WorksPhysics and Astronomy
Observation of time-reversal symmetric Hall effect in graphene-WSe2 heterostructures at room temperature
Abstract
In this letter, we provide experimental evidence of the time-reversal symmetric Hall effect in a mesoscopic system, namely high-mobility graphene/WSe heterostructures. This linear, dissipative Hall effect, whose sign depends on the sign of the charge carriers, persists up to room temperature. The magnitude and the sign of the Hall signal can be tuned using an external perpendicular electric field. Our joint experimental and theoretical study establishes that the strain induced by lattice mismatch, or angle inhomogeneity, produces anisotropic bands in graphene while simultaneously breaking the inversion symmetry. The band anisotropy and reduced spatial symmetry lead to the appearance of a time-reversal symmetric Hall effect. Our study establishes graphene-transition metal dichalcogenide-based heterostructures as an excellent platform for studying the effects of broken symmetry on the physical properties of band-engineered two-dimensional systems.
† These authors contributed equally.
E-mail: [email protected],[email protected]
Introduction
Topological and band geometric effects in two-dimensional systems have attracted significant attention due to their fascinating physics and potential applications in spintronics and novel electronic devices[1, 2, 3, 4, 5]. Graphene-based heterostructures offer one such exciting platform for studying band geometric effects. The coupling to the charge, spin, and valley degrees of freedom in graphene lead to, among other things, a multitude of unconventional Hall effects; some prominent examples include the spin Hall [6, 7, 8, 9], the valley Hall [10, 11, 12, 13, 14, 15], the nonlinear anomalous Hall [16, 17, 18, 19, 20, 21, 22] and the layer contrasted Hall effect [3, 23]. All these phases require a finite Berry curvature and opposite anomalous velocity in the two valleys of graphene to exist [16, 17, 18]. They fall under the category of dissipationless anti-symmetric Hall effect, in which the current always flows perpendicular to the electric field, and the polarity of voltage changes if one interchanges the source and drain contacts. What is under-appreciated is that a time-reversal symmetric system with anisotropic band velocity can also host non-trivial and symmetric Hall effects irrespective of the presence or absence of a finite Berry curvature. The focus of this Letter is the observation of a finite ‘symmetric Hall response’[24, 25, 26]. The symmetric Hall effect can appear without time-reversal symmetry breaking and is unrelated to the so-called quantum geometry of the wave function.
In this letter, we report the observation of a large time-reversal symmetric, linear Hall response originating from asymmetric momentum space band dispersion of the heterostructure.
We find that the flips sign as the Fermi level moves from the valence to the conduction band (mimicking the finite B-field classical Hall signal in graphene) and is observed up to room temperature. We explain our observations using a realistic two-band model for an anisotropic band dispersion with an orthogonal anisotropy orientation in the two bands. We show that such band anisotropy can appear in graphene-WSe heterostructures through realistic model calculations.
Single-layer WSe used as a substrate influences the graphene bands in two significant ways. The first of these is well studied: Graphene on WSe possesses spin-split bands owing to the Ising-like SOC, which gives rise to the spin Hall effect[27, 28, 29]. The second effect, equally vital for our purposes but ill-explored to date, is the appearance of a substantial lateral strain in the graphene layer. We show that the effect of this proximity-induced SOC, lattice mismatch, and small angle rotation can induce gap and asymmetric dispersion in graphene, leading to the appearance of the symmetric Hall signal near the Dirac point. We verify that the symmetric, charge-carrier sign-dependent Hall response is absent in graphene devices without the WSe layer. Note that previous studies on the SLG-WSe heterostructure (or graphene on transition metal dichalcogenides in general) focused primarily on the spin aspects of the transport[27, 28, 30, 31, 32] where a non-local signal is measured as a signature of the spin Hall effect and weak (anti-) localization measurements were used to quantify the spin-orbit coupling strength[29] [33, 34, 35, 36, 37, 38, 39]. Interestingly, these studies did not probe the finite Hall effect without a magnetic field. This makes our observation of the time-reversal symmetric non-trivial Hall effect in this system unique.
Results
Device characteristics
Heterostructures of single-layer graphene (SLG) and single-layer WSe, encapsulated by crystalline hexagonal boron nitrate (hBN), were fabricated using a dry transfer technique[40, 41]. One-dimensional electrical contacts were formed by electron beam lithography, followed by etching (using a mixture of and and deposition of 5 nm/60 nm contacts. A separate deposition was done for the top-gate electrode. See the Supporting Information (SI) for details. A schematic of the device structure is shown in Figure 1(a), and an optical image of the device is shown in Figure 1(b). The dual-gated architecture of the devices allows independent control of the charge-carrier density and the vertical displacement field ; and . Here () is the capacitance per unit area of the back-gate (top-gate), () is the back-gate (top-gate) bias. and are the residual charge carrier density and residual vertical displacement field induced by impurities in the device channel.
Figure 1. Device characteristics and band dispersion: (a) Schematic of the graphene/WSe layers encapsulated in hBN illustrating the sequence of crystal stacking. (b) Optical image of the device. (c) Map of the longitudinal conductance () with varying carrier density and perpendicular magnetic field at mK. The thicker dashed lines correspond to the signature plateaus of single-layer graphene. Thinner lines mark the broken-symmetry phases indicating complete lifting of the spin and valley degeneracies at low-. (d) SdH oscillations versus at V. (e) Fourier spectrum of the SdH oscillations: two peaks are distinctly visible, establishing the presence of two Fermi surfaces. (f) Schematic of the band dispersion of the valley of monolayer graphene (left panel) and graphene on WSe heterostructure (right panel). The WSe layer essentially lifts the spin degeneracy of the low-lying energy bands and opens up a gap at the Fermi energy. (g) Schematic of the band structure of strained graphene on WSe with the orientation of band anisotropy being orthogonal in the conduction and valence bands. This is highlighted by the arrows in the contour plots.
