Room Temperature Electrical Detection of Spin Polarized Currents in

Nov 11, 2015 - (17) However, the direct electrical detection of a current-induced spin ..... Area of the Advance program at Chalmers University of Tec...
2 downloads 3 Views 2MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Room Temperature Electrical Detection of Spin Polarized Currents in Topological Insulators André Dankert,* Johannes Geurs, M. Venkata Kamalakar, Sophie Charpentier, and Saroj P. Dash* Department of Microtechnology and Nanoscience, Chalmers University of Technology, SE 41296 Göteborg, Sweden S Supporting Information *

ABSTRACT: Topological insulators (TIs) are a new class of quantum materials that exhibit a current-induced spin polarization due to spinmomentum locking of massless Dirac Fermions in their surface states. This helical spin polarization in three-dimensional (3D) TIs has been observed using photoemission spectroscopy up to room temperatures. Recently, spin polarized surface currents in 3D TIs were detected electrically by potentiometric measurements using ferromagnetic detector contacts. However, these electric measurements are so far limited to cryogenic temperatures. Here we report the room temperature electrical detection of the spin polarization on the surface of Bi2Se3 by employing spin sensitive ferromagnetic tunnel contacts. The current-induced spin polarization on the Bi2Se3 surface is probed by measuring the magnetoresistance while switching the magnetization direction of the ferromagnetic detector. A spin resistance of up to 70 mΩ is measured at room temperature, which increases linearly with current bias, reverses sign with current direction, and decreases with higher TI thickness. The magnitude of the spin signal, its sign, and control experiments, using different measurement geometries and interface conditions, rule out other known physical effects. These findings provide further information about the electrical detection of current-induced spin polarizations in 3D TIs at ambient temperatures and could lead to innovative spin-based technologies. KEYWORDS: Topological insulators, Bi2Se3, spin-momentum locking, spin polarized surface states, room temperature, ferromagnetic tunnel contacts

T

presence of unpolarized bulk charge currents.17 However, the direct electrical detection of a current-induced spin polarization in 2D and 3D TIs has been so far restricted to cryogenic temperatures,9,10,14−16 which limits further progress in this research field and its application potentials. The room temperature electrical detection of such highly correlated spin systems is not only interesting for fundamental research but also for applications in dissipationless quantum spintronic devices.3,4 Here we demonstrate the room temperature electrical detection of spin polarized currents on the surface of the Bi2Se3 crystals using ferromagnetic tunnel contacts. An applied electric field creates a finite momentum kx⃗ of the charge carriers yielding a perpendicularly locked spin transport S⃗ on the surface, as depicted in Figure 1a. We probe this current-induced spin polarization by using sensitive FM tunnel contacts (see Figure 1b−d and Methods) resulting in a magnetoresistance between the magnetization of the FM and the spin-polarized current on the Bi2Se3 surface. Electrical Characterization of Bi2Se3 and FM Tunnel Contacts. Bi2Se3 flakes were exfoliated from a bulk crystal (see Supplementary Figure S1) onto a SiO2/Si substrate. The TiO2/ Co FM tunnel contacts were prepared by electron beam lithography and evaporation methods. The details about the

he strong spin−orbit (SO) coupling in three-dimensional (3D) topological insulators (TIs) leads to insulating bulk and conducting surface states protected by time reversal symmetry.1−3 Electrons populating these topological surface states (TSS) possess only one spin state per momentum state (spin-momentum locking) in contrast to conventional materials.3,4 The TSS are extremely robust against most perturbations from defects or impurities and can enable the propagation of dissipationless spin currents.2,3 Semiconducting Bi2Se3 is a 3D TI with a single Dirac cone at the Fermi level,1,5 which makes it an ideal prototype to study topological effects. Additionally, its weak electron−phonon coupling allows the persistence of topological surface states up to room temperature. Such spin helicity of the Bi2Se3 surface has been experimentally measured by spin-resolved photoemission spectroscopy up to 300 K,6−8 which has the advantage to probe only the surface effects without bulk contributions. However, the electrical detection of spin polarizations in 3D TIs remained challenging. Undesired doping and low TI bulk band gaps create a parallel conduction channel. Therefore, the electrical quantum spin Hall method, used for detection of spin edge states in two-dimensional (2D) TIs,9,10 is unsuitable for utilization in the 3D case. So far, dynamical methods were employed to couple the TSS to ferromagnetic (FM) contacts creating a spin transfer torque11 or for spin pumping.12,13 Only recently, potentiometric measurements have been used to detect spin-polarized surface currents in 3D TIs probed by a FM contact,14−16 which act as an efficient detector even in the © XXXX American Chemical Society

