# Fileset

[Adv Funct Materials - 2023 - Lebedev - Electrical Interrogation of Thickness‐Dependent Multiferroic Phase Transitions in.pdf](https://mdr.nims.go.jp/filesets/f6245c2f-005d-4297-91fe-6596dae0e520/download)

## Creator

Dmitry Lebedev, Jonathan Tyler Gish, Ethan Skyler Garvey, Teodor Kosev Stanev, Junhwan Choi, Leonidas Georgopoulos, Thomas Wei Song, Hong Youl Park, [Kenji Watanabe](https://orcid.org/0000-0003-3701-8119), [Takashi Taniguchi](https://orcid.org/0000-0002-1467-3105), Nathaniel Patrick Stern, Vinod Kumar Sangwan, Mark Christopher Hersam

## Rights

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

## Other metadata

[Electrical Interrogation of Thickness‐Dependent Multiferroic Phase Transitions in the 2D Antiferromagnetic Semiconductor NiI            <sub>2</sub>](https://mdr.nims.go.jp/datasets/8013f95e-bc45-41cf-9b57-6126aa84c232)

## Fulltext

Electrical Interrogation of Thickness‐Dependent Multiferroic Phase Transitions in the 2D Antiferromagnetic Semiconductor NiI2www.afm-journal.de2212568 (1 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHElectrical Interrogation of Thickness-Dependent Multiferroic Phase Transitions in the 2D Antiferromagnetic Semiconductor NiI2Dmitry Lebedev, Jonathan Tyler Gish, Ethan Skyler Garvey, Teodor Kosev Stanev, Junhwan Choi, Leonidas Georgopoulos, Thomas Wei Song, Hong Youl Park, Kenji Watanabe, Takashi Taniguchi, Nathaniel Patrick Stern, Vinod Kumar Sangwan, and Mark Christopher Hersam*2D magnetic materials hold promise for quantum and spintronic applications. 2D antiferromagnetic materials are of particular interest due to their relative insensitivity to external magnetic fields and higher switching speeds compared to 2D ferromagnets. However, their lack of macroscopic magnetization impedes detection and control of antiferromagnetic order, thus motivating magneto-electrical measurements for these purposes. Additionally, many 2D magnetic materials are ambient-reactive and electrically insulating or highly resistive below their magnetic ordering temperatures, which imposes severe constraints on electronic device fabrication and characterization. Herein, these issues are overcome via a fabrication protocol that achieves electrically conductive devices from the ambient-reactive 2D antiferromagnetic semiconductor NiI2. The resulting gate-tunable transistors show band-like electronic transport below the antiferromagnetic and multiferroic transition temperatures of NiI2, revealing a Hall mobility of 15 cm2 V−1 s−1 at 1.7 K. These devices also allow direct electrical probing of the thickness-dependent multiferroic phase transition temperature of NiI2 from 59 K (bulk) to 28 K (monolayer).DOI: 10.1002/adfm.202212568D. Lebedev, J. T. Gish, J. Choi, L. Georgopoulos, T. W. Song,  H. Y. Park, V. K. Sangwan, M. C. HersamDepartment of Materials Science and EngineeringNorthwestern UniversityEvanston, IL 60208, USAE-mail: m-hersam@northwestern.eduE. S. Garvey, T. K. Stanev, N. P. SternDepartment of Physics and AstronomyNorthwestern UniversityEvanston, IL 60208, USA1. Introduction2D van der Waals magnetic and ferroelec-tric (FE) materials have attracted signifi-cant attention for advanced applications including low-power logic and memory switches.[1,2] Moreover, 2D FE materials have shown layer-dependent ferroelec-tricity[3,4] and FE switching of a 2D metal,[5] while 2D magnetic materials have demon-strated thickness-dependent and stacking-dependent magnetic order,[6,7] electrically controlled magnetism,[8,9] and giant tun-neling magnetoresistance.[10] However, the majority of 2D magnets order antifer-romagnetically.[2] While antiferromagnetic (AFM) materials are more robust against external parasitic magnetic fields and enable higher operating frequencies com-pared to ferromagnets, their lack of macro-scopic magnetization limits detection and control of AFM order.[11] For example, RESEARCH ARTICLE © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH. This is an open access article under the terms of the  Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/adfm.202212568.K. WatanabeResearch Center for Functional MaterialsNational Institute for Materials Science1-1 Namiki, Tsukuba 305-0044, JapanT. TaniguchiInternational Center for Materials NanoarchitectonicsNational Institute for Materials Science1-1 Namiki, Tsukuba 305-0044, JapanM. C. HersamDepartment of ChemistryNorthwestern UniversityEvanston, IL 60208, USAM. C. HersamDepartment of Electrical and Computer EngineeringNorthwestern UniversityEvanston, IL 60208, USAAdv. Funct. Mater. 