betaFeedback

WorksPhysics and Astronomy

Rashba spin-orbit coupling enhanced magnetoresistance in junctions with one ferromagnet

Abstract

We explain how Rashba spin-orbit coupling (SOC) in a two-dimensional electron gas (2DEG), or in a conventional ss-wave superconductor, can lead to a large magnetoresistance even with one ferromagnet. However, such enhanced magnetoresistance is not generic and can be nonmonotonic and change its sign with Rashba SOC. For an in-plane rotation of magnetization, it is typically negligibly small for a 2DEG and depends on the perfect transmission which emerges from a spin-parity-time symmetry of the scattering states, while this symmetry is generally absent from the Hamiltonian of the system. The key difference from considering the normal-state magnetoresistance is the presence of the spin-dependent Andreev reflection at superconducting interfaces. In the fabricated junctions of quasi-2D van der Waals ferromagnets with conventional ss-wave superconductors (Fe0.29_{0.29}TaS2_2/NbN) we find another example of enhanced magnetoresistance where the presence of Rashba SOC reduces the effective interfacial strength and is responsible for an equal-spin Andreev reflection. The observed nonmonotonic trend in the out-of-plane magnetoresistance with the interfacial barrier is an evidence for the proximity-induced equal-spin-triplet superconductivity.

Introduction

The magnetoresistance (MR) is a key figure of merit in spintronics and its enhancement is associated with the major advances in magnetically sensing and storing information [1, 2, 3, 4, 5]. Typically, a large MR is sought in structures with multiple ferromagnetic regions, where the resulting spin-valve effect implies that the resistance of the systems depends of the relative magnetization, M\mathbf{M} orientation of those ferromagnets.

However, as first discovered in 1857 by Lord Kelvin [6], anisotropic MR (AMR) shows that with spin-orbit coupling (SOC) there is a change of the electrical resistivity with the relative direction of the charge current with respect to M\mathbf{M} of a single bulk ferromagnet (F), such as Ni or Fe. Another example of MR with a single F region is the tunneling AMR (TAMR) [7, 8, 9, 10], also a manifestation of the interplay between SOC and M\mathbf{M}. Unfortunately, while a single F region simplify scaled-down devices both AMR and TAMR are limited by their small magnitudes (typically <1%<1\% for in-plane M\mathbf{M} rotation) [10, 11].

In this work we explore a possibility for a much larger MR with a single F region and Bychkov-Rashba, also known as Rashba SOC [12], in both normal and superconducting state. The resulting enhanced MR is not only significant for the potential spintronic applications, but could also distinguish between the trivial and topological states [11], or provide a signature of equal-spin-triplet superconductivity, sought to realize coexistence of ferromagnetism and superconductivity, dissipationless spin currents, and Majorana bound states for fault-tolerant topological quantum computing [13, 14, 15, 16, 17, 18, 19, 20, 21, 22].

To realize an enhanced normal-state MR, we consider

a lateral geometry from Figure 1(a). The resulting TAMR is not determined by highly spin-polarized carriers or a large exchange energy in the F region, but by the high interfacial transmission from the interplay of the Rashba SOC, the barrier strength at the F/2DEG interface, and the proximity-induced exchange field, which can reach tens of meV and extend over tens of nm [23, 24, 25, 26, 27].

Lateral geometry and symmetry operations in a junction with one ferromagnet

Figure 1. (a) Lateral geometry, the current flows in the 2DEG. (b) Action of the time reversal, T\mathcal{T}, and the space inversion operator, P\mathcal{P}, on the incident wave. By applying the PT\mathcal{PT} operator, the incident wave on the left is transformed into a transmitted wave on the right. For a spinless barrier system, the PT\mathcal{PT} symmetry gives the perfect transmission. (c) Action of the PσzT\mathcal{P}\sigma_z\mathcal{T} operator and the PσzT\mathcal{P}\sigma_z\mathcal{T} symmetry generalizes the condition from (b) for an incident wave with an in-plane spin and a barrier with the in-plane magnetization M\mathbf{M}, defined by the polar angle ϕ\phi.

For a simple spinless case of a square barrier with thickness dd and height V0V_0, a standard expression for transmission, TT, with energy, E>V0E>V_0, is [28, 29]

T=(1+V02sin⁡2(kd)4E(E−V0))−1,(1)T = \left(1 + \frac{V_0^2\sin^2(kd)}{4E(E-V_0)}\right)^{-1}, \tag*{(1)}

where k=2m(E−V0)/ℏk = \sqrt{2m(E-V_0)}/\hbar is the wave vector. The perfect transmission, T=1T = 1, requires (i) V0=0V_0 = 0 or (ii) kd=nπkd = n\pi, where n=1,2,3,…n = 1, 2, 3, \ldots, gives the resonance condition. This well-known behavior [28, 29] is usually not connected to the parity-time symmetry PT\mathcal{PT}, depicted in Figure 1(b), which satisfies both cases (i) and (ii). Such a symmetry, where P\mathcal{P} is the parity operator, T=K\mathcal{T} = \mathcal{K} is the time-reversal operator, and K\mathcal{K} is the complex conjugation operator, which ensures that the incoming and transmitted spinless wave are identical, up to a phase.

However, generalization of the perfect transmission for spin-1/2 carriers, where T=−iσyK\mathcal{T} = -i\sigma_y\mathcal{K}, and σy\sigma_y is the Pauli matrix, is much less explored with SOC and magnetic barriers. It was recently found that the perfect transmission emerges when the eigenstates of the F/2DEG Hamiltonian, HF/2DEGH_{\mathrm{F/2DEG}}, satisfy PσzT\mathcal{P}\sigma_z\mathcal{T}, the spin-parity-time symmetry, where [PσzT,HF/2DEG]≠0[\mathcal{P}\sigma_z\mathcal{T}, H_{\mathrm{F/2DEG}}] \ne0 [11]. Intuitively, the emerging perfect transmission for the eigenstates of HF/2DEGH_{\mathrm{F/2DEG}} which satisfy PσzT\mathcal{P}\sigma_z\mathcal{T} can be understood from Figure 1(c). T\mathcal{T} reverses the spin and motion of the incident wave, while Pσz\mathcal{P}\sigma_z inverts both the spin (through σz\sigma_z) and position (through P\mathcal{P}) of the wave. By applying the PσzT\mathcal{P}\sigma_z\mathcal{T} operator the incident wave on the left is transformed to itself, but as a transmitted wave on the right. Therefore, scattering states which are eigenfunctions of PσzT\mathcal{P}\sigma_z\mathcal{T} experience perfect transmission. The resulting in-plane TAMR amplitude

TAMR∥=G(ϕ=0)−G(ϕ=π/2)G(ϕ=π/2),(2)\mathrm{TAMR}_{\parallel} = \frac{G(\phi= 0)-G(\phi= \pi/2)}{G(\phi= \pi/2)}, \tag*{(2)}

where the ϕ\phi-dependent conductance GG is the inverse of the resistance, G=1/RG = 1/R, with ϕ\phi defined in Figure 1(c), can be enhanced by one or two orders of magnitude.

To study an enhanced out-of-plane MR in the superconducting state, we focus on the recent transport experiments in junctions of quasi-2D van der Waals F with conventional ss-wave superconductors (S) (Fe0.29_{0.29}TaS2_2/NbN) [30]. Compared to the normal state MR, the key difference is the presence of Andreev reflection at the F/S interface with Rashba SOC. In this process an incoming electron is reflected as a hole and the Cooper pair enters the S region. With SOC, in addition to the conventional Andreev reflection, where the incoming and reflected particle have an opposite spin, a spin-flip or equal-spin Andreev reflection is also possible in which the incoming and reflected particles have the same spin [31, 32]. While the first process is responsible for the spin-singlet proximity-induced superconductivity, the second yields the spin-triplet counterpart.

The corresponding anisotropic behavior, magnetoan isotropic Andreev reflection (MAAR) can be viewed as the generalization of TAMR [32, 33, 34], which is recovered in the normal state of the F/S junction for a large bias, VV, applied magnetic field, BB, or a high temperature. From a full MAAR calculation for the F/S junction, such a TAMR value can be obtained by taking the

superconducting gap to vanish. Similar to Eq. (2), the out-of-plane MAAR amplitude in F/S junctions is

MAAR⊥=G(θ=0)−G(θ=π/2)G(θ=π/2),(3)\mathrm{MAAR}_{\perp} = \frac{G(\theta= 0)-G(\theta= \pi/2)}{G(\theta= \pi/2)}, \tag*{(3)}

where θ\theta is the angle defined in Figure 2. As we later show, F/S measurements for the MAAR can be orders of magnitude larger than TAMR in the normal state. This enhancement is dominated by the spin-flip Andreev reflection with Rashba SOC at the F/S interface.

Figure 2. Schematic of the geometry including a quasi-2D vdW F/S junction Fe0.29_{0.29}TaS2_2/NbN, with an Al2_2O3_3 insulating barrier. MR measurements are performed for an out-of-plane M rotation, defined by angle θ\theta with respect to the applied magnetic field, B\mathbf{B}. The interfacial resistance is obtained as V/IV/I.

Despite the different geometries for the considered F/2DEG and F/S structures, which also differ by the in-plane vs out-of-plane rotation of M, we find important common features and surprising nonmonotonic trends in TAMR in both systems. Some of these identified trends arise from the role of Rashba SOC which modifies the effective barrier strength and determines the condition for perfect transmission or an enhanced contribution of the spin-flip Andreev reflection.

Following this Introduction, we explore the structures from Figure 1(a). In Sec. II we describe the F/2DEG Hamiltonian, consider the influence of SOC on the scattering states and analyze the conductance which reveals different resonant-like behavior, peaked at different parameters. The calculation of in-plane TAMR in Sec. III reveals different trends as a function of the barrier and SOC strengths as well as the proximity-induced exchange splitting. In Sec. IV we examine the superconducting structures from Figure 2 and describe the corresponding F/S Hamiltonian, followed by the measured and calculated conductance using a simple F/S Hamiltonian. In Sec. V we describe how the predicted influence of the Rashba SOC on the effective barrier strength provides a good description of the measured out-of-plane MR and confirms the evidence for the equal-spin-triplet superconductivity in the considered heterostructures. In Sec. VI we provide additional discussion for the relevance of enenhanced MR in the normal and superconducting state and note some open questions for future work.

F/2DEG HAMILTONIAN AND CONDUCTANCE ANALYSIS

To explore the interplay of the proximity-modified 2DEG and SOC, together with the magnetic anisotropy of the transport properties, we consider a model Hamiltonian of the system represented in Figure 1(a) given by

HF/2DEG=p22m∗+αℏ(σ×p)⋅z^+[V0−Δxc(m⋅σ)]h(x),(4)H_{\mathrm{F/2DEG}} = \frac{p^2}{2m^*}+\frac{\alpha}{\hbar}(\boldsymbol{\sigma}\times\mathbf{p})\cdot\hat{\mathbf{z}}+[V_0-\Delta_{\mathrm{xc}}(\mathbf{m}\cdot\boldsymbol{\sigma})]h(x), \tag*{(4)}

where m∗m^* is the effective mass, α\alpha is the Rashba SOC strength, z^\hat{\mathbf{z}} is the unit vector along the zz-axis, p=(px,py)\mathbf{p}=(p_x,p_y) is the 2D momentum operator, σ\boldsymbol{\sigma} is the vector of Pauli matrices, V0V_0 describes the potential barrier, Δxc\Delta_{\mathrm{xc}} and m\mathbf{m} are the magnitude and direction of the proximity-induced ferromagnetic exchange field. The function h(x)=Θ(d/2+x)Θ(d/2−x)h(x)=\Theta(d/2+x)\Theta(d/2-x) describes a square barrier of thickness dd. We focus on electrons, not holes [35, 36, 37, 38], with the effective barrier region and typical band structure shown in Figure 3. In Figure 3(a) the resulting spin-dependent barrier is shown in a weak SOC limit, αkF≪Δxc\alpha k_{\mathrm{F}}\ll\Delta_{\mathrm{xc}}, where kF=2m∗EF/ℏk_{\mathrm{F}}=\sqrt{2m^*E_{\mathrm{F}}/\hbar} is the Fermi wave vector averaged over the inner and outer contours in the leads, while EFE_{\mathrm{F}} is the Fermi energy. The blue (yellow) contours in Figure 3(c) denote lower (upper) bands. Inside the barrier the upper band is irrelevant since its bottom is above EFE_{\mathrm{F}}. The spin orientations are marked by arrows.

