eduKate Learning Manual
Science | Edge Cases Science | Physical World
Understand → Observe → Explain → Test → Transfer → Go Deeper
Feshbach Resonance
How a Hidden Molecular State Can Make Two Atoms Suddenly Interact Strongly
Wait, What? Two Atoms Can Barely Notice Each Other—Until a State They Cannot Normally Occupy Is Tuned Into Resonance
At ultralow temperature, two atoms can approach one another in a collision channel that is energetically open. Their interaction may be weak and described largely by one low-energy number: the scattering length.
Inside the same two-atom system, however, another internal configuration may support a molecular bound state. That configuration is normally energetically closed to the incoming atoms.
Change an external magnetic field and the energy of that hidden molecular state can move relative to the open collision threshold. Bring it close enough and the open collision channel strongly mixes with the closed molecular channel.
a state that was not directly accessible can reshape the scattering of the channel that is accessible.
This is a Feshbach resonance.
Quick Answer
A Feshbach resonance occurs when an unbound two-particle scattering state in an open channel is coupled to a bound or quasibound state in a different closed channel, and the two become nearly degenerate in energy.
In ultracold atomic gases, a magnetic field often tunes the relative energies because open and closed channels have different magnetic moments.
Near a simple isolated magnetic resonance, the low-energy s-wave scattering length is commonly written
a(B) = abg[1 − Δ/(B − B₀)],
where abg is the background scattering length, B₀ the resonance position and Δ a width parameter under the chosen convention.
As B approaches B₀, the scattering length can become very large and can change sign. This gives experimenters extraordinary control over interaction strength, molecular association, Bose–Einstein condensates, Fermi superfluids and the BEC–BCS crossover.
But “a diverges” does not mean every measurable collision rate becomes infinite. At finite collision energy, unitarity limits the scattering cross-section, and real resonances can include inelastic loss, finite range and multichannel corrections.
Reviews of Modern Physics — Feshbach Resonances in Ultracold Gases →
Physical Review A (2026) — Tunable Three-Body Force Near an Interspecies Feshbach Resonance →
What You Will Learn
- What open and closed scattering channels are.
- How internal spin states create different collision thresholds.
- Why a molecular bound state can influence an unbound collision.
- How magnetic tuning brings channels into resonance.
- What the scattering length means.
- Why scattering length can become very large and change sign.
- Why a pole in a(B) does not mean an infinite physical cross-section at every energy.
- How Feshbach molecules are created.
- How resonance width changes open- versus closed-channel character.
- Why inelastic collisions and three-body loss become important near resonance.
- How Feshbach control enables the BEC–BCS crossover.
- Where one-resonance, zero-range models stop being sufficient.
Part 1 — A Collision Has More Than One Internal Route
Two atoms are not structureless billiard balls. Each atom can carry electron spin, nuclear spin and hyperfine state.
A collision channel specifies the internal states of the separated atoms together with their relative motion.
At a given total energy, some channels are energetically open: the atoms can separate to infinity in those states. Other channels lie above the available asymptotic energy and are closed.
Closed does not mean irrelevant. Interatomic couplings can mix channel wavefunctions at short range.
Part 2 — The Hidden Bound State
A closed channel can contain a molecular bound state just below its own threshold.
Normally that state may lie far from the energy of incoming atoms in the open channel, so its effect on scattering is modest.
When the energies approach, virtual occupation of the closed-channel bound state becomes much more important and the scattering phase shift changes rapidly.
Part 3 — Why Magnetic Field Tunes the Resonance
Different spin configurations usually have different magnetic moments.
An applied magnetic field therefore shifts open-channel and closed-channel energies at different rates through the Zeeman effect.
Scanning B can move the closed-channel molecular level through the open-channel threshold and create a tunable scattering resonance.
Part 4 — Scattering Length as the Low-Energy Control Knob
At ultralow temperature, higher partial waves are often suppressed by centrifugal barriers and s-wave scattering dominates for distinguishable particles or bosons.
At sufficiently low collision energy, the complex microscopic potential can then be summarized approximately by the s-wave scattering length a.
- Small |a|: comparatively weak low-energy interactions.
- Large positive a: strong scattering and often a weakly bound dimer on the resonant side under suitable conditions.
- Large negative a: strong attractive-interaction behaviour without the same universal shallow two-body bound state.
The sign language is useful but must be interpreted with the actual many-body system and resonance convention.
Part 5 — Why the Scattering Length Has a Pole
As a bound state approaches the scattering threshold, the collision phase becomes extremely sensitive to energy and field.
At the resonance position in the ideal zero-energy model, the scattering length passes through a pole and changes sign.
This is not a literal atom–atom distance becoming infinite. The scattering length is a parameter describing the low-energy phase shift.
Part 6 — Unitarity Stops the Cross-Section From Becoming Infinite
A simplistic zero-energy formula suggests σ ≈ 4πa² for distinguishable s-wave scatterers.
But at finite relative wave number k, quantum probability conservation limits each partial-wave scattering amplitude.