2023, 33, 2212568 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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://crossmark.crossref.org/dialog/?doi=10.1002%2Fadfm.202212568&domain=pdf&date_stamp=2023-01-12www.afm-journal.dewww.advancedsciencenews.com2212568 (2 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHcommon techniques for studying AFM materials, such as X-ray magnetic dichroism and neutron scattering, are exceptionally challenging to apply to microcrystals and 2D flakes.[2]Alternatively, charge transport characterization is prefer-able for detecting and manipulating AFM order since it can be straightforwardly applied on nanoscale samples and ena-bles AFM device integration with peripheral electrical circuits. In addition, spin-dependent charge transport measurements allow additional options for coupling magnetic and electronic phenomena, such as magnetoresistance and spin-orbit torque, providing further motivation for electrically probing 2D mag-netic materials.[11] However, since most 2D magnetic mate-rials are chemically unstable in ambient conditions, electronic device fabrication is cumbersome, which constrains accessible architectures. Furthermore, most 2D magnetic materials show undetectably low charge transport below their magnetic ordering temperatures, which is often attributed to carrier local-ization,[12,13] but may also result from poor charge injection due to chemical reactions with electrode materials. Consequently, previous attempts at 2D AFM electrical transport measure-ments have been limited to vertical device geometries, where the 2D AFM layers act solely as tunneling barriers.[8–10,14]NiI2 is an emerging 2D AFM material, whose structure con-sists of layers of edge-sharing [NiI6] octahedra that are sepa-rated by van der Waals gaps (Figure  1a,b). Chemically, NiI2 is highly hygroscopic, resulting in rapid degradation in ambient conditions.[15] In the bulk, NiI2 is a Mott-insulator,[16] which exhibits both long-range magnetic ordering and spontaneous Adv. Funct. Mater. 2023, 33, 2212568Figure 1. a) Top and b) side view on the crystal structure of NiI2 (CdCl2-type). b) Schematic of the helical ground state magnetic order of NiI2 below TN2: spins are perpendicular to Q ≈ (0.138,0,1.457). c) Ultrahigh vacuum scanning tunneling microscopy image of NiI2, taken at 200 mV sample bias and 0.9 nA tunneling current, reveals a hexagonal lattice with a periodicity of 0.40 nm as expected for NiI2. d) Optical and atomic force microscopy images with e) extracted height profiles corresponding to monolayer and bilayer regions. f) Raman spectroscopy of bulk NiI2 at room temperature. g) Photoluminescence and absorption spectra of bulk NiI2 at room temperature. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (3 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHelectrical polarization, known as multiferroicity.[17] Bulk NiI2 undergoes two antiferromagnetic transitions at TN1 = 76 K and TN2 = 59 K, with helical AFM ground state magnetic ordering along the propagation vector Q (Figure 1b).[18] While NiI2 has a centrosymmetric CdCl2-type crystal structure at room tempera-ture, the emerging magnetic order at the TN2 transition breaks inversion symmetry and drives in-plane ferroelectricity due to strong Dzyaloshinskii–Moriya interactions.[19] Recently, this multiferroicity was detected down to the monolayer limit for NiI2 via optical measurements.[19] Since the optical absorption spectrum of NiI2 possesses an excitonic peak at 1.38 eV[20] and the width of the conduction band is estimated at ≈1 eV,[21] NiI2 is expected to show gate-tunable semiconducting electronic transport over a broad temperature range. However, likely due to complications arising from its high chemical reactivity, var-iable-temperature charge transport measurements below the multiferroic transition temperature of atomically thin NiI2 have not yet been reported.Here, we combine a dry flake transfer method with atomic layer deposition (ALD) to fabricate multi-terminal electrical devices from exfoliated NiI2. This fabrication protocol results in ambient-stable electrical contacts to NiI2 that enable charge transport measurements at cryogenic temperatures. In this manner, we confirm that NiI2 is a gate-tunable semiconductor down to monolayer thickness with band-like transport and a Hall mobility of 15 cm2 V−1 s−1 at a temperature of 1.7 K. Var-iable-temperature magnetotransport measurements further reveal an anisotropic magnetoresistance of NiI2, which changes sign at TN2, likely associated with strong spin-orbit coupling. These magnetotransport measurements allow the onset of mul-tiferroicity to be electrically probed, thereby enabling the thick-ness dependence of TN2 to be measured from 59 K in bulk NiI2 to 28 K in monolayer NiI2, which is consistent with significant interlayer exchange interactions in NiI2.