Electrical transport measurements were performed at 10 nA source-drain current using low-frequency lock-in detection techniques. All data were obtained at 20 mK unless specified otherwise. The measurements were performed on multiple devices; the results were similar. In the main manuscript, we present the data from a single device, SW1, unless specified. The data from three other devices (SW2, SG3, SG4) are presented in Fig. S5, Fig. S6, and Fig. S7 of the SI.
A map of the measured longitudinal conductance as a function of charge carrier density and perpendicular magnetic field is shown in Figure 1(c). The appearance of broken symmetry quantum Hall states at low -fields implies a complete lifting of the spin and valley degeneracies in SLG bands. The splitting of the spin-degenerate bands in SLG (shown schematically in Figure 1(f)) is also evident from the beating pattern seen in the Shubnikov de Haas oscillations [Figure 1(d)], and the double periodicity in the corresponding Fourier spectrum [Figure 1(e)]. Figure 1(g) is a schematic representation of the band structure of strained graphene on with the orientation of the band anisotropy in the valence and conduction band being orthogonal (see § 10 of SI for more details). The lifting of spin-and valley degeneracies in the band dispersion (along with the high field-effect mobility of the device) shows that the graphene and interface is atomically clean with significant interfacial coupling and minimal random potential fluctuations.
Room temperature symmetric Hall effect at T
In Figure 2(a), we present the data for the longitudinal resistance, (left-axis, red line), and transverse resistance, (right-axis, blue line) measured at T. We observe a finite Hall signal in a narrow range of charge carrier densities centered about the charge neutrality point. measured for other contact configurations give quantitatively similar results (see Fig. S5 in SI). Strikingly, the features a change in the sign about the charge neutrality point – it is positive for (hole-band) and negative for (electron-band). The current independence of establishes it to be a linear Hall effect [see Figure 2(c)]. The finite Hall response survives at least to room temperature with diminished amplitude as shown in Figure 2(b) and (Figure 2d). Importantly, the transverse signal is symmetric upon exchange of the voltage and current leads; [see Figure 2(e)]. This observation in hBN/graphene/WSe/hBN heterostructures of a room temperature, symmetric Hall effect that changes sign across the charge neutrality point is unique, and it is the central result of this letter.
Figure 2. Sign reversing Symmetric Hall effect (a) Plots of the zero magnetic-field longitudinal resistance (left-axis, red line) and zero magnetic-field transverse resistance (right-axis, blue line) versus ; the data were measured at mK. (b) response as a function of at few representative values of temperature; the AHE persists up to 300 K. (c) Plot of as a function of for two different values of electrical current; the data were taken at K. (d) Plot of the peak value of versus . The dotted line is a guide to the eye. (e) Plot of transverse resistance versus number density in two different configurations for device SW2. Configuration 1 measures and configuration 2 measures . One can see that ; the measured Hall effect is symmetric on interchanging the voltage and current leads. (f) Theoretically calculated symmetric Hall resistivity () for three different temperatures.
A finite zero-magnetic-field Hall response is often seen in graphene devices near the charge neutrality point (see Fig. S7 in SI). This can originate from trivial effects that mix the longitudinal and transverse signals (e.g., misaligned probes or misalignment of the measurement axis with the crystal axis); thus, the sign of this signal is independent of the charge carrier type. We emphasize that our observation is distinct from this trivial case in one very important aspect: In graphene/WSe heterostructures, the hall response flips sign as the carrier type changes from electrons to holes (see SI for a comparison of data in graphene and graphene/WSe devices). The data for were also reproduced in cryostats without a superconducting magnet, ruling out the remnant field of a magnet as the origin of the observed Hall response.
Having ruled out experimental artifacts as the origin of our unusual observation, we develop below a detailed model that explains our essential observations. Asymmetric Hall response can arise from the Drude mechanism in materials with an anisotropic band dispersion. The symmetric Hall conductivity is specified by [24, 42] . Here, is the band velocity, is the energy derivative of the Fermi distribution function, and is the scattering time. However, such a symmetric Hall response usually does not change its sign when going from the valence to the conduction band, contrary to our experimental observations (see Fig. S6 in the SI). We argue below that such sign reversal can arise in particle-hole symmetry broken systems with different anisotropy orientations in the electronic dispersion of the conduction and valence band.
To gain a complete understanding of the unusual Hall effect observed in SLG/WSe, we start by noting that the band structure of this system is anisotropic. This is schematically highlighted in Figure 1(g). The anisotropy axis in the conduction and the valence bands is close to orthogonal (see Fig. S1 of the SI for realistic calculations for graphene-WSe heterostructure). This band structure is reminiscent of that seen in periodically strained bilayer graphene that hosts a time-reversal symmetric linear pseudo-planar Hall effect [24]. Such an anisotropic band dispersion preserves time-reversal symmetry but breaks particle-hole symmetry (see SI for details). Let us consider a symmetric system in which we add a small anisotropy. The anisotropy modifies the band structure to induce a change in the band velocity, . In a simplistic scenario where are independent of the Bloch momentum, we have, . One can show that if the tilt direction of the conduction and the valence band are orthogonal to each other, as it is for SLG/WSe [see Fig. 1(e)], the product and consequently the symmetric Hall response reverses its sign as we go from conduction to valence band. This orthogonal orientation of the band anisotropy in the conduction and the valence band is the origin of the sign reversing symmetric Hall effect. See SI for the generic model of sign reversing symmetric Hall effect and detailed calculations for the graphene-WSe heterostructure. Figure 2(f) presents the symmetric Hall conductivity for an effective two-band model, which qualitatively captures the sign-reversing symmetric Hall effect (also see Fig. S2 of the SI).
