When we use laser light to cool atoms, we get into some problems, because when atom beam slows down the Doppler shift changes the frequency of light in atom's frame of reference, so they can't continue to absorb it and lose speed/cool down.

But with light composed of different wave lengths we could avoid such a problem, because at lower speeds atoms would start absorbing photons with higher frequencies - thus we would eliminate problems caused by Doppler shift and everything would be well. But instead we use Zeeman cooling, Doppler cooling and other complicated techniques, when the simpler solution would suffice.

Why am I wrong and light composed of multiple frequencies can't cool our atoms?

  • 3
    $\begingroup$ Did you try it? Ordinary light is chaotic, observing it to cool down atoms would be an unexpected phenomenon. Laser light is more orderly and can be controlled so it can be used to cool down the atoms, instead of heating them up. Some mixture of laser beams with different wavelengths could work as well. $\endgroup$ Commented Mar 14, 2023 at 14:37
  • $\begingroup$ Lasers are a "simple" solution these days. One can, of course, imitate some properties of laser light with suitable spectral shaping of thermal and atomic/molecular light sources, but the efficiency of such light sources would be horrendous. $\endgroup$ Commented Apr 11, 2023 at 23:17

3 Answers 3


Doppler shift is not a problem, but the essential element in laser cooling mechanism: the laser frequency is chosen in such a way that it is just below the Doppler shifter natural frequency of the atom: thus, the atom absorbs light and then re-emits it at a higher energy, in the process losing some energy and slowing down. Broad spectrum light source would completely eliminate the effect of cooling, as it would work in both ways: heating and cooling atoms simultaneously.

The reason why we have to resort to the Doppler cooling is that we want to get to very low temperatures, which means that variations in energy supplied/subtracted from atoms should be at least of the same order (or lower) than the energy that we try to reach - i.e., lower than the laser line width. Incoherent light sources, of course, have much higher line widths and cannot be used for this purpose.

Perhaps this illustration could be helpful (image source): enter image description here

  • $\begingroup$ Why would atoms be heating? Wouldn't light just pass them if it can't be absorbed? $\endgroup$ Commented Mar 14, 2023 at 17:11
  • 2
    $\begingroup$ @EdwardHenryBrenner if detuning is big, it just passes by. By can we have laser heating as well as laser cooling. $\endgroup$
    – Roger V.
    Commented Mar 14, 2023 at 18:01

The idea is that when an atom moves towards a laser source tuned just to the red of one of its natural transitions, it absorbs a photon and is slowed down. If it's moving away, the laser light isn't absorbed and the atom is unaffected.

With white light, just as much energy would be absorbed by atoms moving towards or away from the light source. This would mean as many atoms were accelerated as decelerated.

The key point is, the laser line needs to be narrower than the equivalent velocity spread of the atoms.


I want to add to the previous answer, that for cooling, one needs to have a closed cooling cycle. If one energy state is not addressed by the light, this becomes an escape channel. E.g. for Rubidium, one typically needs to add another frequency to the cooling laser light to repump the atoms into a state which is addressed by the cooling light.

Laser cooling is very delicate and only works for certain configuration of atoms and laser light. If you have laser light blue detuned to an atomic transition, the atoms will be accelerated by the light, defacto heating the atoms. (Minor remark: Blue detuned magneto optical traps are possible, but harder to achieve)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.