Spin-singlet to triplet Cooper pair converter interface

Conventional theories hold that ferromagnetism and superconductivity are antagonistic because the ferromagnetic exchange field breaks conventional spin-singlet superconducting pair correlations. However, local magnetic order at metallo-molecular multilayers can coexist with superconducting proximity effects, enabling regions where both magnetic exchange and diffusion of superconducting pairs are present. Prior work has shown coexistence of superconductivity and magnetism in designed materials, influence of superconducting order on molecular magnets, and evidence for equal-spin triplet p-wave symmetry in proximity systems with chiral molecules. For super-spintronic devices, generating long-range S_z=±1 triplet components is essential, as these are robust in ferromagnets. Odd-frequency spin-triplet states provide a route to synergy between superconductivity and spintronics. Two key challenges for applications are the need for complex magnetic/superconducting materials that can impair qubit coherence and the limited tunability of superconducting/spin correlations without regrowing samples. Inorganic–molecular interfaces can host emergent electronic and magnetic surface states. This study investigates carbon-based metallo-molecular hybrids to manipulate superconducting correlations without intrinsic magnetic or superconducting materials in the interface, aiming to convert spin-singlet Cooper pairs into odd-frequency triplet states at a Cu/C60 interface and to probe their spatial distribution and Meissner response.

Odd-frequency superconducting correlations have been probed via transport, tunnelling spectroscopy, and low-energy muon spin rotation (LE-µSR), which directly detects paramagnetic Meissner screening generated by odd-frequency triplet pairs. Super-spintronic devices rely on the long-range equal-spin triplet component, immune to pair breaking in ferromagnets. Evidence for unconventional superconductivity induced by chiral molecules indicates molecular control over pair symmetry. Theoretical predictions show that spin-active or ferromagnetic interfaces can convert singlet to triplet correlations and generate odd-frequency pairing. Prior LE-µSR and reflectometry studies have observed anomalous Meissner screening and electromagnetic proximity effects in superconductor/normal or superconductor/ferromagnet systems. These bodies of work motivate probing molecular interfaces as triplet converters and using LE-µSR as a sensitive depth-resolved probe of Meissner responses.

• Sample architectures: Two classes of hybrid thin-film heterostructures were studied: (i) S/M/N: Nb/C60/Au, and (ii) S/M/I_sc/M/N: Nb/C60/Cu/C60/Au, where I_sc denotes a spin-converter interface formed at Cu/C60. Nb is a conventional s-wave superconductor (T_c ~ 9 K). Thicknesses were engineered so that the Au–Nb separation was ≤ the Cooper-pair propagation length in C60 (~30 nm).
• Structural characterization: Transmission electron microscopy (TEM) to assess layer continuity and diffusion; energy-dispersive X-ray analysis to confirm Cu and Au separation by C60 barriers; X-ray reflectivity (XRR) and polarized neutron reflectivity (PNR) to determine layer thickness, roughness, and density. Reflectivity data were modeled with GenX, yielding coherent structural parameters for SiO2/Nb(50)/C60(10)/Cu(3)/C60(10)/Au(20) nm stacks.
• Transport measurements: Low-temperature four-probe transport to measure T_c suppression in Nb(50)/C60(t)/Au(20)/Al(2) stacks, with t = 0–80 nm, alongside a co-grown Nb control on the same substrate for normalization. An exponential decay model (Ginzburg–Landau) was fitted to T_c suppression versus C60 thickness to extract an effective Cooper pair propagation length ξ_eff in C60. Discontinuous C60 (~4 nm islanded) samples served as controls for interface transparency. Upper critical field H_c2⊥(T) was measured to extract Ginzburg–Landau coherence length ξ_GL.
• Magnetometry: SQUID-VSM magnetometry measured Meissner onset and room-temperature hysteresis to quantify emergent interfacial magnetization from Cu/C60.
• Low-energy muon spin rotation (LE-µSR): Performed in transverse field (300 G) with implantation energies tuned to probe specific depths, using two designs: • Single-Probe: Nb(50)/C60(20)/Au(50) and Nb(50)/C60(10)/Cu(3)/C60(10)/Au(50). London model fits to B(x) within Nb yielded temperature-dependent penetration depth λ_L(T). Muon stopping profiles converted B(E) to depth-averaged B(x). • Multi-Probe: S/M/N: Nb(50)/C60(30)/Au1(10)/C60(20)/Au2(10)/C60(20)/Au3(10); S/M/I_sc/M/N: Nb(50)/C60(15)/Cu(4)/C60(15)/Au1(10)/C60(20)/Au2(10)/C60(20)/Au3(10). LE-µSR at 9 keV (probing Au1, no muons in Nb) and 20 keV (probing inside Nb) to separate screening in Au and Nb. Data analysis used time-dependent muon polarization fits to extract local field B(E), depolarization rate, asymmetry, and phase; London equation B(x)=B_y cosh(x/λ_L)/cosh(d_Nb/2λ_L) modeled flux profiles within Nb.
• Quasiclassical (Q-C) modeling: Dirty-limit Usadel equation with Nazarov boundary conditions used to compute Green functions and linear-response screening currents for Nb/C60/Au, comparing S-only and proximitized N/S models, to reproduce B(T) trends and penetration depths.
• Thin film growth: DC magnetron sputtering for metals (Nb, Cu, Au) and in-situ thermal sublimation for C60 at 450 °C under UHV (base ~1e-8 Torr, O2 ~1e-10 Torr). Argon sputter at 2.4 Torr from high-purity targets. Growth rates calibrated by XRR Kiessig fringes. PNR on ISIS PolRef; LE-µSR on PSI µE4.
• Data availability: Raw data DOIs provided in the paper.