We do not find evidence of sign reversing symmetric Hall effect in hBN/graphene/hBN devices without the intervening WSe layer (SI). Thus, the orthogonal asymmetry in the conduction and valence band induced by WSe-graphene combination is essential for this effect.
The band anisotropy in SLG on WSe can be traced back to lattice mismatch-induced strain. The lattice constant of graphene is Å while that of WSe is Å. This large lattice mismatch generates a significant strain across the graphene flake as the heterostructure relaxes to the stable ground state. From Raman spectroscopy, we estimate the magnitude of the strain on the SLG layer in our hBN/SLG/WSe/hBN heterostructure to be (see SI). We find that the strain magnitude decreases with temperature (see SI for details). Our observed symmetric Hall effect survives the combined action of i) decreasing strain and band anisotropy with increasing temperature and ii) the thermal broadening of the Fermi function to be finite at room temperature.
Symmetric Hall response with vertical displacement and magnetic field
Having demonstrated the symmetric Hall effect, we now focus on the dependence of the symmetric Hall response on a perpendicular displacement field D (Figure 3). It is illuminating to map the transverse zero-B-field conductivity data in the plane (Figure 3(a)). The plot shows to be finite only at the band edges, consistent with the idea that the longitudinal conductivity is large inside the band and vanishingly small in the vicinity of the band edges. This can be seen clearly in the line plots of for different values of D shown in Figure 3(b). Note that the plots are vertically offset by 200 for clarity. The measured has an intriguing D dependence; it changes its sign as the direction of D flips [Figure 3(a-Figure 3b)].
Figure 3. Dependence of the transverse resistance on and . (a) A 2-dimensional contour map of plotted in the – plane. (b) Plots of versus for different values of . The data have been vertically shifted by 200 for clarity. The dashed horizontal line for each plot marks the zero of . (c) A representative plot of versus measured at , an arrow marks the value of the Hall resistance.
Measurements in a finite magnetic field B applied perpendicular to the device interface (see SI) reveal the interplay between the classical Hall effect and the B = 0 symmetric Hall effect. The data smoothly crosses over from the symmetric Hall phase at B = 0 to the conventional Hall phase at finite B-field with an anti-crossing feature. This feature resembles the planar Hall effect in corrugated bilayer graphene[24]. A non-zero intercept of the plot of versus B [shown for a fixed in Figure 3(c)] on the B-axis captures the symmetric Hall effect. We note that is non-hsyteretic in the presence of a small non-quantizing magnetic field (see Fig. S6 (d) of SI), ruling out emergent ferromagnetism in the system.
In Figure 4(a), we present a plot of in the – plane measured at . We observe that with increasing , the resistance peak at the charge neutrality point splits into two maxima. This feature can be better appreciated from Figure 4(b), where we show individual plots of versus at several representative values of . At higher values of , we find two distinct peaks in separated by a shallow valley. Such a displacement field-dependent dispersion of the bands near the Dirac point is not captured by the existing models for graphene/ heterostructures [43, 44, 45, 46, 47, 48, 49] [33]. To remedy this, we construct a new model Hamiltonian for the graphene/ system, retaining both the and the graphene Hamiltonian blocks, which allows us to include the impact of a vertical displacement field systematically (see SI for details). Figure 4(c) is a plot of the theoretically calculated as a function of the chemical potential – the panels show the splitting of the conductivity minima into two asymmetric conductivity minima at finite . Our model thus reproduces the prominent features of both at zero displacement field [43, 45] and at a finite , along with the observed symmetric Hall effect.
Figure 4. Dependence of on . (a) A 2-dimensional contour map of plotted in the plane. (b) Plots of versus for different values of . The data have been vertically shifted by 1 k for clarity. The dashed horizontal line for each plot is the zero of the y-axis. (c) Variation of the calculated Drude conductivity with energy () for three different values of the interlayer potential induced by the applied electric field, meV (red line), 0 meV (blue line) and meV (green line), respectively. The values of have been scaled by where .
Discussion
To summarize, we report the first observation of room temperature time-reversal symmetric Hall effect in high-mobility heterostructures of graphene/. Primarily known for their promising spintronics aspects, the charge Hall response of such a heterostructure was expected to be relatively mundane. Contrary to this, we measure a time-reversal symmetric Hall effect at zero magnetic field. More interesting, the observed symmetric Hall effect changes sign across the charge neutrality point and persists till room temperature. We show that the symmetric Hall effect arises from the anisotropy of the electronic band structure. The sign reversal of the observed Hall effect arises from the different orientations of the anisotropy direction in the conduction and the valence band. The symmetric Hall response features a unique perpendicular electric field tunability. Our work establishes graphene- heterostructure as an excellent platform for further exploration of the interplay of charge, spin, and valley responses in band-engineered two-dimensional systems.
AUTHOR INFORMATION
Author Contributions
A.B., P.T., and D.S. conceptualized the study, performed the measurements, and analyzed the data. A.A., A.C., and K.D. performed the theoretical analysis. K.W. and T.T. grew the hBN single crystals. All the authors contributed to preparing the manuscript.
Notes
The authors declare no competing financial interest.
