# Fileset

[Advanced Science - 2024 - Xiao - Polarity‐Reversal of Exchange Bias in van der Waals FePS3 Fe3GaTe2 Heterostructures.pdf](https://mdr.nims.go.jp/filesets/fc0b61bd-738a-4d07-b251-199635237a23/download)

## Creator

Han Xiao, Bingbing Lyu, Mengjuan Mi, Jian Yuan, Xiandong Zhang, Lixuan Yu, Qihui Cui, Chaofan Wang, Jun Song, Mingyuan Huang, Yufeng Tian, Liang Liu, [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Min Liu](https://orcid.org/0000-0001-5671-9653), [Yanfeng Guo](https://orcid.org/0000-0002-9386-4857), Shanpeng Wang, [Yilin Wang](https://orcid.org/0000-0002-6297-0331)

## Rights

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

## Other metadata

[Polarity‐Reversal of Exchange Bias in van der Waals FePS<sub>3</sub>/Fe<sub>3</sub>GaTe<sub>2</sub> Heterostructures](https://mdr.nims.go.jp/datasets/88fee441-55ca-4203-9677-97adfb89299d)

## Fulltext

Polarity‐Reversal of Exchange Bias in van der Waals FePS3/Fe3GaTe2 HeterostructuresRESEARCH ARTICLEwww.advancedscience.comPolarity-Reversal of Exchange Bias in van der WaalsFePS3/Fe3GaTe2 HeterostructuresHan Xiao, Bingbing Lyu, Mengjuan Mi, Jian Yuan, Xiandong Zhang, Lixuan Yu, Qihui Cui,Chaofan Wang, Jun Song, Mingyuan Huang, Yufeng Tian, Liang Liu, Takashi Taniguchi,Kenji Watanabe, Min Liu,* Yanfeng Guo,* Shanpeng Wang,* and Yilin Wang*Exchange bias (EB) in antiferromagnetic (AFM)/ferromagneticheterostructures is crucial for the advancement of spintronic devices and hasattracted significant attention. The common EB effect in van der Waalsheterostructures features a low blocking temperature (Tb) and a singlepolarity. In this work, a significant EB effect with a Tb up to 150 K is observedin FePS3/Fe3GaTe2 heterostructures, and in particular, the EB exhibits anunusual temperature-dependent polarity-reversal behavior. Under a highpositive field-cooling condition (e.g., 𝝁0H ≥ 0.5 T), a negative EB field (HEB) isobserved at low temperatures, and with increasing temperature, the HEBcrosses zero at ≈20 K, subsequently becomes positive and later approacheszero again at Tb. A model composed of a top FePS3/interfacialFePS3/Fe3GaTe2 sandwich structure is proposed. The charge transfer fromFe3GaTe2 to FePS3 at the interface induces net magnetic moments (∆M) inFePS3. The interface favors AFM coupling, and thus the reversal of ∆M of theinterfacial FePS3 leads to the polarity-reversal of EB. Moreover, the EB can beextended to the bare Fe3GaTe2 region of the Fe3GaTe2 flake partially coveredby FePS3. This work provides opportunities for a deeper understanding of theEB effect and opens a new route toward constructing novel spintronic devices.H. Xiao, B. Lyu, M. Mi, L. Yu, M. Liu, Y. WangSchool of Integrated CircuitsShandong Technology Center of Nanodevices and IntegrationState Key Laboratory of Crystal MaterialsShandong UniversityJinan 250100, ChinaE-mail: liumin@sdu.edu.cn; yilinwang@email.sdu.edu.cnJ. Yuan, Y. GuoSchool of Physical Science and TechnologyShanghaiTech UniversityShanghai 201210, ChinaE-mail: guoyf@shanghaitech.edu.cnX. Zhang, J. SongShandong Wanbo Technologies Co. LTDJinan 250100, ChinaThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/advs.202409210© 2024 The Author(s). Advanced Science published by Wiley-VCHGmbH. This is an open access article under the terms of the CreativeCommons Attribution License, which permits use, distribution andreproduction in any medium, provided the original work is properly cited.DOI: 10.1002/advs.2024092101. IntroductionThe magnetic proximity effect, originat-ing from the interfacial coupling of het-erostructures, facilitates a plurality of in-triguing physical phenomena, includingskyrmions,[1,2] magnons,[3,4] exchange bias(EB),[5] etc., and has garnered considerableattention. EB arises from the exchange cou-pling between a ferromagnetic (FM) layerand an antiferromagnetic (AFM) layer,[6,7]which causes the spins of the FM layer tobe pinned by the AFM layer,[8] or appliesa torque to the AFM layer.[9,10] The prop-erties of AFM and FM materials, includ-ing crystal structure, magnetic anisotropy,magnetic domains, as well as interface qual-ity, have a significant impact on the EBeffect.[5,11] Achieving a high-quality inter-face with 3D materials is challenging dueto lattice mismatch and atomic diffusionduring the epitaxial growth process.