Optical molasses surrounds slow chemical atoms with opposing red-detuned laser beams, usually three pairs along perpendicular axes. In the elementary Doppler cooling picture, each pair damps one component of velocity. The result resembles viscous drag, even though the atoms are moving through a vacuum rather than a liquid.
Uniform balanced optical molasses supplies velocity damping, without a restoring force towards a chosen position. An atom at rest has zero mean force wherever it sits within the uniform illuminated region. Without noise or external forces, givesThe atom approaches a position determined by its initial condition, not the beam intersection. Recoil-driven diffusion and gravity eventually remove atoms from finite beams. A long residence time in optical molasses is therefore compatible with the absence of position trapping. An optical trap provides spatial confinement by a different force mechanism; a magneto-optical trap adds a position-dependent magnetic field and suitable beam electromagnetic polarization to create a restoring force.
Steven Chu's Bell Laboratories team prepared sodium atoms by pulsed-laser evaporation of a pellet. A frequency-swept opposing beam slowed them before six crossed laser beams supplied optical molasses. The vacuum chamber used a shield coated with graphite and cooled with liquid nitrogen as a pump. Fluorescence provided the readout; the initial beams were aligned as closely counterpropagating as possible. [ref]
The manipulation of neutral particles by Steven Chu. Steven Chu, The manipulation of neutral particles, 1998 publication of his Nobel lecture, pp. 687–690, Figs. 3–4 and footnote 5: sodium source, precooling, vacuum apparatus, beam geometry and fluorescence detection
.After equilibration, the team switched off the cooling light for a variable interval, then restored it. Faster atoms escaped during the dark interval; slower atoms were recaptured. The recaptured fraction versus delay constrained the velocity distribution and hence temperature. The inference depended on the initial cloud shape and recapture volume: Steven Chu later acknowledged adjusting a geometry correction towards the expected Doppler cooling limit. Delay was the scanned variable, and the cooling beams themselves supplied the release/recapture control; visible fluorescence alone could not establish temperature. [ref]
The manipulation of neutral particles by Steven Chu. Steven Chu, The manipulation of neutral particles, p. 688, release-and-recapture paragraph: measurement sequence and retrospective account of the geometry correction
.