Long-Range Spin-Triplet Supercurrents

• Long-range spin-triplet supercurrents are defined by equal-spin Cooper pairs that convert from singlets via noncollinear magnetizations or spin-active contacts, extending coherence beyond typical limits.
• Experimental platforms like CrO₂ and LSMO show high critical current densities and clear signatures of triplet transport through both ferromagnets and half metals.
• Theoretical and experimental studies highlight tunable mechanisms—including spin-orbit effects, geometric curvature, and helimagnetic textures—that can control triplet generation for superconducting spintronics.

Long-range spin-triplet supercurrents are Josephson supercurrents that traverse ferromagnets, strong ferromagnets, or half metals over distances far exceeding the usual singlet ferromagnetic coherence length because the propagating Cooper pairs are equal-spin triplets rather than conventional opposite-spin singlets. In the standard framework, a conventional $s$-wave superconductor injects spin-singlet pairs, and magnetic inhomogeneity, noncollinear magnetization, or a spin-active interface converts part of that condensate into odd-frequency equal-spin triplet correlations, which are not pair-broken by the exchange field in the same way as singlets and therefore propagate on the long normal-metal or thermal coherence scale (Eschrig, 2015). Experimental and theoretical work has established this phenomenology in metallic multilayers, oxide half metals such as CrO$_2$ and La$_{0.7}$Sr$_{0.3}$MnO$_3$, helimagnet-based junctions, spin-textured single-ferromagnet devices, van der Waals heterostructures, and more recent proposals based on spin-orbit coupling, nonequilibrium control, geometric curvature, and altermagnets (Anwar et al., 2011).

1. Pair symmetry, coherence scales, and the long-range channel

In a homogeneous ferromagnet, conventional spin-singlet Cooper pairs and the opposite-spin triplet component with $S_z=0$ are short-ranged because the exchange field splits the spin bands and causes rapid oscillation and decay. In the dirty limit, one representative scale is

$\xi_F=\sqrt{\hbar D_F/h_{ex}},$

which is only a few nanometers in a ferromagnet with a large exchange field (Anwar et al., 2011). In the review formulation, the ballistic short-range oscillation scale is

$\xi_J \sim \frac{\hbar v_F}{2|J|},$

whereas the long-range equal-spin component decays on the thermal coherence length $\xi_T$, or in the diffusive limit on

$$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$

The central distinction is therefore not whether superconducting correlations exist in a ferromagnet, but which spin sector carries them (Eschrig, 2015).

The long-range channel consists of equal-spin triplets with spin projection $_2$0 along the local magnetization axis. In a half metal, where only one spin band is available at the Fermi level, ordinary singlets cannot propagate in the bulk, while equal-spin triplets can; this is why observation of a supercurrent through a fully spin-polarized half metal is taken as direct evidence for long-range triplet correlations (Junxiang et al., 2023). CrO$_2$1 and LSMO experiments exploit exactly this logic: the supercurrent cannot be attributed to ordinary singlets because the weak link is fully spin-polarized or half-metallic (Anwar et al., 2011).

A major conceptual point is that the long-range equal-spin component in diffusive $_2$2 hybrids is typically odd in frequency and $_2$3-wave in orbital symmetry. The review classifies this as Type D pairing: spin-triplet, odd-frequency, even-parity. Because it is $_2$4-wave, it survives impurity averaging; because it is odd in frequency, it remains compatible with Fermi statistics despite even spatial parity (Eschrig, 2015). This odd-frequency structure is the symmetry basis for long-range triplet transport in most diffusive ferromagnetic Josephson systems.

2. Conversion mechanisms from singlets to long-range triplets

The canonical mechanism is magnetic noncollinearity. In multilayer ferromagnetic Josephson junctions of the form $_2$5, the triplet correlations are created by non-collinear magnetizations between a central Co/Ru/Co synthetic antiferromagnet and the two outer thin $_2$6 layers. Theory identified the optimal condition as orthogonal alignment, and experiment showed that applying a large in-plane magnetizing field can increase the critical current by as much as a factor of $_2$7, consistent with a spin-flop transition in the Co/Ru/Co synthetic antiferromagnet that leaves the Co moments approximately perpendicular to the outer $_2$8 layers (Klose et al., 2011). In this class of devices, the magnetic structure is an active converter, not merely a passive barrier.