Received: August 4, 2015 Revised: October 30, 2015

A

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 1. Schematics of spin-valve device to probe the spin helical topological surface states. (a) Dirac cone of the topological surface state with elevated Fermi level with spins locked perpendicular to its momentum. (b) Schematic of a TI with FM tunnel contacts. The direction of spin current in Bi2Se3 is defined by the charge current direction due to spin-momentum locking (SML). (c) Atomic force microscope image of a Bi2Se3 flake with a thickness of 35 nm(Inset: Line-scan (red) over the flake). (d) Colored electron microscope image of a Bi2Se3 device with FM tunnel contacts (Co/ TiO2). (e) Temperature dependence of the current−voltage characteristics, (f) Resistance-voltage dependence and (g) Temperature dependence of the zero bias resistance (ZBR) of the Co/TiO2 tunnel contacts.

Bi2Se3 crystal and device fabrication are presented in the Methods and Supporting Information. There are several advantages to introduce a tunnel barrier for the detection of spin polarization on the surfaces of TIs. First, direct contacts of FM on TIs can break the time reversal symmetry and change the electronic structure of the TIs, which can be prevented by introducing a thin TiO2 layer. Additionally, this tunnel barrier circumvents the conductivity mismatch problem between the FM and TI and suppresses leakage current and diffusion of spins into the FM.18,19 Recent studies also found that the growth of TiO2 on 2D layered materials to be more uniform than other oxides.20 Finally, such contacts are known to be very sensitive to spin-polarized currents up to room temperature, despite an unpolarized background charge current.21−23 We characterized the tunnelling properties of the FM tunnel contacts in a three-terminal measurement configuration (Figure 1e). The current−voltage characteristics exhibit a nonlinear behavior (Figure 1e and f) typical for tunnelling transport. Furthermore, the temperature dependence shows an increase of the resistance by a factor of 2.5 at low temperature indicating a uniform pinhole free TiO2 tunnel barrier on the Bi2Se3 flakes (Figure 1g).24 This excellent electrical tunnelling behavior of our TiO2 tunnel barrier allows the efficient detection of spinmomentum locking (SML) in Bi2Se3. The electrical properties of the Bi2Se3 flakes were characterized on devices having Ti/Au contacts. The Bi2Se3 channel resistivity shows a metallic behavior with a reduction in sheet resistance by a factor of 2 from R□ = 36 Ω at 300 K to R□ = 18 Ω at 2 K (Figure 2a). This stems from a large charge carrier concentration of 5 × 1019 cm−3 as extracted from Hall measurements. We observe a channel mobility of 2000 cm2 (V s)−1 at room temperature, which almost doubles at 2 K (see Supplementary Figure S2). Since there are at least two parallel conduction channels (bulk and surface) in the TI, an analysis of the Hall data measured at high magnetic field using a two-carrier model could clarify the influence of the bulk carriers. However, the metallic behavior indicates a shift of the Fermi level into the conduction band yielding a dominating bulk conductance and preventing a distinction of surface transport parameters employing the single-carrier model used here.

Figure 2. Electrical and magnetotransport measurements in Bi2Se3. (a) Temperature dependence of the Bi2Se3 channel resistance. (b) Magnetoconductance measurements of the Bi2Se3 with applied perpendicular magnetic field showing weak antilocalization up to 33 K. (c) Temperature dependence of phase coherence length lφ and (d) dimensionality factor |α|.