Acknowledgement
The authors acknowledge stimulating discussions with Sumanta Tewari and Jay Deep Sau. A.B. acknowledges funding from the Department of Science & Technology FIST program, DST fellowship (No. DST/SJF/PSA01/2016-17), and the U.S. Army DEVCOM Indo-Pacific (Project number: FA5209 22P0166). A.C. acknowledges the Indian Institute of Technology, Kanpur, and the Science and Engineering Research Board (SERB) National Postdoctoral Fellowship (PDF/ 2021/ 000346), India, for partial financial support. A.C. acknowledges Alexander Von Humboldt Post-doctoral fellowship, Germany, for funding. A.A. acknowledges the Science and Engineering Research Board for Project No. MTR/2019/001520, and the Department of Science and Technology for Project No. DST/NM/TUE/QM-6/2019(G)-IIT Kanpur of the Government of India for funding. K.W. and T.T. acknowledge support from JSPS KAKENHI (Grant Numbers 19H05790, 20H00354, and 21H05233).
Supporting Information Available
Supporting information contains detailed discussions of (a) Model Hamiltonian of Graphene/WSe heterostructure, (b) Generic low energy model for sign reversing symmetric Hall effect, (c) Drude longitudinal conductivity, (d) Device fabrication, (e) Raman shift and strain (e) Additional Data for symmetric Hall effect on device SW1. (f) Additional Data for symmetric Hall effect on device SW2 (g) Transverse resistance Data on hBN/graphene/hBN device at .
Supplementary Information
Model Hamiltonian of Graphene WSe heterostructure
In this section, we construct the low energy model Hamiltonian of monolayer graphene on a WSe layer. Going beyond the effective graphene model as reported in recent literature [43, 45, 50], we explicitly solve for the composite low energy Hamiltonian for the graphene-WSe heterostructure to capture the effect of perpendicular electric field correctly. We solve the following low-energy Hamiltonian
Here, and are the onsite Hamiltonian for graphene and the WSe respectively. The interaction between graphene and WSe layer has been included through spin and valley conserved off-diagonal hopping (). The effect of the perpendicular electric field is captured through the diagonal matrix .
We consider the monolayer of WSe in the - plane in the presence of intrinsic spin-orbit coupling (SOC) (), spin Zeeman field (). In addition, finite Rashba SOC term () is also considered within the WSe sector??. Including all these effects, the two-dimensional extended Dirac Hamiltonian () of WSe monolayer can be written as
The explicit forms of each term are as follows,
where for and valley respectively. As in the monolayer WSe, two degenerate but inequivalent valleys ( and ) are separated by a large momentum; we can split the total Hamiltonian into two valley-specific parts. Here, we have considered eV.Å as the Fermi velocity of WSe. represents the mass term that breaks the inversion symmetry. Here, and correspond to the SOC strengths of conduction and valence bands. In general, the valence band ( meV) of WSe possesses larger SOC strength than the conduction band ( meV), promoting relatively larger splitting in the valence band [51]. For simplicity of the calculation, we choose the SOC strengths of both the conduction and valence bands to be equal, meV.
We set meV, which induces a large gap between the conduction and valence bands of WSe. To model the low energy physics of graphene, we choose valley-specific Hamiltonian of the following form,
Here, eV.Å is the Fermi velocity of graphene. Eqn. (4) represents a gapless Dirac dispersion for the graphene sector. The coupling between the two layers is captured by the generalized hopping
where .
In a perfectly aligned graphene-WSe heterostructure, the off-diagonal term is expected to be zero. However, in realistic experimental scenarios, misalignment of the heterostructure, lattice mismatch-induced strain, and small angular mismatch can result in a non-zero off-diagonal hopping, which can be complex in general [21, 22, 24]. In our calculation, we set the hopping strength meV and meV. The proximity effect of the WSe layer opens up a gap at the Dirac crossing of the graphene bands. The induced band gap of graphene gets enhanced with increased hopping strength (primarily ). The off-diagonal terms of the generalized hopping matrix induce an asymmetry in the low energy band dispersion (see Figure 5) in the vicinity of the band edges. More importantly, the axis of band structure asymmetry in the conduction band and the valance band is different (almost orthogonal). This band structure is reminiscent of that seen in strained bilayer graphene [24].
Figure 5. Anisotropy of graphene-WSe₂ band structure in the conduction and valence band in both valleys. (a-b) and (e-f) show the energy dispersion of graphene-WSe₂ composite Hamiltonian, in presence of an off-diagonal hopping, along () direction for −valley and −valley respectively. The valley-specific constant energy contours of the valence band (VB) and conduction band (CB) in the vicinity of the band edges ( meV) are shown in (c-d) and (g-h). The opposite orientation of the anisotropy axis in the two valleys ensures the presence of time reversal symmetry. This was also explicitly checked via the relation . The anisotropic band structure plays a crucial role in generating a symmetric Hall effect. The different (close to orthogonal) orientation of the anisotropy axis in the conduction and valence bands ensures that the symmetric Hall response changes sign as one goes from electron-doped to hole-doped region.
Such an asymmetric band structure in the conduction and valance band is analogous to a tilted massive Dirac dispersion, with the tilt direction being orthogonal for the conduction band and the valance band. We have plotted the fixed energy contours near the conduction and valence band edges in Figure 5. We have explicitly checked that the time-reversal symmetry is preserved and the energy dispersion relation satisfies . The symmetric response calculated from the composite graphene- model suggests the Drude Hall resistivity changes sign when we go from the valence to conduction sector, similar to what we obtained in our experimental results. As an additional check, we have also calculated the linear anomalous Hall response of the system, which turns out to be zero (within bounds of numerical accuracy) in presence of time reversal symmetry (see Figure 6).