A closely related interfacial mechanism is spin mixing at engineered spin-active contacts. In CrO$_2$9 junctions on TiO$_{0.7}$0, insertion of a $_{0.7}$1 nm Ni layer and $_{0.7}$2 nm Cu spacer between the half metal and amorphous Mo$_{0.7}$3Ge$_{0.7}$4 leads was interpreted as furnishing a spin-active Ni/Cu/CrO$_{0.7}$5 interface. The intended role of Ni is magnetic spin mixing, while Cu magnetically decouples Ni from CrO$_{0.7}$6 while preserving spin-dependent scattering. The resulting multilayer appears to generate the odd-frequency equal-spin triplet correlations needed to carry supercurrent through the half metal over almost a micrometer (Anwar et al., 2011).

Helical, conical, and domain-wall textures supply a more distributed form of spin rotation. In the Ho$_{0.7}$7Co$_{0.7}$8Ho problem, the spiral Ho layers continuously rotate the spin quantization axis of the injected Cooper pairs and thereby mix singlet and triplet sectors; full numerical diffusive calculations found qualitatively very good agreement with experiment for the behavior of the supercurrent as a function of Co-layer thickness, and similar behavior was obtained when the Ho layers were replaced by domain-wall ferromagnets (Alidoust et al., 2010). In the clean-limit $_{0.7}$9 formulation motivated by Nb/Ho/Co/Ho/Nb, the singlet reaching the right superconductor factorizes into conversion in the first helimagnet, propagation through the ferromagnet, and reconversion in the second helimagnet. The conversion efficiency

$_{0.3}$0

makes the maximum triplet supercurrent strongly dependent on helimagnet thickness and exchange-coupling angle $_{0.3}$1, producing a non-monotonic dependence on $_{0.3}$2 (Halász et al., 2012).

Several later works generalize triplet generation beyond explicit magnetic textures. A diffusive Keldysh-Usadel analysis showed that long-range triplet currents can be generated very generically when an ordinary singlet supercurrent flows parallel to a noncollinear $_{0.3}$3 interface, independent of actual junction geometry (Alidoust et al., 2015). In a lateral junction with a single in-plane ferromagnet contacted through thin heavy normal metals, Rashba spin-orbit coupling in the contacts can convert singlets into long-ranged $_{0.3}$4 triplets using only an in-plane magnetization rotation, without magnetic inhomogeneity or an out-of-plane component (Eskilt et al., 2019). A further extension showed that geometric curvature in a diffusive SFS junction can act as a synthetic spin-rotation mechanism, effectively replacing magnetic misalignment or intrinsic SOC and generating long-range triplets together with a curvature-tunable $_{0.3}$5-$_{0.3}$6 transition (Salamone et al., 2021). Another proposal identified dc supercurrent in the superconducting leads, flowing parallel to the S/F interfaces in the presence of extrinsic impurity SOC, as a generator of long-range triplet correlations via a superconducting spin Hall effect and spin current swapping (Mazanik et al., 2021).

3. Material platforms and junction architectures

Long-range spin-triplet supercurrents are not tied to a single material class or device geometry. They have been reported in half-metallic oxides, transition-metal multilayers, helimagnet/ferromagnet hybrids, single-ferromagnet spin textures, and van der Waals heterostructures. The table summarizes representative platforms and the corresponding triplet generator.

System	Triplet generator	Representative signature
CrO$_{0.3}$7 lateral junction on TiO$_{0.3}$8 (Anwar et al., 2011)	Ni(2 nm)/Cu(5 nm) spin-active multilayer contact	supercurrent over almost $_{0.3}$9m; $_3$0
$_3$1 with Co/Ru/Co SAF (Klose et al., 2011)	field-induced orthogonality between outer $_3$2 layers and SAF	critical current enhancement up to $_3$3
NbTi/LSMO lateral bar, square, and disk junctions (Junxiang et al., 2023)	magnetic inhomogeneity at, or close to, the LSMO/NbTi interface	current densities of order $_3$4 in the disk estimate
Nb/Ho/Co/Ho/Nb and ideal $_3$5 (Halász et al., 2012)	helimagnetic conversion and reconversion at HM/F interfaces	non-monotonic dependence on Ho thickness
NbS$_3$6/Cr$_3$7NbS$_3$8 vdW bilayer (Spuri et al., 2023)	helimagnetic spin texture at the vdW interface	positive jump in $_3$9 near $S_z=0$0
Disk-shaped Nb/Co bilayer junction (Fermin et al., 2021)	vortex texture acting as effective SOC	sub-80 nm rim-localized triplet channels

