Fermionic pairs, from the surface to the bulk

Fermionic Pairs, From the Surface to the Bulk

Pair formation drives collective quantum phenomena. It governs everything from electrons in superconductors to the stability of atomic nuclei. In large, extended systems like ultracold Fermi gases, this process follows the BCS–BEC crossover. This is a smooth transition where pairs evolve from large, overlapping Cooper pairs into tightly bound molecules called dimers. However, in finite systems like nuclei or quantum dots, this simple picture breaks down. In these small, confined spaces, pairing competes with the discrete energy levels of the container. This makes the microscopic structure of these pairs difficult to access.

A new study from the University of Heidelberg provides a high-resolution window into this complexity. By using a tiny, tunable trap of lithium atoms, the researchers imaged exactly where and how pairs form. They discovered that confinement reorganizes pairing. Depending on how the "shells" of the system are filled, pairs can be pushed toward the surface. Alternatively, they can form robust, bulk-like structures even in small clusters.

The breakdown of the bulk paradigm

The standard model for fermionic pairing assumes a "bulk" environment. This assumes a system so large that the edges are negligible and the density is uniform. In this regime, we categorize pairing by the relationship between the pair size ($r_B$) and the distance between particles ($\ell$). If the pairs are small and separate, they are dimers. If they are large and overlapping, they are Cooper pairs. This is the classic BCS–BEC crossover .

This paradigm fails when the system size ($r_F$) is comparable to the pair size. In finite systems, energy levels are not continuous. They are discrete "shells," much like the electron shells in an atom. When a system reaches a "magic number," all energy shells are perfectly filled. At this point, the density of states (the number of available energy levels) at the Fermi energy (the highest energy state occupied by particles) vanishes. This can suppress pairing in the center of the trap. The authors report that this shell-filling effect reshapes correlations in ways that bulk theories cannot predict.

Mapping the hierarchy of length scales

To resolve this, the authors used $^6$Li atoms in a quasi-two-dimensional optical tweezer. Their approach relies on three specific design choices:

1. Single-particle resolution: Using advanced imaging, they can resolve individual atoms and their spins. This allows them to calculate the density-density correlation function $C^{(2)}(r, r)$. This function counts how often a spin-up atom and a spin-down atom occupy the same location.
2. Feshbach resonance tuning: They use magnetic fields to tune the two-body binding energy ($E_B$). This lets them move from weakly interacting to strongly bound molecular regimes.
3. Deterministic particle numbering: They control the exact number of fermions ($N$) in the trap. This enables a move from "few-body" physics to "many-body" physics.

As shown in, the researchers identify three distinct regimes. When $r_B \ll \ell \leq r_F$, the system is in the molecular regime of dimers. When $\ell < r_B \leq r_F$, it enters the bulk pairing regime of Cooper pairs. Finally, when $r_B \gtrsim r_F$, the system enters the confinement-dominated "shell pairing" regime.

Evidence of surface-dwelling pairs

The researchers report striking differences in how pairs distribute themselves.

In the molecular regime ([Figure 2c]), the authors find that pair correlation is proportional to the density. This means dimers form wherever atoms are present. In the intermediate regime ([Figure 2b]), they observe a hybrid state. Near the low-density edges, atoms behave like isolated dimers. Toward the trap center, correlations saturate. This signals the onset of overlapping Cooper pairs.

The most significant finding occurs in the confinement-dominated regime ([Figure 2a]). In "closed-shell" configurations (where energy levels are perfectly filled, such as 6+6 atoms), pairing is suppressed in the high-density center. Instead, correlations peak near the low-density surface. This happens because the discrete level structure of the trap influences the formation of pairs.

This suppression is sensitive to the system's state. The authors show that "open-shell" systems (such as 5+5 atoms, where the top shell is partially filled) do not suffer this suppression. These systems support strong central pairing ([Figure 3a]). Furthermore, they show that increasing the particle number restores the "bulk-like" Cooper-pair profile in the center.

Increasing the number of atoms effectively washes out these shell effects.

Limits of the microscopic lens

This study provides a powerful way to visualize pairing, but it has limits. The authors note that their experimental resolution ($\delta r_{res} = 300$ nm) acts as a physical cutoff. Because they cannot resolve distances smaller than this, they cannot directly extract the "contact." The contact is a parameter describing short-range correlations at nearly zero distance.

The findings are also tied to the quasi-2D geometry of the optical tweezer. While pairing physics is universal, shell structures in two dimensions may differ from three-dimensional systems. For example, the interior of a heavy nucleus is 3D. Practitioners applying these insights to 3D superconducting nanostructures must account for how dimensionality alters the density of states and shell gaps.

The verdict: A new bridge for quantum simulation

The results show that in finite systems, pairing is not just a local phenomenon. It is a global phenomenon shaped by the container. The study bridges the gap between the solvable math of few-body physics and the statistical models of many-body theory.

For those designing quantum simulators or mesoscopic devices, this work is a vital guide. It proves that you cannot rely on "bulk" assumptions when scaling down to a few dozen particles. The shell structure can invert your expectations. It can push correlations to the surfaces of a device rather than its core. This experiment serves as a successful quantum simulator of finite matter.

Figures from the paper