[12] Vander Waals (vdW) magnets facilitate to formQ. Cui, S. WangState Key Laboratory of Crystal MaterialsInstitute of Crystal MaterialsShandong UniversityJinan 250100, ChinaE-mail: wshp@sdu.edu.cnC. Wang, M. HuangDepartment of PhysicsSouthern University of Science and TechnologyShenzhen 518055, ChinaY. Tian, L. LiuSchool of PhysicsShandong UniversityJinan 250100, ChinaT. TaniguchiResearch Center for Materials NanoarchitectonicsNational Institute for Materials ScienceTsukuba 305-0044, JapanK. WatanabeResearch Center for Electronic and Optical MaterialsNational Institute for Materials ScienceTsukuba 305-0044, JapanY. GuoShanghaiTech Laboratory for Topological PhysicsShanghaiTech UniversityShanghai 201210, ChinaAdv. Sci. 2024, 11, 2409210 2409210 (1 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbHhttp://www.advancedscience.commailto:liumin@sdu.edu.cnmailto:yilinwang@email.sdu.edu.cnmailto:guoyf@shanghaitech.edu.cnhttps://doi.org/10.1002/advs.202409210http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/mailto:wshp@sdu.edu.cnhttp://crossmark.crossref.org/dialog/?doi=10.1002%2Fadvs.202409210&domain=pdf&date_stamp=2024-11-04www.advancedsciencenews.com www.advancedscience.coman atomically sharp and ultraclean interface because their sur-faces are free of dangling bonds,[13–16] and thus provide a desiredplatform for the investigation of EB.Significant EB has been broadly observed in variousvdW AFM/FM heterostructures, e.g., CrCl3/Fe3GeTe2,[17]MnPX3/Fe3GeTe2 (X = S, Se),[18,19] CrPS4/Fe3GeTe2,[20]CrOCl/Fe3GeTe2,[21] FePS3/Fe5GeTe2,[22] CrI3/MnBi2Te4,[23]etc. In the FePS3/Fe3GeTe2 heterostructure, the Curie temper-ature (TC) of Fe3GeTe2 was elevated from 150 to 180 K, andthe coercive field was also enhanced by ≈100%.[24] The EBeffect is dependent on the magnitude of the magnetic fieldduring the cooling process. In the CrPS4/(Fe0.74Co0.26)3GeTe2heterostructure, a negative EB was observed with a low coolingfield, while the EB was suppressed or even eliminated with anincreased cooling field.[20] In oxidized-Fe3GeTe2/Fe3GeTe2/CrSeheterostructure, a substantially large exchange bias field (HEB) of≈90 mT was observed at low temperature, and the HEB graduallydiminishes in magnitude as temperature increases until reach-ing the blocking temperature (Tb, the temperature at which theHEB becomes zero).[25] The EB effect can also be influenced byvarious extrinsic stimuli. For instance, in the FePSe3/Fe3GeTe2heterostructure, modulation of interface spacing through pres-sure engineering significantly enhances the HEB and Tb.[11] Inanother example, the application of a solid protonic gate in theFePS3/Fe5GeTe2 heterostructure toggles the presence or absenceof EB by directly introducing protons into the interface.[22]The HEB is sensitive to the cooling field, interface spacing,and carrier density, while the polarity of EB does not changewith varying temperatures. Polarity-reversal of EB is valuablefor the design of novel spintronic devices[26,27] and has beenobserved in certain 3D systems,[28,29] while it is scarce in vdWheterostructures.Among the intrinsic vdW magnets, the recently discoveredferromagnet Fe3GaTe2 has the highest TC (up to 380 K forbulk and 350 K for 9.5 nm nanosheet), and exhibits itiner-ant ferromagnetism with robust large perpendicular magneticanisotropy, positioning it as a promising candidate for practi-cal applications.[30–34] The Ising-type antiferromagnet FePS3 fea-tures a zigzag antiferromagnetic structure and strong out-of-plane magnetic anisotropy,[35] and maintains a constant Néeltemperature (TN) of 115 K regardless of thickness from bulk tomonolayer.[36] In addition, FePS3 has a band gap of 2.18 eV in athin layer and behaves as an insulator,[37] ensuring that currentmainly flows through the metallic (or semiconducting) layer inthe heterostructure through electrical measurements. Thus, thevdW heterostructure of FePS3/Fe3GaTe2 is appealing for explor-ing the EB effect.In this work, a significant EB was observed in FePS3/Fe3GaTe2heterostructures by both anomalous Hall effect (AHE) measure-ments and reflective magnetic circular dichroism (RMCD). Fur-thermore, the EB induced at the interface of the FePS3/Fe3GaTe2heterostructure can laterally extend to the exposed region ofthe same Fe3GaTe2 flake, and such non-local EB effect is at-tributed to the itinerant ferromagnetism and the single-domainstate of Fe3GaTe2. More interestingly, the EB features an unusualpolarity-reversal behavior (from negative EB to positive EB) withincreasing temperature under a suitable high-field cooling con-dition (e.g., 𝜇0H ≥ 0.5 T). An intuitive model consisting of topFePS3/interfacial FePS3/Fe3GaTe2 is proposed, and the reversalof the net magnetic moments of the interfacial FePS3 layer is re-sponsible for the polarity-reversal of EB.2. Results and Discussion2.1. Construction of FePS3/Fe3GaTe2 HeterostructureThin Fe3GaTe2 flakes, FePS3 flakes, and FePS3/Fe3GaTe2 het-erostructures were obtained by mechanical exfoliation and drytransfer techniques in a glove box, and details can be found in theexperimental sections. Figure 1a illustrates anomalous Hall resis-tance (Rxy) of Fe3GaTe2 flake with a thickness of ≈18 nm at dif-ferent temperatures, indicating the absence of EB in the isolatedFe3GaTe2 flakes. The upper and lower panels in Figure 1b showa schematic diagram and an optical image