[6,19]2. Results and DiscussionNiI2 crystals were prepared using the chemical vapor trans-port (CVT) method from elemental sources. Ultrahigh vacuum (UHV) scanning tunneling microscopy (STM) at room tem-perature revealed a hexagonal lattice with a periodicity of 0.40 nm (Figure  1c), consistent with the NiI2 crystal structure (a  = 0.392  nm).[18] The NiI2 crystal was micromechanically exfoliated in an inert-atmosphere N2 glovebox, where few-layer flakes were identified optically. The optical contrast of the resulting NiI2 flakes was correlated with thickness using inert-atmosphere atomic force microscopy (Figure 1d,e and Figure S1,  Supporting Information). NiI2 flakes were encapsulated with 55 nm of alumina grown by ALD (Figure S1, Supporting Infor-mation),[22] which enabled ambient Raman, photoluminescence (PL), and optical absorption measurements (Figure  1f,g) that are consistent with previous literature reports.[15,19,20] To con-firm the ferroelectricity of NiI2, second harmonic generation and linear dichroism measurements were performed, which confirmed the existence of in-plane polarization in NiI2 below TN2 (Figure S2, Supporting Information).To develop the device fabrication protocol, we first studied the stability of exfoliated NiI2 flakes in air and in different solvents commonly used for flake transfer and lithographic pro-cesses (e.g., acetone and chloroform). Unprotected NiI2 flakes were found to degrade in air and chloroform over the course of hours (Figure 2). Monolayers of NiI2 even degrade in an inert atmosphere (N2 glovebox) within a few days when in contact with tape residue (Figure S3, Supporting Information), further emphasizing the need for effective passivation. In contrast, we find that 1,2-dichloroethane (1,2-DCE) does not degrade NiI2 (Figure S4, Supporting Information). Hence, 1,2-DCE was used as the solvent for subsequent device fabrication.NiI2 devices were assembled via a dry pick-up and transfer method using few-layer hexagonal boron nitride (hBN) to sand-wich NiI2 flakes, which provided a well-defined dielectric envi-ronment for charge transport measurements (see Experimental Section for details). In addition, few-layer graphene was used as electrical contacts since its chemical inertness minimizes the chance of unwanted interfacial reactions at NiI2 contacts. Fabricating NiI2 field-effect transistors (FETs) down to mono-layer thickness presents another challenge as large few-layer flakes are often connected to thicker (>20 nm) flakes (Figure 3a and Figures S5 and S6, Supporting Information). Therefore, a PC/PDMS (polydimethylsiloxane) polymer stamp was used to sequentially remove surrounding thicker NiI2 flakes, while keeping the few-layer NiI2 flakes intact (Figure  3a–c). If left unremoved, large neighboring NiI2 flakes prevent homo-geneous contact between the top hBN and few-layer NiI2, resulting in tearing, folding, and/or formation of bubbles trapped at the interface during the stamping process, all of which negatively affect electrical transport. Overall, this pro-cedure enables the fabrication of high-quality NiI2 FETs with semiconducting channels of different thicknesses (Figure S6, Supporting Information).To achieve multi-terminal devices, the hBN/graphene/NiI2/hBN heterostructures were patterned using electron-beam lithography followed by reactive ion etching (RIE) (see Experi-mental Methods for details and Figure 3d,e). It should be noted that RIE exposes the edges of the semiconducting material in the heterostructure stack, which has limited this procedure to ambient-stable 2D semiconductors in the past.[23] Indeed, NiI2 flakes with exposed edges degrade in both ambient conditions and acetone (Figure 2c,d), similar to unprotected flakes. Charge transport in transition metal halides is highly sensitive to chem-ical degradation, with FET electrical characteristics decaying irreversibly even for flakes that appear optically intact.[22] To avoid this degradation, the resist was stripped using 1,2-DCE, and then the entire hBN/graphene/NiI2/hBN stack was encap-sulated with 55 nm of ALD alumina. In this manner, the final device structures were stable in ambient conditions, which allowed for handling in air for wire bonding, as well as elec-trical and optical measurements (Figure 3f and Figure S7, Sup-porting Information).The resulting devices reveal that NiI2 is a gate-tunable, ambipolar semiconductor with electron (hole) field-effect mobilities of ≈1 cm2 V−1 s−1 (0.01 cm2 V−1 s−1) and on/off ratios of 105 (103) at room temperature (Figure  4a). Temperature-dependent charge transport measurements with a back-gate voltage VG = 60 V revealed a clear peak in resistance at TN2 = 59 K for bulk NiI2, which corresponds to the helical magnetic and structural transition of NiI2, as well as the onset of the Adv. Funct. Mater. 2023, 33, 2212568 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (4 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHmultiferroic order (Figure 4b).