The Spinterface Mechanism for CISS: A Critical Perspective

A 2025 perspective article by Subhajit Sarkar (Shiv Nadar Institution of Eminence), Amos Sharoni (Bar Ilan University), Oliver L. A. Monti (University of Arizona) and Yonatan Dubi (Ben Gurion University of the Negev), published on arXiv (2510.21061), evaluates the "spinterface model" as a candidate unifying explanation for the Chirality-Induced Spin Selectivity (CISS) effect. The paper is unusual in the CISS literature for two reasons: it openly catalogues the model's unresolved weaknesses alongside its strengths, and it claims that the spinterface approach is the only CISS theory to date that has quantitatively reproduced raw experimental data such as current–voltage curves.

What the spinterface model proposes

The central hypothesis is that fluctuating magnetic moments exist at the surface of the metal electrode used in CISS experiments. In the absence of a chiral molecule these moments have no preferred orientation. When a chiral molecule is adsorbed or contacted to the surface, electron motion through the molecule generates a small "solenoid" magnetic field aligned with the molecular chiral axis. This field is far too weak to account for observed spin polarisations on its own — the authors estimate it is roughly seven orders of magnitude too small — but the model assigns it the role of a symmetry-breaking seed.

The amplification comes from a feedback loop. The solenoid field tilts the surface moments slightly, and these tilted moments interact with electrons crossing the interface through spin-orbit coupling or exchange interactions, creating a spin-dependent shift in molecular orbital energies. The resulting spin imbalance in the molecule exerts further torque on the surface moments, tilting them more, which in turn increases the spin imbalance. Dissipation — attributed to coupling with the electron reservoir, phonons, or other degrees of freedom — damps the coherent oscillations and stabilises the moments in their tilted orientation. The steady state produces a measurable magnetoresistance.

The paper distinguishes two variants. The "dynamical" spinterface model applies to transport experiments: a bias voltage drives current, and the feedback loop runs during measurement. The "displacement" variant explains photoemission and magnetisation experiments: spontaneous displacement currents flow when a chiral molecule first contacts a metal surface, stabilising the surface moments even without applied bias.

Why the authors argue other theories fall short

The perspective identifies three requirements any CISS theory must satisfy: it must incorporate molecular chirality, include interactions beyond single-particle spin-orbit coupling, and provide a mechanism for time-reversal symmetry breaking (which the authors attribute to dissipation).

Two competing frameworks are examined in detail. The Guo–Sun tight-binding model, which combines double-stranded molecular geometry, discretised spin-orbit interactions, and Büttiker-probe dephasing, is criticised on two grounds. First, the quantity it calculates — the spin polarisation of differential conductance — is not the quantity measured in transport experiments; as Fransson showed, the relationship between spin polarisation and CISS magnetoresistance is non-trivial, and one can vanish while the other is large. Second, the model treats electrons as non-interacting. The authors argue, following a formal demonstration in their earlier work, that any non-interacting theory confines the CISS polarisation to at most the polarisation of the ferromagnetic electrode — typically around 30 % for nickel — while experiments routinely exceed 90 %.

Fransson's electron–chiral-phonon theory, which adds vibrational interactions to amplify spin-orbit effects, is acknowledged as qualitatively successful across several experimental contexts but is said to lack quantitative fits to raw data. The authors point out one case where the chiral-phonon model reproduces the temperature dependence of the polarisation ratio but predicts currents that decrease with temperature while the measured currents increase — a discrepancy they regard as significant.

Quantitative fits to experiment

The paper presents fits to four experimental data sets, each using a different transport model appropriate to the system but the same spinterface parameters for the interface interaction:

1. DNA oligomers of 26, 40 and 50 base pairs, measured in the Naaman group. A classical multi-level hopping model is used for the bare current–voltage curves. The spinterface parameters — electrode spin-orbit coupling and exchange interaction — are held constant across all three lengths; only the molecular hopping parameters change. The resulting theoretical curves follow the experimental data closely.

2. Single maleimide molecule in a molecular junction between a ferromagnetic and a paramagnetic electrode, measured by Yang et al. (2023). The spinterface parameters are fitted at 2 K only; the model then reproduces the current–voltage curves and CISS polarisation at 10–150 K without additional fitting.

3. Chiral molecules intercalated in layered 2D crystals, measured by Qian et al. The authors report that the original experimenters' claim that the CISS effect increases with temperature rests on a misapplication of the Jullière formula, and that the raw data, when fitted by the spinterface model, actually show a decrease.

4. Helicenes on cobalt nano-islands, measured by Safari et al. (2023). Here the dynamical spinterface model is applied, using a single-level Landauer model with a constant spin-dependent energy shift χ ≈ 0.1 eV.