• Proximity through C60: T_c suppression versus C60 thickness in Nb(50)/C60(t)/Au(20) yields an effective Cooper-pair propagation length in C60 of ξ_eff = 30 ± 8 nm. Discontinuous C60 (~3.9 ± 0.1 nm islands) shows almost no T_c suppression, indicating reduced interface transparency and disrupted Andreev reflection when C60 is not continuous.
• Single-Probe LE-µSR: London-model fits to B(x) in Nb show penetration depth λ(2.5 K) = 218 nm for S/M/N versus λ(2.5 K) = 313 nm for S/M/I_sc/M/N, signifying a strong modification of screening in the presence of the Cu/C60 spin-converter. Despite this, superconductivity is not suppressed (slightly higher T_c with I_sc), suggesting added paramagnetic contribution rather than pair breaking.
• Magnetism at Cu/C60: Room-temperature magnetic hysteresis indicates emergent interfacial magnetization for samples with Cu/C60: moment per area m = 27 ± 1 µemu cm^-2 in Single-Probe and 24 ± 1 µemu cm^-2 in Multi-Probe, versus background 2 ± 1 µemu cm^-2 and 0.4 ± 0.2 µemu cm^-2, respectively, without Cu.
• GL coherence length: From H_c2⊥(T), ξ_GL ≈ 9.34 ± 0.02 nm (without I_sc) and 9.63 ± 0.02 nm (with I_sc), typical of sputtered Nb in the dirty limit; T_c ~ 8.7 K in both sample sets.
• Multi-Probe LE-µSR depth resolution: • Control S/M/N: At 20 keV (Nb), Meissner screening reduces internal field by 2.41 ± 0.08 G between 10 K and 2.7 K. At 9 keV (Au1), a small diamagnetic reduction 0.21 ± 0.08 G at 2.7 K indicates even-frequency correlations leaking into Au. • With spin-converter S/M/I_sc/M/N: At 20 keV (Nb), Meissner reduction is suppressed to 1.0 ± 0.1 G at 2.7 K. At 9 keV (Au1), a paramagnetic increase of 0.3 ± 0.1 G at 2.7 K is observed, evidencing odd-frequency triplet correlations dominating in Au.
• Temperature dependence: With I_sc, the reduction of B in Nb sets in only below ~6 K (~0.7 T_c), unlike the control which follows conventional behavior; Q-C modeling reproduces control trends when including proximitized C60 but cannot account for the anomalous response with I_sc without invoking odd-frequency triplet contributions.
• Alternative mechanisms: Possible sources of paramagnetic signals (Andreev bound state instabilities, lattice inhomogeneity, pinned flux, orbital paramagnetism) are ruled out by controls, morphology consistency (TEM/XRR), lack of increased muon depolarization below T_c, and applied transverse field (300 G) suppressing orbital effects. The spatial localization of the paramagnetic response in Au supports a triplet origin.

The experiments demonstrate that a continuous C60 layer can transmit superconducting phase information over tens of nanometers via proximity, consistent with interfacial metallization and charge transfer altering the electronic structure of C60. Introducing a Cu/C60 spin-converter interface creates emergent interfacial magnetism and converts singlet Cooper pairs into odd-frequency triplet correlations. LE-µSR reveals two hallmarks of such conversion: (i) suppression of diamagnetic Meissner screening in Nb (increased λ and reduced flux expulsion) and (ii) a local paramagnetic Meissner signal in the adjacent Au layer, where the triplet density outweighs the singlet contribution (j ∝ n_s − n_t). The depth-resolved Multi-Probe measurements cleanly separate the Au-layer response from Nb, confirming that the paramagnetic susceptibility is concentrated in the normal metal proximitized through C60. Quasiclassical modeling with Usadel equations captures the control sample behavior only when including proximitized C60, validating that Cooper pairs leak through C60. However, the anomalous temperature dependence and paramagnetic signature in the spin-converter samples require odd-frequency triplet correlations generated at the Cu/C60 interface. Control experiments and structural characterization exclude alternative sources of paramagnetism (pinned vortices, orbital effects, inhomogeneities), strengthening the interpretation that metallo-molecular interfaces can act as singlet-to-triplet pair converters. This establishes a route to engineer super-spintronic functionality without incorporating bulk ferromagnets directly into the superconducting path, potentially reducing decoherence and enabling tunability via interface chemistry.

Superconducting correlations can traverse thin-film C60 barriers over an effective length of ~30 nm, enabling proximity coupling across molecular layers. A Cu/C60 metallo-molecular interface induces spin order and converts spin-singlet pairs into odd-frequency triplet states, as evidenced by suppressed diamagnetic Meissner screening in Nb and a paramagnetic Meissner response localized in Au detected by LE-µSR. These findings reveal unconventional superconducting states in hybrid metal/molecule systems and demonstrate that metallo-molecular interfaces can function as singlet-to-triplet converters absent in the constituent materials alone. Future directions include integrating a spin-selective ferromagnetic layer to filter and generate S_z = 1 long-range triplets for super-spintronic devices, and exploiting electrical or optical control of interfacial charge transfer to switch and modulate odd-frequency correlations for quantum computing, coherence, and interferometry applications.