Zapperz, back from vacation, notes that Willis Lamb passed away recently. The Lamb shift, the energy splitting of the 2s and 2p states of the hydrogen atom, was a huge confirmation of quantum electrodynamics and garnered him the Nobel prize, and you can read more about that here. But that’s not the only effect named after him. Another artifact is the Lamb dip.

The Lamb dip is not a sauce, nor is it related to sheep dip. It arises in a certain geometry of spectroscopy: if you pass a near-resonant laser through an atomic vapor, some of the light will be absorbed. If the laser’s frequency is scanned, you will map out an absorption profile of the atoms, but because they are moving, the absorption depends not only on the transition frequency, but also on the motion of the atoms, which causes a Doppler shift. So your absorption profile is really a representation of the thermal motion of the atoms. At any one frequency the light will be absorbed by those atoms whose motion places them in resonance with the light.

If a counterpropagating beam is added (perhaps by reflecting the incident beam), the two beams will be interacting with different atoms in the vapor — an atom can’t be moving toward (or away from) both laser beams at the same time, so the Doppler shift can only put it into resonance with one beam at a time . . . except for atoms moving perpendicular to the beam. Those atoms have zero velocity with respect to the beam, so there is no Doppler shift. What happens is that one of the beams (the “pump” beam) excites the atoms, which “burns” a hole, i.e. it puts a number of the atoms into an excited state. The other beam (the “probe”), since it “sees” fewer atoms in the ground state which might absorb the light, suffers less absorption, and there is a dip in the absorption profile when you detect that light. The Lamb dip. You can see a representation of this type of setup, known as saturated-absorption Doppler-free spectroscopy (scroll down to section 5.2)

“Burning” the hole to the point that it’s noticable “saturates” the transition, in the jargon of atomic physics; at very low light levels the amount of absorption rises linearly with intensity. At the point where this stops being linear, the transition has saturated.

How is this useful? The dip’s minimum is at the resonance of the transition, so you know what that frequency is. There are ways in which you can basically take the derivative of this signal (which involves some tricks like modulating the laser and using a lock-in amplifier, and is a little too involved for this post), but that leaves you with a signal that is basically linear over a small range and passes through zero: ideal for servo-locking. And so this is what you do, you set up a servo loop to drive your laser back toward the zero crossing should it ever start drifting away from that point. Now you have a stable laser locked to a known frequency, and there are loads of things you can do with that, including precision spectroscopy and trapping atoms.

(Edit: removed dead link and added my own spectroscopy picture)