Condensed Matter
Andreev Reflection: How an Electron Turns Into a Hole at a Superconductor Interface
Send a single electron toward a superconductor with less than about 1 milli-electron-volt of spare energy, and something strange happens: it cannot enter. There are no single-particle states inside the superconducting gap to receive it. Instead, the electron grabs a partner from the metal, dives into the superconductor as half of a Cooper pair, and leaves behind a hole that retraces the electron's path backward like a movie run in reverse. This charge-swapping bounce is Andreev reflection.
Named after Soviet physicist Alexander F. Andreev, who described it in 1964, Andreev reflection is the fundamental process by which charge crosses a normal-metal–superconductor (N–S) interface at energies below the superconducting gap Δ. Because one electron in and one hole out transfers a net charge of 2e, Andreev reflection is what carries dissipationless supercurrent into a normal conductor, doubles the sub-gap conductance of a clean N–S contact, and underlies the physics of proximity effects, Josephson weak links, and topological Majorana experiments.
- TypePhase-coherent electron-to-hole scattering at N–S interface
- DiscoveredA. F. Andreev, 1964 (Sov. Phys. JETP 19, 1228)
- Charge transferred2e per reflection (one Cooper pair enters S)
- Energy regimeE < Δ (sub-gap); Δ ≈ 0.18 meV in Al, ≈ 1.5 meV in Nb
- Key modelBlonder–Tinkham–Klapwijk (BTK), 1982, barrier parameter Z
- SignatureZero-bias conductance up to 2× normal value for a transparent contact
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What Andreev Reflection Is: The Physical Setup
Picture a clean junction between an ordinary metal (N) — say gold or copper — and a superconductor (S) such as aluminum or niobium. Cooled below the superconductor's critical temperature (T_c ≈ 1.2 K for Al, 9.3 K for Nb), the electrons in S bind into Cooper pairs and open an energy gap Δ around the Fermi level. No single-electron states exist for energies within ±Δ of E_F.
Now drive an electron from N toward S with energy E measured from the Fermi level, where E < Δ. Ordinary transmission is forbidden — there is no state to land in. Ordinary reflection is possible, but for a transparent interface it is not what dominates. Instead the electron pairs up with a second electron of opposite spin and momentum, the two enter S together as a Cooper pair, and the absence of that second electron propagates back into N as a hole.
- In: one electron, charge −e, spin up
- Out: one hole, charge +e, spin down (the missing electron was spin down)
- Net into S: charge −2e as a Cooper pair
The result is a coherent conversion of a particle into an antiparticle-like excitation at the interface.
The Mechanism: Bogoliubov–de Gennes and Retroreflection
The rigorous description uses the Bogoliubov–de Gennes (BdG) equations, which couple an electron amplitude u and a hole amplitude v through the pair potential Δ. Inside the gap the only propagating solutions in S are evanescent, decaying over the coherence length ξ; in N, an incoming electron at energy E is matched to an outgoing hole at energy −E.
The reflected hole is remarkable because it is a time-reversed electron. Its group velocity points back toward the source, so it retraces the incident trajectory — this is retroreflection, in contrast to the mirror-like specular reflection of a wall. Momentum is very nearly, but not exactly, reversed: the electron sits slightly above E_F at wavevector k_e = k_F + E/(ħv_F), while the hole sits slightly below at k_h = k_F − E/(ħv_F).
- Momentum mismatch: δk = 2E/(ħv_F)
- Coherence length: ξ = ħv_F/(πΔ) (clean limit)
- Phase picked up on reflection: an electron-to-hole bounce adds −arccos(E/Δ) minus the superconductor's phase
That inherited phase is exactly what stitches together Andreev bound states and lets a supercurrent leak into the normal region — the proximity effect.
Key Quantities and a Worked Example
Let's put numbers to it for a gold–niobium contact at T → 0.
- Gap: Nb has Δ ≈ 1.5 meV = 2.4×10⁻²² J (from Δ = 1.76·k_B·T_c with T_c = 9.3 K).
- Fermi velocity of gold: v_F ≈ 1.4×10⁶ m/s.
- Coherence length: ξ = ħv_F/(πΔ) = (1.05×10⁻³⁴ · 1.4×10⁶)/(π · 2.4×10⁻²²) ≈ 195 nm.
- Momentum mismatch at E = Δ: δk = 2Δ/(ħv_F) ≈ 3×10⁶ m⁻¹, tiny next to k_F ≈ 1.2×10¹⁰ m⁻¹ — hence near-perfect retroreflection.
The transport fingerprint is conductance doubling. For a perfectly transparent contact (BTK barrier parameter Z = 0), the sub-gap differential conductance is
G(eV < Δ) = 2·G_N,
because each Andreev event moves 2e instead of e. As the barrier grows (larger Z), this crossover flips: for Z ≫ 1 the junction becomes a tunnel junction whose conductance peaks sharply at eV = ±Δ, mapping out the BCS density of states. Measuring the whole G(V) curve and fitting Z, Δ, and a broadening Γ is the basis of point-contact Andreev-reflection spectroscopy.
How It Is Observed, Measured, and Used
The workhorse experiment is point-contact spectroscopy: a sharp normal tip is pressed onto a superconductor (or vice versa), forming a nanometer-scale contact, and the differential conductance dI/dV is recorded versus bias voltage V. The Blonder–Tinkham–Klapwijk (BTK) model of 1982 provides the fitting formula, with the dimensionless barrier strength Z tuning between the two limits above.
- Gap spectroscopy: the width of the enhanced-conductance region gives Δ directly; the zero-bias enhancement gives interface transparency.