Figure 6. Drude Hall resistivity at K calculated from the graphene-WSe composite model in absence of perpendicular electric field. The Hall resistivity changes its sign from the valence () to the conduction () sector. The black dashed line is the calculated anomalous Hall resistivity of the system. The vanishing AHE indicates that the system preserves time-reversal symmetry.
The effect of the external perpendicular electric field is introduced by adding a diagonal Hamiltonian.
Figure 7 shows the band dispersion evolution with a perpendicular electric field. The band dispersion essentially undergoes an insulator-to-metal transition with the electric field (see Figure 7).
Figure 7. Impact of the electric field on the band structure of graphene/WSe heterostructure. (a), (b) and (c) show the band dispersion in the presence of electric field values meV, meV, and meV, respectively. The external electric field changes the low energy band dispersion of the composite graphene-WSe heterostructure, inducing a metal-insulator transition.
Generic low energy model for sign reversing symmetric Hall effect
The observed zero field Hall effect in our graphene-WSe devices has two important features: i) it is symmetric in nature, and ii) it changes sign across the charge neutrality point. To understand these, we note that in crystals with reduced symmetries, an anisotropic band dispersion can give rise to a symmetric Hall conductivity with . This is typically interpreted as the Drude Hall conductivity or as the pseudo-planar Hall conductivity[24]. An elegant way to see this is by considering a system with an anisotropic longitudinal Drude conductivity , for a given set of coordinates (say ). This anisotropic conductivity can be induced either by strain, heterostructure misalignment, or small moiré angle inhomogeneity. Rotating this conductivity matrix ( for a 2D system) to find the conductivity matrix in the device coordinate system (say ), we have . As long as this rotation is not by integer multiples of , there will always be a finite in the device coordinate system.
To show the existence of the symmetric Drude Hall effect in graphene-based devices, let us consider the case of gapped graphene with a tilted dispersion. It can be described by the following two-band model Hamiltonian for each valley (denoted by ),
Here, the symbols have their usual meaning. The second term induces gap-opening between the bands, and the third term causes opposite orientations of the tilt in the two valleys. The opposite tilt in the two valleys preserves time-reversal symmetry. The tilt in each valley of the Hamiltonian in Eq. (7) breaks particle-hole symmetry. For this tilted massive Dirac Hamiltonian, we can calculate the symmetric Drude Hall conductivity to be[24, 42]
Here, denotes the sum over bands, is the valley index to be summed over, and is the Fermi function. Note that in deriving the second part of Eq. (8), we have retained only the lowest order contribution in and after systematically expanding the , and the Fermi function up to second order contributions, and taken the limit. This establishes that graphene-based systems with anisotropic bands structure can support a symmetric Hall effect. Such an anisotropy can be induced by strain, substrate effects, or by angle misalignment in the case of heterostructures such as graphene-WSe. We note the following here: i) Without tilt (or band anisotropy), the Drude mechanism can not generate a symmetric Hall conductivity in massive graphene. ii) We obtain the same Drude Hall conductivity for both valleys with opposite tilt orientation. iii) Most importantly, the Hall conductivity generated in this manner does not change its sign as we move from the valence to the conduction band. However, in our experiment, the Hall conductivity changes sign across the charge neutrality point. This mandates us to go beyond the tilted Dirac model described by Eq. (7).
To understand the origin of sign reversing symmetric from Eq. (8), we note that it can arise if either of or have a different sign in the conduction band and the valance band. As an example, this can be achieved by considering the following set of band dispersion
Here, the superscript () refers to the conduction (valence) band. These band dispersions have orthogonal tilts or orthogonal anisotropy directions in the conduction and valence bands. For these set of bands, we have .
These simple calculations show that i) the anisotropy of the electronic dispersion can give rise to a symmetric Hall effect (or the Drude Hall effect) and, more importantly, ii) the sign reversal can arise from the conduction band and valance bands being anisotropic along different axes. The scenario in our experiment is similar. The graphene-WSe band structure has an anisotropy, which is similar to the example presented above. Specifically, the anisotropy axis in the conduction band and the valence band are orthogonal to each other (see Figure 5) and mimic the setup above with orthogonal tilt directions in the conduction and valance bands as described by Eq. (9)–(10). Consequently, the measured Drude Hall conductivity is opposite when we move from the valence to the conduction band.
Drude longitudinal conductivity
We calculate using the expression of the Drude conductivity[42]
The band velocity is defined as , where is the band index. The longitudinal conductivity (), which follows the density of states (DOS), shows a -like pattern with an increase in the electric field. The calculated captured the qualitative nature of the inverse of the experimental resistivity () plot of Fig. 4(a) of the main manuscript. The pseudo gap within the first and second valence (conduction) bands promotes the low conducting dips below (above) the Fermi energy, whereas for a finite electric field, the substantial DOS at Fermi energy promotes the metallic nature indicated by a peak at the of Fig. 4(c) of the main manuscript.
Device fabrication
Thin flakes of WSe, hBN, and single-layer graphene (SLG) were mechanically exfoliated on Si/SiO substrates. The thickness of the flakes was initially estimated from the color contrast under an optical microscope and later confirmed using Raman spectroscopy. This was followed by sequential pickup of each flake using Polycarbonate (PC) film at 90°C. The assembled heterostructure was transferred on a new Si/SiO substrate[40, 41]. The heterostructure is then cleaned in chloroform, acetone, and IPA to remove the PC residue. The heterostructure was then annealed at 250°C for 3 hours. Electron beam lithography was used to define the contact and top gate electrodes. We used reactive ion etching (mixture of CHF and O gas) to etch top hBN to make one-dimensional edge contacts to graphene?. For making the electrical contacts, Cr/Au (5 nm/60 nm) was deposited, followed by liftoff in hot acetone and cleaning in IPA. The unwanted hBN and graphene were removed using E-beam lithography and dry etching to define the Hall bar. We transferred an hBN top of the device and fabricated a metallic top gate using lithography and thermal deposition.