of the constructedFePS3/Fe3GaTe2 heterostructure (Device 1), respectively, wherea thin FePS3 layer (5 nm) is stacked on a Fe3GaTe2 layer (25 nm),and the Fe3GaTe2 layer is in direct contact with the pre-patteredelectrodes. The FePS3/Fe3GaTe2 heterostructure was protectedby covering a h-BN layer. Raman scattering techniques were em-ployed to assess the quality of the FePS3/Fe3GaTe2 heterostruc-ture, as shown in Figure 1c, the Raman signatures of the het-erostructure region display a superposition of the Raman peaksof individual Fe3GaTe2 and FePS3 flakes,[24,30,36] and no additionalpeaks are observed.2.2. Temperature- and Hcool-Dependent Polarity-Reversal of EBThe interfacial exchange interaction of FePS3/Fe3GaTe2 het-erostructure was initially investigated by AHE measurement.Figure 2a illustrates the typical temperature-dependent Rxy of De-vice 1 with a positive cooling field (Hcool) of 1 T. The Hcool, perpen-dicular to the surface of the device, was applied at 320 K, afterwhich the device was cooled down to a preset temperature formeasurements. An obvious asymmetrical hysteresis loop withrespect to the zero magnetic field was observed at low tempera-tures, indicating the emergence of EB. HEB, which characterizesthe strength of EB in a system, is defined by a relation HEB =(HC+ + HC−)/2 where HC+ and H−C represent the positive coer-cive field and negative coercive field, respectively.[7,20] At 2 K, theHEB is negative, indicating that the polarity of EB is negative. Astemperature increases, the HEB crosses zero and gradually be-comes positive, demonstrating that the polarity of EB reversesfrom negative to positive ≈20 K, and remains positive until Tb of150 K. To establish a direct comparison, the hysteresis loops ofDevice 1 at 10 and 20 K are shown in Figure 2b.When the direction of the cooling field is reversed, the signof HEB is also reversed. As shown in Figure 2c, when a nega-tive Hcool of −1 T was applied, the HEB was positive at 2 K, andas the temperature increased, the HEB crossed zero and grad-ually became negative, also demonstrating the polarity-reversalbehavior of EB with increasing temperature. Figure 2d com-pares the temperature-dependent HEB with Hcool = 1 T andHcool = −1 T. The overall trends of the temperature-dependentHEB for Hcool = ± 1 T are nearly symmetric, and both exhibitpolarity-reversal behavior of EB at ≈20 K. The amplitude of |HEB|exhibits a non-monotonic relationship with temperature: |HEB|Adv. Sci. 2024, 11, 2409210 2409210 (2 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 1. a) Temperature-dependent Rxy for isolated Fe3GaTe2 flake (18 nm). Inset: the optical image of the isolated Fe3GaTe2 device (scale bar20 μm). b) Schematic of the vdW FePS3/Fe3GaTe2 heterostructure covered with a h-BN layer (upper panel), and optical image of Device 1 with FePS3(5 nm)/Fe3GaTe2 (25 nm) heterostructure (lower panel, scale bar 20 μm). c) Raman spectra of FePS3/Fe3GaTe2 heterostructure, individual Fe3GaTe2flake, and individual FePS3 flake.initially decreases until ≈20 K, then increases to reach a max-imum value at 40–60 K, and later gradually decreases until fi-nally disappears at ≈150 K. Moreover, the slight asymmetry ob-served in HEB under Hcool = ± 1 T, especially at 40 K, maybe attributed to the pinning effect induced by interfacial de-fects formed during the fabrication of the heterostructure. Thepinning effect may be dependent on specific conditions, underHcool = ± 0.5 T, the temperature-dependent HEB appears moresymmetric, as shown in Figure S2 (Supporting Information). TheTb in FePS3/Fe3GaTe2 heterostructure significantly surpassesthat of FePS3/Fe5GeTe2 (Tb ≈30 K)[22] and FePSe3/Fe3GeTe2 (Tb≈20 K).[11] The substantially high Tb is attributed to the strongmagnetic coupling at the interface between the ultra-thin FePS3and Fe3GaTe2 layers, which is further demonstrated by the large|HEB| greater than 50 mT. Moreover, the Tb is higher than the TNof FePS3, which may arise from possible short-range spin corre-lation in FePS3.[38]According to the Meiklejohn–Bean model,[39] EB is also af-fected by the magnitude of the cooling field. The temperature-dependent Rxy of Device 1 under Hcool varied from 0.002, 0.2,0.5, and 3 T were further investigated. As shown in Figure 3a,under Hcool = 0.2 T, the negative EB is observed and theFigure 2. Significant EB effect in Device 1. a) Temperature-dependent Rxy under Hcool = 1 T. b) Comparison of Rxy at 10 and 20 K under Hcool = 1 T. c)Temperature-dependent Rxy under Hcool = −1 T. d) Temperature-dependent HEB under Hcool = ± 1 T.Adv. Sci. 2024, 11, 2409210 2409210 (3 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 3. a) Temperature-dependent Rxy for Device 1 under Hcool = 0.2 T. b) Temperature-dependent HEB for Device 1 under applied Hcool varied from0.002, 0.2, 0.5, 1 and 3 T. c) Temperature-dependent