Schematic of the spin-dependent barrier and corresponding Fermi contours

Figure 3. (a) Schematic of the spin-dependent barrier. Its effective strength is modified by the proximity-induced exchange field Δxc\Delta_{\mathrm{xc}}, shown in a weak SOC limit. (b) 2DEG band structure with the Fermi energy EFE_{\mathrm{F}} and the effective barrier region (middle) of height V0V_0 and Δxc\Delta_{\mathrm{xc}}. (c) Corresponding Fermi contours, arrows denote the spin orientation. Dashed lines: the range of a conserved wave vector kyk_y in the scattering states. For incident angles exceeding φc\varphi_c backscattering is suppressed.

Due to Rashba SOC, the wave functions can be classified by the helicity index, where λ=1\lambda=1 (−1-1) refers to the inner (outer) Fermi contour as depicted in Figure 3(c). The scattering states for the finite square barrier model can be written as ψλ(x,y)=(1/2A)eikyyϕλ(x)\psi_\lambda(x,y)=(1/\sqrt{2A})e^{ik_y y}\phi_\lambda(x), with sample area AA and the conserved parallel component of the wave vector kyk_y in the ballistic transport. Right from the barrier, ϕλ(x)\phi_\lambda(x) is a combination of the two plane waves with transmission coefficient, tλλt_{\lambda\lambda} and tλλˉt_{\lambda\bar{\lambda}}, where λˉ=−λ\bar{\lambda}=-\lambda, to describe intraband and interband scattering processes. By matching ϕλ(x)\phi_\lambda(x) and dϕλ(x)/dxd\phi_\lambda(x)/dx in the regions, x<−d/2x<-d/2, −d/2<x<d/2-d/2<x<d/2, and x>d/2x>d/2, we obtain tλλt_{\lambda\lambda}, tλλˉt_{\lambda\bar{\lambda}}, and the particle current density of the λ\lambda channel

jλ=Re⁡[vA(∣tλλ∣2cos⁡φλ+∣tλλˉ∣2cos⁡φλˉ)],(5)j_\lambda=\operatorname{Re}\left[\frac{v}{A}\left(|t_{\lambda\lambda}|^2\cos\varphi_\lambda+|t_{\lambda\bar{\lambda}}|^2\cos\varphi_{\bar{\lambda}}\right)\right], \tag*{(5)}

here, the group velocity of the scattered particle, v=(α/ℏ)2+2E/m∗v=\sqrt{(\alpha/\hbar)^2+2E/m^*}, has the same magnitude for the two bands. This current contains contributions from the intra- and inter-channel transmission, where φλ\varphi_\lambda is the incident and φλˉ\varphi_{\bar{\lambda}} the propagation angle of the cross-channel wave with the conservation of the kyk_y component and φλˉ\varphi_{\bar{\lambda}}

is related to φλ\varphi_\lambda by cos⁡φλˉ=kλ2−ky2sin⁡2φλ/kλˉ\cos\varphi_{\bar{\lambda}}=\sqrt{k_\lambda^2-k_y^2\sin^2\varphi_\lambda}/k_{\bar{\lambda}}, where

kλ=(αm∗ℏ2)2+2m∗Eℏ2−λαm∗ℏ2.(6)k_\lambda=\sqrt{\left(\frac{\alpha m^*}{\hbar^2}\right)^2+\frac{2m^*E}{\hbar^2}}-\frac{\lambda\alpha m^*}{\hbar^2}. \tag*{(6)}

In the low-bias limit, i.e. ∣eV∣≪EF|eV|\ll E_{\mathrm{F}}, we get the expression for the conductance GλG_\lambda in the λ\lambda channel [39]

Gλ=e2hD2π∫−π2π2dφλkFλTλcos⁡φλ,(7)G_\lambda=\frac{e^2}{h}\frac{D}{2\pi}\int_{-\frac{\pi}{2}}^{\frac{\pi}{2}}d\varphi_\lambda k_{\mathrm{F}}^\lambda T_\lambda\cos\varphi_\lambda, \tag*{(7)}

where DD is the sample width, kFλk_{\mathrm{F}}^\lambda the λ\lambda-channel Fermi wave vector [E=EF[E=E_{\mathrm{F}} in Eq. (6)] and the transmission is

Tλ=Re⁡[∣tλλ∣2+∣tλλˉ∣2(cos⁡φλˉcos⁡φλ)].(8)T_\lambda=\operatorname{Re}\left[|t_{\lambda\lambda}|^2+|t_{\lambda\bar{\lambda}}|^2\left(\frac{\cos\varphi_{\bar{\lambda}}}{\cos\varphi_\lambda}\right)\right]. \tag*{(8)}

The total conductance is the sum of the two channels,

G=∑λ=±1Gλ.(9)G=\sum_{\lambda=\pm1}G_\lambda. \tag*{(9)}

To explore the evolution of the conductance as a function of the proximity-induced field and its direction, we consider its normalized value in Eq. (9) expressed in terms of the Sharvin conductance for the 2DEG system

G02D=e2h2Dπ(αm∗ℏ2)2+2m∗EFℏ2.(10)G_0^{\mathrm{2D}}=\frac{e^2}{h}\frac{2D}{\pi}\sqrt{\left(\frac{\alpha m^*}{\hbar^2}\right)^2+\frac{2m^*E_{\mathrm{F}}}{\hbar^2}}. \tag*{(10)}

As shown in Figure 1(c), the PσzTP\sigma_z\mathcal{T} symmetry leads to perfect transmission and influences the conductance for the 2DEG system. Specifically, scattering states which satisfy PσzTψ(x,y)=ξψ(x,y)P_{\sigma_z}\mathcal{T}\psi(x,y)=\xi\psi(x,y), with eigenvalues of the form ξ=eiη\xi=e^{i\eta}, include the perfect transmission either due to effectively vanishing barrier or the resonances expected from the standing-wave condition. However, PσzTP_{\sigma_z}\mathcal{T} symmetry is not an intrinsic symmetry of the system and PσzTP_{\sigma_z}\mathcal{T} does not commute with HF/2DEGH_{\mathrm{F/2DEG}} in Eq. (4). Instead, the PσzTP_{\sigma_z}\mathcal{T} symmetry emerges only for certain specific system parameters and scattering states satisfying,

[HF/2DEG,PσzT]ψR(x,y)=0,(11)[H_{\mathrm{F/2DEG}},P_{\sigma_z}\mathcal{T}]\psi_R(x,y)=0, \tag*{(11)}

where the index RR emphasizes that the relation is restricted only to the cases of the perfect transmission (for vanishing of the effective barrier or at resonances). This symmetry generalizes a simple case of the perfect transmission in a potential barrier (or a spinless) system [28].

To understand the relation between conductance, barrier parameters, and the perfect transmission, we perform a Fermi contour analysis illustrated in Figure 4. In the spinless case, with the better matching between Fermi contours in the lead and barrier a higher transmission is expected. For a simple 1D geometry (normal incidence or a fixed kyk_y) and δ\delta-barrier, the transmission is [40]

Normalized conductance and Fermi contours

Figure 4. Normalized conductance, GG, as a function of Δxc\Delta_{\mathrm{xc}}, for EF=5 meVE_F = 5\,\mathrm{meV}, V0=10 meVV_0 = 10\,\mathrm{meV}, α=0.093 eVA˚\alpha= 0.093\,\mathrm{eV}\mathring{\mathrm{A}}, and d=26 nmd = 26\,\mathrm{nm}. Fermi contours at different Δxc\Delta_{\mathrm{xc}}, m∥x\mathbf{m}\parallel x: Blue and red circles are in the lead and the dashed contours in the barrier (only one band exists at the given EE). Each of the three peaks in GG satisfies the PσzTP_{\sigma_z}\mathcal{T} symmetry for the perfect transmission.

T=11+Z2,(12)T = \frac{1}{1+Z^2}, \tag*{(12)}

where ZZ is the effective barrier strength which combines the influence of a native barrier and the Fermi wave vectors mismatch in the two regions [31, 33, 41, 42]. Since the difference between the lead and barrier Fermi contours is associated with the effective barrier potential, a larger mismatch between Fermi contours corresponds to a larger effective ZZ and thus to a lower transmission.

We now generalize this situation to a spin-dependent case in the presence of Δxc\Delta_{\mathrm{xc}} and α\alpha, as depicted in Figure 3(b) and (Figure 3c). The effective barrier at the F/2DEG interface becomes spin-dependent since the energy bands in the barrier are split by the exchange field and spinless ZZ becomes Z±effZ_{\pm}^{\mathrm{eff}} [33]. For the situation from Figure 3, we can decompose the incident spinor in the basis constructed by the corresponding barrier eigenspinor (kx,yk_{x,y} is the same as in the incident state) for the lower band χ↑\chi_{\uparrow} and its antiparallel partner from the upper band χ↓\chi_{\downarrow}, i.e., ∣χin⟩=⟨χ↑∣χin⟩∣χ↑⟩+⟨χ↓∣χin⟩∣χ↓⟩|\chi_{\mathrm{in}}\rangle=\langle\chi_{\uparrow}|\chi_{\mathrm{in}}\rangle|\chi_{\uparrow}\rangle+\langle\chi_{\downarrow}|\chi_{\mathrm{in}}\rangle|\chi_{\downarrow}\rangle. The first term undergoes a weak effective barrier Z−effZ_{-}^{\mathrm{eff}} while the second term experiences a strong barrier Z+effZ_{+}^{\mathrm{eff}}.

The effective interfacial barrier is inequivalent for two helicities (for outer/inner Fermi contours) leading to an important influence of spin mismatch on transmission, not just the mismatch of the Fermi contours [33]. Considering that the direction of the proximity-induced exchange field, m\mathbf{m} [see Eq. (4)] is determined by ϕ\phi, for an in-plane M\mathbf{M} as depicted in Figure 1(c), the effective barrier for ∣χin⟩|\chi_{\mathrm{in}}\rangle from the outer band of the lead is

Z±eff∝V0∓(αkx+Δxcsin⁡ϕ)2+(αky−Δxccos⁡ϕ)2−αk.(13)Z_{\pm}^{\mathrm{eff}}\propto V_0\mp\sqrt{(\alpha k_x+\Delta_{\mathrm{xc}}\sin\phi)^2+(\alpha k_y-\Delta_{\mathrm{xc}}\cos\phi)^2}-\alpha k. \tag*{(13)}

Equivalently, for ∣χin⟩|\chi_{\mathrm{in}}\rangle from the inner band of the lead is

Z±eff∝V0∓(αkx+Δxcsin⁡ϕ)2+(αky−Δxccos⁡ϕ)2+αk.(14)Z_{\pm}^{\mathrm{eff}}\propto V_0\mp\sqrt{(\alpha k_x+\Delta_{\mathrm{xc}}\sin\phi)^2+(\alpha k_y-\Delta_{\mathrm{xc}}\cos\phi)^2}+\alpha k. \tag*{(14)}

In Eqs. (13) and (14) the impact of the reflected states is neglected, and this treatment is most accurate for small effective ZZ, i.e., for a good matching of Fermi contours. We note that the effective barrier becomes energy-dependent when Δxc≠0\Delta_{\mathrm{xc}}\ne0 and α≠0\alpha\ne0, opening the possibility for resonant tunneling to occur when the energy of the incident carrier is such that Z−eff→0Z_{-}^{\mathrm{eff}}\to0 (or Z+eff→0Z_{+}^{\mathrm{eff}}\to0).

