. L., Cao, T. & Louie, S. G. Physical origin of giant exci-tonic and magneto-optical responses in two-dimensional ferro-magnetic insulators. Nat. Commun. 10, 2371 (2019).23. Huang,B. et al. Layer-dependent ferromagnetism in a vanderWaalscrystal down to the monolayer limit. Nature 546, 270–273 (2017).24. Hejazi, K., Luo, Z. X. & Balents, L. Noncollinear phases in moirémagnets. Proc. Natl Acad. Sci. USA 117, 10721–10726 (2020).25. Xie, H. et al. Twist engineering of the two-dimensionalmagnetism indouble bilayer chromium triiodide homostructures. Nat. Phys. 18,30–36 (2021).26. Xie, H. et al. Evidence of noncollinear spin texture in magneticmoiré superlattices. arXiv preprint, arXiv:2204.01636 (2022).27. Cheng, G. et al. Electrically tunable moiré magnetism in twisteddouble bilayer antiferromagnets. arXiv preprint, arXiv:2204.03837(2022).28. Cai, X. H. et al. Atomically thin CrCl3: an in-plane layered anti-ferromagnetic insulator. Nano Lett. 19, 3993–3998 (2019).29. Sun, Q. C. et al. Magnetic domains and domain wall pinning inatomically thin CrBr3 revealed by nanoscale imaging. Nat. Com-mun. 12, 1989 (2021).30. Ma, Z. W. et al. Micro-MOKEwith optical interference in the study of2DCr2Ge2Te6 nanoflake basedmagnetic heterostructures. AIP Adv.9, 125116 (2019).31. Jiang, S. W., Shan, J. & Mak, K. F. Electric-field switching of two-dimensional vanderWaalsmagnets.Nat.Mater. 17, 406–410 (2018).32. Lu, X., Fei, R. & Yang, L. Curie temperature of emerging two-dimensional magnetic structures. Phys. Rev. B 100, 205409 (2019).33. Gong, C. et al. Discovery of intrinsic ferromagnetism in two-dimensional van der Waals crystals. Nature 546, 265–269 (2017).34. Lado, J. L. & Fernández-Rossier, J. On the origin of magnetic ani-sotropy in two dimensional CrI3. 2D Mater. 4, 035002 (2017).35. Cenker, J. et al. Direct observation of two-dimensional magnons inatomically thin CrI3. Nat. Phys. 17, 20–25 (2021).36. Rustagi, A., Solanki, A. B., Tserkovnyak, Y. & Upadhyaya, P. Coupledspin-charge dynamics in magnetic van der Waals heterostructures.Phys. Rev. B 102, 094421 (2020).37. Birss, R. R. Symmetry and Magnetism (North-Holland, 1966).38. Baek, S. H. C. et al. Complementary logic operation based onelectric-field controlled spin-orbit torques. Nat. Electron. 1,398–403 (2018).39. Lee, H., Ebrahimi, F., Amiri, P. K. & Wang, K. L. Low-power, high-density spintronic programmable logicwith voltage-gated spinHalleffect in magnetic tunnel junctions. IEEE Magn. Lett. 7,3102505 (2016).40. Liu, Y. & Petrovic, C. Three-dimensional magnetic critical behaviorin CrI3. Phys. Rev. B 97, 014420 (2018).AcknowledgementsWe thank Di Xiao, Wenguang Zhu for helpful discussions and Adam W.Tsen for help with crystals. The first-principles calculations have beendone on the supercomputing system in the Supercomputing Center ofthe University of Science and Technology of China. We acknowledgepartial support of the work from WPI-AIMR, Center for Science andInnovation in Spintronics, JSPS KAKENHI Basic Science A (18H03858),New Science (18H04473 and 20H04623), Tohoku University FRiDuoprogram, US Department of Defense (DOD) Multidisciplinary UniversityResearch Initiatives (MURI) program (FA9550-20-1-0322), US Departmentof Energy (DOE) Office of Science through the Quantum Science Center(QSC, a National Quantum Information Science Research Center), andVillum Foundation. For crystal synthesis: Z.M. acknowledges the supportby the US DOE under grants DE-SC0019068. H.L. acknowledges thesupport by National Key R&D Program of China (2018YFE0202600,2022YFA1403800), Beijing Natural Science Foundation (Z200005), andNational Natural Science Foundation of China (12274459). K.W. and T.T.acknowledge support from the Elemental Strategy Initiative conductedby the MEXT, Japan (JPMXP0112101001) and JSPS KAKENHI (19H05790,20H00354 and 21H05233).Author contributionsG.C. and Y.P.C. conceived the project. G.C. fabricated the devices andperformed experiments, assisted by A.L.A. M.M.R., Z.H., A.R., K.A.S. andArticle https://doi.org/10.1038/s41467-022-34812-6Nature Communications |         (2022) 13:7348 5P.U. performed supporting theoretical modeling. Y.Z. and Z.M. providedbulk CrI3 crystals. S.Y., S.T. and H.L. provided bulk CrCl3 crystals. K.W.and T.T. provided bulk hBN crystals. Y.P.C. supervised the project. G.C.,M.M.R., Z.H., P.U. and Y.P.C. wrote the manuscript with input from allauthors.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-022-34812-6.Correspondence and requests for materials should be addressed toYong P. Chen.Peer review information Nature Communications thanks Ping KwanWong and theother, anonymous, reviewer(s) for their contribution to thepeer review of this work.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons license, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons license and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2022Article https://doi.org/10.1038/s41467-022-34812-6Nature Communications |         (2022) 13:7348 6https://doi.org/10.1038/s41467-022-34812-6http://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/ Emergence of electric-field-tunable interfacial�ferromagnetism in 2D antiferromagnet�heterostructures Results and discussion Methods Crystal growth Device fabrication MOKE microscopy Electrical control of the dual-gated device Data availability Code availability References Acknowledgements Author contributions Competing interests Additional information