RMCD hysteresis loops of FePS3 (5 nm)/Fe3GaTe2 (9 nm). Inset: optical image of FePS3/Fe3GaTe2heterostructure (scale bar 10 μm). d) Thickness-dependent Tb of the FePS3 (5 nm)/Fe3GaTe2 (t nm) heterostructures. Error bars represent deviation.polarity remains negative until Tb of 150 K. The non-monotonicrelationship between |HEB| and temperature is also observed,where |HEB| first increases and then decreases as the tempera-ture increases. The temperature-dependent Rxy of Device 1 underother cooling fields is illustrated and discussed in supporting in-formation (Figures S3, S4, Supporting Information). Figure 3bsummarizes the temperature-dependent HEB under differentmagnitudes of Hcool. As the temperature increases, the polarityof EB reverses from negative to positive at ≈20 K for a suitablehigh cooling field of Hcool ≥ 0.5 T, while remaining negative for arelatively low cooling field of Hcool ≤ 0.2 T, evidencing that thetemperature-dependent polarity-reversal of EB requires a suit-able field-cooling condition. In addition, the maximum values of|HEB| slightly vary depending on the magnitudes of Hcool, andreach a maximum |HEB| of 76 mT for Hcool ≤ 0.2 T. The Tb isunaffected by the magnitudes of Hcool, and the EB effect undervarying cooling fields disappears at ≈150 K. Further, as shownin Figure S5 (Supporting Information), the similar temperature-dependent polarity-reversal behavior of EB was also observed inother devices with different Fe3GaTe2 thickness (30, 15, 13 and10 nm), and the temperature at which the polarity reverses de-crease as the Fe3GaTe2 thickness decreases. Therefore, it is ev-ident that the temperature-dependent polarity reversal of EB inFePS3/Fe3GaTe2 heterostructures is robust and repeatable.The EB effect in FePS3/Fe3GaTe2 heterostructure was also ob-served by RMCD, which reveals the interfacial coupling at the mi-croscopic scale.[21] The temperature-dependent RMCD hysteresisloops of isolated Fe3GaTe2 (10 nm) are symmetric with respect tothe zero magnetic field (Figure S6, Supporting Information), re-vealing the absence of EB, which is also consistent with the AHEmeasurements. Figure 3c illustrates the temperature-dependentRMCD hysteresis loops of the FePS3/Fe3GaTe2 (5 nm/9 nm) het-erostructure, and the corresponding optical image is shown inthe inset of Figure 3c. A negative EB with an absolute value |HEB|of 126 mT was observed at 2 K, and |HEB| gradually decreasesas the temperature increases and becomes zero until Tb of 40 K.Moreover, because of the zero field-cooling condition for RMCDmeasurements, the polarity-reversal behavior of EB was not ob-served.Figure 3d summarizes the variation of Tb as a function ofthe Fe3GaTe2 thickness. The Tb monotonically increases from40 to 170 K as the FM layer thickness increases from 9 to30 nm with the FePS3 thickness fixed at 5 nm. The maxi-mum absolute value of HEB also monotonically increases from≈38 to ≈120 mT with increasing the Fe3GaTe2 layer thick-ness, as shown in Figure S7 (Supporting Information). Such asimilar relationship between the strength of EB and FM layerthickness was also observed in other heterostructures such asCrCl3/Fe3GeTe2,[17] CrPS4/(Fe0.74Co0.26)3GeTe2[20] and oxidized-Fe3GeTe2/Fe3GeTe2.[40] The enhancement of EB as the FM layerthickness increases is attributed to that the FM flakes with largerthickness have higher volume magnetization[17,20,40] and favorimproved interface quality due to the reduction of surface rough-ness of the FM layer.[41,42]Adv. Sci. 2024, 11, 2409210 2409210 (4 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 4. a) Schematic of the half-covered FePS3/Fe3GaTe2 heterostructure measured by RMCD (upper panel). RMCD hysteresis loops ofFePS3/Fe3GaTe2 region (region A) and exposed Fe3GaTe2 region (region B) at 2 K (lower panel). b) Schematic of the half-covered FePS3/Fe3GaTe2device measured by AHE (upper panel). Temperature-dependent Rxy of exposed Fe3GaTe2 region under Hcool = 0.5 T, which is connected to the regionof the same Fe3GaTe2 flake (15 nm) covered by FePS3 (5 nm) in Device 2 (lower panel).2.3. Non-Local EB EffectParticularly, the non-local EB effect was observed inFePS3/Fe3GaTe2 heterostructures. The FePS3/Fe3GaTe2 het-erostructures with varied interfacial configurations were con-structed, where the interfacial configuration is adjusted byselectively exposing or covering part of the Fe3GaTe2 flakes, asillustrated in the upper panels of Figures 4a,b. As shown inFigure 4a, the RMCD was conducted on two separated regions(labeled A and B in Figure 4a) of the same Fe3GaTe2 flake,region A was directly covered by FePS3 flake while region B wasexposed. RMCD hysteresis loops of both A and B regions exhibitsimilar features, where the negative EB and a substantially large|HEB| even up to 200 mT at 2 K were observed. In addition, thetemperature-dependent