Bi/Ni Bilayer: Interfacial Superconductivity
- Bi/Ni bilayer is a thin-film heterostructure combining Bi’s strong spin–orbit coupling and Ni’s ferromagnetic order to generate interface-induced superconductivity.
- Experimental studies demonstrate that growth methods, epitaxial orientation, and interdiffusion critically influence Tc (≈3–4 K) and superconducting anisotropy.
- Transport and spectroscopic measurements reveal complex pairing phenomena and mixed-state vortex dynamics, prompting diverse theoretical models.
Bi/Ni bilayer denotes a thin-film heterostructure formed by elemental bismuth and nickel, realized in both Bi/Ni and Ni/Bi stacking sequences and, in surface-science usage, as Bi films supported on Ni surfaces. The system is studied because Bi contributes strong spin–orbit coupling and interfacial inversion-symmetry breaking, whereas Ni provides ferromagnetic order and exchange proximity; together they generate a superconducting platform with around $3$–$4$ K even though bulk Bi becomes superconducting only below mK and Ni is not superconducting (Nanda et al., 4 May 2026). Across the literature, Bi/Ni bilayers have been interpreted in two markedly different ways: as an intrinsic spin–orbit-coupled ferromagnetic heterostructure hosting unconventional or triplet superconductivity, and as an interdiffused NiBi-containing system whose low-temperature response can be conventional and -wave (Gong et al., 2015).
1. Materials platform and experimental realizations
The experimental literature reports several distinct Bi/Ni platforms. Epitaxial heterostructures were realized as Ni/Bi/Cu/MgO and Bi/Ni/MgO, with rhombohedral Bi(110) and fcc Ni(001), and superconductivity was reported to be independent of growth order; in a later MBE study, superconductivity was observed only for Bi(110)/Ni(100), whereas Bi(111)/Ni(100) did not superconduct (Gong et al., 2015). A separate transport-oriented platform used electron-beam evaporation of Ni and resistive-source evaporation of Bi in ultra-high vacuum, typically with Ni nm and Bi nm, followed by in situ AlO capping, Hall-bar patterning by laser scribing, and indium Ohmic contacts (Nanda et al., 4 May 2026).
A second experimental family consists of sputtered or room-temperature-processed stacks on oxidized Si. In those systems, the nominal bilayer often evolves into a chemically more complex Ni/NiBi/Bi or NiBi$3$0/Bi structure, with superconducting properties that depend strongly on Bi thickness, annealing history, and interdiffusion (Sant'ana et al., 2023). MBE growth at 300 K on Al$3$1O$3$2(001) and MgO(001) has also produced Bi(10 nm)/Ni(1 nm) bilayers in which depth profiling and diffraction indicate an ultrathin NiBi$3$3 layer beneath the Bi overlayer, capped by an approximately 1 nm BiNi interfacial layer (Sant'ana et al., 18 Sep 2025).
| Platform | Structural features | Reported superconducting picture |
|---|---|---|
| Epitaxial Bi(110)/Ni(100) or Ni/Bi on MgO/Si | Growth-order independent for Bi(110)/Ni(100); sharp interfaces; orientation-sensitive | Interface-triggered superconductivity with triplet interpretations (Gong et al., 2015, Zhao et al., 2018) |
| Evaporated Ni $3$4 nm / Bi $3$5 nm Hall bars | Atomically sharp interface; no discernible interdiffusion within resolution | Quasi-2D bilayer-spanning superconductivity with vortex-dominated transport (Nanda et al., 4 May 2026) |
| Sputtered or annealed Ni/Bi films | Interdiffusion stabilizes NiBi$3$6; residual Bi and sometimes residual Ni remain | Conventional or NiBi$3$7-based superconductivity (Vaughan et al., 2019, Sant'ana et al., 2023) |
| MBE Bi(10)/Ni(1) at 300 K | Ultrathin NiBi$3$8 plus BiNi interfacial layer | Quasi-2D superconductivity with enhanced $3$9 (Sant'ana et al., 18 Sep 2025) |
This experimental diversity is central to the subject. The phrase “Bi/Ni bilayer” does not identify a single structural state; rather, it names a family of thin-film systems whose interface sharpness, epitaxial orientation, and degree of alloying can differ substantially.