[18,24] Plotting ln(R) versus inverse T shows a kink at 76 K (inset of Figure 4b), which corresponds to TN1 of bulk NiI2.[18,24] Moreover, photocurrent measurements reveal the bulk photovoltaic effect, which is induced by the spontaneous electrical polarization of ferroelectric materials, further confirming the multiferroic nature of NiI2 below TN2 (Figure S8, Supporting Information). The use of graphene con-tacts allowed electrical measurements down to 1.7 K, but the current–voltage characteristics became nonlinear at low tem-peratures (Figure S9, Supporting Information), which suggests Adv. Funct. Mater. 2023, 33, 2212568Figure 2. a,b) Degradation of unprotected NiI2 flakes in ambient atmosphere and chloroform, respectively. c,d) Edge-initiated degradation of an etched hBN/NiI2/hBN stack in acetone and ambient atmosphere, respectively. e) Schematic of the etched hBN/NiI2/hBN stack with the edge exposed to ambient species and/or solvents. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (5 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHa Schottky barrier is present that could potentially be resolved by subsequent contact engineering (e.g., electrostatic doping of contact regions). In contrast, direct metal contacts on NiI2 FETs resulted in highly resistive devices with 100-fold lower mobility at low temperatures, thus illustrating the advantages of graphene contacts for NiI2 (Figure S10, Supporting Informa-tion). To demonstrate low-voltage operation, we also fabricated local-gate NiI2 FETs, which show on/off ratios of 105 and oper-ating voltages below 3 V both at room temperature and at 5 K (Figure S11, Supporting Information).Electrical interrogation of the NiI2 multiferroic transition facilitates thickness-dependent measurements of TN2 down to the monolayer limit (Figure 4c,d). Even monolayer NiI2 showed gate-tunable conductivity at low temperatures, although it was more resistive than thicker layers (Figure S9, Supporting Information). Electrical measurements revealed multiferroic transition temperatures of ≈28, 40, and 48 K for monolayer, bilayer, and trilayer NiI2, respectively (Figure 4d and Figure S12, Supporting Information). These values are marginally higher than reports based on optical measurements,[15,19,25] which can likely be explained by differences in growth conditions and/or differing levels of ambient exposure. The strong thickness dependence of the multiferroic transition temperature indicates the presence of significant interlayer exchange interactions in NiI2.[6,19]To gain further insight into magnetotransport properties, Hall bars of several-layer NiI2 were tested at temperatures above and below TN2. Magnetoresistance is calculated based on four-probe (Rxx) measurements as MR = ∆R/R × 100% = (RH − R0)/R0 × 100%. With an out-of-plane magnetic field and VG = 60 V, Adv. Funct. Mater. 2023, 33, 2212568Figure 3. a,b) Schematic demonstrating removal of large and thick flakes preceding c) fabrication of few-layer NiI2 electrical devices. Unless removed, thick NiI2 flakes compromise the quality of the heterostructure due to large height differences that cause poor flake-to-flake contact and thus gas bubble incorporation. d–f) Schematic of the NiI2 (>10 layers) Hall bar fabrication, which consists of flake pick-up and transfer to obtain the hBN/graphene/NiI2/hBN stack, patterning and etching, and encapsulation via alumina atomic layer deposition. Although the graphene flakes are overlapping after the heterostructure assembly, subsequent etching defines the required Hall bar geometry. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (6 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHa small (<1%) negative magnetoresistance was measured above TN2, which can be assigned to field-induced suppression of electron scattering by magnetic fluctuations (Figure  4e).[12,26,27] At TN2, the magnetoresistance changed sign to positive (Figure  4e and Figures S12 and S13, Supporting Information), and its magnitude increased with decreasing temperature, reaching 7% at 8 T and 1.7 K. The same trend was observed for the trilayer NiI2, where the sign change occurred at lower temperature, in agreement with the lower TN2 (46–50  K for trilayer NiI2, Figure S14, Supporting Information). A positive magnetoresistance has recently been observed in layered AFM materials, such as CrSBr[12,26,27] and non-collinear Co1/3NbS2.[28] For both in-plane and out-of-plane orientations of the magnetic field, no kinks were observed in the magnetoresistance versus Adv. Funct. Mater. 