Raman shift and strain
We used low-temperature Raman spectroscopy in graphene WSe stack to estimate the strain in graphene. High-quality SLG has two prominent Raman active modes, the G-mode (1580 cm) and the 2D-mode (2690 cm). In the presence of a uniaxial strain , the shift in 2D peak has been measured to be ?. Figure 8a shows a comparison of the temperature-dependence of the Raman shift of the 2D band measured for graphene and for graphene on WSe . In Figure 8b, we show a plot of the T-dependence of . One can see that the difference in the Raman shift of the 2D peak increases rapidly with a decrease in ; the positive value of indicates that the strain is compressive. The temperature dependence of the strain in graphene was extracted from the data in Figure 8b; its magnitude is plotted in Figure 8c. The data shows that SLG on single layer WSe undergoes a significant compressive strain of about 0.2% at 4 K.
Figure 8. Raman shift in the 2D band of graphene (a) Temperature variation of the measured Raman shift of the 2D peak of SLG (blue filled circles) and of SLG on single-layer WSe (red filled circles). (b) Plot of versus for SLG on single-layer WSe. (c) Plot of the - dependence of the magnitude of the strain in SLG on single-layer WSe. (d) Plots of Raman intensity versus Raman shift in hBN/SLG/WSe/hBN. Four data points are from four different locations. (e) Plots of Raman intensity versus Raman shift in the hBN/SLG/hBN region of the device. The data were taken at the four different locations in the sample.
To check the charge and strain homogeneity in our devices, we measured the Raman spectra of the comparative device consisting of hBN/SLG/hBN and hBN/SLG/WSe/hBN (these devices have common hBN and SLG). We measured the Raman spectra at several places in both devices. In Figure 8d, we show the Raman shift of hBN/SLG/WSe/hBN device at four representative points. In Figure 8e, we show the Raman shift of the hBN/SLG/hBN device at four representative points. We observed uniform intensity and no peak shift in these two individual regions attesting to the charge and strain homogeneity of the stacks. The spectra in each region are identical, again confirming the homogeneity of the device.
Additional Data for symmetric Hall effect on device SW1
The data from the device named SW1 measured across two pairs of transverse leads are compared in Figure 9 – the data in both cases show a finite zero-magnetic-field transverse signal with sign depending on whether the charge carriers are electrons or holes. The for this device was estimated to be from the measured longitudinal resistance. The changes sign for over range of which is much larger than the . So the study reported in this manuscript is outside the range of electron-hole puddles and, we believe, unaffected by it.
Figure 9. Data on device SW1. Plot of transverse resistance versus number density in two different configurations for device SW1. The measurement in both configurations shows a sign reversing symmetric Hall effect.
Additional Data for symmetric Hall effect on device SW2
Figure 10(a) shows the data for zero-field longitudinal and transverse resistance in device SW2; one can see the appearance of a finite that changes its sign near the Dirac point. Interestingly, the seems to have contribution from , probably due to misaligned probes. The Figure 2(e) of the main text presents the transverse signal measured in two different configurations, configuration 1 measures while configuration 2 measures . The two signals overlap exactly with each other. Note that this is one expects from the Onsager relation for .
Figure 10. Data on device SW2. (a) Plot of longitudinal and transverse resistivity versus number density for device SW2. (b) Plot of at small magnetic field values measured for device SW2. (c) A 2D map of the transverse resistance in the plane; the data shows a finite Hall signal at T. (d) Plot of magnetoresistance in a small magnetic field at V/nm displacement field. The data were taken at m. The red open circles are the data for -sweep from positive to negative values; the filled circles are the data for -sweep from negative to positive values. The arrows indicate the direction of the magnetic field sweep.
Figure 10(b) shows the line plots of the transverse signal measured in device SW2 in the presence of a small perpendicular magnetic field. The data show smooth evolution of the time-reversal symmetric Hall effect (at ) into the classical Hall signal (to ). This can be better appreciated from Figure 10(c), which is a 2D map of the transverse signal in the - plane.
The measured magnetoresistance in our devices is non-hysteretic (Figure 10(d)). This is clear evidence of the absence of ferromagnetism in the system.
Transverse resistance Data on hBN/graphene/hBN device at
To complement the data presented in the main manuscript, this section presents the data of zero-magnetic field transverse resistance in two hBN/graphene/hBN devices – SG3 and SG4. The data have been plotted in Figure 11(a) and (b). In both plots, the transverse resistance is plotted in red, and the longitudinal resistance is plotted in black. In both devices, we find a finite around the charge neutrality point. This has been seen previously by several groups and is usually attributed to misaligned transverse voltage probes or to impurity scattering. We have also measured the in a configuration where the probes measuring the transverse voltage are intentionally misaligned by a large amount (see Figure 11(c)). In all these cases, the mimics the (with a diminished amplitude) signal; it does not change sign at the charge neutrality point.
Figure 11. Comparison of transverse resistance in SLG and SLG/WSe devices. Plots of measured (black line) and (red line) versus number density for (a) device SG3 and (b) device SG4. The insets in (a) show the measurement configurations common to both panels. (c) Plots of measured (black line) and with intentionally misaligned transverse probes (red line) versus number density for device SG4. The measurement configuration is shown in the inset. (d) Comparison of measured in bare SLG device (red line) and SLG/WSe heterostructure device (black line). The data for the SLG device has the same sign for electrons and holes while that for the SLG/WSe changes sign as the Fermi energy moves from the conduction band (positive ) to the valence band (negative ).