Magnetotransport Measurements in Bi2Se3. Magnetotransport measurements show weak antilocalization (WAL) behavior up to 33 K indicating a strong SO coupling in Bi2Se3 (Figure 2b and Supplementary Figure S3).25 From the WAL measurements we extract a dimensionality factor α of about −1.25 below 4 K and about −0.5 at higher temperatures (Figure 2d). The abrupt change at 4 K indicates a transition from a multichannel to a one 2D-channel conductance due to coupling of the coherent transport in the Bi2Se3 surface states at high temperature.26 This transition coincides with a constant phase coherence length lφ ≈ 400 nm at low temperatures and a lφ ∝ T−0.54 dependence for higher temperatures (Figure 2c). B

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 3. Electrical detection of SML in Bi2Se3 at room temperature. (a) Schematics of the magnetoresistance measurements between currentinduced spins (S⃗) on the Bi2Se3 surface and magnetization (M⃗ ) of the detector FM contact. A positive charge current is applied between contact I and III to create S↑, and a voltage signal is detected with a FM contact (II) against a reference contact (IV). (b) The FM contact (II) measures the difference in spin chemical potential of the channel (Δμ = μ↑ − μ↓ = −eΔV > 0), whereas M↑ detects S↓ and vice versa. (c) Reversing the current direction yields a change in spin orientation S↓ and (d) results in a spin potential configuration with Δμ = μ↑ − μ↓ = −eΔV < 0. (e) Spin signal of Dev1 according to the situation described in a and b showing a hysteretic switching at 300 K for a bias current of +50 μA. (f) Spin signal obtained for a negative current −50 μA as discussed in c and d. The arrows show the field sweep directions up (blue) and down (red). (g) Bias current dependence of the spin signal amplitude ΔV = (μ↑ − μ↓)/−e measured at 300 K with a linear fit (red line).

electrode. By sweeping the in-plane magnetic field we obtain a hysteretic spin signal. The switching field corresponds to the coercive field of our Co detector electrode (II), as verified by anisotropic magnetoresistance measurements (see Supplementary Figure S7). Reversing the current direction (−50 μA) locks the spins in the opposite direction resulting in an inverted hysteretic behavior of the measured signal (Figure 3f). We subtracted a linear background from the data stemming from the contribution of the charge electrochemical potential (see Supplementary Figure S4). A similar switching behavior has been measured at different bias currents at room temperature. We observe a point symmetry around zero bias and a linear dependence of the spin signal ΔV (Figure 3g). This is expected, since the spin polarization magnitude should be current independent and thus the spin density scales linearly with the current density.32 The measured spin valve signal is found to be very reproducible at different thicknesses and temperatures. Figure 4a shows the spin valve signals for Dev1 (40 nm Bi2Se3) measured at different temperatures with a bias current of 100 μA. The signals exhibit a clear switching and a signal amplitude of about 7 μV corresponding to a spin resistance of RS = 70 mΩ. Dev2 with 70 nm Bi2Se3 shows a similar spin valve signal with an amplitude of about 4 μV persistent up to room temperature, measured at a bias current of 1 mA (Figure 4b). Assuming the bulk contribution increases proportionally with the number of quintuple layers14 N and taking the different flake geometries of both devices into account, we observe a scaling of the spin-resistance-area product RSA = ΔV/I ∝ 1/N for both of our devices. The amplitudes of both signals are very weakly dependent on the temperature (Figure 4c). This is in agreement with previous spin torque11 and spin-charge conversion measurements,33 which also observed a spin signal with little decay up to room temperature. More recently,

Both observations indicate that our samples is a high SO coupling material with parallel surface and bulk conduction channels with high mobility.27,28 However, it has been proposed and demonstrated that the SML can still be probed with spin-sensitive FM tunnel contacts despite a large bulk contribution, since the scattering between bulk and surface states is suppressed.29,30 Room Temperature Electrical Detection of SpinMomentum Locking. Figure 3a−d shows the measurement principle of our multiterminal devices to detect SML on Bi2Se3 using Co/TiO2 FM tunnel contacts. The application of a positive electric bias between two contacts (I and III) results in a charge carrier flow with their spins locked perpendiculars to their momentum (Figure 3a). This results in a spin-polarized current with a spin potential splitting (μ↑ > μ↓) on the surfaces of Bi2Se3 (Figure 3b). The FM electrode II is used to probe the spin potentials in Bi2Se3 with respect to a reference contact (IV) placed outside the electric potential. A FM detects the spin potential in the TI with the same orientation as the majority spins in the FM, which is opposite to its magnetization direction. Consequently, if the spin polarization S⃗ on the Bi2Se3 surface is parallel to the FM magnetization M⃗ , a higher voltage can be detected than for the antiparallel configuration resulting in a voltage difference ΔV = (μ↑ − μ↓)/−e < 0 with the electron charge e (Figure 3b).16,31 The magnetization of the FM detector can be switched by an in-plane magnetic field, whereas the spin orientation in Bi2Se3 can be flipped by inverting the current direction (Figure 3c). This results in a reversed spin potential (μ↑ < μ↓) yielding a measured voltage difference ΔV > 0 (Figure 3d). Figure 3e shows the magnetic field dependence of the detected voltage signal at room temperature with a Bi2Se3 flake thickness of 40 nm (Dev1). At a fixed applied bias current of I = +50 μA, we observe a change in voltage signal with the change in magnetization direction of the FM detector C