2023, 33, 2212568Figure 4. a) Transfer characteristics of a bulk NiI2 FET at different temperatures (VD = 4 V) with the inset showing the same data on a log-linear scale. b) Resistance versus temperature of bulk NiI2 FET (VG = 60 V, VD = 4 V) showing a peak at TN2 ≈ 59 K and a kink at TN1 ≈ 76 K (inset shows ln(R) versus 1/T). c) Transfer characteristics of a bilayer NiI2 FET at different temperatures (VD = 2 V) with the inset showing the same data on a log-linear scale. d) Resistance of various few-layer NiI2 FETs versus temperature, showing a decrease of TN2 with decreasing number of layers. e) Four-probe magnetoresistance and f) electron mobility and carrier concentration of bulk NiI2 (>10 layers, shown in Figure 3f) at different temperatures, VG = 60 V. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (7 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHAdv. Funct. Mater. 2023, 33, 2212568magnetic field curves, which indicates that higher magnetic fields are needed to achieve spin-flip/spin-flop transitions in NiI2. The magnetoresistance remained positive below TN2 when the magnetic field was applied in-plane, and its magnitude was higher compared to the out-of-plane direction, reaching 12% at 8 T and 1.7 K (Figure S13, Supporting Information), implying significant magnetoresistive anisotropy. Anisotropic magne-toresistance has also been observed in the layered AFM mate-rial CrSBr, which was associated with spin-flip transitions and quasi-1D transport.[12,27] However, the anisotropic magnetore-sistance of NiI2 is more likely to be associated with large spin-orbit interactions.[19,29]By fitting the transverse Rxy data to the standard Hall effect model, the number of carriers at VG  = 60 V decreased from N ≈ 2 × 1019 cm−3 at 83 K to N ≈ 3 × 1018 cm−3 below TN2 and then did not vary significantly with further decreasing tem-perature (Figure  4f and Figure S13, Supporting Information). On the other hand, the Hall mobility increased at these low temperatures below TN2, ultimately reaching 15 cm2 V−1 s−1 at 1.7 K (Figure  4f and Figure S13, Supporting Information). By comparing the Hall mobilities with the field-effect mobilities extracted from the transistor data for the same device, we find that the Hall mobilities are approximately two times smaller (Figure S13, Supporting Information), which is most likely associated with the presence of localized states below the band edge.[30]3. ConclusionIn summary, we have designed and implemented a protocol for fabricating electrical devices for ambient-reactive 2D NiI2. These devices enable probing of the gate-tunable semicon-ducting properties of NiI2 down to cryogenic temperatures, which allows extraction of the thickness dependence of the NiI2 multiferroic phase transition down to the monolayer limit. This fabrication protocol also yields ambient-stable multi-terminal devices with a Hall mobility of 15 cm2 V−1 s−1 at 1.7 K. The results of this work will facilitate further exploration of the magneto-electric effects and gate-tunable properties of NiI2, including the incorporation of NiI2 into heterostructures with other van der Waals materials. Since the methodology demon-strated here can be broadly applied to other ambient-reactive 2D materials, it has the potential to enable electrical interroga-tion of diverse magnetic, quantum, and spintronic phenomena at the atomically thin limit.4. Experimental SectionNiI2 Crystal Growth and Exfoliation on SiO2/Si: NiI2 crystals were grown by the chemical vapor transport method.[18] A stoichiometric amount of nickel (99.99%, 200-mesh) and iodine (99.99%) powders were sealed in a quartz ampoule (15 mm × 250 mm, <10−3 Torr) that was placed in a horizontal tube furnace. The reaction mixture was heated to 750 °C for 5 days, while the cold end of the ampoule was held at room temperature. The NiI2 crystals were removed from the ampoule and stored in an inert nitrogen atmosphere glovebox (O2 and H2O < 0.1  ppm). Using Scotch tape, few-layer flakes were micromechanically exfoliated from the bulk crystal onto 300  nm SiO2/Si substrates inside the inert nitrogen atmosphere glovebox. The thickness of few-layer NiI2 on SiO2/Si was identified based on optical contrast (red color channel, Figure S1, Supporting Information) and verified using inert-atmosphere atomic force microscopy (Figure 1).Atomic Force Microscopy: An Asylum Cypher atomic force microscope with a Si cantilever (resonant frequency ≈320–340  kHz) was used for all atomic force microscopy analysis. The image resolution was 512 ×  512 pixels at a scanning rate of 1.0–1.5 Hz. The images were all taken in a sealed UHP argon environmental cell.Ultrahigh Vacuum Scanning Tunneling Microscopy: Room-temperature UHV-STM measurements (10−10 Torr) were performed on a homebuilt STM with a Lyding-type scanner[31] interfaced with a SPECS Nanonis controller. Mechanically cut Pt-Ir tips (Bruker) were used for the measurements. Single-crystal NiI2 was used for the UHV-STM measurements. To avoid ambient exposure, the NiI2 crystal was mounted on the