Thus, there is a very important distinction between the observed in bare graphene and the observations on strained SLG/WSe reported in the main manuscript. In the data presented in Figure 11(a-Figure 11c), the measured has the same sign for both types of carriers. This is in stark contrast to the in SLG/WSe devices where the measured for electrons and holes have opposite signs (Figure 11(d)).
References
- [1]Xiao, D.; Chang, M.-C.; Niu, Q. Berry phase effects on electronic properties. Rev. Mod. Phys. 2010, 82, 1959–2007.DOI
- [2]Ahn, J.; Guo, G.-Y.; Nagaosa, N.; Vishwanath, A. Riemannian geometry of resonant optical responses. Nature Physics 2022, 18, 290–295.DOI
- [3]Gao, A. et al. Layer Hall effect in a 2D topological axion antiferromagnet. Nature 2021, 595, 521–525.
- [4]Bhalla, P.; Das, K.; Culcer, D.; Agarwal, A. Resonant Second-Harmonic Generation as a Probe of Quantum Geometry. Phys. Rev. Lett. 2022, 129, 227401.DOI
- [5]Han, W.; Kawakami, R. K.; Gmitra, M.; Fabian, J. Graphene spintronics. Nature Nanotechnology 2014, 9, 794–807.DOI
- [6]Sinova, J.; Valenzuela, S. O.; Wunderlich, J.; Back, C.; Jungwirth, T. Spin hall effects. Reviews of modern physics 2015, 87, 1213.DOI
- [7]Hirsch, J. Spin hall effect. Physical review letters 1999, 83, 1834.
- [8]Bernevig, B. A.; Zhang, S.-C. Quantum spin Hall effect. Physical review letters 2006, 96, 106802.pubmed.ncbi.nlm.nih.gov/16605772
- [9]Tiwari, P.; Jat, M. K.; Udupa, A.; Narang, D. S.; Watanabe, K.; Taniguchi, T.; Sen, D.; Bid, A. Experimental observation of spin-split energy dispersion in high-mobility single-layer graphene/WSe2 heterostructures. npj 2D Materials and Applications 2022, 6, 68.DOI
- [10]Xiao, D.; Liu, G.-B.; Feng, W.; Xu, X.; Yao, W. Coupled Spin and Valley Physics in Monolayers of MoS₂ and Other Group-VI Dichalcogenides. Phys. Rev. Lett. 2012, 108, 196802.DOI
- [11]Cresti, A.; Nikolić, B. K.; García, J. H.; Roche, S. Charge, spin and valley Hall effects in disordered graphene. La Rivista del Nuovo Cimento 2016, 39, 587–667.
- [12]Mak, K. F.; McGill, K. L.; Park, J.; McEuen, P. L. The valley Hall effect in MoS2 transistors. Science 2014, 344, 1489–1492.DOI
- [13]Lee, J.; Mak, K. F.; Shan, J. Electrical control of the valley Hall effect in bilayer MoS2 transistors. Nature nanotechnology 2016, 11, 421–425.DOI
- [14]Liu, J.; Ma, Z.; Gao, J.; Dai, X. Quantum valley Hall effect, orbital magnetism, and anomalous Hall effect in twisted multilayer graphene systems. Physical Review X 2019, 9, 031021.DOI
- [15]Qiao, Z.; Yang, S. A.; Feng, W.; Tse, W.-K.; Ding, J.; Yao, Y.; Wang, J.; Niu, Q. Quantum anomalous Hall effect in graphene from Rashba and exchange effects. Phys. Rev. B 2010, 82, 161414.DOI
- [16]Shimazaki, Y.; Yamamoto, M.; Borzenets, I. V.; Watanabe, K.; Taniguchi, T.; Tarucha, S. Generation and detection of pure valley current by electrically induced Berry curvature in bilayer graphene. Nature Physics 2015, 11, 1032–1036.DOI
- [17]Sui, M.; Chen, G.; Ma, L.; Shan, W.-Y.; Tian, D.; Watanabe, K.; Taniguchi, T.; Jin, X.; Yao, W.; Xiao, D.; Zhang, Y. Gate-tunable topological valley transport in bilayer graphene. Nature Physics 2015, 11, 1027–1031.DOI
- [18]Wallbank, J. R. et al. Tuning the valley and chiral quantum state of Dirac electrons in van der Waals heterostructures. Science 2016, 353, 575–579.DOI
- [19]Sodemann, I.; Fu, L. Quantum Nonlinear Hall Effect Induced by Berry Curvature Dipole in Time-Reversal Invariant Materials. Phys. Rev. Lett. 2015, 115, 216806.DOI
- [20]Du, Z. Z.; Wang, C. M.; Li, S.; Lu, H.-Z.; Xie, X. C. Disorder-induced nonlinear Hall effect with time-reversal symmetry. Nature Communications 2019, 10, 3047.
- [21]Sinha, S.; Adak, P. C.; Chakraborty, A.; Das, K.; Debnath, K.; Sangani, L. D. V.; Watanabe, K.; Taniguchi, T.; Waghmare, U. V.; Agarwal, A.; Deshmukh, M. M. Berry curvature dipole senses topological transition in a moiré superlattice. Nature Physics 2022, 18, 765–770.DOI
- [22]Chakraborty, A.; Das, K.; Sinha, S.; Adak, P. C.; Deshmukh, M. M.; Agarwal, A. Nonlinear anomalous Hall effects probe topological phase-transitions in twisted double bilayer graphene. 2D Materials 2022, 9, 045020.DOI
- [23]Zhai, D.; Chen, C.; Xiao, C.; Yao, W. Layer-Contrasted Hall Effect in Twisted Bilayers with Time Reversal Symmetry. 2022; https://arxiv.org/abs/2207.14644.