Due to the spin mismatch, we need to include a correction of ∣⟨χ↑∣χin⟩∣2|\langle\chi_{\uparrow}|\chi_{\mathrm{in}}\rangle|^2 in the transmission, i.e., T≈∣⟨χ↑∣χin⟩∣2Tχ↑T\approx|\langle\chi_{\uparrow}|\chi_{\mathrm{in}}\rangle|^2T_{\chi_{\uparrow}}, where Tχ↑T_{\chi_{\uparrow}} is the transmission without the spin mismatch. We can see from Figure 1(c) that the spin mismatch is much smaller for the incident state from the outer lead band (λ=−1\lambda=-1), i.e., ∣⟨χ↑∣χin,λ=−1⟩∣2≫∣⟨χ↑∣χin,λ=1⟩∣2|\langle\chi_{\uparrow}|\chi_{\mathrm{in},\lambda=-1}\rangle|^2\gg|\langle\chi_{\uparrow}|\chi_{\mathrm{in},\lambda=1}\rangle|^2, which indicates that the maximum of GG is contributed by the incident states from the outer lead band. When the state inside the barrier is the same (up to a phase) as that in the lead, perfect transmission is achieved.

With vanishing Rashba SOC, α→0\alpha\to0, outer and inner lead Fermi contours coincide and the spin-dependent barrier, V±=V0±ΔxcV_{\pm}=V_0\pm\Delta_{\mathrm{xc}}, from Figure 3(a) is recovered. We focus on V0>0V_0>0, V−=0V_{-}=0 leads to the perfect transmission, Tχ↑=1T_{\chi_{\uparrow}}=1, from our discussion above. From Eqs. (13) and (14), Z±eff∝V0±ΔxcZ_{\pm}^{\mathrm{eff}}\propto V_0\pm\Delta_{\mathrm{xc}}, so one can still get Z−eff→0Z_{-}^{\mathrm{eff}}\to0 when V0=ΔxcV_0=\Delta_{\mathrm{xc}}. We will further address the importance of vanishing V−V_{-} in F/S junctions. At a fixed kyk_y, the eigenstate for the perfect transmission T(ky)=1T(k_y)=1, precisely satisfies the PσzTP_{\sigma_z}\mathcal{T} symmetry: PσzTψ(x,y)=ξψ(x,y)P_{\sigma_z}\mathcal{T}\psi(x,y)=\xi\psi(x,y), with eigenvalue ξ=eiη\xi=e^{i\eta} and an arbitrary phase η\eta.

This analysis and the relevance of the PσzTP_{\sigma_z}\mathcal{T} symmetry can be applied to Figure 4 with parameters of a typical InGaAs/InAlAs 2DEG, m∗=0.05m0m^{*}=0.05m_{0}, where m0m_{0} is the free-electron mass, and α=0.093 eVA˚\alpha=0.093\ \mathrm{eV}\mathring{\mathrm{A}} [43]. It is crucial to note that PσzTP\sigma_{z}T symmetry arguments for a perfect transmission pertain to a single kyk_{y} (single channel), while the geometry from Figure 1(a) and Figure 3 corresponds to a 2D system where the total GG reflects the contribution from all allowed kyk_{y}. In the spinless case, maximizing GG means an overall contour matching, while even with adding spin our Fermi contour analysis provide a valuable tool to visualize and understand various trends in GG.

With conserved kyk_{y}, from Figure 3(c) we can recognize many similarities with the Snell’s law [42]. We may get propagating or decaying (evanescent) states in different regions, depending on the incident angle, φ\varphi and the particular contour (outer, inner) considered. For example, for carriers from the λ=−1\lambda=-1 band (outer contour) and incident angles φ\varphi exceeding the critical angle,

φc=±arcsin⁡(k1/k−1),(15)\varphi_{c}=\pm\arcsin\left(k_{1}/k_{-1}\right), \tag*{(15)}

where kλ=±1k_{\lambda}=\pm1 is given by Eq. (6), the transmission and reflection to the λ=1\lambda=1 band are not allowed because there are no such propagating states. In this regime, back scattering is suppressed while Tλ=−1T_{\lambda=-1} is enhanced.

The overall maximum in G(Δxc)G(\Delta_{\mathrm{xc}}) is related to the perfect transmission due to the Fermi contour matching depicted by the example (2) in Figure 4. As shown for examples (1) and (3) a small change in Δxc\Delta_{\mathrm{xc}} leads to worse Fermi contour matching and reduced GG. The origin of the overall peak in GG near Δxc=V0\Delta_{\mathrm{xc}}=V_{0} can be further understood in the small SOC limit. The barrier contour reduces to a shifted circle with radius kF(EF+Δ−V0)/EFk_{\mathrm{F}}\sqrt{(E_{\mathrm{F}}+\Delta-V_{0})/E_{\mathrm{F}}}, where we recall that kFk_{\mathrm{F}} is average of the outer and inner lead circles. The barrier contour can then be approximated by

[kx−(m∗α/ℏ2)sin⁡ϕ]2+[ky+(m∗α/ℏ2)cos⁡ϕ]2≈(2m∗/ℏ2)(EF+Δ−V0).(16)\left[k_{x}-\left(m^{*}\alpha/\hbar^{2}\right)\sin\phi\right]^{2}+\left[k_{y}+\left(m^{*}\alpha/\hbar^{2}\right)\cos\phi\right]^{2} \approx\left(2m^{*}/\hbar^{2}\right)\left(E_{\mathrm{F}}+\Delta-V_{0}\right). \tag*{(16)}

In the region near the Δxc=V0\Delta_{\mathrm{xc}}=V_{0}, both G(ϕ=0)G(\phi=0) and G(ϕ=π/2)G(\phi=\pi/2), which correspond to m∥x\mathbf{m}\parallel x and m∥y\mathbf{m}\parallel y, reach their maxima because of the best matching of Fermi contours between lead and barrier. The shift of barrier circle is always ⊥\perp to m\mathbf{m} and is of the first order in α\alpha, while the deformation is a higher-order correction.

Remarkably, the other local maxima in GG, for larger Δxc\Delta_{\mathrm{xc}} in Figure 4, also satisfy the PσzTP\sigma_{z}T symmetry. However, instead of the contour matching, they are due to standing de Broglie waves in the barrier and the constructive interference. With outer/inner lead contours, there are four eigenstates in the leads with ±kxλ,±kxλ~\pm k_{x\lambda},\pm k_{x\tilde{\lambda}}, we can distinguish cases: (i) 4 (4), (ii) 4 (2), (iii) 2 (4), and (iv) 2 (2) propagating states in the leads (barrier) [11].

In cases (i) and (ii) [arising for an incident state from the inner band of for the outer band and φ<φc\varphi<\varphi_{c}], the perfect transmission means that no reflected wave should exist in the left lead, while the state in the right lead should match, up to a phase, the corresponding one in the left lead. For case (i), under PσzTP\sigma_{z}T symmetry operation, up to a phase, all the propagating states inside the barrier, with xx component of the wave vector, k~j\tilde{k}_{j}, j=1,…,4j=1,\ldots,4, remain the same. By matching φλ(x)\varphi_{\lambda}(x) and dφλ(x)/dxd\varphi_{\lambda}(x)/dx, as noted in Sec. II, the resonance condition is satisfied when

(k~j−k~1)d=2πnj,j=2,3,4,(17)\left(\tilde{k}_{j}-\tilde{k}_{1}\right)d=2\pi n_{j},\qquad j=2,3,4, \tag*{(17)}

where nj=1,2,…n_{j}=1,2,\ldots However, for the system we consider, k~jd∼10\tilde{k}_{j}d\sim10 and Eq. (17) only holds in few special cases.

In case (ii) the resonance condition similar to Eq. (17) is not possible, unless for the normal incidence when m∥y\mathbf{m}\parallel y. In that case, the coefficients of the decaying barrier states vanish and the spins of all scattering states become parallel to each other which makes the system “spinless” and perfect transmission becomes possible.

In cases (iii) and (iv), the presence of the decaying states in the leads can be understood for the incident particle from the λ=−1\lambda=-1 band at an angle φ>φc\varphi>\varphi_{c}. Since such decaying states in the leads do not carry any current [42], as can be understood from the Snell’s law, the perfect transmission and the resonance conditions are still possible. In case (iv), under the symmetry PσzTP\sigma_{z}T operation, up to a phase, the two propagating states inside the barrier remain the same, while the two decaying states in the barrier become, up to a phase, the two decaying states in the leads. The resonance condition is

(k~2−k~1)d=2nπ−δ,n=1,2,3…,(18)\left(\tilde{k}_{2}-\tilde{k}_{1}\right)d=2n\pi-\delta,\qquad n=1,2,3\ldots, \tag*{(18)}

where for the xx component of the propagating wave vectors in the barrier we assume k~2>k~1\tilde{k}_{2}>\tilde{k}_{1}, while the correction δ∝(α/ℏ)m∗/Δxc≪1\delta\propto(\alpha/\hbar)\sqrt{m^{*}/\Delta_{\mathrm{xc}}}\ll1. Since the resonances occur at Δxc>EF\Delta_{\mathrm{xc}}>E_{\mathrm{F}}, the magnitude of the propagating barrier wave vector is much larger than kFk_{\mathrm{F}}. Therefore, the xx component of the propagating barrier wave vector is almost the same for all the incoming states with different incident angles, which means the transmission resonances occur for all the incoming states almost simultaneously when the resonance condition is satisfied and thus the maximum GG is reached.

The two local maxima in GG near Δxc=20 meV\Delta_{\mathrm{xc}}=20\ \mathrm{meV} and 55 meV55\ \mathrm{meV}, shown in Figure 4, both correspond to the case (iv) and a large Δxc\Delta_{\mathrm{xc}} regime, where for m∥x\mathbf{m}\parallel x the propagating wave vectors are kx1,2≈∓2m∗Δ/ℏk_{x1,2}\approx\mp\sqrt{2m^{*}\Delta}/\hbar. Applying the resonance condition from Eq. (18) the first (second) of these maxima corresponds to n=1n=1 (n=2n=2). However, with large Δxc\Delta_{\mathrm{xc}}, the correction δ\delta in Eq. (18) is very small such that the resonance condition is accurately described by the spinless case where δ=0\delta=0.

F/2DEG In-Plane TAMR

A large value of TAMR, as defined in Eq. (2), has the origin in the difference for the maximum G(ϕ=0)G(\phi=0) and G(ϕ=π/2)G(\phi= \pi/2), that is for the m∥x\mathbf{m} \parallel\mathbf{x} and m∥y\mathbf{m} \parallel\mathbf{y} orientations. In Figure 4, for G(ϕ=0)G(\phi= 0) and shown examples (1)–(3), we see the expected barrier Fermi contour (broken line) shifted downward (⊥x\perp\mathbf{x}). This broken kyk_y symmetry leads to an unusual tunneling planar Hall effect, where the transverse voltage is maximum for the M∥x\mathbf{M} \parallel\mathbf{x}, while other Hall effects would vanish [44]. Due to the same asymmetry, the perfect match for both upper and lower half of the circle cannot be achieved simultaneously. We see in examples (1)–(3) that the upper (lower) half of the barrier contour tends to match the inner (outer) lead contour since the spin mismatch is smaller for these states. The best possible simultaneous match of Fermi contours yields the large transmission at Δxc=V0\Delta_{\mathrm{xc}} = V_0, resulting in the first peak in GG, even for large α\alpha when the barrier contour is strongly deformed from a circle, as in Figure 3(c).

As we see from Figure 4 with multiple peaks in GG, there also exist several resonances which could influence the TAMR. In the limit of α→0\alpha\to0, the condition for such resonances can be derived from Eq. (18) with δ=0\delta= 0

Δxc=V0−EF+π2ℏ2n22m∗d2,(19)\Delta_{\mathrm{xc}} = V_0 - E_F + \frac{\pi^2\hbar^2 n^2}{2m^{*}d^2}, \tag*{(19)}

for both G(ϕ=0)G(\phi= 0) and G(ϕ=π/2)G(\phi= \pi/2) which correspond to m∥x\mathbf{m} \parallel\mathbf{x} and m∥y\mathbf{m} \parallel\mathbf{y}, where n=1,2,…n = 1, 2, \ldots describes the order of the resonance that pertains to the case (iv) in Sec. II. In this situation, of vanishing Rashba SOC, the conductance no longer depends on the in-plane m\mathbf{m} and we expect a vanishing TAMR, as inferred from Eq. (2).