The CrO$S_z=0$1 and LSMO half-metal devices are especially important because they test triplet transport in the strongest spin-filtering limit. In CrO$S_z=0$2, lateral Josephson junctions with $S_z=0$3, $S_z=0$4, and $S_z=0$5 nm gaps on unstructured $S_z=0$6 nm films produced supercurrents in $S_z=0$7 out of $S_z=0$8 devices, with critical current densities about $S_z=0$9 times larger than earlier sapphire-based CrO$\xi_F=\sqrt{\hbar D_F/h_{ex}},$0 junctions and of similar magnitude to the earlier landmark CrO$\xi_F=\sqrt{\hbar D_F/h_{ex}},$1 result of Keizer et al. (Anwar et al., 2011). In LSMO, the electrode separation is only about $\xi_F=\sqrt{\hbar D_F/h_{ex}},$2 nm, placing the system at the boundary of short and mesoscopic behavior while still yielding robust Josephson coupling and very high current density, which the authors interpret as evidence for an efficient triplet generator at the interface (Junxiang et al., 2023).

The materials palette now extends to low-dimensional magnets and superconductors. In NbS$\xi_F=\sqrt{\hbar D_F/h_{ex}},$3/Cr$\xi_F=\sqrt{\hbar D_F/h_{ex}},$4NbS$\xi_F=\sqrt{\hbar D_F/h_{ex}},$5, the helimagnetic state of the magnetic layer is tuned by an in-plane field that unwinds the helix into a homogeneous ferromagnet above $\xi_F=\sqrt{\hbar D_F/h_{ex}},$6, about $\xi_F=\sqrt{\hbar D_F/h_{ex}},$7 T in the experiment. The resulting field dependence of the bilayer $\xi_F=\sqrt{\hbar D_F/h_{ex}},$8 is used as evidence for long-ranged triplet generation across a van der Waals interface (Spuri et al., 2023). This suggests that strong covalent bonding is not a prerequisite for triplet proximity if the interface transparency is sufficiently good.

4. Experimental signatures and diagnostics

Transport signatures are the first line of evidence. In the CrO$\xi_F=\sqrt{\hbar D_F/h_{ex}},$9 devices, clear zero-voltage branches were observed: for example, one junction showed a zero-resistance branch up to about $\xi_J \sim \frac{\hbar v_F}{2|J|},$0 mA at $\xi_J \sim \frac{\hbar v_F}{2|J|},$1 K before switching to an Ohmic branch with $\xi_J \sim \frac{\hbar v_F}{2|J|},$2 m$\xi_J \sim \frac{\hbar v_F}{2|J|},$3, and the critical current was defined by a $\xi_J \sim \frac{\hbar v_F}{2|J|},$4V criterion. Using the bridge geometry and $\xi_J \sim \frac{\hbar v_F}{2|J|},$5 nm CrO$\xi_J \sim \frac{\hbar v_F}{2|J|},$6 thickness, the estimated critical current density at $\xi_J \sim \frac{\hbar v_F}{2|J|},$7 K was about $\xi_J \sim \frac{\hbar v_F}{2|J|},$8, while a normal-state resistance estimate based on the bridge dimensions was much larger than the few m$\xi_J \sim \frac{\hbar v_F}{2|J|},$9 residual resistance below $\xi_T$0, supporting the interpretation that the supercurrent traverses the CrO$\xi_T$1 bridge rather than shorting through the contacts (Anwar et al., 2011).

Temperature dependence can discriminate between competing regimes. In the LSMO/NbTi disk device, the critical current followed

$\xi_T$2

with $\xi_T$3 K for the proximized device. The authors contrast this with the approximately linear $\xi_T$4 expected near $\xi_T$5 for ordinary short Josephson junctions with a normal-metal interlayer, and interpret the quadratic law as the clean half-metal triplet result predicted near $\xi_T$6. They also estimated $\xi_T$7 nm, $\xi_T$8 nm, and $\xi_T$9 meV, larger than $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$0 meV, to argue that the weak link is not a long diffusive SNS junction (Junxiang et al., 2023).