2. Structural motifs, epitaxy, and bilayer chemistry
On the surface-science side, first-principles calculations for Bi on Ni(111) identify a strongly bound adsorbed Bi surface monolayer followed by bulk-like rhombohedral bilayers of Bi, rather than the previously proposed HCP films. The energetically preferred structure is a bulk-like rhombohedral (hR2) (001)-oriented bilayer stacking on top of the surface monolayer, and the odd-layer sequence $4$0 arises naturally from a “monolayer + $4$1 bilayers” motif rather than from free-electron quantum size effects (Gao et al., 2013). At total coverage $4$2 Bi/Ni in a $4$3 cell, the 5-layer hR2 film is lower in energy than the 4-layer HCP film by $4$4 eV, and the driving force is covalent bonding within the Bi bilayers rather than confinement (Gao et al., 2013).
The same emphasis on bonding appears in grain-boundary calculations. For Ni grain boundaries, Bi bilayers are predicted to be unstable on $4$5 twist CSL boundaries such as $4$6 and $4$7, but stable on most $4$8 twist CSL boundaries and on high-energy $4$9 boundaries in the Bi-rich limit (Gao et al., 2013). When present, those bilayers strongly reduce the work of separation, so the Bi bilayer is treated not only as a structural motif but also as a segregation-driven embrittling phase at Ni boundaries (Gao et al., 2013).
In superconducting thin films, structural characterization separates two broad cases. In the sharp-interface evaporated Ni/Bi Hall bars, cross-sectional TEM/EDAX show laterally uniform, chemically distinct layers and an atomically sharp interface, with no discernible interdiffusion within resolution; that observation was used to support interfacial electronic reconstruction rather than bulk intermetallic phases as the origin of superconductivity (Nanda et al., 4 May 2026). By contrast, sputtered and annealed stacks, as well as some room-temperature-grown MBE stacks, exhibit interdiffusion that stabilizes NiBi0 and sometimes BiNi, thereby shifting the physical meaning of the bilayer from a clean elemental heterostructure toward an intermetallic-containing multilayer (Vaughan et al., 2019).
3. Superconducting phenomenology and dimensionality
The central empirical fact is that superconductivity appears although neither rhombohedral Bi nor Ni is superconducting in isolation under the same conditions. In epitaxial Bi/Ni heterostructures, 1 is reported near 2 K, and the superconducting state is highly sensitive to layer thicknesses (Gong et al., 2015). With Bi fixed at 15 nm, increasing Ni thickness from 2 to 4 nm suppresses superconductivity and destroys it at 4 nm Ni; with Ni fixed at 4 nm, increasing Bi thickness reintroduces superconductivity and restores 3 K, implying that the superconducting condensate is triggered by the interface yet develops away from the region of strongest exchange field (Gong et al., 2015).
Dimensionality is a recurrent theme. In evaporated Ni/Bi bilayers with 4 K and transition width 5 K, the resistive transition is fitted by a BKT-like form,
6
with 7 K, while two-dimensional GL fits describe 8 well and a 9 crossover appears for 0 at 1 K (Nanda et al., 4 May 2026). The extracted coherence lengths are 2 nm and 3 nm, giving anisotropy 4, so the Bi thickness is comparable to the superconducting length scales (Nanda et al., 4 May 2026).
Critical-field analyses in earlier epitaxial work gave 5 T for Bi(15 nm)/Ni(2 nm), corresponding to 6 nm, while the in-plane 7 exceeded the Pauli limit (Gong et al., 2015). Time-domain THz spectroscopy on epitaxial Bi(110)/Ni(100) found that the entire 11 nm bilayer becomes fully gapped below 8 K, with 9 meV, 0 meV, and 1; the inferred coherence length was 2 nm and the penetration depth 3 (Chauhan et al., 2018). These results place Bi/Ni bilayers in a regime where quasi-2D geometry, finite thickness, and mixed orbital–paramagnetic pair breaking all matter.