Rxy measured on a bare region of theFe3GaTe2 flake (Device 2, the other region of the same Fe3GaTe2flake was covered by FePS3) also reveals the non-local EB effect.As shown in the lower panel of Figure 4b, a negative EB, asubstantially large |HEB| up to 40 mT, and a Tb of ≈90 K wereobserved. The Tb and negative EB of Device 2 with Fe3GaTe2partially covered by FePS3 are consistent with those of Device 4with Fe3GaTe2 having a similar thickness and fully covered byFePS3 (Figure S8, Supporting Information). Such phenomenaindicate that EB can effectively propagate in the transversedirection at the micrometer scale. Manipulating magnetizationat lateral scales provides a promising way for constructingnovel spin logic devices such as magnetoelectric spin-orbit logic(MESO).[43] The non-local EB effect is attributed to the itinerantferromagnetic properties and the single magnetic domain ofFe3GaTe2 flakes.[11,44]2.4. Physical Mechanism of Polarity-Reversal of EBThe observed polarity-reversal of EB as the temperature in-creases under a suitable high Hcool is intriguing and rarely re-ported in other vdW heterostructures. For FePS3/Fe3GaTe2 het-erostructure, FePS3 possesses a higher work function (4.9 eV,insulating)[45] compared to Fe3GaTe2 (≈4.4 eV, metallic),[46,47]such that charge transfer from Fe3GaTe2 to FePS3 occurs at theinterface, as illustrated in the upper panel of Figure 5a. Elec-tron doping causes the magnetic ground state of FePS3 to trans-form from AFM order to ferrimagnetic (FIM) order, as demon-strated by the intercalation of organic cations in FePS3.[48] Inthe FIM order of electron-doped FePS3, the two ferromagneticzigzag chains of Fe atoms possess unequal magnetic moments,producing net magnetic moments (∆M), as shown in the lowerpanel of Figure 5a. The FIM-FePS3 retains the large coercivefield of FePS3, which prevents ∆M from being fully alignedwithin the sweeping magnetic field range during EB measure-ments at low temperatures, such that an unusual temperature-dependent hysteresis loop of FIM-FePS3 was observed.[48] Asshown in Figure S9 (Supporting Information), as temperature in-creases to 60 K, the hysteresis loops become more pronounced,and ∆M gradually increases, which is ascribed to that thermalkinetic energy promotes the alignment of magnetic moments.Thermal fluctuations would weaken the exchange coupling as thetemperature increases, therefore the trade-off between the mag-nitude of ∆M and thermal fluctuations lead to the maximum EBin a range of 40–60 K as the temperature increases.The mechanisms of the negative EB under low Hcooland temperature-dependent polarity-reversal (from negative toAdv. Sci. 2024, 11, 2409210 2409210 (5 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 5. a) Upper panel: schematic of electron transfer from Fe3GaTe2 to FePS3 at the interface. Lower panel: the zigzag AFM order of FePS3 and theFIM order with ∆M of doped FePS3, where arrows indicate the orientation and magnitude of the magnetic moments of Fe atoms. b,c) Upper panels:schematic of a model for the magnetic configurations of top-FePS3/inter-FePS3/Fe3GaTe2 under low Hcool b) and high Hcool c), respectively. Arrowsindicate the orientation and magnitude of interfacial Fe3GaTe2 or ∆M of inter-FePS3 and top-FePS3. Lower panels: negative EB under low Hcool, negativeEB under high Hcool at a low temperature c-i), and positive EB under high Hcool at a high temperature (c-iii), respectively. (c-ii) Schematic diagramshowing the upward rotation of ∆M of the inter-FePS3 as temperature increases.positive) of EB under high Hcool were then discussed. If the inter-facial exchange coupling is FM, a negative EB is always formedregardless of the field-cooling conditions. If the interfacial ex-change coupling is AFM, the polarity of EB is dependent on thefield-cooling conditions.[7,49] Therefore, the observation of posi-tive EB under high Hcool in FePS3/Fe3GaTe2 heterostructures in-dicates that their interfacial exchange coupling is AFM.Moreover, the temperature-dependent polarity-reversal of EBcannot be explained by a simple AFM coupling interface, and arelatively complex model composed of multiple structures is nec-essary to understand such phenomena. As shown in Figure 5b,the FePS3/Fe3GaTe2 heterostructure can be divided into threeportions[19,50]: the bottom Fe3GaTe2 portion, the interfacial FePS3portion in direct contact with Fe3GaTe2 (inter-FePS3), and thetop FePS3 portion (top-FePS3). The magnetic configurations ofthe FePS3/Fe3GaTe2 heterostructure are determined by the com-petition between the energy of the interfacial exchange couplingat the interfacial-FePS3/Fe3GaTe2 interface (Eex), the interlayermagnetic interaction between the top-FePS3 and inter-FePS3portions of doped FePS3 layer (Einter), and the Zeeman energy foraligning the magnetic moments (EZ).