Magnetic-field interference patterns are similarly diagnostic, but they are often non-Fraunhofer in triplet devices. The CrO$$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$1 junctions showed no clear Fraunhofer oscillations; one junction dropped rapidly below about $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$2 mT, another showed a shoulder near $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$3 mT, and the supercurrent was not quenched even at $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$4 mT, which was interpreted as evidence for residual magnetic inhomogeneity in the Ni/Cu/CrO$$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$5 structure (Anwar et al., 2011). In LSMO, by contrast, the perpendicular-field $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$6 patterns were strongly geometry dependent: bar-shaped devices showed a conventional single-junction Fraunhofer-like pattern, whereas disk-shaped devices exhibited a two-channel-like pattern with slowly decaying oscillations, consistent with current concentrated near the rim. Dynes-Fulton reconstruction yielded a channel width of about $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$7 nm in the disk (Junxiang et al., 2023).

Direct magnetic and structural probes have been decisive in separating transport phenomenology from microscopic origin. In the Co/Ru/Co synthetic-antiferromagnet junctions, polarized neutron reflectometry and scanning electron microscopy with polarization analysis supplied direct evidence for a spin-flop transition. After applying a $$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$8 T in-plane field, spin-flip scattering increased strongly while non-spin-flip splitting nearly disappeared, and SEMPA showed that remanent Co magnetizations tilt away from the applied field in a manner consistent with the relaxed scissor-like spin-flop state that optimizes triplet generation (Klose et al., 2011). In LSMO/NbTi, STEM-EELS revealed a sharp but disordered interface with a small oxygen peak at the interface in two out of three samples, while the Mn valence stayed essentially constant through the LSMO layer up to the interfacial region; the authors therefore locate the triplet generator at, or close to, the interface rather than in the bulk LSMO (Junxiang et al., 2023).

A complementary diagnostic dispenses with direct Josephson transport and tracks how magnetic inhomogeneity drains pairs from the superconductor. In the NbS$$\xi_T \sim \sqrt{\frac{\hbar D}{2(\pi k_B T+\alpha_s)}}.$$9/Cr$_2$00NbS$_2$01 heterostructure, the key observation is a positive jump in $_2$02 when the helimagnet is driven through $_2$03 into a homogeneous ferromagnet. The authors argue that this behavior is opposite to what conventional short-ranged singlet proximity or stray fields would produce, and use a control device with an hBN spacer to separate proximity effects from stray-field effects (Spuri et al., 2023). This establishes $_2$04 itself as a triplet diagnostic in low-dimensional superconductor/helimagnet systems.

5. Alternative regimes, superharmonics, and points of dispute

Although the dominant interpretation of long-range supercurrents invokes equal-spin triplets, the literature also contains nonstandard regimes in which spin and charge transport separate or in which the long-range current is attributed to a different pairing sector. A quasiclassical study of $_2$05 with noncollinear ferromagnets found that the spin current through $_2$06 obeys

$_2$07

while the Josephson charge current is

$_2$08

In that formulation, the long-range part of the spin current is carried mainly by equal-spin spin-triplet Cooper pairs, whereas the charge current is carried only by singlet pairs, so that the spin current can remain finite even where the Josephson charge current is practically zero (Hikino et al., 2013). A related clean-limit study of asymmetric $_2$09 junctions found a dominant second harmonic,

$_2$10

generated by equal-spin triplets in the thick $_2$11 layer; in the half-metal limit of $_2$12, the Josephson current is prohibited but $_2$13 remains present throughout $_2$14, which the authors describe as spin-charge separation (Meng et al., 2014).

Several nonequilibrium proposals move the triplet generator from the ferromagnet into adjacent regions. In one SFS proposal, spin-polarized quasiparticles are injected into the superconducting leads, producing a spin accumulation

$_2$15

and an effective Zeeman splitting

$_2$16

This exchange-renormalized nonequilibrium spin polarization generates triplet pairing in the $_2$17-wave superconductors, and the long-range Josephson current scales as

$_2$18

with sign reversal and $_2$19- to $_2$20-junction switching controlled by the injected spin directions (Mal'shukov et al., 2012). In a different clean $_2$21 construction, antiparallel half-metal electrodes create spin-dependent quasiparticle distributions in the normal metals,

$_2$22

which endow Cooper pairs with an extra momentum

$_2$23

Equal biases on both $_2$24 layers yield a long-range triplet current, reversing one bias reverses the current direction, and biasing only one $_2$25 layer produces a superharmonic $_2$26-periodic response with $_2$27 (Meng et al., 2013).