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters

Figure 4. Temperature and thickness dependence of the spin signal. (a) Spin-valve measurements of Dev1 (40 nm Bi2Se3) at different temperatures for 100 μA bias current. (b) Spin-valve measurements of Dev2 (70 nm Bi2Se3) at different temperatures for 1 mA bias current. (c) Temperature dependence of the spin signal amplitude ΔV for two different thicknesses (Dev1 and Dev2) of Bi2Se3.

larger than commonly observed spin Hall effect in strong SO coupled materials, such as platinum (RSH ≈ 0.4 mΩ).30 Also, due to the short diffusion length in bulk Bi2Se3 the spin Hall effect contributes less than 1% to the surface polarization.13,30 Any contribution of spin injections from other FM electrodes should be suppressed by the SML in the surface and the large SO coupling in the bulk states. This is supported by the absence of multiple magnetoresistance switchings in a larger magnetic field sweep range (see Supplementary Figure S7). Furthermore, the linear bias dependence also rules out any heating related effects by the applied bias current, which would be independent of the sign.37 We also rule out any similar contribution from the anomalous Hall effect and other anisotropic ferromagnetic effects by using control devices with 5 nm nonmagnetic Ti layer at the interface together with Co electrodes (see Supplementary Figures S5 and S6). Finally, applying the magnetic field parallel to the current direction, i.e., the spins in the surface states of the TI perpendicular to the magnetization of the detector, we did not observe any step like switching or hysteresis (see Supplementary Figure S10).14,16 These control experiments suggest that the observed signals are due to the spin polarization on the TI surface, which is detected by the FM tunnel contacts. Spin Polarization. From the SML characteristics obtained in our devices, we can analytically evaluate the results to estimate the spin polarization in the Bi2Se3 surface states. The spin-valve signal ΔV, induced by a current IS = ηI, is directly related to the surface spin polarization PS:

potentiometric measurements demonstrated a similar weak temperature dependent behavior in a similar configuration (Edelstein effect) up to 125 K and its inverse setup (spingalvanic effect) up to 200 K.30 Such a behavior has been expected since ARPES measurements confirmed that the surface states are thermally stable and accessible up to room temperature.3,7 Contrasting previous studies, which observed a strong temperature dependence, attributed it to the preparation of the interface and the FM tunnel contacts.14,16 However, intrinsic transport parameters of the TI, including the temperature dependence of the charge carrier concentration and current distribution between surface and bulk, could significantly influence the behavior of the spin signal at low temperatures, which need to be established in further studies. Additionally, the in situ surface cleaning and preparation of the devices, as well as the measurement in a high vacuum system ensure the high quality of the interfaces and the FM tunnel contact, and prevent further degradation over time. Such high quality tunnel barriers with an appropriate resistance to overcome the conductivity mismatch provide a high sensitivity to detect spin accumulations at the interface.19,30 The localized measurement directly under the contact can detect the currentinduced spin polarization in spite of a large bulk channel contribution or counter propagating spin current at the bottom surface of the TI flake. This is in particular relevant at high temperatures, where the coherence length is significantly reduced.8 These results underline the reproducibility of the measurement and indicate that the signal originates from a spin polarization on the TI surface. Control Experiments. From above results and several control experiments we can rule out any artifacts in the spin signals. The observed polarity of the magnetoresistance could stem from the SML in the TSS or spin Hall effect, but rules out Rashba SO coupling as a dominating contribution, since it is expected to exhibit the opposite polarity.32,34,35 We can further rule out the spin Hall effect, since recent studies using spin transfer torque measurements on Bi2Se3 proposed that it is suppressed in the surface and bulk of the TI by the surface spin polarization.11,36 The amplitude of our observed magnetoresistance of up to RS = 70 mΩ is 1−2 orders of magnitude