STM sample holder in a N2 glovebox and transferred to the load-lock of the UHV chamber using a custom-made transfer flask. NiI2 was cleaved in UHV to expose a fresh crystal surface prior to STM characterization.Raman, Photoluminescence, and Absorption Spectroscopies: Raman and photoluminescence spectroscopies were conducted using a Horiba XPloRA confocal setup (532  nm laser with a spot size of ≈1 um with a 100× objective). For absorption spectroscopy (Cary-5000 UV/Vis Spectrophotometer with an integration sphere), NiI2 was exfoliated on sapphire. All measurements were performed after alumina ALD to encapsulate the flakes and prevent ambient degradation.Device Fabrication: Device fabrication was performed by combining polymer-assisted flake pick-up and clean-room lithography methods. Flake pick-up and transfer were performed using a 2D crystal manipulation system (Graphene Industries) located in the nitrogen glovebox. To prepare the polymer stamp, a dome of polydimethylsiloxane (PDMS) was first prepared on a glass slide by curing a droplet of the PDMS curing agent and base mixture (1:10). Next, a film of poly(bisphenol A carbonate) (PC, Sigma–Aldrich) was prepared by doctor-blading a solution of PC in chloroform, which was then stretched over the PDMS dome using Scotch tape. The resulting PC/PDMS stamp was heated on a hot plate at 120 °C for 10 min. Hexagonal boron nitride (hBN) and graphite were exfoliated onto SiO2/Si using Scotch tape, and flakes of desired thickness and size were identified prior to device assembly. Device assembly was started by picking up the top hBN flake, followed by sequential pick-up of other flakes, and landing the heterostructure on prepatterned metal contacts by melting the PC at a temperature exceeding 150  °C. The prepatterned metal contacts on 300 nm thick SiO2/Si substrates were fabricated using a Maskless Aligner (Heidelberg MLA150) with positive resist followed by reactive ion etching (Samco RIE-10NR) and metal evaporation.Hall bar devices were fabricated by first picking up two or three few-layer graphene flakes with the top hBN flake (which defines the area for the NiI2 flake), followed by pick up of the NiI2 flake and bottom hBN. The stack was then patterned (electron-beam lithography, FEI Quanta 600F) and etched via reactive-ion etching. A thick poly(methyl methacrylate) (PMMA, A8 950) resist (≈1 µm) prepared by spin coating at a rate of 4000  rpm was used for electron-beam lithography. Since reactive ion etching exposes the edges of the NiI2 flake, the devices were quickly moved into a metallic vessel and evacuated (<10 s of ambient exposure), followed by transfer into the inert-atmosphere glovebox. Resist stripping was performed in the glovebox using air-free and water-free 1,2-dichloroethane (1,2-DCE), followed by masking the metal contacts with Kapton tape and encapsulating the device structure via atomic layer deposition of alumina (Anric AT 400).Sequential Flake Removal: A polymer stamp of PC/PDMS was brought into contact with the substrate/NiI2 flakes at 100 °C. The substrate was heated to 130 °C with careful control of the PC contact front to prevent any contact with the few-layer sample while maximizing contact with the thicker flakes. After leaving the PC stamp in contact with the flakes of interest for a few minutes following turn off of the heater, the PC film was slowly detached from the surface, thus tearing off the undesired bulk NiI2 flakes. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (8 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHAdv. Funct. Mater. 2023, 33, 2212568Reactive Ion Etching: Reactive ion etching was used to fabricate prepatterned metal contacts and to etch the devices to the Hall bar shape. For the prepatterned metal contacts, the etching of SiO2 prior to metal evaporation was performed using CF4 gas (20 sccm, 3.3 Pa, 100 W, etch rate 0.5 nm s−1) using a Samco RIE-10NR instrument. Etching of the van der Waals heterostructures was performed in two steps: top hBN was etched with CF4 (20 sccm, 3.3 Pa, 100 W, etch rate 2.6 nm s−1) and the rest of the stack (NiI2, graphene) was etched with Ar plasma (20 sccm, 3.3 Pa, 60 W, 2–4 min).Atomic Layer Deposition: Alumina was grown in a commercial Anric AT 400 ALD reactor using alternating pulses of trimethyl aluminum (TMA, Sigma–Aldrich) and water. The ALD reactor is connected to a nitrogen glovebox (O2 and H2O < 1  ppm) to load samples without ambient exposure. The ALD growth was performed at 100 °C for 500 cycles, which corresponds to ≈55 nm of alumina.Electrical Transport and Bulk Photovoltaic Effect Measurements: Device measurements were performed in a Lakeshore CRX 4K and Lakeshore CRX-VF probe stations, variable-temperature cryostat (Attocube AttoDRY2100), and Dynacool PPMS (Quantum Design). The