[8] Take the heterostructurebeing field-cooled along a positive direction as an example.The EZ tends to make the magnetic moments of Fe3GaTe2and the ∆M of both inter-FePS3 and top-FePS3 point to thedirection of the external magnetic field (upward); the Eex tendsto make the ∆M of inter-FePS3 point to a direction (downward)opposite to that of the magnetic moments of Fe3GaTe2 due tothe interfacial AFM coupling; and the Einter tends to make the∆M of inter-FePS3 and the ∆M of top-FePS3 align in the samedirection.When Hcool is low, EZ is substantially small, causing the ∆M ofinter-FePS3 to point downward driven by the Eex, and the ∆M oftop-FePS3 to point downward driven by the Einter, such that thecorresponding magnetic configuration of the top-FePS3/inter-FePS3/Fe3GaTe2 is initially formed as ↓↓/↓↓/↑↑, as illustrated inthe upper panel of Figure 5b. To overcome the AFM interfacialcoupling, a larger magnitude of HC- H−C is required to reversethe magnetic moments of Fe3GaTe2 during the backward pro-cess, while a smaller magnitude of HC+ H+C is needed during theforward process, and thus a negative EB is observed, as shown inthe lower panel of Figure 5b.When enhancing Hcool to a high level, EZ increases, andthe ∆M of top-FePS3 points upward driven by the EZ, whilethe direction of ∆M of the inter-FePS3 is determined by thecompetition between Eex, EZ, and Einter. At low temperatures,Eex dominates the magnetic configuration of the inter-FePS3(Eex > EZ + Einter), and thus the ∆M of inter-FePS3 point down-ward, and the corresponding magnetic configuration of the top-FePS3/inter-FePS3/Fe3GaTe2 is initially formed as ↑↑/↓↓/↑↑, asillustrated in the upper panel of Figure 5c (i). Because the ex-change interactions are short-range interactions, the magneticconfiguration of inter-FePS3 dominates the EB effect. In view ofthis, similar to the case of the heterostructure cooled with a lowHcool, a negative EB is observed, as shown in the lower panel ofFigure 5c (i). As the temperature increases, thermal fluctuationincreases and the Eex decreases, the Eex is no longer sufficientAdv. Sci. 2024, 11, 2409210 2409210 (6 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comenough to pin the ∆M of inter-FePS3 pointing downward, andthe ∆M of inter-FePS3 gradually rotates upward driven by the EZand Einter, as illustrated in Figure 5c (ii). Therefore, at a suitablehigh temperature, the ∆M of inter-FePS3 points upward, and thecorresponding magnetic configuration of the top-FePS3/inter-FePS3/Fe3GaTe2 is formed as ↑↑/↑↑/↑↑, as illustrated in the up-per panel of Figure 5c (iii). Benefiting from AFM interfacial cou-pling, a smaller magnitude of HC- H−C is required to reverse themagnetic moments of Fe3GaTe2 during the backward process,while a larger magnitude of HC+ H+C is needed during the for-ward process, and thus a positive HEB shown in the lower panel ofFigure 5c (iii) is observed. Micromagnetic simulations based onthe proposed sandwich structure were also conducted to simulatethe process of gradually rotating ∆M in the inter-FePS3 layer astemperature increases under the competition between Einter andEex, as shown in Figure S10 (Supporting Information). The re-sults of micromagnetic simulations reveal the possible reversalof ∆M of the inter-FePS3 as the temperature increases, whichleads to the temperature-dependent polarity reversal of EB.3. ConclusionIn summary, an obvious and considerable EB effect was observedin the FePS3/Fe3GaTe2 heterostructure, and in particular, the EBeffect exhibits non-local properties and temperature-dependentpolarity-reversal behavior. Due to the itinerant ferromagnetismand single magnetic domain state of Fe3GaTe2, for the Fe3GaTe2flake partially covered with FePS3, the bare region exhibits sim-ilar EB characteristics as the covered region. The temperature-dependent polarity-reversal behavior of EB is cooling field de-pendent, where with increasing temperature, the negative EB re-verses to the positive EB under suitable high cooling field condi-tions, while the negative polarity of EB remains unchanged underrelatively low cooling field conditions. An intuitive model com-posed of top-FePS3/inter-FePS3/Fe3GaTe2 sandwich structure isproposed to understand the polarity-reversal behavior, where thecharge transfer from Fe3GaTe2 to FePS3 at the interface induces∆M in FePS3, and the orientations of ∆M of the interfacial FePS3layer, which is determined by the competition between Eex, Einterand EZ, dominates the polarity of EB. The findings of this workprovide new insights for understanding the EB effect in vari-ous vdW heterostructures and offer emerging strategies for con-structing innovative spin logic devices.4. Experimental SectionDevice Fabrication: Fe3GaTe2 and FePS3 bulk crystals were grownby the flux method. To minimize the oxidation and contamination, thinFe3GaTe2 and FePS3 