- Spin polarization: in a ferromagnet, Andreev reflection needs an opposite-spin partner, so a spin-polarized current suppresses the doubling. The zero-bias conductance falls from 2 toward 2(1−P), letting experimenters extract the transport spin polarization P — a standard tool since Soulen and Upadhyay (1998).
- Proximity effect: repeated Andreev reflections induce a mini-gap and dissipationless transport in an attached normal metal.
- Order-parameter symmetry: zero-bias conductance peaks from Andreev bound states reveal d-wave and unconventional pairing.
Technologically, Andreev processes govern Josephson junctions, superconducting qubits, and Andreev/Majorana qubit proposals.
Andreev Reflection Versus Its Cousins
Several related processes are easy to confuse, and the distinctions carry real physics.
- Normal reflection conserves charge and merely mirrors the electron; Andreev reflection converts charge and retroreflects the trajectory.
- Quasiparticle tunneling transfers a single electron and only above the gap (eV > Δ); it dominates the large-Z limit and produces the classic BCS conductance peaks.
- Specular Andreev reflection is a variant predicted by C. W. J. Beenakker in 2006 for graphene: when the Fermi energy is smaller than Δ, the reflected hole leaves the conduction band for the valence band and reflects specularly like light off a mirror, not retro. This is a signature of Dirac (relativistic-like) quasiparticles.
- Crossed Andreev reflection (CAR) splits a Cooper pair into two electrons that exit through two separate normal leads — a solid-state source of entangled electrons.
The unifying idea: normal scattering is diagonal in the electron/hole basis, while Andreev scattering is off-diagonal, coupling electrons to holes through the pair potential.
Significance, Famous Cases, and Open Questions
Andreev reflection is one of the load-bearing concepts of modern condensed-matter physics. It explains why a normal metal in good contact with a superconductor can carry supercurrent (the proximity effect), why S–N–S junctions host discrete Andreev bound states that carry the Josephson current, and how charge and heat cross N–S boundaries.
A prominent modern application is the hunt for Majorana zero modes. A Majorana bound state at the end of a proximitized semiconductor nanowire should produce a quantized zero-bias conductance peak of exactly 2e²/h via resonant Andreev reflection. Reports of such peaks (from ~2012 onward) drove enormous interest, but distinguishing a true Majorana from trivial Andreev-bound-state peaks or disorder artifacts remains actively contested — several high-profile claims were later retracted or reinterpreted.
- Open questions: unambiguous Majorana signatures beyond conductance peaks; Andreev physics in twisted and Dirac materials; equal-spin (triplet) Andreev reflection at ferromagnet–superconductor interfaces.
From a 1964 thermal-conductivity puzzle to today's quantum-computing roadmaps, the electron-turns-into-a-hole bounce keeps proving central.
| Process | Charge into S | Energy range | Conductance signature |
|---|---|---|---|
| Andreev reflection (transparent, Z=0) | 2e (Cooper pair) | E < Δ | G rises to 2·G_N below the gap |
| Normal (specular) reflection | 0 | any | Electron bounces back, no transfer |
| Single-quasiparticle tunneling (Z≫1) | e | E > Δ only | BCS density-of-states peak at eV = Δ |
| Crossed Andreev reflection (CAR) | 2e split across two leads | E < Δ | Non-local negative conductance |
| Josephson supercurrent (S–N–S) | 2e, phase-coherent | E = 0 equilibrium | Zero-voltage current via Andreev bound states |
Frequently asked questions
What is Andreev reflection in simple terms?
It is what happens when an electron hits a superconductor with too little energy to enter (below the gap Δ). Rather than passing through or bouncing back normally, the electron drags a second electron into the superconductor as a Cooper pair and reflects back into the metal as a hole. Net result: charge 2e crosses the interface even though single-particle transmission is forbidden.
Why does the hole travel backward along the electron's path?
Because the reflected hole is essentially a time-reversed electron: it has nearly opposite momentum but the same trajectory line, so its velocity points back toward where the electron came from. This is called retroreflection, and it is the opposite of the mirror-like specular reflection you get from a wall. The tiny momentum mismatch is δk = 2E/(ħv_F), which is negligible compared with the Fermi wavevector.
What is conductance doubling and why does it happen?
For a transparent normal-metal–superconductor contact, the differential conductance at voltages below the gap (eV < Δ) rises to twice the normal-state value, G = 2·G_N. It happens because each Andreev event transfers 2e instead of the usual 1e, so the same number of scattering events carries double the charge. A barrier (large BTK parameter Z) suppresses this and instead produces sharp conductance peaks at eV = ±Δ.
Who discovered Andreev reflection and when?
Alexander F. Andreev described it in 1964 while explaining the thermal conductivity of superconductors in their intermediate state (Sov. Phys. JETP 19, 1228). The quantitative transport theory used to fit experiments today is the Blonder–Tinkham–Klapwijk (BTK) model, published in 1982, which introduced the dimensionless interface barrier parameter Z.
What is the difference between retro and specular Andreev reflection?
In ordinary metals the reflected hole retraces the incident electron's path (retroreflection) because the Fermi energy is huge compared to the gap. In graphene near the Dirac point, where the Fermi energy is smaller than Δ, the hole can leave the conduction band for the valence band and reflect specularly, like light off a mirror. Specular Andreev reflection was predicted by Beenakker in 2006 and is a hallmark of Dirac quasiparticles.
How is Andreev reflection connected to Majorana fermions?
A Majorana zero mode at the end of a proximitized nanowire should cause resonant Andreev reflection that produces a zero-bias conductance peak quantized at exactly 2e²/h. This makes Andreev-reflection spectroscopy the main probe in the Majorana search. However, trivial Andreev bound states and disorder can mimic the same peak, so a quantized peak alone is not considered conclusive proof.