ΔV = ISRBPSPFM

(1)

where PFM is the polarization of the FM detector and RB is the ballistic resistance.32 The former has been found in previous studies to have an upper limit of PFM = 20% for our Co/TiO2 contacts.19 The ballistic conductance σB = 1/RB can be calculated by the unit conductance q2/h times the number of propagating modes kFW/π, where W is the width of the conductance channel. The Fermi wavenumber kF can be derived from the charge carrier concentration as kF = (3π2n)1/3 for three-dimensional bulk conductance. However, the main interest lies in the spin polarized transport of the surface charge D

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters carrier concentration n2D, which is associated with kF = (4πn2D)1/2.14,38 This means, a better understanding of the conduction distribution between surface and bulk is necessary to analyze the experimental data. Introducing a correction factor η takes the ratio between TSS and bulk contribution into account, which is in particular necessary for highly doped TIs presenting a large conduction background in the bulk. So far, Shubnikov−de Haas oscillations reveal a surface contribution of about 0.003 at 0.3 K in Bi2Te3,39 whereas for Bi2Se3 a value of about 0.06 has been found.29,40−42 Consequently, solving eq 1 provides a lower limit for the surface polarization PS of 0.36 ± 0.03 and 0.27 ± 0.03 for Dev1 and Dev2, respectively. These values are comparable to previous studies which obtained polarizations between 0.2 and 0.4,7,14,35 but lower than 0.75 obtained in low temperature ARPES measurements.6 Furthermore, these results represent the spin polarization in the surface states themselves and are independent of the parallel bulk conduction channel. Nevertheless, several factors can yield a reduction of the observed SML signal, such as increased interstate scattering at high temperatures29 or the Rashba effect.14,32,35,43 Further studies are required to provide a better understanding of the charge carrier distribution between bulk and surface channel, the influence of other magnetotransport effects, as well as the limiting scattering mechanisms. In summary, we have demonstrated the room temperature electrical detection of spin polarized surface currents on thin exfoliated Bi2Se3 flakes by FM tunnel contacts. High-quality Bi2Se3 crystals and FM tunnel contacts allowed the observation of a hysteretic magnetoresistance signal reproducibly up to room temperature representing a magnetoresistance of up to 70 mΩ. The large magnitude of the signal, its sign, and several control experiments rule out any other possible effect and suggests the SML in the TSS as the origin. We expect that further control over the spin polarization can be achieved by accessing the dominant surface transport regime of the 3D TIs, for example, by increasing the surface-to-volume ratio, compensation doping, or electric gating.3,7 Positioning the Fermi level in the bandgap and controlling the surface carrier density will allow the electrical tuning of the spin polarization of the surface states.17 Our results will pave the way for using TIs as spin polarized sources for spintronic devices at ambient temperatures. The possibility of coupling TIs to other materials for spin injection19,44,45 opens up novel avenues in spintronic device design for energy efficient spin-logic applications. Methods. Fabrication. The Bi2Se3 flakes were exfoliated from a bulk crystal, using the conventional cleavage technique, onto a clean SiO2 (285 nm)/highly doped n-type Si substrate. The crystal was obtained from Miracrys, grown from a melt using a high vertical Bridgeman method (see Supplementary Figure S1). The flakes were identified using a combination of optical and atomic-force microscopy (Figure 1c). We used multilayer Bi2Se3 with a thickness in the range of 15−100 nm and widths of 1−5 μm. Electrodes were patterned by electron beam lithography. The contact deposition was performed in an ultrahigh vacuum electron beam evaporator after an in situ interface cleaning using low power argon ion milling for 10 s. Electrodes with widths of 0.3−1 μm and channel length of 0.2− 1 μm are used. As contact material we used Ti/Au for the devices for Hall and WAL measurements and TiO2/Co for the detection of spin-momentum locking. The ≈1.5 nm TiO2 tunnel barrier for the latter was deposited by electron beam