4. Pairing symmetry, spectroscopic evidence, and theoretical interpretations
The pairing symmetry of Bi/Ni bilayers is the most disputed aspect of the subject. Point-contact Andreev reflection on Bi/Ni samples reported a robust single zero-bias conductance peak without the double-peak structure characteristic of conventional 4-wave superconductors, and the peak remained strong even for highly spin-polarized injection from an LSMO tip; this was taken as evidence against singlet 5-wave pairing and in favor of triplet 6-wave pairing (Gong et al., 2015). A subsequent ARS study identified the superconducting state more specifically with the ABM state, reporting 7–8 meV, 9, a donut-like gap magnitude 0, and a nodal axis parallel to the Ni magnetization (Zhao et al., 2018).
Those claims were linked to other probes of broken time-reversal symmetry. Polar Kerr measurements on Bi/Ni were interpreted as evidence for spontaneous TRS breaking, and a specially designed SQUID detected an anomalous hysteretic magnetic response in the superconducting state that was attributed to chiral superconducting domains (Wang et al., 2016). TDTS, however, emphasized a fully gapped, nodeless bulk electrodynamic response over the entire bilayer and argued that dominant odd-frequency pairing is disfavored by the absence of subgap absorption (Chauhan et al., 2018). The coexistence of a nodeless bulk response and spectroscopic zero-bias anomalies therefore became a central interpretive problem.
Theoretical work has produced several mutually competing resolutions. One line of analysis argued that conventional 1-wave superconductivity in alloyed Bi/Ni can project to an effective 2 channel at the Bi surface because the strong spin–orbit-coupled Bi surface state experiences an in-plane exchange field from Ni; on that view, 3-wave-like Andreev signatures do not require intrinsically unconventional bulk pairing (Chao, 2018). A separate BTK-type analysis concluded that the ABM triplet state best fits the directional PCAR spectra, whereas another Andreev-spectroscopy study argued that only an ABM state mixed with a 4-wave component can explain both the PCAR data and the nodeless THz response (He et al., 2021). Yet another proposal considered a chiral 5 state in which disorder and in-plane Zeeman coupling generate gap-filling states and Bogoliubov Fermi surfaces, thereby accounting for zero-bias anomalies in an otherwise nodeless TRSB superconductor (1901.10132).
Recent transport work has pushed the discussion in a more conservative direction. In evaporated Ni/Bi bilayers, the combined BKT-like broadening, 2D GL scaling, flux-flow transport, and vortex Hall response were argued to be consistently described by a conventional 6-wave order parameter, with any unconventional contributions described as subtle rather than necessary (Nanda et al., 4 May 2026). The pairing problem in Bi/Ni bilayers therefore remains unsettled not because of a lack of data, but because the data can be organized into distinct materials regimes and distinct modeling assumptions.
5. Mixed-state transport, upper critical fields, and magnetic control
Transport near 7 has revealed unusually rich mixed-state behavior. In Ni/Bi Hall bars, 8 evolves from pinned-vortex behavior at low field to flux flow at intermediate field and then to normal-state transport near 9, while 0 exhibits pronounced odd-in-field peaks that reverse sign under 1 and appear only in the mixed state (Nanda et al., 4 May 2026). After subtraction of the ordinary Hall background, the vortex contribution is described by a force-balance model,
2
with
3
leading to
4
Within that framework, the antisymmetric transverse peaks arise from the competition between Magnus and viscous forces (Nanda et al., 4 May 2026).
Magnetic control experiments indicate that the superconducting state is not restricted to an atomically thin interface. In Ni/Bi/EuS trilayers, the EuS ferromagnetic insulator introduces exchange proximity without additional normal conduction, and the resulting hysteresis in 5 and in the transition-region magnetoresistance shows that magnetic configuration tunes the superconductivity (Nanda et al., 4 May 2026). Because the superconducting coherence length is comparable to the Bi thickness, that hysteretic EuS control was interpreted as evidence that superconductivity extends across the bilayer rather than remaining confined to an ultrathin sheet (Nanda et al., 4 May 2026).
