When I talk to family and friends about laser cooling I always hear the repsonse: "but lasers heat things, don't they?". Well, normally they do, but when you get to the level of single photons interacting resontantly with single atoms it begins to look a lot like dodgeball...
Imagine you are in a gym hall and at either end there are teams with an infinite amount of balls. As you move towards one side, that team throws the balls at you. As you catch the ball you stagger back a little as the ball is heavy (it has transferred momentum). You then throw the ball away in a random direction and recoil a little bit at the ball leaves your hands. If you move in the opposite direction the same thing happens but from the opposite team. If the teams continually throw enough balls at you then you find it hard to reach either end of the hall as you move so slowly. If you are completely surrounded by ball-throwers then all of your movements are slowed down. This is essentially the effect of 'optical molasses' in which thousands of photons in a laser beam impart momentum to the atoms which slows (and therefore cools) them down. The directionality is induced via the Doppler effect so that only atoms moving towards detuned lasers beams are in resonance and absorb the photon momenta. In the anology, this is as if you are a bit clumsy and can only catch a ball if it is thrown at just the right speed and therefore your motion with respect to the ball will determine whether you can catch it or not (this is your 'resonance'). We can trap the atoms with magnetic fields which affect the rate the atoms absorb momentum by shifting the resonance using the Zeeman effect. In our dodgeball analogy, the balls are thrown at you harder the further you move from the centre spot, thus driving you back to the same position. Collectively, the cooling and trapping system is called a 'Magneto-Optical Trap (MOT)'.
Magneto Optical TrapA typical MOT is formed of 6 mutually orthogonal beams intersecting at the zero point of a magnetic field gradient. This requires careful alignement and polarization control of the beams into a vacuum chamber. This geometry is fragile and not practical for devices which could be depolyed in vehicles, spacecraft, down mineshafts, or under the sea. Different geometries have been considered such as the Mirror-MOT, Pyramid-MOT, and Tetrahedral-MOT. The latter has been developed into the planar structure using 2D gratings (known as the Grating-MOT) which is high suited for miniature MOT devices and requires only a single beam. We are working with the University of Strathclyde in developing these structures for 'MicroMOTs'. We are also developing compact four beam (or two retroreflected) geometries that require only a single vacuum viewport for situations where one requires several wavelengths, or pure optical wavefronts, in which diffractive grating a disadvantageous. More details on those will follow as we publish the results.
The aim is to trap as many atoms as possible in the shorted volume. The scaling laws of MOTs have been explored extensively and there is a clear correlation between laser beam diameter and atom number. One usually requires at least 106 atoms to ensure adequate signal to noise in sensitive measurement, and this requires ~5mm beams, and thus cm scale packages. To generate degenerate gases one requires several orders of magnitude higher atom numbers to maintain a suitable phase space density during evaporative cooling. There are some trick to increasing trap density such as atom beam loading, dark spot traps and bichromatic forces. The Atomic Devices and Instrumentation Group at NIST in the US are currently exploring the limits of MOT dimensions.