devices were wire-bonded to the chip carriers using a homebuilt In-Au bonder. Keithley Source Meter 2400 units and a Keithley Nanovoltmeter 2182 unit were used to probe charge transport.Bulk photovoltaic effect measurements were performed using a Lakeshore CRX-VF probe station (with a custom lid allowing for use of magnifying optics) and a SuperK Extreme EXR-20 laser (NKT Photonics) at 530  nm (LLTF-VIS monochromator) at a power of 100 µW. The laser was modulated with a mechanical chopper before being linearly polarized and focused on the channel of the device using a long working distance 50× objective. The signal was detected using an SR830 lock-in amplifier (Stanford Research).Second Harmonic Generation and Linear Dichroism Measurements: All low-temperature optical measurements were carried out with samples mounted in a closed cycle variable temperature cryostat. Second harmonic generation measurements were performed using a Ti:sapphire laser with a pulse repetition rate of 76 MHz and pulse width of 150 fs. The 1.49  eV output from the laser was focused onto the sample with a long working distance 50× objective using a homebuilt microscope setup. The second harmonic generation signal was collected in reflection geometry and directed to a spectrometer with an Andor CCD camera for detection.Linear dichroism measurements were performed using a CW 2.33  eV laser. The laser was modulated with a mechanical chopper before being linearly polarized and sent through a photo-elastic modulator (PEM). The PEM was set to have a maximum retardance of 0.5λ with a fast axis at 45 degrees with respect to the input polarization. A half waveplate was used to rotate the modulated polarization with respect to the crystal axes before being focused onto the sample with a 100× objective. The linear dichroism signal was collected in reflection geometry and directed to a Thorlabs avalanche photodiode for lock-in detection.Supporting InformationSupporting Information is available from the Wiley Online Library or from the author.AcknowledgementsThis research was primarily supported by the National Science Foundation Division of Materials Research (NSF DMR-2004420). In addition, D.L. acknowledges support from the Swiss National Science Foundation for an Early PostDoc Mobility Fellowship (P2EZP2_181614) and the Materials Research Science and Engineering Center of Northwestern University (NSF DMR-1720139) for charge transport measurements. H.Y.P. acknowledges support from the Department of Energy (DOE DE-SC0019356) for UHV-STM characterization, and E.S.G. acknowledges support from the National Science Foundation Division of Materials Research (NSF DMR-1905986) for variable-temperature cryostat measurements. K.W. and T.T. acknowledge support from JSPS KAKENHI (Grant Numbers 19H05790, 20H00354, and 21H05233) for hBN synthesis. This work made use of the Northwestern University NUANCE Center and the Northwestern University Micro/Nano Fabrication Facility (NUFAB), which have received support from the SHyNE Resource (NSF ECCS-1542205), the International Institute for Nanotechnology, and the Northwestern University MRSEC program (NSF DMR-1720139). This work also utilized the Northwestern University Magnet, Low Temperature, and Optical Facility, which is supported by the Northwestern University MRSEC program (NSF DMR-1720139). The Lakeshore CRX-VF probe station, SuperK Extreme EXR-20 laser (NKT Photonics), and 2D crystal manipulation system (Graphene Industries) used in this work were supported by an Office of Naval Research DURIP grant (ONR N00014-19-1-2297).Conflict of InterestThe authors declare no conflict of interest.Author ContributionsD.L. and J.T.G. contributed equally to this work. D.L., J.T.G., and M.C.H. devised the principal objectives of the project. D.L. grew the NiI2 crystals by chemical vapor transport method. J.T.G., D.L., J.C., L.G., and T.W.S. exfoliated the NiI2 flakes for optical measurements and fabricated the electrical devices. L.G. performed atomic force microscopy measurements. H.Y.P. performed scanning tunneling microscopy measurements. D.L. and J.T.G. performed the magnetotransport measurements with the help of E.S.G. and T.K.S. under the supervision of V.K.S. E.S.G. performed low-temperature optical measurements under the supervision of N.P.S. K.W. and T.T. provided the hBN crystals. M.C.H. supervised the project. D.L., J.T.G., and M.C.H. wrote the manuscript with input from all authors.Data Availability StatementThe data that support the findings of this study are available from the corresponding author upon reasonable request.Keywords2D magnets, 2D multiferroicity, 2D semiconductors, band-like transport, helical antiferromagnetsReceived: October 29, 2022Published online: January 12, 2023[1] H. Kurebayashi, J. H. Garcia, S. Khan, J. Sinova, S. Roche, Nat Rev Phys 2022, 4, 150.