- [24]Ho, S.-C.; Chang, C.-H.; Hsieh, Y.-C.; Lo, S.-T.; Huang, B.; Vu, T.-H.-Y.; Ortix, C.; Chen, T.-M. Hall effects in artificially corrugated bilayer graphene without breaking time-reversal symmetry. Nature Electronics 2021, 4, 116–125.DOI
- [25]Nandy, S.; Sharma, G.; Taraphder, A.; Tewari, S. Chiral anomaly as the origin of the planar Hall effect in Weyl semimetals. Physical review letters 2017, 119, 176804.DOI
- [26]Kumar, N.; Guin, S. N.; Felser, C.; Shekhar, C. Planar Hall effect in the Weyl semimetal GdPtBi. Physical Review B 2018, 98, 041103.DOI
- [27]Avsar, A.; Tan, J. Y.; Taychatanapat, T.; Balakrishnan, J.; Koon, G.; Yeo, Y.; Lahiri, J.; Carvalho, A.; Rodin, A.; O’Farrell, E., et al. Spin-orbit proximity effect in graphene. Nature communications 2014, 5, 1–6.
- [28]Ghiasi, T. S.; Kaverzin, A. A.; Blah, P. J.; van Wees, B. J. Charge-to-spin conversion by the Rashba–Edelstein effect in two-dimensional van der Waals heterostructures up to room temperature. Nano letters 2019, 19, 5959–5966.DOI
- [29]Tiwari, P.; Srivastav, S. K.; Ray, S.; Das, T.; Bid, A. Observation of Time-Reversal Invariant Helical Edge-Modes in Bilayer Graphene/WSe2 Heterostructure. ACS Nano 2021, 15, 916–922, PMID: 33378173.DOI
- [30]Herling, F.; Safeer, C. K.; Ingla-Aynés, J.; Ontoso, N.; Hueso, L. E.; Casanova, F. Gate tunability of highly efficient spin-to-charge conversion by spin Hall effect in graphene proximitized with WSe2. APL Materials 2020, 8, 071103.DOI
- [31]Dastgeer, G.; Afzal, A. M.; Jaffery, S. H. A.; Imran, M.; Assiri, M. A.; Nisar, S. Gate modulation of the spin current in graphene/WSe2 van der Waals heterostructure at room temperature. Journal of Alloys and Compounds 2022, 919, 165815.DOI
- [32]Lee, S.; de Sousa, D. J. P.; Kwon, Y.-K.; de Juan, F.; Chi, Z.; Casanova, F.; Low, T. Charge-
- [40]Pizzocchero, F.; Gammelgaard, L.; Jessen, B. S.; Caridad, J. M.; Wang, L.; Hone, J.; Bøggild, P.; Booth, T. J. The hot pick-up technique for batch assembly of van der Waals heterostructures. Nature Communications 2016, 7, 1–10.DOI
- [41]Wang, L.; Meric, I.; Huang, P.; Gao, Q.; Gao, Y.; Tran, H.; Taniguchi, T.; Watanabe, K.; Campos, L.; Muller, D., et al. One-dimensional electrical contact to a two-dimensional material. Science 2013, 342, 614–617.DOI
- [42]Ashcroft, N. W.; Mermin, N. D. Solid State Physics; Holt-Saunders, 1976.
- [43]Gmitra, M.; Kochan, D.; Högl, P.; Fabian, J. Trivial and inverted Dirac bands and the emergence of quantum spin Hall states in graphene on transition-metal dichalcogenides. Phys. Rev. B 2016, 93, 155104.DOI
- [44]Offidani, M.; Milletarì, M.; Raimondi, R.; Ferreira, A. Optimal Charge-to-Spin Conversion in Graphene on Transition-Metal Dichalcogenides. Phys. Rev. Lett. 2017, 119, 196801.
- [45]Cummings, A. W.; Garcia, J. H.; Fabian, J.; Roche, S. Giant Spin Lifetime Anisotropy in Graphene Induced by Proximity Effects. Phys. Rev. Lett. 2017, 119, 206601.DOI
- [46]Garcia, J. H.; Vila, M.; Cummings, A. W.; Roche, S. Spin transport in graphene/transition metal dichalcogenide heterostructures. Chemical Society Reviews 2018, 47, 3359–3379.
- [47]Li, Y.; Koshino, M. Twist-angle dependence of the proximity spin-orbit coupling in graphene on transition-metal dichalcogenides. Phys. Rev. B 2019, 99, 075438.DOI
- [48]Zubair, M.; Vasilopoulos, P.; Tahir, M. Influence of interface induced valley-Zeeman and spin-orbit couplings on transport in heterostructures of graphene on WSe₂. Phys. Rev. B 2020, 101, 165436.DOI
- [49]Kumar, A.; Maiti, S.; Maslov, D. L. Zero-field spin resonance in graphene with proximity-induced spin-orbit coupling. Phys. Rev. B 2021, 104, 155138.DOI
- [50]Gmitra, M.; Fabian, J. Graphene on transition-metal dichalcogenides: A platform for proximity spin-orbit physics and optospintronics. Phys. Rev. B 2015, 92, 155403.DOI
- [51]Tahir, M.; Vasilopoulos, P. Magneto-optical transport properties of monolayer WSe₂. Phys. Rev. B 2016, 94, 045415.DOI