evaporation and in situ oxidation using a pure oxygen atmosphere. Measurement. The devices were measured with a Keithley 2400 sourcemeter using direct current (DC). Dev1 was measured in a liquid nitrogen cryostat between 75 and 300 K. Dev2 was measured in a liquid 4He cryostat with a temperature range of 1.5−300 K.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.5b03080. Crystallographic information on Bi2Se3, Hall measurement data, details on WAL measurement and background subtraction, as well as the data of several control experiments (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the support of colleagues at the Quantum Device Physics Laboratory and Nanofabrication Laboratory at Chalmers University of Technology. We are grateful for the supply of the crystals and the active discussion and support by Miracrys. We thank Dr. Tapati Sarkar for the XRD measurement, and Dr. Thilo Bauch, Dr. Ion Garate, and Dr. Ching-Tzu Chen for insightful discussions. This research is financially supported by the Nano Area of the Advance program at Chalmers University of Technology, EU FP7Marie Curie Career Integration grant, the Swedish Research Council (VR) Young Researchers Grant and an EMM Nano Consortium scholarship.

■

REFERENCES

(1) Xia, Y.; Qian, D.; Hsieh, D.; Wray, L.; Pal, A.; Lin, H.; Bansil, A.; Grauer, D.; Hor, Y. S.; Cava, R. J.; et al. Nat. Phys. 2009, 5, 398−402. (2) Ando, Y. J. Phys. Soc. Jpn. 2013, 82, 102001. (3) Hasan, M. Z.; Kane, C. Rev. Mod. Phys. 2010, 82, 3045−3065. (4) Pesin, D.; MacDonald, A. H. Nat. Mater. 2012, 11, 409−416. (5) Zhang, H.; Liu, C. X.; Qi, X. L.; Dai, X.; Fang, Z.; Zhang, S. C. Nat. Phys. 2009, 5, 438−442. (6) Pan, Z. H.; Vescovo, E.; Fedorov, a. V.; Gardner, D.; Lee, Y. S.; Chu, S.; Gu, G. D.; Valla, T. Phys. Rev. Lett. 2011, 106, 257004. (7) Hsieh, D.; Xia, Y.; Qian, D.; Wray, L.; Dil, J. H.; Meier, F.; Osterwalder, J.; Patthey, L.; Checkelsky, J. G.; Ong, N. P.; et al. Nature 2009, 460, 1101−1105. (8) Pan, Z. H.; Fedorov, a. V.; Gardner, D.; Lee, Y. S.; Chu, S.; Valla, T. Phys. Rev. Lett. 2012, 108, 187001. (9) König, M.; Wiedmann, S.; Brüne, C.; Roth, A.; Buhmann, H.; Molenkamp, L. W.; Qi, X. L.; Zhang, S. C. Science 2007, 318, 766− 770. (10) Brüne, C.; Roth, a.; Novik, E. G.; König, M.; Buhmann, H.; Hankiewicz, E. M.; Hanke, W.; Sinova, J.; Molenkamp, L. W. Nat. Phys. 2010, 6, 448−454. (11) Mellnik, A. R.; Lee, J. S.; Richardella, A.; Grab, J. L.; Mintun, P. J.; Fischer, M. H.; Vaezi, A.; Manchon, A.; Kim, E. A.; Samarth, N.; et al. Nature 2014, 511, 449−451.