However, already for a small and nonvanishing α\alpha the maximum G(ϕ=π/2)G(\phi= \pi/2) is achieved for a different condition when the barrier circle shares the same size as the outer circle of the lead [11]. Up to the first order in α\alpha the condition leading to the maximum G(ϕ=π/2)G(\phi= \pi/2) is given by Δxc=V0+αkF\Delta_{\mathrm{xc}} = V_0 + \alpha k_F. When such a condition is satisfied, the barrier Fermi contour matches the outer lead contour instead of the inner one, because the main GG contribution is from the incident particles on the outer Fermi contour.

Correspondingly, near Δxc=V0\Delta_{\mathrm{xc}} = V_0, due to contour matching and the perfect transmission, we expect enhanced TAMR, as can be seen from Figure 5.

Dependence of the TAMR amplitude on the Rashba SOC $\alpha$ and barrier strength $V_0$ for a 2DEG system

Figure 5. (a) Dependence of the TAMR amplitude on the Rashba SOC α\alpha and the barrier strength V0V_0 for a 2DEG system with a d=13d = 13 nm thick barrier, EF=10E_F = 10 meV, and proximity-induced exchange splitting Δxc=15\Delta_{\mathrm{xc}} = 15 meV. (b) Dependence of the TAMR amplitude on Δxc\Delta_{\mathrm{xc}} and V0V_0, for the same dd and EFE_F as in (a), with α=0.093\alpha= 0.093 eVÅ for a typical InGaAs/InAlAs 2DEG. (c) The same as (b), but α=0.93\alpha= 0.93 eVÅ. The blue and red dashed lines in (a) are for the same parameters as the ones in (b) and (c).

Furthermore, when α≠0\alpha\ne0, k⃗2−k⃗1\vec{k}_2 - \vec{k}_1, which determine the resonance condition in Eq. (18), varies as M\mathbf{M} is rotated. This means the maxima for G(ϕ=0)G(\phi= 0) and G(ϕ=π/2)G(\phi= \pi/2) will be achieved at different Δxc\Delta_{\mathrm{xc}}. Therefore, similar to the situation near Δxc=V0\Delta_{\mathrm{xc}} = V_0, an enhanced TAMR will arise from the small difference in the peak conditions. With fixed EFE_F and V0V_0, up to the lowest-order correction of α\alpha, we can derive the difference in Δxc\Delta_{\mathrm{xc}} at the same order maxima η\eta of G(ϕ=0)G(\phi= 0) and G(ϕ=π/2)G(\phi= \pi/2)

Δxcdiff∝α2m∗ℏ2.(20)\Delta_{\mathrm{xc}}^{\mathrm{diff}} \propto\frac{\alpha^2 m^*}{\hbar^2}. \tag*{(20)}

Unlike the perfect transmission condition due to contour matching, the difference in now quadratic in α\alpha.

With this conductance analysis and nonmonotonic G(Δxc)G(\Delta_{\mathrm{xc}}) in Figure 4, a consequence of the collective contributions of multiple T≈1T \approx1 states corresponding to different propagation directions of the tunneling carriers, we expect various nonmonotonic TAMR trends and a strong influence of Rashba SOC. We can see from Figure 5(a) that increasing α\alpha can enhance both TAMR and the range of V0V_0 for such enhancement. However, if α\alpha gets too large (>1.5> 1.5 eVÅ), the absolute value of TAMR decreases, for a large range of V0V_0.

The reason for this unexpected trend in TAMR(α\alpha) can be understood from calculated results in Figure 5(b) and (c), which are shown for two different values of α\alpha, as indicated by blue and red vertical lines in Figure 5(a). From the previous discussion of a small α\alpha limit, we know that GG peaks near Δxc=V0\Delta_{\mathrm{xc}} = V_0 and such a peak will be shifted by αkF\alpha k_F when M\mathbf{M} is rotated. The sensitive dependence of GG on the M\mathbf{M} orientation leads to a large TAMR near the diagonal line in the parameter space, shown in Figure 5(b) and (c). Since such a shift is proportional to α\alpha, by comparing Figure 5(b) and (c) there is a wider range of the enhanced TAMR for a larger $\alpha`.

However, if α\alpha is too large (compared to V0/kFV_0/k_F and Δxc/kF\Delta_{xc}/k_F), the impact of the M orientation becomes negligible, and TAMR starts to decrease. This can be seen at the lower-left corner in Figure 5c, where the resonant TAMR is much smaller when both V0/kFV_0/k_F, Δxc/kF<α\Delta_{xc}/k_F < \alpha. Even for a fixed V0V_0, for example at 5 meV, we see that with α\alpha the absolute value of TAMR is nonmonotonic and TAMR also changes sign.

The calculated in-plane TAMR reveals various other peculiar trends. For a fixed α\alpha [Figure 5a] or Δxc/kF\Delta_{xc}/k_F [Figure 5b and (Figure 5c)], we can see that TAMR can be nonmonotonic in V0V_0 and even change its sign. Furthermore, there is as a clear nonmonotonic amplitude with Δxc\Delta_{xc} at a fixed V0V_0 in Figure 5b. Different barrier thickness in Figure 5 (26 nm) from that in Figure 4 (13 nm) modify the respective Δxc\Delta_{xc} values for the perfect transmission. These different TAMR trends in Figure 5 were primarily related to the best contour matching, near Δxc=V0\Delta_{xc}=V_0 [as the first peak in GG, seen in Figure 4)], while the role of n=1n=1 resonance can be seen in the lower-right corner in Figure 5(c).

From the angular dependence of TAMR, it is possible to obtain valuable information about the interfacial crystallographic symmetry [23]. For a 2DEG system, in the limit and Δxc/V0≪1\Delta_{xc}/V_0 \ll1, the leading contribution to the in-plane angular dependence of GG from the two incoming channels with helicity λ=±1\lambda=\pm1, is ±sin⁡ϕ\pm\sin\phi. However, with their opposite signs, these leading contributions cancel in the total GG, which becomes significantly smaller, quadratic in the small parameter, and has a different angular dependence, resulting in

TAMRF/2DEG(ϕ)∼(ΔxcV0)2sin⁡2ϕ.(21)\mathrm{TAMR}_{\mathrm{F}/2\mathrm{DEG}}(\phi) \sim\left(\frac{\Delta_{xc}}{V_0}\right)^2\sin^2\phi. \tag*{(21)}

While this angular dependence also coincides with the results from the surface states of 3D topological insulators, their junctions with ferromagnets do not have a resonant TAMR behavior of a 2DEG and thus lead to different trends in Δxc\Delta_{xc} and V0V_0 [11].

Another striking signature of the underlying resonant behavior of the calculated TAMR is its magnitude. In commonly considered vertical tunneling devices, the in-plane TAMR is rather small (typically <1%<1\%) even for large carrier spin polarization [7] and exchange energies >eV> \mathrm{eV} [8, 45, 46]. In contrast, with much smaller proximity-induced exchange fields, Δxc∼10\Delta_{xc}\sim10 meV, we find a much larger TAMR in our lateral (planar) structures, >10%>10\% even for SOC for a typical InAs-based 2DEG.

As we next focus on the F/S junctions and their superconducting out-of-plane TAMR analog [recall Eq. (3)], we will see that some of the calculated nonmonotonic trends, similar to those we discussed for F/2DEG systems, are experimentally verified, while the analysis of effective barriers allows us to give simple estimates of the enhanced MR which can even reach 100% in the measured samples.

F/S HAMILTONIAN AND CONDUCTANCE

Motivated by the recent experiments demonstrating a large MR enhancement in F/S junctions [30], depicted in Figure 2, we consider a simple model for F/S Hamiltonian, HF/SH_{\mathrm{F/S}}. It shares several similarities with HF/2DEGH_{\mathrm{F}/2\mathrm{DEG}} in Eq. (4), used to analyze in-plane TAMR. The key difference in the superconducting state is the need to simultaneously include the presence of electrons and holes and the pair potential Δ\Delta, which couples electronlike and holelike components of the underlying wavefunction.

Additionally, while Rashba SOC was inherent for whole 2DEG region, in F/S junction it is only an interfacial contribution, along with the potential barrier, between F and S region. Now we study vertical transport and out-of-plane MR [recall Eq. (3)], which is dominated by the process of Andreev reflection, introduced in § 1 and it is the microscopic mechanism for the superconducting proximity effects [23].

Before specifying HF/SH_{\mathrm{F/S}} and the resulting equation for quasiparticle states, it is instructive to note a similarity between the two-component transport in normal metal (N)/S junctions (for electronlike and holelike quasiparticles) and F/N junctions (for spin ↑,↓\uparrow,\downarrow), which both lead to current conversion, accompanied by additional boundary resistance [3]. In the N/S junction Andreev reflection is responsible for the conversion between the normal and the supercurrent, characterized by the superconducting coherence length, while in the F/N case a conversion between spin-polarized and unpolarized current is characterized by the spin diffusion length. In the N/S charge transport, Δ\Delta, which couples electronlike and holelike states, plays a similar role of the spin-flip potential which couples ↑,↓\uparrow,\downarrow states in the F/N spin transport.

We consider ballistic F/S junction, depicted in Figure 2, having a flat interface at z=0z=0 with potential and Rashba SOC scattering. Similar to the approach in Sec. II of matching the wavefunctions for the scattering states in different regions, we generalize the Griffin-Demers-Blonder-Tinkham-Klapwijk formalism [40, 47, 48, 49] to solve Bogoliubov-de Gennes equation for quasiparticle states HF/SΨ(r)=EΨ(r)H_{\mathrm{F/S}}\Psi(\mathbf{r})=E\Psi(\mathbf{r}), with energy EE [32], where

HF/S=(HeΔΘ(z)I2×2Δ∗Θ(z)I2×2Hh),(22)H_{\mathrm{F/S}}= \begin{pmatrix} H_e & \Delta\Theta(z)I_{2\times2}\\ \Delta^*\Theta(z)I_{2\times2} & H_h \end{pmatrix}, \tag*{(22)}

with Δ\Delta the ss-wave pair potential which, for a homogeneous S region, is also the value of the superconducting gap, and the single-particle Hamiltonian for electrons is

He=−ℏ22∇1m(z)∇−μ(z)−Δxc(m⋅σ)Θ(−z)+HB,(23)H_e=-\frac{\hbar^2}{2}\nabla\frac{1}{m(z)}\nabla-\mu(z)-\Delta_{xc}(\mathbf{m}\cdot\boldsymbol{\sigma})\Theta(-z)+H_B, \tag*{(23)}

while for holes Hh=−σyHe∗σyH_h=-\sigma_y H_e^*\sigma_y. In Eq. (22) m(z)m(z) is the effective mass and μ(z)\mu(z) is the chemical potential. The interfacial scattering is modeled by delta-like potential barrier, with the effective barrier height V0V_0 and width $d and the interfacial Rashba SOC of strength α\alpha has different units than Rashba SOC in F/2DEG [8]

HB=[V0d+α(kyσ^x−kxσ^y)]δ(z).(24)H_B = [V_0d + \alpha(k_y\hat{\sigma}_x - k_x\hat{\sigma}_y)]\delta(z). \tag*{(24)}

As in Sec. II, Δxc\Delta_{\mathrm{xc}} is the exchange spin splitting and we denote the orientation of the magnetization, M\mathbf{M}, by m\mathbf{m}, but we now also consider its out-of-plane rotation m=(sin⁡θcos⁡ϕ,sin⁡θsin⁡ϕ,cos⁡θ)\mathbf{m} = (\sin\theta\cos\phi,\sin\theta\sin\phi,\cos\theta).