When |a| becomes much larger than 1/k, the cross-section saturates at the unitarity scale rather than increasing indefinitely.
a mathematical pole in a low-energy parameter is not a license to predict an infinite finite-energy observable.
Part 7 — Open-Channel and Closed-Channel Character
Not all Feshbach resonances are alike.
Broad, open-channel-dominated resonances often show universal behaviour controlled strongly by scattering length over a useful field range.
Narrow, closed-channel-dominated resonances can retain stronger molecular-channel character and important effective-range corrections.
The labels refer to the dressed resonance physics, not to a hard wall separating two independent physical molecules.
Part 8 — Making Feshbach Molecules
Sweep the magnetic field sufficiently slowly across a suitable resonance and pairs of ultracold atoms can be converted coherently into weakly bound dimers.
These Feshbach molecules are often very large and weakly bound near resonance.
They can be stepping stones toward more deeply bound ultracold molecules using coherent optical transfer.
The sweep is not simply “two atoms stick whenever a becomes large.” Conversion depends on density, statistics, sweep rate, resonance width and available pair states.
Part 9 — BEC–BCS Crossover
In a two-component ultracold Fermi gas, a broad Feshbach resonance can tune the interaction continuously from a regime of weakly bound molecular pairs to a regime of large overlapping Cooper pairs.
Near unitarity, |a| is much larger than the interparticle length scales and the scattering length no longer supplies a small perturbation parameter.
This strongly interacting unitary Fermi gas is one of the central many-body systems made experimentally accessible by Feshbach control.
Part 10 — The Price of Strong Interaction: Loss
Strong tunability is not free.
Near resonances, three-body recombination can become strong. Three atoms collide, two form a molecule and the binding energy is carried away as kinetic energy, often ejecting particles from the trap.
Inelastic two-body channels can also appear depending on species and internal states.
Therefore “maximum |a|” is not automatically the most experimentally useful operating point.
Part 11 — Feshbach Resonance vs Shape Resonance
A shape resonance can occur within a single scattering channel when a quasibound state is trapped behind a centrifugal or potential barrier.
A Feshbach resonance specifically involves coupling between channels, with a bound state associated with a channel that is closed asymptotically.
Both are scattering resonances, but their microscopic bookkeeping is different.
Part 12 — Feshbach Resonance vs Efimov Physics
Efimov physics concerns universal three-body bound states when two-body scattering lengths are very large compared with interaction range.
Feshbach resonances are one experimental way to tune a into that regime.
The Feshbach resonance therefore supplies the tunable two-body interaction; the Efimov effect is a distinct three-body phenomenon that may emerge nearby. The two should not be merged.
Part 13 — 2026: Resonances as Ingredients for New Effective Forces
Current research continues to use closed-channel resonances not only to tune ordinary two-body scattering but to engineer effective interactions in more complex many-body environments.
A 2026 Physical Review A study showed theoretically how a tunable three-body force can emerge among impurities in a Bose condensate near an interspecies Feshbach resonance, with the closed-channel level participating in the effective interaction.
The result does not redefine Feshbach resonance itself. It demonstrates how the resonance can become one reusable interaction component inside a more elaborate many-body system.
Failed Model → Better Model
| Naive model | Why it fails | Better model |
|---|---|---|
| Only energetically open channels matter. | A closed-channel bound state can mix virtually with the open collision. | Use coupled-channel scattering. |
| A pole in scattering length means infinite collision rate. | Finite-energy scattering is constrained by unitarity and loss. | Use energy-dependent amplitudes and unitarity bounds. |
| Every resonance is one universal a(B) curve. | Background scattering, resonance width, effective range and inelasticity differ. | Characterize the actual resonance. |
| Stronger interaction is always better. | Three-body and inelastic losses can rise. | Optimize interaction strength against lifetime and task. |
How Do We Know?
- Prepare atoms in known hyperfine states.
- Scan magnetic field and measure collision loss or thermalization.
- Measure molecular binding energy near threshold.
- Use radio-frequency spectroscopy to probe dressed molecular states.
- Measure cloud expansion or collective-mode shifts as interaction strength changes.
- Perform field sweeps across resonance and quantify molecule conversion.
- Map three-body loss separately from elastic interaction.
- Fit a coupled-channel model to resonance position, width and background scattering.
- Repeat at different temperatures to identify finite-energy/unitarity effects.
Observation vs Inference
- Observation: ultracold collision rates, cloud dynamics and bound-state spectra can change sharply with magnetic field.
- Measurement: a near-threshold molecular state shifts relative to the open-channel threshold.
- Inference: open and closed channels are resonantly coupled.
- Low-energy model: tunable scattering length with finite width/effective-range corrections.
- Boundary: multichannel structure, inelasticity, finite energy and many-body effects can invalidate the simplest pole formula.
