Evidence for charge transfer and proximate magnetism in graphene–$\alpha \text{−}{\mathrm{RuCl}}_3$ heterostructures

We report a study of electronic transport in van der Waals heterostructures composed of flakes of the antiferromagnetic Mott insulator $\alpha \text{−}{\mathrm{RuCl}}_3$ placed on top of monolayer graphene Hall bars. While the zero-field transport shows a strong resemblance to that of isolated graphene, we find a consistently $p$-type Hall effect suggestive of multiband conduction, along with a nonmonotonic and gate-voltage-dependent excursion of the resistivity at low temperatures that is reminiscent of transport in the presence of a magnetic phase transition. We interpret these data as evidence for charge transfer from graphene to $\alpha \text{−}{\mathrm{RuCl}}_3$ in an inhomogeneous device yielding both highly and lightly doped regions of graphene, while the latter shows a particular sensitivity to magnetism in the $\alpha \text{−}{\mathrm{RuCl}}_3$. Thus proximity to graphene is a means to access magnetic properties of thin layers of magnetic insulators.

Figure 1

Transport at zero magnetic field. Though similar in appearance to the usual electronic transport in graphene-on-oxide, the minimum conductivity is much larger than expected. (a),(b) Resistivity vs temperature and gate voltage in two representative devices, D1 and D2, respectively. (c),(d) Constant-temperature line cuts of (a) and (b), replotted as the conductivity, $\sigma =1/\rho$. Inset to (c) are images of device D1 showing the monolayer graphene Hall bar before (left) and after (right) being covered by a $\sim 10$-nm-thick $\alpha \text{−}{\mathrm{RuCl}}_3$ flake. Circular features are remnant spots from a polycarbonate layer used in transferring the flake.

Figure 2

Magnetotransport of graphene–$\alpha \text{−}{\mathrm{RuCl}}_3$ devices. (a) Shubnikov–de Haas oscillations from device D3 for a range of gate voltages on either side of the minimum conductivity (here at $V_g=25$ V). (b) The SdH oscillation frequency $B_F$ for two devices. The corresponding charge density $n=g{B}_f/{\varphi }_0$ (where we assume $g=4$ degrees of freedom as for graphene and ${\varphi }_0$ is the magnetic-flux quantum) is marked on the right axis. Dashed lines show the expected graphene density, $n=\alpha (V_g-V_{\text{DP}})$, based on prior measurements of isolated graphene flakes on the same substrates, with $\alpha =7.2\times {10}^{10}\mathrm{carriers}/{\mathrm{cm}}^2/\mathrm{V}$ and $V_{\text{DP}}$ is the location of the zero-field conductivity minimum. (c) Main panel: Hall resistance for D3 acquired over a range of gate voltages [converted to carrier densities by the calibration in (b)]. The colors correspond to those in (a), and include three additional traces at densities close to charge neutrality that did not exhibit SdH oscillations. Inset: the same data, vertically offset for clarity, with fits (black lines) to a two-band model of magnetotransport. (d) Carrier densities of the second band determined from two-band model fits.

Figure 3

Transport signatures of magnetic transitions. The resistivity of graphene–$\alpha \text{−}{\mathrm{RuCl}}_3$ heterostructures shows characteristics of magnetic transitions [42, 43, 44, 45, 46]. These data are for the same two devices as in Fig. 1: D1 (a)–(c) and D2 (d)–(f). (a),(d) Color maps showing $d\rho /dT$ vs the graphene charge-carrier density in each device, normalized by the maximum value of $d\rho /dT$ below 60 K at each density. This highlights the evolution in temperature at which peaks in $d\rho /dT$ are observed to occur. (b)–(e) Line cuts of $d\rho /dT$ from (a) and (d), offset vertically and not normalized in order to show the variation in amplitude of the peaks and dips. Blue-to-red shading follows the transition from $p$- to $n$-type doping of the graphene, with charge neutrality at the blue/red border. The inset to (e) shows a $12\times$ magnified view of the top $p$-type traces. Short black lines along the left axis mark the location of $d\rho /dT=0$ for each trace. (c),(f) Temperatures of the $d\rho /dT$ peak maxima (open circles) and dip minima (filled circles). For D1 (a)–(c) two peaks can be discerned for $p$-type graphene. In both devices, the peak temperatures show a sharp variation at charge neutrality in graphene.