Typical fit parameters are α₀ (solenoid field strength) of 0.1–1 eV and α₁ (spin-transfer torque coefficient) of 1–100 meV, which the authors describe as physically realistic.

GMR-versus-TMR geometry reversal

The perspective highlights an observation by Nguyen et al. (2024) that the authors consider strong evidence for an electrode-surface effect. In a "GMR device" where chiral molecules sit on the ferromagnetic layer and are probed by a gold AFM tip, the majority-current magnetisation direction is one way. In a "TMR device" where molecules coat the AFM tip and are pressed onto the ferromagnetic surface, the majority direction reverses — as though the molecular chirality had been flipped, even though the molecules are identical. The spinterface model explains this by noting that the displacement current flows in the same direction relative to the molecule in both cases, but because the molecules are oriented oppositely relative to the surface, the solenoid field and hence the stabilised surface magnetisation point in opposite directions.

Open questions the authors acknowledge

The paper is forthright about several unresolved issues:

• Origin of surface magnetic moments. The model requires fluctuating magnetic moments at metal surfaces, but their microscopic origin is unknown. The authors note that magnetism in metallic nanostructures, particularly gold, is experimentally well-documented but theoretically unexplained.

• Non-helical chiral molecules. The solenoid-field picture is intuitive for helical molecules but less clear for chiral molecules without a helical backbone. The CISS effect has been measured in such molecules, and the spinterface model fits those data, but the authors acknowledge that computational work is still needed to confirm that helical current-density pathways exist in non-helical chiral structures.

• Temperature dependence. The model predicts that the CISS effect should decrease with increasing temperature, because thermal fluctuations destabilise the surface magnetisation. Some experiments show the opposite trend. The authors suggest that apparent temperature increases may reflect temperature-dependent changes in the bare transport properties rather than in the CISS mechanism itself, but concede this is not fully resolved.

• Electrode dependence. Most CISS experiments use gold electrodes, which have strong spin-orbit coupling. The spinterface model includes both spin-orbit and exchange interactions at the interface, and the authors argue that even with weak spin-orbit coupling, exchange interactions may sustain the effect — but distinguishing the two experimentally is described as very difficult.

• Electrode-free systems. Eckvahl et al. (2023) reported CISS-like effects in donor–chiral-bridge–acceptor molecules probed by electron paramagnetic resonance, with no metal electrode. The authors argue that the immobile electron at the donor ground state could play the role of a "localised surface moment," enabling a spinterface-like mechanism, but note that the Eckvahl experiments do not fully resolve whether the observed effect is CISS or magneto-chiral anisotropy, because left- and right-handed enantiomers gave essentially identical signals.

• Time-reversal symmetry breaking. The model assumes dissipation breaks time-reversal symmetry but does not derive this from first principles.

Falsifiable predictions

The perspective lists several predictions intended to be experimentally testable, including: that CISS polarisation should decrease with increasing temperature for any system whose bare transport is temperature-independent; that the GMR/TMR geometry reversal should be reproducible across different molecular systems; and that the surface magnetisation stabilised by displacement currents should be detectable by independent magnetic probes.

What the paper does not establish

The spinterface model does not calculate its interface parameters from first principles — the spin-orbit and exchange coupling strengths are extracted from fits to existing data. The authors describe this as a limitation shared by all CISS theories but one that must eventually be addressed. The model also does not explain why metallic electrodes appear necessary for the CISS effect in most experimental contexts, beyond noting that metal surfaces carry the fluctuating magnetic moments the theory requires. The relationship between the spinterface mechanism and the electrostatic contact-potential mechanism proposed by Tirion and van Wees is not discussed.

References

• Sarkar, S.; Sharoni, A.; Monti, O. L. A.; Dubi, Y. "The spinterface mechanism for the chiral-induced spin selectivity effect: A Critical Perspective." arXiv:2510.21061v1, 24 October 2025.
• Dubi, Y. "Spinterface chirality-induced spin selectivity effect in bio-molecules." Chemical Science 13, 10878–10883 (2022).
• Yang, C.; Li, Y.; Zhou, S.; Guo, Y.; Jia, C.; Liu, Z.; Houk, K. N.; Dubi, Y.; Guo, X. Nature Chemistry 15, 1–8 (2023).
• Alwan, S.; Sarkar, S.; Sharoni, A.; Dubi, Y. "Spinterface origin for the chirality-induced spin-selectivity effect." Journal of Chemical Physics 159, 014106 (2023).
• Nguyen, T. N. H. et al. Chemical Science 15, 14905 (2024).
• Eckvahl, H. J. et al. "Direct observation of chirality-induced spin selectivity in electron donor–acceptor molecules." Science 382, 197–201 (2023).