E

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX

Letter

Nano Letters (12) Shiomi, Y.; Nomura, K.; Kajiwara, Y.; Eto, K.; Novak, M.; Segawa, K.; Ando, Y.; Saitoh, E. Phys. Rev. Lett. 2014, 113, 196601. (13) Deorani, P.; Son, J.; Banerjee, K.; Koirala, N.; Brahlek, M.; Oh, S.; Yang, H. Phys. Rev. B: Condens. Matter Mater. Phys. 2014, 90, 094403. (14) Li, C.; van’t Erve, O.; Robinson, J.; Liu, Y.; Li, L.; Jonker, B. T. Nat. Nanotechnol. 2014, 9, 218−224. (15) Tian, J.; Childres, I.; Cao, H.; Shen, T.; Miotkowski, I.; Chen, Y. P. Solid State Commun. 2014, 191, 1−5. (16) Tang, J.; Chang, L. T. L.; Kou, X.; Murata, K.; Choi, E. S.; Lang, M.; Fan, Y.; Jiang, Y.; Montazeri, M.; Jiang, W.; et al. Nano Lett. 2014, 14, 5423−5429. (17) Burkov, A. A.; Hawthorn, D. G. Phys. Rev. Lett. 2010, 105, 066802. (18) Dankert, A.; Dulal, R. S.; Dash, S. P. Sci. Rep. 2013, 3 (3196), 1−8. (19) Kamalakar, V. M.; Groenveld, C.; Dankert, A.; Dash, S. P. Nat. Commun. 2015, 6, 6766. (20) Dankert, A.; Langouche, L.; Kamalakar, V. M.; Dash, S. P. ACS Nano 2014, 8, 476−482. (21) Fert, A. Rev. Mod. Phys. 2008, 80, 1517−1530. (22) Dash, S. P.; Sharma, S.; Patel, R. S.; de Jong, M. P.; Jansen, R. Nature 2009, 462, 491−494. (23) Dankert, A.; Kamalakar, M. V.; Bergsten, J.; Dash, S. P. Appl. Phys. Lett. 2014, 104, 192403. (24) Jönsson-Åkerman, B. J.; Escudero, R.; Leighton, C.; Kim, S.; Schuller, I. K.; Rabson, D. A. Appl. Phys. Lett. 2000, 77, 1870. (25) Hikami, S.; Larkin, A.; Nagaoka, Y. Prog. Theor. Phys. 1980, 63, 707−710. (26) Kim, D.; Syers, P.; Butch, N. P.; Paglione, J.; Fuhrer, M. S. Nat. Commun. 2013, 4 (2040), 1−5. (27) He, H. T.; Wang, G.; Zhang, T.; Sou, I. k.; Wong, G. K. L.; Wang, J. n.; Lu, H. Z.; Shen, S. Q.; Zhang, F. C. Phys. Rev. Lett. 2011, 106, 166805. (28) Altshuler, B. L.; Aronov, A. G.; Khmelnitsky, D. E. J. Phys. C: Solid State Phys. 1982, 15, 7367−7386. (29) Qu, F.; Zhang, C.; Du, R. R.; Lu, L. J. Low Temp. Phys. 2013, 170, 397−402. (30) Liu, L.; Richardella, A.; Garate, I.; Zhu, Y.; Samarth, N.; Chen, C. T. Phys. Rev. B: Condens. Matter Mater. Phys. 2015, 91, 235437. (31) Tian, J.; Miotkowski, I.; Hong, S.; Chen, Y. P. Sci. Rep. 2015, 5, 14293. (32) Hong, S.; Diep, V.; Chen, Y. P.; Datta, S. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 085131. (33) Jamali, M.; Lee, J.; Lv, Y.; Zhao, Z. arXiv:1407.7940, 2014. (34) Hammar, P.; Johnson, M. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61, 7207−7210. (35) Yazyev, O. V.; Moore, J. E.; Louie, S. G. Phys. Rev. Lett. 2010, 105, 266806. (36) Wang, Y. Phys. Rev. Lett. 2015, 114, 257202. (37) Dankert, A.; Dash, S. P. Appl. Phys. Lett. 2013, 103, 242405. (38) Lee, J. S.; Richardella, A.; Hickey, D. R.; Mkhoyan, K. A.; Samarth, N. arXiv:1507.07063, 2015. (39) Qu, D.; Hor, Y.; Xiong, J.; Cava, R.; Ong, N. Science 2010, 329, 821−824. (40) Cao, H.; Tian, J.; Miotkowski, I.; Shen, T.; Hu, J.; Qiao, S.; Chen, Y. P. Phys. Rev. Lett. 2012, 108, 216803. (41) Petrushevsky, M.; Lahoud, E.; Ron, A.; Maniv, E.; Diamant, I.; Neder, I.; Wiedmann, S.; Guduru, V. K.; Chiappini, F.; Zeitler, U.; et al. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 045131. (42) Taskin, A. A.; Sasaki, S.; Segawa, K.; Ando, Y. Phys. Rev. Lett. 2012, 109, 066803. (43) Liu, L.; Chen, C. t.; Sun, J.; Hu, G.; Worledge, D. Nat. Phys. 2014, 10, 561. (44) Kamalakar, M. V.; Dankert, A.; Bergsten, J.; Ive, T.; Dash, S. P. Sci. Rep. 2014, 4, 6146. (45) Modak, S.; Sengupta, K.; Sen, D. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 205114.

F

DOI: 10.1021/acs.nanolett.5b03080 Nano Lett. XXXX, XXX, XXX−XXX