flakes were obtained by mechanical exfoliation,and the heterostructures were prepared by dry transfer techniques withpoly(dimethylsiloxane) (PDMS) stamps onto the pre-pattered electrodesin a glove box filled with Ar gas, where H2O and O2 levels were maintainedbelow 0.1 ppm. Before performing characterization and measurements,the devices were encapsulated by h-BN.Transport Measurements and RMCD Measurements: Electrical trans-port measurements were performed in a commercial physical propertymeasurement system (PPMS, Quantum Design Dynacool-9) with lock-intechniques (SR830). The RMCD measurements were conducted using astandard lab-made RMCD setup and the wavelength of HeNe laser was633 nm.Characterization: Raman measurements were performed via an inViaconfocal Raman microscope (Renishaw) with a 532 nm laser at room tem-perature. The laser power was kept below 0.05 mW to avoid local heating.Atomic force microscopy (AFM) images were carried out by a Benyuan sys-tem (CSPM5500). The thickness was determined by AFM and color con-trast analysis based on optical images.Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsH.X. and B.L. contributed equally to this work. This work was sup-ported by the National Natural Science Foundation of China (Grant Nos.92065206, 52372011, U22A20123, 12304042), the National Key R&D Pro-gram of China (Grant No. 2022YFA1602704), the Natural Science Foun-dation of Shandong Province (Grant Nos. ZR2023ZD10, ZR2022MF228),Guangdong Provincial Quantum Science Strategic Initiative (Grant No.GDZX2301008), Basic and Applied Basic Research Foundation of Guang-dong Province (Grant Nos. 2023A1515110508), Postdoctoral FellowshipProgram of CPSF (Grant No. GZC20231434), Postdoctoral InnovationProgram of Shandong Province (Grant No. SDCX-ZG-202400329), theOpen Project of Guangdong Provincial Key Laboratory of MagnetoelectricPhysics and Devices (Grant No. 2022B1212010008), Open Project of Chi-nese Academy of Science Sharing Service Platform of CAS Large ResearchInfrastructures (Grant No. 2023SECUFPT001399), the Shanghai Scienceand Technology Innovation Action Plan (Grant No. 21JC1402000) and theDouble First-Class Initiative Fund of ShanghaiTech University.Conflict of InterestThe authors declare no conflict of interest.Data Availability StatementThe data that support the findings of this study are available from the cor-responding author upon reasonable request.Keywordsexchange bias, non-local manipulation, polarity-reversal, van der WaalsheterostructureReceived: August 6, 2024Revised: October 16, 2024Published online: November 4, 2024[1] Y. Wu, S. Zhang, J. Zhang, W. Wang, Y. L. Zhu, J. Hu, G. Yin, K. Wong,C. Fang, C. Wan, X. Han, Q. Shao, T. Taniguchi, K. Watanabe, J. Zang,Z. Mao, X. Zhang, K. L. Wang, Nat. Commun. 2020, 11, 3860.[2] Y. Wu, B. Francisco, Z. Chen, W. Wang, Y. Zhang, C. Wan, X. Han, H.Chi, Y. Hou, A. Lodesani, G. Yin, K. Liu, Y.-T. Cui, K. L. Wang, J. S.Moodera, Adv. Mater. 2022, 34, 2110583.[3] K. Li, L. Wang, Y. Wang, Y. Guo, S. Lv, Y. He, W. Lin, T. Min, S. Hu,S. Yang, D. Xue, A. Zheng, S. Yang, X. Ding, Adv. Mater. 2024, 36,2312137.Adv. Sci. 2024, 11, 2409210 2409210 (7 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.com[4] I. A. Golovchanskiy, N. N. Abramov, V. S. Stolyarov, V. V. Bolginov,V. V. Ryazanov, A. A. Golubov, A. V. Ustinov, Adv. Funct. Mater. 2018,28, 1802375.[5] L. Pan, A. Grutter, P. Zhang, X. Che, T. Nozaki, A. Stern, M. Street, B.Zhang, B. Casas, Q. L. He, E. S. Choi, S. M. Disseler, D. A. Gilbert, G.Yin, Q. Shao, P. Deng, Y. Wu, X. Liu, X. Kou, S. Masashi, X. Han, C.Binek, S. Chambers, J. Xia, K. L. Wang, Adv. Mater. 2020, 32, 2001460.[6] W. H. Meiklejohn, C. P. Bean, Phys. Rev. 1957, 105, 904.[7] J. Nogués, I. K. Schuller, J. Magn. Magn. Mater. 1999, 192, 203.[8] F. Wang, D. Xiao, W. Yuan, J. Jiang, Y.-F. Zhao, L. Zhang, Y. Yao, W.Liu, Z. Zhang, C. Liu, J. Shi, W. Han, M. H. W. Chan, N. Samarth,C.-Z. Chang, Nano Lett. 2019, 19, 2945.[9] T. M. J. Cham, R. J. Dorrian, X. S. Zhang, A. H. Dismukes, D. G. Chica,A. F. May, X. Roy, D. A. Muller, D. C. Ralph, Y. K. Luo, Adv. Mater. 2024,36, 2305739.[10] S. Peng, D. Zhu, W. Li, H. Wu, A. J. Grutter, D. A. Gilbert, J. Lu, D.Xiong, W. Cai, P. Shafer, K. L. Wang, W. Zhao, Nat. Electron. 2020, 3,757.[11] X. Huang, L. Zhang, L. Tong, Z. Li, Z. Peng, R. Lin, W. Shi, K.-H. Xue,H. Dai, H. Cheng, D. de Camargo Branco, J. Xu, J. Han, G. J. Cheng,X. Miao, L. Ye, Nat. Commun. 2023, 14, 2190.[12] H. Aboulfadl, F. Seifried, M. Stueber, F. Muecklich, Mater. Lett. 2019,236, 92.[13] Y. Liu, Y. Huang, X. Duan, Nature. 2019, 567, 323.[14] K. S. Novoselov, A. K. Geim, S. V. Morozov, D.-e. Jiang, Y. Zhang, S.V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science. 2004, 306, 666.[15] C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, W. Bao, C. Wang, Y.Wang, Nature. 