We use a van der Waals pickup technique to fabricate different heterostructures containing WSe 2 (WS 2 ) and graphene. The heterostructures were structured by plasma etching, contacted by one-dimensional edge contacts, and a top gate was deposited. For graphene/WSe 2 /SiO 2 samples we observe mobilities of ∼12 000 cm 2 V −1 s −1 . Magnetic-field-dependent resistance measurements on these samples show a peak in the conductivity at low magnetic fields. This dip is attributed to the weak antilocalization (WAL) effect, stemming from spin-orbit coupling. Samples where graphene is encapsulated between WSe 2 (WS 2 ) and hexagonal boron nitride show a much higher mobility of up to ∼120 000 cm 2 V −1 s −1 . However, in these samples no WAL peak can be observed. We attribute this to a transition from the diffusive to the quasiballistic regime. At low magnetic fields a resistance peak appears, which we ascribe to a size effect due to boundary scattering. Shubnikov-de Haas oscillations in fully encapsulated samples show all integer filling factors due to complete lifting of the spin and valley degeneracies.

No abstract

We compare different methods to measure the anisotropy of the spin lifetime in graphene. In addition to out-of-plane rotation of the ferromagnetic electrodes and oblique spin precession, we present a Hanle experiment where the electron spins precess around either a magnetic field perpendicular to the graphene plane or around an in-plane field. In the latter case, electrons are subject to both in-plane and out-of-plane spin relaxation. To fit the data, we use a numerical simulation that can calculate precession with anisotropies in the spin lifetimes under magnetic fields in any direction. Our data show a small, but distinct anisotropy that can be explained by the combined action of isotropic mechanisms, such as relaxation by the contacts and resonant scattering by magnetic impurities, and an anisotropic Rashba spin-orbit based mechanism. We also assess potential sources of error in all three types of experiment and conclude that the in-plane/out-of-plane Hanle method is most reliable.

We fabricated a non-local spin valve device with Co-MgO injector/detector tunnel contacts on a graphene spin channel. In this device, the spin polarization of the injector contact can be tuned by both the injector current bias and the gate voltage. The spin polarization can be turned off and even inverted. This behavior enables a spin transistor where the signal is switched off by turning off the spin injection using the field-effect. We propose a model based on a gate-dependent shift of the minimum in the graphene density of states with respect to the tunneling density of states of cobalt, which can explain the observed bias and gate dependence.Spintronics expands electronics from using only the charge of the electron to also utilizing its spin property 1 . So far, spintronic devices have only been used for data storage, but concepts exist for also building spin based logic circuitry 2-4 . The paradigmatic device, the spin field effect transistor was proposed by Datta and Das in 1990 2 . Here, spins are injected from a ferromagnetic electrode into a non-magnetic channel, and a spin-dependent signal is detected at a second, ferromagnetic electrode. Spins are rotated along the way by a gate-tunable spin-orbit interaction. While this device allows for an all-electric control of spin signals, in contrast to magnetic switching of the electrodes, the channel needs to have strong and tunable spin-orbit interaction and, at the same time, a long spin lifetime. Because of these conflicting requirements, an attempt to fully realize the Datta-Das transistor was only presented more than two decades after the original proposal 5 . On the other hand, when the polarization of spin injection can be manipulated electrically, a transistor device can be realized in a device with long spin lifetime in the channel, such as graphene 6 . Electric tunability of spin injection has been demonstrated in magnetic tunnel junctions 7-11 , or Si based devices 12 . For graphene spintronics, bias-dependent spin polarization, including signal inversion, was reported for hexagonal boron nitride (hBN) 13,14 or MgO 15,16 as a tunnel barrier. In a MoS 2 /graphene heterostructure, a gatedependent suppression of the spin signal was reported 17 . However, no gate-controlled signal inversion was reported for tunneling spin injection in graphene devices.In this work, we report on a gate and bias-tunable spin polarization in a Co/MgO/graphene device. Importantly, the sign of spin polarization can be reversed, and spin injection can even be turned off by gate control, thus enabling a true three-terminal spintronic device. We find that at an elevated negative injector bias U inj , the spin polarization vanishes and afterwards changes sign. At this bias setting, which we call spin neutrality point, both sign and magnitude of spin polarization can be controlled by a voltage V g applied to the back gate of the sample. Fig. 1 shows a schematic picture of the graphene FIG. 1. Sample schematic showing a graphene flake with contacts in the non-local spin valve meas...

The weak spin-orbit interaction in graphene was predicted to be increased, e.g., by hydrogenation. This should result in a sizable spin Hall effect (SHE). We employ two different methods to examine the spin Hall effect in weakly hydrogenated graphene. For hydrogenation we expose graphene to a hydrogen plasma and use Raman spectroscopy to characterize this method. We then investigate the SHE of hydrogenated graphene in the H-bar method and by direct measurements of the inverse SHE. Although a large nonlocal resistance can be observed in the H-bar structure, comparison with the results of the other method indicate that this nonlocal resistance is caused by a non-spin-related origin.arXiv:1809.10475v1 [cond-mat.mes-hall]