Atomic Physics
Magneto-Optical Trap: Six Beams and a Quadrupole Field
Ten million sodium atoms, held for two full minutes in a dark vacuum chamber, compressed into a glowing ball less than half a millimeter across and chilled to below one thousandth of a degree above absolute zero. That was the 1987 debut of the magneto-optical trap (MOT), and today essentially every atomic-physics lab on Earth starts its experiments this way.
A MOT is a device that simultaneously cools and confines neutral atoms using six red-detuned laser beams — three counter-propagating pairs along the x, y and z axes — crossed with a static magnetic quadrupole field generated by a pair of anti-Helmholtz coils. The lasers supply velocity-dependent friction (Doppler cooling), while the position-dependent Zeeman shift from the quadrupole field turns that friction into a genuine restoring force, so an atom that drifts off-center is pushed back. The result is a robust, self-loading trap that routinely produces clouds at ~100 μK, the launch pad for Bose-Einstein condensates, atomic clocks, and quantum sensors.
- TypeNeutral-atom laser cooling & trapping device
- SubfieldAtomic, molecular & optical (AMO) physics
- First demonstrated1987, Raab, Prentiss, Cable, Chu & Pritchard (sodium)
- Key ingredients6 red-detuned beams + magnetic quadrupole (anti-Helmholtz)
- Typical temperature~100–300 μK (Doppler limit for Rb-87 ≈ 146 μK)
- Restoring forceF ≈ −β·v − κ·r (damped harmonic oscillator)
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.
The physical setup: three beam pairs and two coils
A MOT is built from two independent pieces of hardware that only work in concert. First, six laser beams arranged as three mutually orthogonal counter-propagating pairs along x, y and z. All six share a single frequency, tuned slightly below (red of) an atomic resonance — for rubidium-87 the 5²S₁/₂ (F=2) → 5²P₃/₂ (F=3) cycling transition at 780 nm. In practice a second 'repump' beam is added to recover atoms that leak into the wrong hyperfine ground state.
- Optical molasses: the six beams alone provide viscous, velocity-dependent damping in all three dimensions but no position information — atoms diffuse away.
- Quadrupole field: a pair of coils in anti-Helmholtz configuration (currents opposed) produces a magnetic field that is zero at the center and grows linearly outward, B(r) ≈ B'·(x, y, −2z), with gradient B' typically 10–15 G/cm.
The beams are also given specific circular polarizations (σ⁺/σ⁻). It is the interplay of polarization, Zeeman shift and detuning that converts pure friction into a trap.
The mechanism: Zeeman shift plus red detuning makes a spring
Consider one dimension, z, and a simplified atom with ground state J=0 and excited state J=1 (sublevels m=−1,0,+1). Moving outward, the quadrupole field lifts the excited-state degeneracy: the m=−1 level shifts down on the +z side and the m=+1 level shifts down on the −z side, by the Zeeman energy ΔE = μ_B·g·m·B'·z.
Now add polarization. The beam coming from +z is σ⁻ (drives Δm=−1); from −z it is σ⁺ (Δm=+1). An atom displaced to +z has its m=−1 transition Zeeman-shifted toward the red-detuned laser, so it preferentially scatters photons from the σ⁻ beam — which pushes it back toward center. The scattering force is:
- F ≈ −β·v − κ·r, a damped harmonic oscillator, where β is the damping coefficient and κ = (μ_B·g·B'/ℏk)·β is the spring constant.
- β = −4ℏk²·s·(2δ/Γ)/[1+s+(2δ/Γ)²]², with detuning δ<0, wavevector k, saturation parameter s=I/I_sat, and linewidth Γ.
Velocity damping cools; position-dependent scattering traps. Same photons, two jobs.
Key quantities and a worked example (Rb-87)
Take rubidium-87 as the canonical case. The D2 linewidth is Γ/2π = 6.07 MHz (excited-state lifetime τ = 1/Γ ≈ 26.2 ns), and a common operating point is detuning δ = −Γ (≈ −6 MHz, i.e. −1 linewidth) with intensity per beam s ≈ 1.
- Doppler cooling limit: T_D = ℏΓ/2k_B. For Rb-87 this is 146 μK. Real MOTs often reach below this (~30–100 μK) because sub-Doppler 'Sisyphus' polarization-gradient cooling operates simultaneously.
- Recoil limit: T_r = ℏ²k²/(m·k_B) = 362 nK — the floor for a single-photon process; reaching it needs evaporative or Raman cooling.
- Capture velocity: ~10–30 m/s, so a MOT loads directly from room-temperature vapor (only the slow tail) or a Zeeman slower.
- Trap frequency: ω = √(κ/m) gives oscillation periods of order milliseconds; the damping time β⁻¹·m is ~100 μs, so the motion is strongly overdamped.
A typical Rb MOT holds 10⁶–10⁹ atoms at densities ~10¹⁰ cm⁻³ in a ~1 mm cloud.
How it's observed, measured and used
A MOT announces itself visually: atoms scatter cooling light continuously, so the cloud glows as a bright point you can photograph with an ordinary camera. Quantitative diagnostics include:
- Fluorescence imaging: total scattered power gives atom number N (each atom scatters at rate up to Γ/2 ≈ 1.9×10⁷ photons/s).
- Absorption imaging: a resonant probe casts a shadow; optical depth yields column density and, after release, time-of-flight expansion gives the temperature directly (σ(t)² = σ₀² + (k_B·T/m)·t²).