Since the in-plane wave vector k∥\mathbf{k}_{\parallel} is conserved, the scattering states for incident spin σ\sigma electron are given by Ψσ(r)=eik∥⋅r∥ψσ(z)\Psi_{\sigma}(\mathbf{r}) = e^{i\mathbf{k}_{\parallel}\cdot\mathbf{r}_{\parallel}}\psi_{\sigma}(z) in a four-component basis, they include Andreev and specular reflection, each without and with spin flip, given by the amplitudes aσa_{\sigma}, bσb_{\sigma}, aˉσ\bar{a}_{\sigma}, and bˉσ\bar{b}_{\sigma} [31].

In the F region, the eigenspinors for electrons and holes are χσe=(χσ,0)T\chi_{\sigma}^{e} = (\chi_{\sigma},0)^{T} and χσh=(0,χ−σ)T\chi_{\sigma}^{h} = (0,\chi_{-\sigma})^{T} with

χσ=12(σ1+σcos⁡θe−iϕ,1−σcos⁡θ)T,(25)\chi_{\sigma} = \frac{1}{\sqrt{2}}\left(\sigma\sqrt{1+\sigma\cos\theta}e^{-i\phi},\sqrt{1-\sigma\cos\theta}\right)^{T}, \tag*{(25)}

where σ=1(−1)\sigma= 1(-1) or ↑(↓)\uparrow(\downarrow) refer to spin parallel (antiparallel) to M\mathbf{M} and the zz components of the wave vector are

kσe(h)=kF2+2mFℏ2[(−)E+σΔxc]−k∥2,(26)k_{\sigma}^{e(h)} = \sqrt{k_F^2 + \frac{2m_F}{\hbar^2}\left[(-)E+\sigma\Delta_{\mathrm{xc}}\right]-k_{\parallel}^2}, \tag*{(26)}

with a spin-averaged Fermi wave vector, kFk_F [42]. At an interface, with conserved k∥\mathbf{k}_{\parallel}, the HBH_B barrier eigenspinors in the helicity basis are given by [10, 50],

χ±=12(1∓ieiγ),(27)\chi_{\pm} = \frac{1}{\sqrt{2}}\begin{pmatrix}1\\ \mp i e^{i\gamma}\end{pmatrix}, \tag*{(27)}

where γ\gamma is the angle between k∥\mathbf{k}_{\parallel} and k^x\hat{\mathbf{k}}_x.

In the S region, coherence factors, uu, vv, satisfy u2=1−v2=(1+E2−Δ2/E)/2u^2 = 1-v^2 = (1+\sqrt{E^2-\Delta^2}/E)/2, while the zz-components of the wave vector are

qe(h)=qF2+(−)2mSℏ2E2−Δ2−k∥2,(28)q^{e(h)} = \sqrt{q_F^2 + (-)\frac{2m_S}{\hbar^2}\sqrt{E^2-\Delta^2}-k_{\parallel}^2}, \tag*{(28)}

with qFq_F the Fermi wave vector. Similar to Snell’s law [42], for a large k∥k_{\parallel} these zz-components can become imaginary representing evanescent states which carry no net current. The vanishing of evanescent states at infinity requires Im⁡[kσh]<0\operatorname{Im}[k_{\sigma}^{h}] < 0, so the sign of the zz-component of the wave vectors needs to be chosen correctly.

From the charge current conservation, the zero-temperature differential conductance at applied bias, VV, can be expressed as [32]

G(V)=∑σ∫dk∥2πkF2[1+Rσh(−eV)−Rσe(eV)].(29)G(V) = \sum_{\sigma}\int\frac{d k_{\parallel}}{2\pi k_F^2}\left[1+R_{\sigma}^{h}(-eV)-R_{\sigma}^{e}(eV)\right]. \tag*{(29)}

Here G(V)G(V) is normalized by the Sharvin conductance [3]

G03D=e2kF2A2πℏ,(30)G_0^{3D} = \frac{e^2k_F^2A}{2\pi\hbar}, \tag*{(30)}

where AA is the interfacial area, while the form is different than the previous expression in 2D [recall Eq. (10)], where SOC is present in the whole 2D region, not just at an interface. Only the probability amplitudes from the F region are needed, for Andreev Rσh=Re⁡[(k−σh/kσe)∣aσ∣2+(kσh/kσe)∣aˉσ∣2]R_{\sigma}^{h} = \operatorname{Re}[(k_{-\sigma}^{h}/k_{\sigma}^{e})|a_{\sigma}|^2 + (k_{\sigma}^{h}/k_{\sigma}^{e})|{\bar{a}}_{\sigma}|^2] and specular reflection Rσe=Re⁡[∣bσ∣2+(k−σe/kσe)∣bˉσ∣2]R_{\sigma}^{e} = \operatorname{Re}[|b_{\sigma}|^2 + (k_{-\sigma}^{e}/k_{\sigma}^{e})|{\bar{b}}_{\sigma}|^2]. A finite-temperature correction is straightforward by adding Fermi functions in Eq. (29), which smear the calculated G(V)G(V) at T=0T = 0 K. The integration kernel in Eq. (29), [1+Rσh−Rσe][1+R_{\sigma}^{h}-R_{\sigma}^{e}] can be viewed as the effective transmission.

We now turn to the measured bias-dependence for differential conductance, dI/dVdI/dV, for a quasi-2D vdW F/S junction [recall its geometry in Figure 2], shown in Figure 6(a) for several TT. Some trends are expected, dI/dVdI/dV peaks reflect the peaks in the density of states, near ∣eV∣=Δ|eV| = \Delta [40], moving inward with TT as Δ(T)\Delta(T) is also suppressed with TT. While the perfect transmission in 1D N/S geometry near V=0V = 0 implies the doubling of the normal-state conductance [40], since an incident electron through Andreev reflection contributes to the transfer of an electron pair across the N/S interface, the F/S transmission should be diminished, both by the interfacial barrier, parametrized by ZZ [recall Eq. (12)] as well by the spin polarization, P=Δxc/μF≈0.4−0.5P = \Delta_{xc}/\mu_F \approx0.4 - 0.5, expected for Fe0.29_{0.29}TaS2_2 and measured in a similar Fe0.26_{0.26}TaS2_2 ferromagnet [51]. Only a fraction of the incident majority spin electrons can find the opposite spin partner for Andreev reflection [42, 52].

Measured and calculated finite-temperature differential conductance

Figure 6. (a) Measured and (b) calculated finite-temperature and B=0B = 0 differential conductance for Fe0.29_{0.29}TaS2_2/NbN junction, normalized to its value at V=5V = 5 mV, corresponding to the normal-state result. For T=2T = 2 K comparison, in our calculations we choose T/Tc=0.16T/T_c = 0.16, P=0.42P = 0.42, and Δ=2.7\Delta= 2.7 meV.

With the NbN critical temperature, Tc=12.5T_c = 12.5 K, we focus on dI/dVdI/dV at the lowest T=2T = 2 K and theoretically explore the role of interfacial properties using dimensional parameters for barrier strength and Rashba SOC

Z=V0dmFmSℏ2kFqF,(31)Z = \frac{V_0d\sqrt{m_Fm_S}}{\hbar^2\sqrt{k_Fq_F}}, \tag*{(31)}
λ=2αmFmSℏ2,(32)\lambda= \frac{2\alpha\sqrt{m_Fm_S}}{\hbar^2}, \tag*{(32)}

since we are interested in identifying trends with these interfacial parameters, rather than obtaining the best fit for dI(V)/dVdI(V)/dV, we further simplify this approach by considering the case for mF=mS=mm_F = m_S = m and kF=qFk_F = q_F.

Our calculated results are given in Figure 6(b) and reflect these simplifications. For example, assuming that, kF=qFk_F = q_F and that mm is the free electron mass, the chosen dimensionless barrier strenght Z=0.8Z = 0.8 corresponds to kF=1.4×1010k_F = 1.4 \times10^{10} m−1^{-1}, consistent with NbN values [53], while V0=0.5V_0 = 0.5 eV, and d=1.7d = 1.7 nm is an average thickness among our studied F/S samples. The value of Rashba SOC, α=5.7\alpha= 5.7 eVÅ2^2, which follows from the choice of λ=1.5\lambda= 1.5, is also consistent with separate fits of the angular dependence of MR measured near zero bias in Ref. [30].

By comparing λ=0\lambda= 0 and λ=1.5\lambda= 1.5 results at a fixed Z=0.8Z = 0.8 above the gap the changes are very small. However, the inclusion of Rashba SOC can strongly enhance the low-bias conductance. To understand the origin of this SOC effect, and similar trends from Sec. II, we note that in the normal state for the barrier region the dispersion is SOC split and shifted by the barrier potential (assuming V0>0V_0 > 0). As in Sec. II, a spinor of an incident electron with k∥k_{\parallel} can be decomposed into barrier eigenspinors, ∣χσ⟩=⟨χ+∣χσ⟩∣χ+⟩+⟨χ−∣χσ⟩∣χ−⟩|\chi_\sigma\rangle= \langle\chi_+|\chi_\sigma\rangle|\chi_+\rangle+ \langle\chi_-|\chi_\sigma\rangle|\chi_-\rangle, where χ±\chi_\pm from Eq. (27) has helicity ±1\pm1. The two helicities have inequivalent effective barriers [recall Eqs. (13) and (14)]

Z±=Z±λk∥/2kF.(33)Z_\pm= Z \pm\lambda k_{\parallel}/2k_F. \tag*{(33)}

Since Z,λk∥/(2kF)≥0Z,\lambda k_{\parallel}/(2k_F) \ge0, for positive helicity the barrier is enhanced, Zeff+≥ZZ_{\mathrm{eff}}^+ \ge Z. However, for negative helicity, at Z=λk∥/(2kF)Z = \lambda k_{\parallel}/(2k_F), Z−Z_- becomes effectively completely transparent, such k∥k_{\parallel} can be viewed as “open channels” and are responsible for a strongly increased dI/dVdI/dV.

Some peculiar conductance trends can already be understood at V=0V = 0 and T=0T = 0 K, where the charge transport in F/S junctions is determined by conventional and spin-flip Andreev reflection with opposite and equal spin projection of the incident and reflected particle, respectively, corresponding to the spin-singlet and spin-triplet interfacial superconducting correlations. For G(V=0)G(V = 0) plotted in the plane defined by Z>0Z > 0 and λ>0\lambda> 0, a striking behavior was found for ∣P∣<1|P| < 1 in the triangular region approximately delimited with lines T1 and T2 [33]

T1:λ=2Z1−P,T2:λ=2Z,(34)\mathrm{T1}: \lambda= \frac{2Z}{\sqrt{1-P}}, \qquad\mathrm{T2}: \lambda= 2Z, \tag*{(34)}

where GG is dominated by the spin-flip Andreev reflection and proximity-induced spin-triplet superconductivity.

Before showing next that a region in GG delimited by the lines T1 and T2 also pertains to the enhanced calculated MAAR, we comment on relaxing the assumption of equal masses and Fermi velocities in the F and S regions. In a simple 1D N/S case, without SOC, different Fermi velocities, Fv=(vS/vN)≠1F_v = (v_S/v_N) \ne1, are known to increase the effective barrier strength, Zeff2→Z2+(1−Fv)2/4Fv>Z2Z_{\mathrm{eff}}^2 \to Z^2+(1-F_v)^2/4F_v > Z^2 [41]. While this argument is often used to also ignore a Fermi velocity mismatch in the F/S case [52], as it is accounted for by simply enhancing ZZ, some subtleties exist [31, 42]. The mismatch of effective masses, Fm=mS/mFF_m = m_S/m_F, together with the mismatch of the Fermi wave vectors, Fk=qF/kFF_k = q_F/k_F, can then be viewed as determining FvF_v and enhancing the effective ZZ. In our studies with SOC, Fk≠1F_k \ne1 implies that Eq. (33) should be generalized by replacing kFk_F with kFqF\sqrt{k_Fq_F}, including in the expression for open channels and a vanishing Z−Z_-. Even with such a wave vector mismatch, the spin-triplet contribution remains enhanced within a triangular region, as long as Fk>1−PF_k > \sqrt{1-P} i.e., qF>k↓q_F > k_{\downarrow}. The delimiting lines T1 and T2, for both Fk<1F_k < 1 and Fk>1F_k > 1, have slopes modified by FkF_k as compared to those given by Eq. (34).