2017, 546, 265.[16] B. Huang, G. Clark, E. Navarro-Moratalla, D. R. Klein, R. Cheng, K.L. Seyler, D. Zhong, E. Schmidgall, M. A. McGuire, D. H. Cobden,Nature. 2017, 546, 270.[17] R. Zhu, W. Zhang, W. Shen, P. K. J. Wong, Q. Wang, Q. Liang, Z. Tian,Y. Zhai, C.-w. Qiu, A. T. S. Wee, Nano Lett. 2020, 20, 5030.[18] H. Dai, H. Cheng, M. Cai, Q. Hao, Y. Xing, H. Chen, X. Chen, X. Wang,J.-B. Han, ACS Appl. Mater. Interfaces. 2021, 13, 24314.[19] G. Hu, Y. Zhu, J. Xiang, T.-Y. Yang, M. Huang, Z. Wang, Z. Wang, P. Liu,Y. Zhang, C. Feng, D. Hou, W. Zhu, M. Gu, C.-H. Hsu, F.-C. Chuang,Y. Lu, B. Xiang, Y.-L. Chueh, ACS Nano. 2020, 14, 12037.[20] K. Gu, X. Zhang, X. Liu, X. Guo, Z. Wu, S. Wang, Q. Song, W. Wang,L. Wei, P. Liu, J. Ma, Y. Xu, W. Niu, Y. Pu, Adv. Sci. 2024, 11, 2307034.[21] T. Zhang, Y. Zhang, M. Huang, B. Li, Y. Sun, Z. Qu, X. Duan, C. Jiang,S. Yang, Adv. Sci. 2022, 9, 2105483.[22] S. Albarakati, W.-Q. Xie, C. Tan, G. Zheng, M. Algarni, J. Li, J. Partridge,M. J. S. Spencer, L. Farrar, Y. Xiong, M. Tian, X. Wang, Y.-J. Zhao, L.Wang, Nano Lett. 2022, 22, 6166.[23] Z. Ying, B. Chen, C. Li, B. Wei, Z. Dai, F. Guo, D. Pan, H. Zhang, D.Wu, X. Wang, S. Zhang, F. Fei, F. Song, Nano Lett. 2023, 23, 765.[24] L. Zhang, X. Huang, H. Dai, M. Wang, H. Cheng, L. Tong, Z. Li, X.Han, X. Wang, L. Ye, J. Han, Adv. Mater. 2020, 32, 2002032.[25] Y. Wu, W. Wang, L. Pan, K. L. Wang, Adv. Mater. 2022, 34, 2105266.[26] A. van den Brink, G. Vermijs, A. Solignac, J. Koo, J. T. Kohlhepp, H. J.M. Swagten, B. Koopmans, Nat. Commun. 2016, 7, 10854.[27] Y.-W. Oh, S.-h. C. Baek, Y. Kim, H. Y. Lee, K.-D. Lee, C.-G. Yang, E.-S.Park, K.-S. Lee, K.-W. Kim, G. Go, Nat. Nanotechnol. 2016, 11, 878.[28] Deepak, A. K., S. M. Yusuf, Phys. Rev. Mater. 2021, 5, 124402.[29] F. Hong, Z. Cheng, J. Wang, X. Wang, S. Dou, Appl. Phys. Lett. 2012,101, 121913.[30] G. Zhang, F. Guo, H. Wu, X. Wen, L. Yang, W. Jin, W. Zhang, H. Chang,Nat. Commun. 2022, 13, 5067.[31] M. Wang, B. Lei, K. Zhu, Y. Deng, M. Tian, Z. Xiang, T. Wu, X. Chen,npj 2D Mater. Appl. 2024, 8, 22.[32] W. Jin, G. Zhang, H. Wu, L. Yang, W. Zhang, H. Chang, ACS Appl.Mater. Interfaces. 2023, 15, 36519.[33] W. Li, W. Zhu, G. Zhang, H. Wu, S. Zhu, R. Li, E. Zhang, X. Zhang,Y. Deng, J. Zhang, L. Zhao, H. Chang, K. Wang, Adv. Mater. 2023, 35,2303688.[34] S. N. Kajale, T. Nguyen, C. A. Chao, D. C. Bono, A. Boonkird, M. Li,D. Sarkar, Nat. Commun. 2024, 15, 1485.[35] A. R. Wildes, D. Lançon, M. K. Chan, F. Weickert, N. Harrison, V.Simonet, M. E. Zhitomirsky, M. V. Gvozdikova, T. Ziman, H. M.Rønnow, Phys. Rev. B. 2020, 101, 024415.[36] J.-U. Lee, S. Lee, J. H. Ryoo, S. Kang, T. Y. Kim, P. Kim, C.-H. Park, J.-G.Park, H. Cheong, Nano Lett. 2016, 16, 7433.[37] Z. Cheng, T. A. Shifa, F. Wang, Y. Gao, P. He, K. Zhang, C. Jiang, Q.Liu, J. He, Adv. Mater. 2018, 30, 1707433.[38] A. Bhutani, J. L. Zuo, R. D. McAuliffe, C. R. dela Cruz, D. P. Shoemaker,Phys. Rev. Mater. 2020, 4, 034411.[39] A. P. Malozemoff, Phys. Rev. B. 1987, 35, 3679.[40] H. K. Gweon, S. Y. Lee, H. Y. Kwon, J. Jeong, H. J. Chang, K.-W. Kim,Z. Q. Qiu, H. Ryu, C. Jang, J. W. Choi, Nano Lett. 2021, 21, 1672.[41] Z. Ye, A. Balkanci, A. Martini, M. Z. Baykara, Phys. Rev. B. 2017, 96,115401.[42] J. C. Spear, J. P. Custer, J. D. Batteas, Nanoscale. 2015, 7, 10021.[43] S. Manipatruni, D. E. Nikonov, C.-C. Lin, T. A. Gosavi, H. Liu, B.Prasad, Y.-L. Huang, E. Bonturim, R. Ramesh, I. A. Young, Nature.2019, 565, 35.[44] H. Dai, M. Cai, Q. Hao, Q. Liu, Y. Xing, H. Chen, X. Chen, X. Wang,H.-H. Fu, J. Han, ACS Nano. 2022, 16, 12437.[45] M. Ramos, M. Gadea, S. Mañas-Valero, C. Boix-Constant, E.Henríquez-Guerra, M. A. Díaz-García, E. Coronado, M. R. Calvo,Nanoscale Adv. 2024, 6, 1909.[46] E. Ko, Sci. Rep. 2022, 12, 5101.[47] Y. Zhang, X. Li, J. Sheng, S. Yu, J. Zhang, Y. Su, Appl. Phys. Lett. 2023,123, 192402.[48] M.-J. Mi, L.-X. Yu, H. Xiao, B.-B. Lü, Y.-L. Wang, Acta. Phys. Sin. 2024,73, 057501.[49] J. Nogués, D. Lederman, T. Moran, I. K. Schuller, Phys. Rev. Lett. 1996,76, 4624.[50] W. Qi, F. Fei, Z. Zhang, B. Chen, H. Xie, B. Wei, S. Zhang, F. Song,Appl. Phys. Lett. 2024, 125, 023101.Adv. Sci. 2024, 11, 2409210 2409210 (8 of 8) © 2024 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2024, 48, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202409210 by Cochrane Japan, Wiley Online Library on [28/12/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.com Polarity-Reversal of Exchange Bias in van der Waals FePS3/Fe3GaTe2 Heterostructures 1. Introduction 2. Results and Discussion 2.1. Construction of FePS3/Fe3GaTe2 Heterostructure 2.2. Temperature- and Hcool-Dependent Polarity-Reversal of EB 2.3. Non-Local EB Effect 2.4. Physical Mechanism of Polarity-Reversal of EB 3. Conclusion 4. Experimental Section Supporting Information Acknowledgements Conflict of Interest Data Availability Statement Keywords