Applications are everywhere in modern physics: MOTs are the first cooling stage for every Bose-Einstein condensate and degenerate Fermi gas; they feed optical-lattice and fountain atomic clocks (the primary Cs standard, and optical Sr/Yb clocks good to ~10⁻¹⁸); they seed cold-atom interferometers for gravimetry and inertial navigation; and single-atom MOTs in optical tweezers underpin neutral-atom quantum computers.
How it compares to related cooling and trapping schemes
The MOT sits in a family of techniques, each with a distinct job:
- Optical molasses is a MOT minus the magnetic field: it cools but does not confine, because it lacks a restoring force. Molasses reaches the same temperatures but the cloud slowly diffuses away.
- Magnetic (Ioffe-Pritchard, TOP) traps confine via the Zeeman potential of the magnetic field alone — no light — so they don't heat via photon scattering and are the workhorse for the final evaporative cooling stage to BEC. But they only trap 'low-field-seeking' states and don't cool.
- Optical dipole traps use far-off-resonant light and the AC Stark shift to make a conservative well; they trap any spin state and reach nanokelvin, but need atoms pre-cooled by a MOT first.
- Zeeman slower / MOT loading: a slower decelerates an atomic beam before the MOT captures it — complementary, not competing.
The MOT's unique virtue is that it does both jobs at once — cool and confine — from a hot, undirected vapor, which is why it is universally the first stage.
Significance, limits and open frontiers
The MOT is arguably the single most enabling apparatus of modern AMO physics. Steven Chu, Claude Cohen-Tannoudji and William Phillips shared the 1997 Nobel Prize in Physics for laser cooling and trapping; the 2001 prize (Cornell, Wieman, Ketterle) for Bose-Einstein condensation rested on MOTs as the first stage. Every optical atomic clock and cold-atom quantum computer traces back to this 1987 idea by Raab and colleagues, built on Jean Dalibard's proposal to add the magnetic field.
Known limits and frontiers:
- Density is capped at ~10¹¹ cm⁻³ by re-absorption of scattered photons (radiation trapping) and light-assisted collisions; 'dark SPOT' and compressed-MOT tricks push higher.
- Molecular MOTs: extending the technique to molecules (CaF, SrF, YO) is hard because they lack a closed cycling transition — first achieved in 2014 (DeMille group, SrF), enabling ultracold chemistry and searches for the electron's electric dipole moment.
- Chip-scale MOTs using nano-fabricated diffraction gratings (a single beam replacing six) are miniaturizing quantum sensors toward field-deployable devices.
| Quantity | Rubidium-87 | Sodium-23 | Cesium-133 |
|---|---|---|---|
| Cooling wavelength | 780 nm | 589 nm | 852 nm |
| Natural linewidth Γ/2π | 6.07 MHz | 9.79 MHz | 5.22 MHz |
| Saturation intensity I_sat | 1.67 mW/cm² | 6.26 mW/cm² | 1.10 mW/cm² |
| Doppler limit T_D | 146 μK | 235 μK | 125 μK |
| Recoil temperature T_r | 362 nK | 2.40 μK | 198 nK |
| Typical field gradient B' | 10–15 G/cm | 10 G/cm | 10–20 G/cm |
Frequently asked questions
Why does a magneto-optical trap need exactly six laser beams?
To cool and confine in all three spatial dimensions you need a counter-propagating beam pair along each of the x, y and z axes — that is 2 × 3 = 6 beams. Each pair provides velocity damping and, thanks to the magnetic field, a restoring force along its axis. Some modern 'grating MOTs' cheat by using one incoming beam and a micro-fabricated diffraction grating to generate the other five effective beams.
What does the quadrupole magnetic field actually do?
By itself it does not trap the atoms. Its role is to make the atomic transition frequency depend on position through the Zeeman effect: the field is zero at center and grows linearly outward. This shifts an off-center atom's resonance so it preferentially scatters photons from the beam pointing back toward the center, converting the pure velocity-damping of optical molasses into a genuine position-dependent restoring force.
How cold does a MOT get, and what sets the limit?
A basic MOT reaches the Doppler cooling limit, T_D = ℏΓ/2k_B, which is 146 μK for rubidium-87 and 235 μK for sodium. In practice polarization-gradient (Sisyphus) sub-Doppler cooling operates alongside it, so real clouds often reach 10–100 μK. The ultimate single-photon floor is the recoil temperature (362 nK for Rb-87); getting below that requires evaporative or Raman cooling.
What is the difference between a MOT and optical molasses?
Optical molasses is the six-beam laser configuration without the magnetic field. It provides viscous, velocity-dependent friction that cools atoms but supplies no position information, so the cloud slowly diffuses and falls away under gravity. Adding the quadrupole field turns molasses into a MOT by creating a restoring force, so atoms are both cooled and trapped at a fixed point in space.
Who invented the magneto-optical trap and when?
It was first demonstrated in 1987 by E. L. Raab, M. Prentiss, A. Cable, Steven Chu and David Pritchard, who trapped up to 10 million sodium atoms. The key idea of adding a magnetic field to Doppler cooling was proposed by Jean Dalibard. The broader laser-cooling program earned Chu, Cohen-Tannoudji and Phillips the 1997 Nobel Prize in Physics.
Why must the laser be red-detuned rather than on resonance?
Red detuning (laser frequency below resonance) is what creates velocity-selective absorption via the Doppler effect: an atom moving toward a beam sees it Doppler-shifted up toward resonance, so it scatters more photons from the beam opposing its motion, producing net friction. On resonance the two beams would scatter equally and there would be no cooling. Typical detunings are one to a few natural linewidths, e.g. −6 to −20 MHz for rubidium.