The most explicit departure from the standard equal-spin narrative comes from clean SFS analyses that attribute the long-range current mainly to singlet or opposite-spin-triplet interference. One model of an $_2$28 junction with a noncollinear magnetic domain claims that equal-spin triplets are generated only locally in the middle domain and do not diffuse into the outer ferromagnets; instead, the long-range effect arises because spin flips in the middle domain cancel the exchange phases acquired in the two symmetric outer ferromagnets, leaving an additional $_2$29 phase shift and a supercurrent that mostly arises from singlet Cooper pairs (Meng et al., 2014). Another clean $_2$30 study argues that the enhancement for antiparallel, equal-thickness ferromagnets is due primarily to destructive interference between opposite-spin triplet correlations generated at the two $_2$31 interfaces, rather than equal-spin triplets (Meng et al., 2016). These proposals do not overturn the broader equal-spin framework, but they show that “long-range” does not automatically specify a unique microscopic carrier in every clean-limit geometry.

6. Spintronics significance and emerging directions

Long-range spin-triplet supercurrents underpin what the review terms super-spintronics: the combination of macroscopic superconducting phase coherence with microscopic exchange interactions, spin selectivity, and spin transport (Eschrig, 2015). The practical attraction is dissipationless spin-polarized transport through strongly spin-polarized materials, with potential consequences for superconducting spin valves, cryogenic memory, magnetic Josephson logic, phase batteries, and spin-current-controlled magnetization dynamics. Experiments in half metals sharpen this prospect because they combine large critical current density with extreme spin selectivity; the LSMO/NbTi work explicitly frames order-$_2$32 current densities as promising for superconducting spintronics (Junxiang et al., 2023).

A central trend is the replacement of static magnetic inhomogeneity by tunable control parameters. Field history reconfigures the Co/Ru/Co synthetic antiferromagnet into an orthogonal state that maximizes triplet generation (Klose et al., 2011). Supercurrent in the superconducting leads can act as a control knob in the presence of extrinsic impurity SOC, where reversing the lead supercurrent switches the ground state between $_2$33 and $_2$34, realizing a new physical principle of the $_2$35-$_2$36 shifter (Mazanik et al., 2021). Geometric curvature offers a purely structural tuning parameter for long-range triplet generation and dynamic $_2$37-$_2$38 switching in diffusive SFS junctions (Salamone et al., 2021). Vortex texture in a single Co disk produces rim-localized triplet channels that can be reconfigured into $_2$39-like or $_2$40-$_2$41-like states by a weak in-plane field (Fermin et al., 2021).

Low-dimensional and magnetization-free routes broaden the design space further. The van der Waals NbS$_2$42/Cr$_2$43NbS$_2$44 system indicates that long-ranged triplet generation can survive across a weakly bonded two-dimensional interface when the magnetic layer is helimagnetic and the interface transparency is good (Spuri et al., 2023). A more radical proposal identifies nodeless altermagnets as a platform for long-range spin-polarized Josephson transport with zero net magnetization. In that scenario, spin-valley-locked, spin-split Fermi surfaces support a pure triplet supercurrent in a $_2$45-aligned junction, the sign of the critical current is controlled by the product $_2$46 of interface Rashba couplings, and rotating the junction to $_2$47 yields a crossover from pure triplet to mixed singlet-triplet transport (Li et al., 17 Sep 2025). This suggests that long-range spin-triplet supercurrents need not remain tied to ferromagnetism in the conventional sense.

The field therefore contains two simultaneously valid narratives. One is mature and experimentally established: equal-spin odd-frequency triplets generated by magnetic inhomogeneity or spin-active scattering carry Josephson currents through ferromagnets and half metals over anomalously long distances (Eschrig, 2015). The other is still expanding: current bias, voltage bias, SOC, curvature, spin texture, helimagnetism, and valley-selective magnetic order all function as new control handles on how triplet supercurrents are generated, distributed, and phase-programmed. A plausible implication is that future devices will be classified less by whether they host long-range triplets and more by which conversion mechanism, control knob, and readout geometry they implement.