Common Misconceptions
| Misconception | Better model |
|---|---|
| The closed channel is physically impossible. | It is closed only as an asymptotic separated-atom channel at the chosen energy; it can still mix at short range. |
| The atoms must occupy the molecule permanently to feel it. | Virtual coupling to the molecular state can strongly alter scattering. |
| The magnetic field simply pushes atoms together. | It tunes the relative internal-state energies and resonance condition. |
| Infinite a means infinite force. | a is a low-energy scattering parameter and observables remain bounded by finite-energy physics. |
Checkpoint Questions
- What makes a scattering channel open or closed?
- How can a closed-channel state affect open-channel scattering?
- Why can a magnetic field tune the resonance?
- What does the scattering length describe?
- Why can a(B) change sign?
- Why does the physical cross-section not become infinite at finite energy?
- What is a Feshbach molecule?
- Why do broad and narrow resonances behave differently?
- What loss mechanisms become important near resonance?
- How is Feshbach resonance different from Efimov physics?
Answer Key
Open after attempting the questions
- Whether separated particles in that internal configuration are energetically allowed at the collision energy.
- Interchannel coupling allows virtual mixing with a bound state near threshold.
- Different channels have different magnetic moments and therefore different Zeeman shifts.
- The leading low-energy s-wave scattering phase behaviour.
- A bound state crossing threshold changes the scattering phase through resonance.
- Partial-wave unitarity limits the scattering amplitude.
- A weakly bound dressed dimer associated with the resonance.
- Their open/closed-channel character and effective-range scales differ.
- Three-body recombination and species-dependent inelastic collisions.
- Feshbach is a tunable two-body channel resonance; Efimov is a universal three-body bound-state phenomenon.
Primary Science Bridge
- a hidden route can affect a visible route;
- changing an external condition can change how strongly two objects interact;
- a model parameter can become very large without every real quantity becoming infinite;
- stronger interactions can bring both useful effects and new losses;
- one phenomenon can be a tool that enables another without owning the second phenomenon.
Secondary and JC Bridge
| Core idea | Higher-resolution route |
|---|---|
| Collision | Quantum scattering channel |
| Bound state | Closed-channel molecule |
| Magnetic field | Differential Zeeman tuning |
| Interaction | Scattering length and phase shift |
| Strong limit | Unitarity |
| Many body | BEC–BCS crossover and universal gases |
Unfamiliar Transfer Challenge
An ultracold gas shows a sharp atom-loss feature at 742 G. A researcher declares that a Feshbach resonance has been found.
What would strengthen the claim? Measure interaction-sensitive observables on both sides of the feature, identify a near-threshold bound state spectroscopically, fit coupled-channel theory and separate elastic scattering changes from purely inelastic loss resonances.
Deep Science Window — Pole and Zero of Scattering Length
A simple magnetic Feshbach resonance contains both a pole, where |a| diverges in the zero-energy parameterization, and a nearby zero crossing where a = 0. Experiments can therefore tune not merely “weak to strong” but also the sign of low-energy interactions and, near the zero crossing, suppress the leading s-wave mean-field interaction.
Deep Science Window — Hidden-State Coupling as a General Primitive
Feshbach resonance teaches a reusable systems idea: an observed channel can change dramatically when it couples to a hidden state whose energy is tuned nearby. The hidden state need not become a permanently occupied final state; near-resonant virtual coupling can be enough to reshape the visible response.
Evidence Boundaries
- Feshbach resonance ≠ every scattering resonance.
- Large scattering length ≠ infinite finite-energy cross-section.
- Magnetic tuning ≠ magnetic force physically squeezing atoms together.
- Closed channel ≠ completely irrelevant or inaccessible internally.
- Feshbach resonance ≠ Efimov effect.
- One-pole a(B) formula ≠ exact for overlapping, lossy or strongly energy-dependent resonances.
Manual Summary — KNOW → CONNECT → EXPLAIN → APPLY → CHECK
KNOW: open channel, closed channel, molecular bound state, scattering length, magnetic tuning, resonance width, unitarity.
CONNECT: magnetic field to channel detuning, channel detuning to resonant mixing, and resonant mixing to tunable scattering.
EXPLAIN: how a hidden molecular state can radically change interactions between atoms that enter through another channel.
APPLY: decide whether a new magnetic-field collision feature is genuinely Feshbach-like.
CHECK: require bound-state, elastic-scattering, field-dependence and loss evidence rather than naming a resonance from one loss peak.
Teaching Guide for Parents, Tutors and Teachers
Teach this first as two routes with different thresholds. Keep one route open and one route closed, then move a bound state in the closed route toward the open-channel energy. The learner should see why “closed” does not mean “cannot influence.”
- Review collision energy.
- Introduce internal-state channels.
- Separate open and closed thresholds.
- Place a bound molecular state in the closed channel.
- Tune it with magnetic field.
- Introduce scattering length and its pole.
- Add unitarity and loss limits.
- Finish with molecule formation and BEC–BCS transfer.
Independent check: later give a new scattering feature and ask learners which measurements distinguish a Feshbach resonance from a generic inelastic or shape resonance.
Safety boundary: authentic Feshbach-resonance experiments use ultrahigh vacuum, high-current magnetic coils, lasers and ultracold gases. Use simulations and published collision data outside specialist laboratories.