F/S Out-Of-Plane MAAR

Just as in the F/2DEG structures, the interplay between exchange field and Rashba SOC will lead to various nonmonotonic trends [30, 32, 33, 54]. Such trends can be inferred from Eqs. (33) and (34), as well as in the MR, which in F/S junctions corresponds to the superconducting counterpart of TAMR, and its term, MAAR (recall § 1), identified the key role of the Andreev reflection.

Similar to the significance of the condition for perfect transmission and contour matching in understanding the origin of an enhanced TAMR, for MAAR it is important to identify the influence of open channels [33]. From Eq. (33), we see that Z,λ>0Z,\lambda> 0 yields vanishing Z−Z_- with open channels located at the Fermi contour of radius

k∥open=(2Z/λ)kF.(35)k_{\parallel}^{\mathrm{open}} = (2Z/\lambda)k_F. \tag*{(35)}

Open channels give the dominant contribution to GG and are also expected to determine the amplitude of MAAR.

Since evanescent state for large k∥k_{\parallel} do not contribute to GG they should be excluded in open channels. From the Snell’s law [42], for the incident majority spin electron this implies that to have propagating scattering states, k∥≤k↓k_{\parallel} \le k_{\downarrow} is required for conventional Andreev reflections and k∥≤qFk_{\parallel} \le q_F for dominant spin-flip Andreev reflections [31]. The maximum of the total GG is achieved when the amount of the open channels ∝k∥\propto k_{\parallel}, is maximized. For kF=qFk_F = q_F, in the two limiting case k∥open=k↓k_{\parallel}^{\mathrm{open}} = k_{\downarrow} and k∥open=kFk_{\parallel}^{\mathrm{open}} = k_F we recover the conditions for lines T1 and T2 from Eq. (34), such that the maximum spin-flip Andreev reflection is located near k∥open=kFk_{\parallel}^{\mathrm{open}} = k_F i.e., λ=2Z\lambda= 2Z.

From the calculated MAAR in Figure 7 we see that the triangular region delimited by lines T1, T2 is also identifying the region of an enhanced MAAR and confirming the influence of open channels. Similar to our contour matching in TAMR, as noted in Sec. II., for a full picture the role of spin mismatch also needs to be included. Two characteristic features are easy to see from Figure 7: (i) a large MAAR enhancement, which can reach 100 % and (ii) MAAR is nonmonotonic with both ZZ and λ\lambda, when the other parameter is kept constant.

Amplitude of out-of-plane magnetoanisotropic Andreev reflection as a function of interfacial parameters $Z$ and $\lambda$

Figure 7. Amplitude of out-of-plane magnetoanisotropic Andreev reflection (MAAR), given by Eq. (3) calculated at zero bias and T=0T = 0 K, shown as a function of interfacial parameters ZZ and λ\lambda for P=0.47P = 0.47, consistent with Fe0.29_{0.29}TaS2_2. Lines T1 and T2 are defined in Eq. (34).

To understand (i), we recall that in the triangular region, the dominant contribution to the total GG is from the open channels that is associated with spin-flip Andreev reflection and spin-triplet pairing ∣↑↑⟩|\uparrow\uparrow\rangle. Therefore, the total conductance can be approximately written as

G(V=0)≈∫k∥openk∥ dk∥∫02πdγπkF2Rσ=1h(k∥,γ)≈k∥openΔk∥∫02πdγπkF2Rσ=1h(k∥open,γ)(36)\begin{aligned} G(V = 0) \approx\int_{k_{\parallel}^{\mathrm{open}}} k_{\parallel} \, \mathrm{d}k_{\parallel} \int_{0}^{2\pi} \frac{\mathrm{d}\gamma}{\pi k_{F}^{2}} R_{\sigma=1}^{h}(k_{\parallel},\gamma) \\ &\approx k_{\parallel}^{\mathrm{open}}\Delta k_{\parallel}\int_{0}^{2\pi}\frac{\mathrm{d}\gamma}{\pi k_{F}^{2}}R_{\sigma=1}^{h}(k_{\parallel}^{\mathrm{open}},\gamma) \tag*{(36)} \end{aligned}

where Δk∥\Delta k_{\parallel} is the width of the open channels in k∥k_{\parallel}-plane, γ\gamma is the angle between k∥k_{\parallel} and x^\hat{x}, recall Eq. (27). For out-of-plane MM, Rσ=1h(k∥open,γ)≡Rσ=1h,openR_{\sigma=1}^{h}(k_{\parallel}^{\mathrm{open}},\gamma)\equiv R_{\sigma=1}^{h,\mathrm{open}} does not depend on γ\gamma due to the rotational symmetry and we can write

G(θ=0)≈2k∥openΔk∥kF2Rσ=1h,open.(37)G(\theta= 0) \approx\frac{2k_{\parallel}^{\mathrm{open}}\Delta k_{\parallel}}{k_{F}^{2}}R_{\sigma=1}^{h,\mathrm{open}} . \tag*{(37)}

For in-plane MM, considering the interplay between the incident electron spin and the barrier eigenspinor, the

spin-flip Andreev reflection of an incident ↑\uparrow electron is given by Rσ=1h(k∥open,γ)≈Rσ=1h,opencos⁡2(γ−ϕ)R_{\sigma=1}^{h}(k_{\parallel}^{\mathrm{open}},\gamma)\approx R_{\sigma=1}^{h,\mathrm{open}}\cos^{2}(\gamma-\phi) [33]. As a result, for in-plane MM, we obtain

G(θ=π2)≈k∥openΔk∥kF2Rσ=1h,open=G(θ=0)2.(38)G\left(\theta= \frac{\pi}{2}\right)\approx\frac{k_{\parallel}^{\mathrm{open}}\Delta k_{\parallel}}{k_{F}^{2}}R_{\sigma=1}^{h,\mathrm{open}}=\frac{G(\theta=0)}{2}. \tag*{(38)}

From Eq. (3), this means that MAAR in the triangular region is 100%, which explains that the previous prediction [32] can be attributed to spin-triplet superconductivity and is also consistent with the largest values in the experimental results for Fe0.29_{0.29}TaS2_2/NbN junctions with a perpendicular magnetic anisotropy [30].

Despite our simple theoretical model to include the interplay of exchange field and SOC on the proximity-induced superconductivity in F/S junctions, such a description provides MAAR, which is in a good agreement with the measured angular dependence of MR in Figure 8. Remarkably, just as the variation of ZZ shows a nonmonotonic MAAR in Figure 7 and a huge increase of MAAR at intermediate ZZ values, which with Rashba SOC can lead to a vanishing Z−Z_{-}, we see that this trend in Figure 8(d), (Figure 8e), and (Figure 8h) is also retained in the measured results obtained by varying the Al2_2O3_3 barrier thickness (∼1−2.5 nm\sim1 - 2.5\,\mathrm{nm}).

Atomic force microscope images, barrier schematics, and corresponding magnetoresistance measurements

Figure 8. (a) and (b) Atomic force microscope images of 2 nm Al2_2O3_3 on Fe0.29_{0.29}TaS2_2 thin flake with RMS roughness of ∼0.270\sim0.270 nm indicate that Al2_2O3_3 is relatively flat and without obvious pin holes or discontinuities. (c), (e), and (g) schematic of the incident spin-polarized electrons tunneling across different effective interfacial barriers with Rashba SOC into the proximity-induced spin-triplet superconductivity. (d), (f), and (h) the corresponding magnetoresistance (MR) [recall θ\theta in Figure 2], for different junction resistance-area products RJSR_J S, a proxy for dimensionless barrier strength ZZ. Red lines: theoretically predicted two-fold symmetry. The measurements were at T=2T = 2 K, B=9B = 9 T, and V=1V = 1 mV.

While we cannot rule out a non-MAAR contribution in the measured low-bias MR for Fe0.29_{0.29}TaS2_2/NbN junctions, we can still examine its magnitude from several reported control measurements [30]. For example, by increasing TT above TcT_c, MR is drastically reduced from its maximum at ∼100%\sim100\% in Figure 8(f) and we can attribute the main MR part to a superconducting response. By considering MR in Al/Al2_2O3_3/NbN samples, we no longer have the interplay between the ferromagnetism and SOC, responsible for MAAR, but the resulting small MR of just few % at T=2 KT = 2\,\mathrm{K} is encouraging that the presence of vortices at B=9 TB = 9\,\mathrm{T} has only a modest MR effect. Taken together, these findings confirm that MAAR is indeed the dominant source of the measured MR in Figure 8, which further indicates that Fe0.29_{0.29}TaS2_2/NbN junctions, provide an important platform to realize elusive equal-spin-triplet superconductivity.

Discussion and Conclusions

We have shown that Rashba SOC can lead to strongly enhanced magnetoresistance in junctions with one ferromagnet. Both in the normal-state and the superconducting response the calculated magnetoresistance is characterized by various nonmonotonic trends. While some of these trends and an enhanced magnetoresistance have been measured in superconducting junctions [30, 55], experiments in the normal state are largely unexplored.

To realize magnetic proximity effects for the in-plane transport, magnetic insulators are desirable [23, 56, 57, 58]. This precludes current flow in the more resistive F region [Fig. 1(a)] and minimizes hybridization with the 2DEG to enable a gate-tunable proximity-induced exchange splitting in their respective states. However, as shown in graphene [59, 60, 61] for tunable magnetic proximity effects one could also employ ferromagnetic metals, separated by an insulating layer from the 2DEG.

For a suitable materials platform, which would support large magnetoresistive effects, we could extend our focus on simple Rashba SOC to a growing class of van der Waals (vdW) materials. Their heterostructure offer both transport in materials with strong SOC as well 2D ferromagnets [62, 63, 64] with atomically-sharp interfaces and not limited to lattice-matching constraints. For example, transition-metal dichalcogenides in addition to their inherent SOC also provide spin-orbit proximity [65, 66, 67, 68, 69, 70, 71] and thereby alter spin textures and expected TAMR, while 2D ferromagnets support a versatile gate control [72, 73, 74].

While we have focused on a longitudinal transport in a very simple system, the behavior emerging from a spin parity-time symmetry of the scattering states with perfect transmission is important not just in explaining a surprisingly large TAMR, but also as a sensitive probe to distinguish between trivial and topological states [11]. It would be interesting to know if our predicted non-monotonic trends with interfacial parameters for TAMR could be also relevant for other transport phenomena in junctions with a single ferromagnet, such as spin-orbit torque and spin-Hall magnetoresistance [5, 75].

In the superconducting state, the observed large magnetoresistance has important implications as it provides an alternative probe for spin-triplet or mixed singlet-triplet superconductivity [76] and such a large signal is possible to realize even systems with only a modest SOC and negligible normal-state magnetoresistance [55]. Since this work shows that spin-triplet superconductivity, desirable both for superconducting spintronics and Majorana bound states [14], is feasible in simple structures with a single F and S region, it would be important to extend its analysis to Josephson junctions, where enhanced magnetoresistance was predicted [77], but not connected to the spin-triplet superconductivity, which was extensively studied in other Josephson junctions [13, 14, 78, 79, 80, 81].

In our work, the nonmonotonic trends with the interfacial barrier strength were observed by comparing samples with different thickness of the (Al2_2O3_3). It would be desirable, to realize systems in which such changes, as well as the tunability of the Rashba SOC strength could be controlled in the same sample. A desirable progress is realized in Josephson junctions where currently there are no ferromagnetic regions, but the Zeeman splitting is due to an applied magnetic field [14, 82, 83, 84, 85, 86]. Related experimental support in junctions with Al as a superconductor and InAs-based 2DEG already suggests an observation of topological spin-triplet superconductivity [84]. In the same sample that both a reentrant superconductivity with an applied magnetic field and the jump in the superconducting phase are measured, but an enhanced magnetoresistance as another signature of spin-triplet superconductivity has not yet been studied.