Field-anisotropy studies on interdiffused systems reveal a related but not identical phenomenology. In MBE-grown Bi(10 nm)/Ni(1 nm) with an ultrathin NiBi6 layer, 7 at 8 K shows the cusp expected from the Tinkham thin-film formula, 9 has pronounced upward curvature near 0, the extracted superconducting thickness is 1–2 nm, and 3 reaches approximately 4 at reduced temperatures 5, exceeding the Rashba 6 bound within a BCS singlet picture (Sant'ana et al., 18 Sep 2025). In sputtered Ni(8)/Bi(7) films, thick samples behave approximately like bulk NiBi8, with 9, 0, and 1, whereas very thin samples require 2, 3, and show 4 (Sant'ana et al., 2023). These results place Bi/Ni bilayers among the experimentally tunable systems in which thickness, exchange, and spin–orbit scattering reshape both orbital and paramagnetic limits.
6. Intermetallic formation, competing origins, and present status
A major controversy concerns whether superconductivity is intrinsic to the elemental Bi/Ni interface or instead originates from NiBi5 formed by interdiffusion. The strongest evidence for the intermetallic scenario comes from sputtered bilayers that are initially nonsuperconducting but become superconducting after modest thermal exposure. In Bi/Ni/Ta stacks on oxidized Si, as-grown films did not display a superconducting transition, but after storage at room temperature superconductivity emerged after about 14 days with 6 up to 7 K; controlled annealing at 8 caused the NiBi9 layer to thicken dramatically, while XRD, PNR, SQUID magnetometry, and transport showed that the onset of superconductivity coincided with the formation of ordered NiBi$3$00 (Vaughan et al., 2019). The same study extracted an activation energy of $3$01 eV for the annealing process and concluded that gentle heating causes formation of conventional $3$02-wave NiBi$3$03, making pristine Bi/Ni interfaces difficult to preserve (Vaughan et al., 2019).
Thickness-dependent sputtered Ni/Bi studies support that interpretation. With fixed 8 nm Ni and variable Bi, RBS and XRD indicate trilayer Ni/NiBi$3$04/Bi structures for thinner Bi and effectively NiBi$3$05/Bi structures for thicker Bi; $3$06, transition width, coherence-length anisotropy, and dirty-limit WHH parameters all track that structural evolution (Sant'ana et al., 2023). In that regime, thick films show $3$07 K and parameters close to NiBi$3$08 single crystals, whereas thin films retain stronger anisotropy and finite spin–orbit scattering (Sant'ana et al., 2023).
Room-temperature MBE growth yields a more nuanced intermetallic picture. In Bi(10 nm)/Ni(1 nm), MEIS and GIXRD indicate an ultrathin orthorhombic NiBi$3$09 layer beneath the Bi overlayer and a thin BiNi interfacial layer; the superconductivity is quasi-2D and the in-plane upper critical field violates the spin-paramagnetic limit by about $3$10 at high reduced temperature, yet WHH, KLB spin–orbit scattering, and simple Rashba-enhanced BCS models do not quantitatively explain the data (Sant'ana et al., 18 Sep 2025). That work therefore leaves open the possibility that an ultrathin NiBi$3$11 layer in the 2D limit may host superconductivity beyond the standard spin-singlet framework (Sant'ana et al., 18 Sep 2025).
The sharp-interface evaporated Ni/Bi Hall-bar studies point in the opposite direction. There, low-temperature growth, in situ capping, and cross-sectional TEM/EDAX were used to argue against intermetallic formation, and the EuS control experiments were used to infer bilayer-spanning superconductivity rather than a purely interfacial sheet (Nanda et al., 4 May 2026). Taken together, the literature indicates that “Bi/Ni bilayer” names a strongly process-dependent platform rather than a single superconducting phase. Growth temperature, epitaxial orientation, thickness ratio, microstructure, and post-growth thermal budget determine whether the low-temperature physics is dominated by sharp-interface SOC-plus-exchange superconductivity, by NiBi$3$12-based intermetallic superconductivity, or by an intermediate regime in which both descriptions remain relevant.