[2] H. Li, S. Ruan, Y. J. Zeng, Adv. Mater. 2019, 31, 1900065.[3] C. J.  Cui, W. J.  Hu, X. G.  Yan, C.  Addiego, W. P.  Gao, Y.  Wang, Z. Wang, L. Z. Li, Y. C. Cheng, P. Li, X. X. Zhang, H. N. Alshareef, T. Wu, W. G. Zhu, X. Q. Pan, L. J. Li, Nano Lett. 2018, 18, 1253.[4] K.  Chang, J.  Liu, H.  Lin, N.  Wang, K.  Zhao, A.  Zhang, F.  Jin, Y. Zhong, X. Hu, W. Duan, Q. Zhang, L.  Fu, Q. K. Xue, X. Chen, S. H. Ji, Science 2016, 353, 274. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 Licensewww.afm-journal.dewww.advancedsciencenews.com2212568 (9 of 9) © 2023 The Authors. Advanced Functional Materials published by Wiley-VCH GmbHAdv. Funct. Mater. 2023, 33, 2212568[5] Z. Y.  Fei, W. J.  Zhao, T. A.  Palomaki, B. S.  Sun, M. K.  Miller, Z. Y. Zhao, J. Q. Yan, X. D. Xu, D. H. Cobden, Nature 2018, 560, 336.[6] 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, W.  Yao, D.  Xiao, P.  Jarillo-Herrero, X.  Xu, Nature 2017, 546,  270.[7] T. P.  Kaloni, K.  Chang, B. J. Miller, Q. K.  Xue, X.  Chen, S. H.  Ji, S. S. P. Parkin, S. Barraza-Lopez, Phys. Rev. B 2019, 99, 134108.[8] B. Huang, G. Clark, D. R. Klein, D. MacNeill, E. Navarro-Moratalla, K. L.  Seyler, N.  Wilson, M. A.  McGuire, D. H.  Cobden, D.  Xiao, W. Yao, P. Jarillo-Herrero, X. Xu, Nat. Nanotechnol. 2018, 13, 544.[9] S. Jiang, J. Shan, K. F. Mak, Nat. Mater. 2018, 17, 406.[10] T.  Song, X.  Cai, M. W.  Tu, X.  Zhang, B.  Huang, N. P.  Wilson, K. L.  Seyler, L.  Zhu, T.  Taniguchi, K.  Watanabe, M. A.  McGuire, D. H. Cobden, D. Xiao, W. Yao, X. Xu, Science 2018, 360, 1214.[11] T.  Jungwirth, X. Marti, P. Wadley, J. Wunderlich, Nat. Nanotechnol. 2016, 11, 231.[12] F.  Wu, I.  Gutierrez-Lezama, S. A.  Lopez-Paz, M.  Gibertini, K. Watanabe, T. Taniguchi, F. O. von Rohr, N. Ubrig, A. F. Morpurgo, Adv. Mater. 2022, 34, 2109759.[13] N.  Mounet, M.  Gibertini, P.  Schwaller, D.  Campi, A.  Merkys, A.  Marrazzo, T.  Sohier, I. E.  Castelli, A.  Cepellotti, G.  Pizzi, N. Marzari, Nat. Nanotechnol. 2018, 13, 246.[14] J. B. Wu, H. Y. Chen, N. Yang, J. Cao, X. D. Yan, F. X. Liu, Q. B. Sun, X. Ling, J. Guo, H. Wang, Nat. Electron. 2020, 3, 466.[15] H.  Liu, X. Wang, J. Wu, Y. Chen, J. Wan, R. Wen, J.  Yang, Y.  Liu, Z. Song, L. Xie, ACS Nano 2020, 14, 10544.[16] M. P.  Pasternak, R. D.  Taylor, A.  Chen, C.  Meade, L. M.  Falicov, A. Giesekus, R. Jeanloz, P. Y. Yu, Phys. Rev. Lett. 1990, 65, 790.[17] T. Kurumaji, S. Seki, S. Ishiwata, H. Murakawa, Y. Kaneko, Y. Tokura, Phys. Rev. B 2013, 87, 014429.[18] S. R. Kuindersma, J. P. Sanchez, C. Haas, Physica B Condens Matter 1981, 111, 231.[19] Q.  Song, C. A.  Occhialini, E.  Ergecen, B.  Ilyas, D.  Amoroso, P. Barone, J. Kapeghian, K. Watanabe, T.  Taniguchi, A. S. Botana, S. Picozzi, N. Gedik, R. Comin, Nature 2022, 602, 601.[20] S. Son, Y. Lee, J. H. Kim, B. H. Kim, C. Kim, W. Na, H. Ju, S. Park, A.  Nag, K. J.  Zhou, Y. W.  Son, H.  Kim, W. S.  Noh, J. H.  Park, J. S.  Lee, H. Cheong, J. H. Kim, J. G. Park, Adv. Mater. 2022, 34, 2109144.[21] Y. An, H. Wang, J. Liao, Y. Gao, J. Chen, Y. Wu, Y. Li, G. Xu, C. Ma, Phys. E Low Dimens. Syst. Nanostruct. 2022, 115262.[22] J. T.  Gish, D.  Lebedev, T. K.  Stanev, S.  Jiang, L.  Georgopoulos, T. W. Song, G. Lim, E. S. Garvey, L. Valdman, O. Balogun, Z. Sofer, V. K. Sangwan, N. P. Stern, M. C. Hersam, ACS Nano 2021, 15, 10659.[23] D. A. Bandurin, A. V. Tyurnina, G. L. Yu, A. Mishchenko, V. Zólyomi, S. V.  Morozov, R. K.  Kumar, R. V.  Gorbachev, Z. R.  Kudrynskyi, S. Pezzini, Z. D. Kovalyuk, U. Zeitler, K. S. Novoselov, A. Patanè, L. Eaves, I. V. Grigorieva, V. I. Fal’Ko, A. K. Geim, Y. Cao, Nat. Nano-technol. 2017, 12, 223.[24] T. Kurumaji, S. Seki, S. Ishiwata, H. Murakawa, Y. Kaneko, Y. Tokura, Phys. Rev. B 2013, 87, 014429.[25] H.  Ju, Y.  Lee, K. T.  Kim, I. H.  Choi, C. J.  Roh, S.  Son, P.  Park, J. H. Kim, T. S. Jung, J. H. Kim, K. H. Kim, J. G. Park, J. S. Lee, Nano Lett. 2021, 21, 5126.[26] P. Majumdar, P. B. Littlewood, Nature 1998, 395, 479.[27] E. J.  Telford, A. H.  Dismukes, K.  Lee, M.  Cheng, A.  Wieteska, A. K.  Bartholomew, Y. S.  Chen, X.  Xu, A. N.  Pasupathy, X.  Zhu, C. R. Dean, X. Roy, Adv. Mater. 2020, 32, 2003240.[28] G.  Tenasini, E.  Martino, N.  Ubrig, N. J.  Ghimire, H.  Berger, O. Zaharko, F. Wu, J. F. Mitchell, I. Martin, L. Forró, A. F. Morpurgo, Phys. Rev. Res. 2020, 2, 023051.[29] A. O. Fumega, J. L. Lado, 2D Mater. 2022, 9, 025010.[30] B. W.  Baugher, H. O.  Churchill, Y.  Yang, P.  Jarillo-Herrero, Nano Lett. 2013, 13, 4212.[31] E. T.  Foley, N. L.  Yoder, N. P. Guisinger, M. C. Hersam, Rev. Sci. Instrum. 2004, 75, 5280. 16163028, 2023, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/adfm.202212568 by Cochrane Japan, Wiley Online Library on [18/03/2023]. 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 License