Acknowledgments

We thank Thomas Vezin for useful discussions. We had a privilege of first meeting Professor Emmanuel Rashba at the Spintronics 2001: International Conference on Novel Aspects of Spin-Polarized Transport and Spin Dynamics in Washington, D.C. (chaired by one of us [87]). He gave an insightful talk sharing his personal experience about relevant historic developments as well as a vision for the future opportunities in spintronics, conveyed further in his article [88]. Since then, through his valuable comments, we were able to better understand our own work and its broader context. His advice and guidance was not limited to semiconductor systems, and we again benefited from his comments which came soon after posting online our recent manuscript on spin-triplet superconductivity [30]. This work is supported by U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award No. DE-SC0004890, the UB Center for Computational Research, National Basic Research Programs of China (No. 2019YFA0308401), and National Natural Science Foundation of China (No. 11974025).

References

  1. [1]S. Maekawa and U. Gafvert, “Electron tunneling between ferromagnetic films,” IEEE Trans. on Magn. 18, 707 (1982).DOI
  2. [2]J. S. Moodera, Lisa R. Kinder, Terrilyn M. Wong, and R. Meservey, “Large Magnetoresistance at Room Temperature in Ferromagnetic Thin Film Tunnel Junctions,” Phys. Rev. Lett. 74, 3273 (1995).
  3. [3]I. Žutić, J. Fabian, and S. Das Sarma, “Spintronics: Fundamentals and applications,” Rev. Mod. Phys. 76, 323 (2004).DOI
  4. [4]S. S. P. Parkin, C. Kaiser, A. Panchula, P. M. Rice, B. Hughes, M. Samant, and S.-H. Yang, “Giant tunneling magnetoresistance at room temperature with MgO (100) tunnel barriers,” Nat. Mater. 3, 862 (2004).pubmed.ncbi.nlm.nih.gov/15516928
  5. [5]E. Y. Tsymbal and I. Žutić, eds., Spintronics Handbook Spin Transport and Magnetism, 2nd Edition (CRC Press, Taylor & Francis, Boca Raton, FL, 2019).
  6. [6]W. Thomson, “On the electro-dynamic qualities of metals:—Effects of magnetization on the electric conductivity of nickel and of iron,” Proc. R. Soc. London 8, 546–550 (1857).DOI
  7. [7]C. Gould, C. Rüster, T. Jungwirth, E. Girgis, G. M. Schott, R. Giraud, K. Brunner, G. Schmidt, and L. W. Molenkamp, “Tunneling Anisotropic Magnetoresistance: A Spin-Valve-Like Tunnel Magnetoresistance Using a Single Magnetic Layer,” Phys. Rev. Lett. 93, 117203 (2004).DOI
  8. [8]J. Moser, A. Matos-Abiague, D. Schuh, W. Wegscheider, J. Fabian, and D. Weiss, “Tunneling Anisotropic Magnetoresistance and Spin-Orbit Coupling in Fe/GaAs/Au Tunnel Junctions,” Phys. Rev. Lett. 99, 056601 (2007).pubmed.ncbi.nlm.nih.gov/17930774
  9. [9]A. N. Chantis, K. D. Belashchenko, E. Y. Tsymbal, and M. van Schilfgaarde, “Tunneling Anisotropic Magnetoresistance Driven by Resonant Surface States: First-Principles Calculations on an Fe(001) Surface,” Phys. Rev. Lett. 98, 046601 (2007).pubmed.ncbi.nlm.nih.gov/17358794
  10. [10]J. Fabian, A. Matos-Abiague, C. Ertler, P. Stano, and I. Žutić, “Semiconductor spintronics,” Acta Phys. Slov. 57, 565 (2007).DOI
  11. [11]C. Shen, T. Leeney, A. Matos-Abiague, B. Scharf, J. E. Han, and I. Žutić, “Resonant tunneling anisotropic magnetoresistance induced by magnetic proximity,” Phys. Rev. B 102, 045312 (2020).DOI
  12. [12]Yu. A. Bychkov and E. I. Rashba, “Properties of a 2D with lifted spectral degeneracy,” Zh. Eksp. Teor. Fiz. Pisma Red. 39, 66 (1984), [JETP Lett. 39, 78 (1984)].
  13. [13]R. Cai, I. Žutić, and W. Han, “Superconductor/ferromagnet heterostructures for superconducting spintronics and quantum computation,” Adv. Quant. Tech. 6, 220080 (2023).arxiv.org/abs/2211.08625
  14. [14]M. Amundsen, J. Linder, J. W. A. Robinson, I. Žutić, and N. Banerjee, “Colloquium: Spin-orbit effects in superconducting hybrid structures,” arXiv:2210.03549 (2022).arxiv.org/abs/2210.03549
  15. [15]M. Eschrig, “Spin-polarized supercurrents for spintronics: A review of current progress,” Rep. Prog. Phys. 78, 104501 (2015).DOI
  16. [16]J. Linder and J. W. A. Robinson, “Superconducting spintronics,” Nat. Phys. 11, 307 (2015).DOI
  17. [17]A. I. Buzdin, “Proximity effects in superconductor-ferromagnet heterostructures,” Rev. Mod. Phys. 77, 935 (2005).
  18. [18]F. S. Bergeret, A. F. Volkov, and K. B. Efetov, “Odd triplet superconductivity and related phenomena in superconductor-ferromagnet structures,” Rev. Mod. Phys. 77, 1321 (2005).DOI
  19. [19]A. Yu Kitaev, “Unpaired Majorana fermions in quantum wires,” Phys.-Usp. 44, 131 (2001).DOI
  20. [20]D. Aasen, M. Hell, R. V. Mishmash, A. Higginbotham, J. Danon, M. Leijnse, T. S. Jespersen, J. A. Folk, C. M. Marcus, K. Flensberg, and J. Alicea, “Milestones toward Majorana-based quantum computing,” Phys. Rev. X 6, 031016 (2016).DOI
  21. [21]U. Güngördü and Alexey A. Kovalev, “Majorana bound states with chiral magnetic textures,” J. Appl. Phys. 132, 041101 (2022).
  22. [22]K. Laubscher and J. Klinovaja, “Majorana bound states in semiconducting nanostructures,” J. Appl. Phys. 130, 081101 (2021).DOI
  23. [23]I. Žutić, A. Matos-Abiague, B. Scharf, H. Dery, and K. Belashchenko, “Proximitized materials,” Mater. Today 22, 85 (2019).arxiv.org/abs/1805.07942
  24. [24]I. A. Akimov, M. Salewski, I. V. Kalitukha, S. V. Poltavtsev, J. Debus, D. Kudlaciк, V. F. Sapega, N. E. Kopteva, E. Kirstein, E. A. Zhukov, D. R. Yakovlev, G. Karczewski, M. Wiater, T. Wojtowicz, V. L. Korenev, Yu. G. Kusrayev, and M. Bayer, “Direct measurement of the long-range p-d exchange coupling in a ferromagnet-semiconductor Co/CdMgTe/CdTe quantum well hybrid structure,” Phys. Rev. B 96, 184412 (2017).DOI
  25. [25]V. L. Korenev, I. V. Kalitukha, I. A. Akimov, V. F. Sapega, E. A. Zhukov, E. Kirstein, O.S. Ken, D. Kudlaciк G. Karczewski, M. Wiater, T. Wojtowicz, N. D. Ilyinskaya, N.M. Lebedeva, T.A. Komissarova, Yu. G. Kusrayev, D.R. Yakovlev, and M. Bayer, “Low voltage control of exchange coupling in a ferromagnet-semiconductor quantum well hybrid structure,” Nat. Commun. 10, 2899 (2019).DOI
  26. [26]K. Takiguchi, L. D. Anh, T. Chiba, T. Koyama, D. Chiba, and M. Tanaka, “Giant gate-controlled proximity magnetoresistance in semiconductor-based ferromagnetic/non-magnetic bilayers,” Nat. Phys. 15, 1134 (2019).DOI
  27. [27]C. Betthausen, T. Dollinger, H. Saarikoski, V. Kolkovsky, G. Karczewski, T. Wojtowicz, K. Richter, and D. Weiss, “Spin-transistor action via tunable Landau-Zener transitions,” Science 337, 324 (2012).DOI
  28. [28]D. J. Griffiths, Introduction to Quantum Mechanics, Second Edition (Pearson Prentice Hall, Upper Saddle River, NJ, 2005).
  29. [29]J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, Second Edition (Addison-Wesley, San Francisco, CA, 2011).
  30. [30]B. Cai, Y. Yao, P. Lv, Y. Ma, W. Xing, B. Li, Y. Ji, H. Zhou, C. Shen, S. Jia, X. C. Xie, I. Žutić, Q.-F. Sun, and W. Han, “Evidence for anisotropic spin-triplet Andreev reflection at the 2D van der Waals ferromagnet/superconductor interface,” Nat. Commun. 12, 6725 (2021).DOI
  31. [31]I. Žutić and S. Das Sarma, “Spin-polarized transport and Andreev reflection in semiconductor/superconductor hybrid structures,” Phys. Rev. B 60, R16322 (1999).DOI
  32. [32]P. Högl, A. Matos-Abiague, I. Žutić, and J. Fabian, “Magnetoanisotropic Andreev Reflection in Ferromagnet-Superconductor Junctions,” Phys. Rev. Lett. 115, 116601 (2015).DOI
  33. [33]T. Vezin, C. Shen, J. E. Han, and I. Žutić, “Enhanced spin-triplet pairing in magnetic junctions with s-wave superconductors,” Phys. Rev. B 101, 014515 (2020).
  34. [34]P. Lv, Y.-F. Zhou, N.-X. Yang, and Q.-F. Sun, “Magnetoanisotropic spin-triplet Andreev reflection in ferromagnet-Ising superconductor junctions,” Phys. Rev. B 97, 144501 (2018).
  35. [35]R. Winkler, Spin-Orbit Coupling Effects in Two-Dimensional Electron and Hole Systems (Springer, New York, 2003).
  36. [36]I. V. Rozhansky, N. S. Aver'kiev, and E. Lähderanta, “Resonant tunneling between two-dimensional layers accounting for spin-orbit interaction,” Phys. Rev. B 93, 195405 (2016).DOI
  37. [37]H. Liu, E. Marcellina, A. R. Hamilton, and D. Culcer, “Strong Spin-Orbit Contribution to the Hall Coefficient of Two-Dimensional Hole Systems,” Phys. Rev. Lett. 121, 087701 (2018).
  38. [38]D. P. Arovas and Y. Lyanda-Geller, “Non-Abelian geometric phases and conductance of spin-3/2 holes,” Phys. Rev. B 57, 12302 (1998).DOI
  39. [39]S. M. Girvin and K. Yang, Modern Condensed Matter Physics (Cambridge University Press, Cornwall, UK, 2019).DOI
  40. [40]G. E. Blonder, M. Tinkham, and T. M. Klapwijk, “Transition from metallic to tunneling regimes in superconducting microconstrictions: Excess current, charge imbalance, and supercurrent conversion,” Phys. Rev. B 25, 4515 (1982).DOI
  41. [41]G. E. Blonder and M. Tinkham, “Metallic to tunneling transition in Cu-Nb point contacts,” Phys. Rev. B 27, 112 (1983).DOI
  42. [42]I. Žutić and O. T. Valls, “Tunneling spectroscopy for ferromagnet/superconductor junctions,” Phys. Rev. B 61, 1555 (2000).DOI
  43. [43]B. K. Nikolić, L. P. Zârbo, and S. Souma, “Mesoscopic spin Hall effect in multiprobe ballistic spin-orbit-coupled semiconductor bridges,” Phys. Rev. B 72, 075361 (2005).DOI
