How Science Works | Radio-Frequency Spectroscopy — How Spin Flips Reveal Atomic Interactions

Find your reading route

Atoms have magnetic sublevels and hyperfine states whose separations depend on magnetic field. An oscillating field can drive an allowed transition when its photon energy matches the relevant energy difference.

In a nearly non-interacting reference gas, that resonance supplies a baseline. In an interacting gas, moving an atom into another state changes its interactions with neighbours. The spectral shift therefore includes the energy cost of changing the many-body environment, not simply an isolated atomic splitting.

Contents · Next chapter

A common sequence prepares atoms in two occupied states and leaves a third state empty. A weak radio-frequency pulse transfers some atoms from one occupied state into the empty state. State-selective imaging then counts the transferred population.

Repeating the sequence across frequency produces a spectrum. The pulse must be weak enough for the intended response model, and the detection must discriminate the states accurately. Background transitions and ordinary loss are measured in control runs.

Contents · Previous chapter · Next chapter

Suppose a calibrated spectral feature lies 12 kHz above the isolated-atom reference. The corresponding photon-energy difference is h×12,000 ≈ 7.95×10⁻³⁰ joule. Dividing by Boltzmann’s constant gives an energy scale of about 0.576 μK.

This conversion expresses energy in temperature units; it does not establish that the gas temperature is 0.576 μK. Nor does it automatically identify the shift as a pairing gap. Initial-state correlations, final-state interactions and inhomogeneity must be included before assigning a microscopic interpretation.

Contents · Previous chapter · Next chapter

For a bound molecule, transferring one constituent can dissociate the pair, creating a threshold and a distribution of outgoing kinetic energies. In a strongly interacting Fermi gas, correlated pairing can similarly alter the radio-frequency response.

The observed line depends on both the state before excitation and the state after it. Final-state interactions can move spectral weight and produce bound-bound transitions. A measured peak position is therefore not always equal to a bare binding energy or a simple universal gap parameter.

Contents · Previous chapter · Next chapter

Chin and colleagues studied an ultracold two-component lithium-6 gas using radio-frequency excitation spectra. The 2004 study reported an energy gap associated with pairing in the strongly interacting gas.

The primary experiment demonstrates how internal-state transfer accesses many-body energies. Pairing evidence should be assessed together with the magnetic-field dependence, temperature dependence and theoretical response model. Pair formation alone is not a standalone proof of long-range superfluid coherence.

Contents · Previous chapter · Next chapter

Magnetic-field drift shifts the reference transition. Strong pulses broaden lines, while short pulses impose Fourier broadening. Density variation inside the trap adds a mixture of local interaction energies. Imperfect state preparation introduces additional components.

Researchers calibrate the isolated transition, vary pulse duration and amplitude, scan density and field, and compare full line shapes rather than one point. The wider scientific lesson is that the probe changes the system: interpreting the energy absorbed requires understanding both sides of that change.

Primary research: Chin and colleagues, Observation of the Pairing Gap in a Strongly Interacting Fermi Gas (2004).

Continue through the Atomic Physics learning hub or return to the How Science Works master index.

Contents · Previous chapter · Atomic Physics hub