  44. [44]B. Scharf, A. Matos-Abiague, J. E. Han, E. M. Hankiewicz, and I. Žutić, “Tunneling Planar Hall Effect in Topological Insulators: Spin Valves and Amplifiers,” Phys. Rev. Lett. 117, 168606 (2016).DOI
  45. [45]B. G. Park, J. Wunderlich, D. A. Williams, S. J. Joo, K. Y. Jung, K. H. Shin, K. Olejnik, A. B. Shick, and T. Jungwirth, “Tunneling Anisotropic Magnetoresistance in Multilayer-(Co/Pt)/AlOx/Pt Structures,” Phys. Rev. Lett. 100, 087204 (2008).DOI
  46. [46]T. Uemura, Y. Imai, M. Harada, K. Matsuda, and M. Yamamoto, “Tunneling anisotropic magnetoresistance in epitaxial CoFe/n-GaAs junctions,” Appl. Phys. Lett. 94, 182502 (2009).DOI
  47. [47]A. Griffin and J. Demers, “Tunneling in the normal-metal-insulator-superconductor geometry using the Bogoliubov equations of motion,” Phys. Rev. B 4, 2202 (1971).DOI
  48. [48]S. Kashiwaya and Y. Tanaka, “Tunneling effects on surface bound states in unconventional superconductors,” Rep. Prog. Phys. 63, 1641 (2000).DOI
  49. [49]S. Wu and K. V. Samokhin, “Tunneling conductance of ferromagnet/noncentrosymmetric superconductor junctions,” Phys. Rev. B 80, 014516 (2009).DOI
  50. [50]A. Matos-Abiague and J. Fabian, “Anisotropic tunneling magnetoresistance and tunneling anisotropic magnetoresistance: Spin-orbit coupling in magnetic tunnel junctions,” Phys. Rev. B 79, 155303 (2009).DOI
  51. [51]X.-Y. Yu, H.-L. Feng, G.-X. Gu, Y.-H. Liu, Z.-L. Li, T.-S. Xu, and Y.-Q. Li, “Andreev reflection spectroscopy of ferromagnetic Fe0.26TaS2 with layered structure,” Acta Phys. Sin. 68, 247201 (2019).DOI
  52. [52]R. J. Soulen Jr., J. M. Byers, M. S. Osofsky, B. Nadgorny, T. Ambrose, S. F. Cheng, P. R. Broussard, C. T. Tanaka, J. Nowak, J. S. Moodera, A. Barry, and J. M. D. Coey, “Measuring the spin polarization of a metal with a superconducting point contact,” Science 282, 85 (1998).pubmed.ncbi.nlm.nih.gov/9756482
  53. [53]S. P. Chockalingam, M. Chand, J. Jesudasan, V. Tripathi, and P. Raychaudhuri, “Superconducting properties and Hall effect of epitaxial NbN thin films,” Phys. Rev. B 77, 214503 (2008).DOI
  54. [54]Y. Mizuno, T. Yokoyama, and Y. Tanaka, “Theory of charge transport in ferromagnetic superconductor/s-wave superconductor junction,” Phys. Rev. B 80, 195307 (2009).DOI
  55. [55]I. Martinez, P. Högl, C. González-Ruano, J. P. Casales, C. Tiusan, Y. Lu, H. Hehn, A. Matos-Abiague, J. Fabian, I. Žutić, and F. G. Aliev, “Interfacial Spin-Orbit Coupling: A Platform for Superconducting Spintronics,” Phys. Rev. Applied 13, 014030 (2020).DOI
  56. [56]Z. Jiang, C. Z. Chang, C. Tang, P. Wei, J. S. Moodera, and J. Shi, “Independent tuning of electronic properties and induced ferromagnetism in topological insulators with heterostructure approach,” Nano Lett. 15, 5835 (2015).DOI
  57. [57]P. Wei, F. Katmis, B. A. Assaf, H. Steinberg, P. Jarillo-Herrero, D. Heiman, and J. S. Moodera, “Exchange-Coupling-Induced Symmetry Breaking in Topological Insulators,” Phys. Rev. Lett. 110, 186807 (2013).pubmed.ncbi.nlm.nih.gov/23683236
  58. [58]A. T. Lee, M. J. Han, and K. Park, “Magnetic proximity effect and spin-orbital texture at the Bi2Se3/EuS interface,” Phys. Rev. B 90, 155103 (2014).DOI
  59. [59]P. Lazić, K. D. Belashchenko, and I. Žutić, “Effective gating and tunable magnetic proximity effects in two-dimensional heterostructures,” Phys. Rev. B 93, 241401(R) (2016).DOI
  60. [60]K. Zollner, M. Gmitra, T. Frank, and J. Fabian, “Theory of Proximity-Induced Exchange Coupling in Graphene on hBN/(Co,Ni),” Phys. Rev. B 94, 155441 (2016).DOI
  61. [61]J. Xu, S. Singh, J. Katoch, G. Wu, T. Zhu, I. Žutić, and R. K. Kawakami, “Spin inversion in graphene spin valves by gate-tunable magnetic proximity effect at one-dimensional contacts,” Nat. Commun 9, 2869 (2018).
  62. [62]C. Gong, L. Li, Z. Li, H. Ji, A. Stern, Y. Xia, T. Cao, W. Bao, C. Wang, Y. Wang, Z. Q. Qiu, R. J. Cava, S. G. Louie, J. Xia, and X. Zhang, “Discovery of intrinsic ferromagnetism in two-dimensional van der Waals crystals,” Nature 546, 265 (2017).
  63. [63]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, and X. Xu, “Layer-dependent ferromagnetism in a van der Waals crystal down to the monolayer limit,” Nature 546, 270 (2017).DOI
  64. [64]C. Gong and X. Zhang, “Two-Dimensional Magnetic Crystals and Emergent Heterostructure Devices,” Science 363, eaav4450 (2019).DOI
  65. [65]A. Avsar, J. Y. Tan, T. Taychatanapat, J. Balakrishnan, G. K. W. Koon, Y. Yeo, J. Lahiri, A. Carvalho, A. S. Rodin, E. C. T. O’Farrell, G. Eda, A. H. Castro Neto, and B. Özyilmaz, “Spin-orbit proximity effect in graphene,” Nat. Commun. 5, 4875 (2014).
  66. [66]W. Yan, O. Txoperena, R. Llopis, H. Dery, L. E. Hueso, and F. Casanova, “A two-dimensional spin field-effect switch,” Nat. Commun. 7, 13372 (2016).DOI
  67. [67]W. Han, R. K. Kawakami, M. Gmitra, and J. Fabian, “Graphene spintronics,” Nat. Nanotechnol. 9, 794 (2014).DOI
  68. [68]A. Dankert and S. P. Dash, “Electrical gate control of spin current in van der Waals heterostructures at room temperature,” Nat. Commun. 8, 16093 (2017).DOI
  69. [69]Z. Wang, D.-K. Ki, H. Chen, H. Berger, A. H. MacDonald, and A. F. Morpurgo, “Strong interface-induced spin-orbit interaction in graphene on WS₂,” Nat. Commun. 6, 8839 (2015).DOI
  70. [70]L. Antonio Benitez, J. F. Sierra, W. Savero Torres, M. Timmermans, M. V. Costache, and S. O. Valenzuela, “Investigating the spin-orbit interaction in van der Waals heterostructures by means of the spin relaxation anisotropy,” APL Mater. 7, 120701 (2019).DOI
  71. [71]K. Douli, M. D. Petrović, K. Zollner, P. Plecháč, J. Fabian, and B. K. Nikolić, “Proximity spin-orbit torque on a two-dimensional magnet within van der Waals heterostructure: Current-driven antiferromagnet-to-ferromagnet reversible nonequilibrium phase transition in bilayer CrI₃,” Nano Lett. 20, 2288 (2020).
  72. [72]K. S. Burch, D. Mandrus, and J.-G. Park, “Magnetism in two-dimensional van der Waals materials,” Nature 563, 47 (2018).DOI
  73. [73]Y. Deng, Y.Yu, Y. Song, J. Zhang, N. Z. Wang, Z. Sun, Y. Yi, Y. Z. Wu, S. Wu, J. Zhu, J. Wang, X. H. Chen, and Y. Zhang, “Gate-tunable room-temperature ferromagnetism in two-dimensional Fe₃GeTe₂,” Nature 563, 94 (2018).
  74. [74]W. Xing, Y. Chen, P. M. Odenthal, X. Zhang, W. Yuan, T. Su, Q. Song, T. Wang, J. Zhong, S. Jia, X. C. Xie, Y. Li, and W. Han, “Electric field effect in multilayer Cr₂Ge₂Te₆: A ferromagnetic 2D material,” 2D Mater. 4, 024009 (2017).DOI
  75. [75]K. D. Belashchenko, A. A. Kovalev, and M. van Schilfgaarde, “Interfacial contributions to spin-orbit torque and magnetoresistance in ferromagnet/heavy-metal bilayers,” Phys. Rev. B 101, 020407(R) (2020).DOI
  76. [76]L. P. Gor’kov and E. I. Rashba, “Superconducting 2D system with lifted spin degeneracy: Mixed single-triplet state,” Phys. Rev. Lett 87, 037004 (2001).
  77. [77]A. Costa, P. Högl, and J. Fabian, “Magnetoanisotropic Josephson effect due to interfacial spin-orbit fields in superconductor/ferromagnet/superconductor junctions,” Phys. Rev. B 95, 024514 (2017).
  78. [78]T. S. Khaire, M. A. Khasawneh, W. P. Pratt, Jr., and N. O. Birge, “Observation of spin-triplet superconductivity in Co-based Josephson junctions,” Phys. Rev. Lett. 104, 137002 (2010).pubmed.ncbi.nlm.nih.gov/20481906
  79. [79]J. W. A. Robinson, J. D. S. Witt, and M. G. Blamire, “Controlled injection of spin-triplet supercurrents into a strong ferromagnet,” Science 329, 59 (2010).DOI
  80. [80]N. Banerjee, J. W. A. Robinson, and M. G. Blamire, “Reversible control of spin-polarised supercurrents in ferromagnetic Josephson junctions,” Nat. Commun. 5, 4771 (2014).DOI
  81. [81]O. T. Valls, Superconductor/Ferromagnet Nanostructures (World Scientific, Hackensack, NJ, 2022).DOI
  82. [82]A. Fornieri, A. M. Whiticar, F. Setiawan, E. Portolés, A. C. C. Drachmann, A. Keselman, S. Gronin, C. Thomas, T. Wang, R. Kallaher, G. C. Gardner, E. Berg, M. J. Manfra, A. Stern, C. M. Marcus, and F. Nichele, “Evidence of topological superconductivity in planar Josephson junctions,” Nature 569, 89 (2019).DOI
  83. [83]H. Ren, F. Pientka, S. Hart, A. T. Pierce, M. Kosowsky, L. Lunczer, R. Schlereth, B. Scharf, E. M. Hankiewicz, L. W. Molenkamp, B. I. Halperin, and A. Yacoby, “Topological superconductivity in a phase-controlled Josephson junction,” Nature 569, 93 (2019).
  84. [84]M. C. Dartiailh, W. Mayer, J. Yuan, K. S. Wickramasinghe, A. Matos-Abiague, I. Žutić, and J. Shabani, “Phase signature of topological transition in Josephson junctions,” Phys. Rev. Lett. 126, 036802 (2021).DOI
  85. [85]T. Zhou, M. C. Dartiailh, K. Sardashti, J. E. Han, A. Matos-Abiague, J. Shabani, and I. Žutić, “Fusion of Majorana bound states with mini-gate control in two-dimensional systems,” Nat. Commun. 13, 1738 (2022).DOI
  86. [86]A. Banerjee, O. Lesser, M. A. Rahman, H.-R. Wang, M.-R. Li, A. Kringhøj, A. M. Whiticar, A. C. C. Drachmann, C. Thomas, T. Wang, M. J. Manfra, E. Berg, Y. Oreg, A. Stern, and C. M. Marcus, “Signatures of a topological phase transition in a planar Josephson junction,” arXiv:2201.03453 (2022).DOI
  87. [87]I. Žutić, “Novel aspects of spin-polarized transport and spin dynamics,” J. Supercond. 15, 5 (2002).
  88. [88]E. I Rashba, “Spintronics: Sources and challenge. Personal perspective,” J. Supercond. 15, 13 (2002).DOI

Paper details

vlnc:W_nzewephx
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents