Ultracold atoms and lasers seem to be opposites. Yet, “laser cooling” is one of the most common methods low-temperature physicists use to cool substances to within 1 degree of absolute zero.
Light amplification by stimulated emission of radiation (LASER) usually heats objects, and many industries employ lasers to cut - or rather, burn - through steel plates. The confusion with laser cooling is that most people mix up lasers’ ability to emit light of a specific frequency with lasers just being one color. Technically, laser cooling refers to a variety of techniques, but it most commonly refers to Doppler cooling, so named for its use of the Doppler Effect.
To understand how Doppler cooling works, let’s take a brief journey into quantum mechanics. Atoms consist of a positive nucleus surrounded by an electron cloud. Electrons are restricted to certain energy levels, but can be forced to change energy levels if bombarded with photons. If the electron is hit by a photon of the correct energy, the photon will be absorbed and the electron will jump up a few energy levels. Eventually, the electron will drop back down and emit photons until it returns to the lowest possible energy level. Additionally, photons have some momentum, even though they can, in most scenarios, be considered massless. Therefore, absorbing a photon actually slightly changes the motion of the electron and, by extension, the entire atom.
Given a sample of an element, if we can bombard it with photons of the correct energy, we can change the motion of its atoms, and thereby control those atoms’ kinetic energy. Since temperature is equivalent to average kinetic energy, being able to control kinetic energy means being able to control temperature. So in order to control any element’s temperature, we need a supply of photons of a similar energy. And coincidentally, since the energy of a photon is a function of its frequency, lasers, which output a single, consistent frequency, should be able to control temperature.
Now that we understand the basics, let’s dive into more technical details. A laser lets us change the direction of the atoms, but it can only add momentum in one dimension - the direction it’s pointing. Because of that, it can only affect one component of velocity. The solution is fairly simple; just point lasers at the sample from as many angles as are possible to keep everything centrally balanced. Additionally, in order to supercool, we want to add momentum in such a way that the speed of the atoms is reduced. However, most would think that if every atom absorbed photons, we would just knock around all the atoms. This is where the Doppler Effect comes into play.
In order to restrict the absorption of photons so that absorption only happens when atoms move in a certain direction, we can take advantage of the motion of each atom. In general, all the atoms are traveling in random directions. Some atoms happen to be moving towards the laser as they collide with an emitted photon. In this head on collision, the atom observes a higher frequency for the photon. This is the Doppler Effect - the same effect that explains why police sirens sound higher pitched coming towards you than they do going away from you. The Doppler Effect’s magnitude is dependent on the velocity of each individual atom. So atoms with different velocities will observe different frequencies for the same photon. Now, remember that electrons can only absorb particular frequencies because of their restricted energy levels. So for a given frequency, only atoms with a particular velocity will be able to absorb the photons. The other atoms will observe the incorrect frequency. Since velocity is speed and direction, if we choose a particular frequency, only atoms traveling in a particular direction can absorb photons of that frequency. The temperature gives us the average kinetic energy. However, not all the atoms will have the same kinetic energy because a range of velocities are present in the sample. Mathematical models can describe how many atoms have any particular velocity. That information allows us to determine the correct frequency for restricting photon absorption to only the desired atoms.
We have successfully established how we can influence the velocity of the atoms, but only the atoms we want to. But there is one step we have neglected. The motion of the atom is changed by adding energy to the electrons. However, that energy doesn’t stay stored there forever. As electrons eventually drop down to lower energy states, photons must be released to satisfy the conservation of energy. But these released photons also have momentum, so the atom feels a recoil whenever the electrons emit photons. How can we cool atoms if they regain part of their momentum after release? The answer is surprisingly simple. Whenever an electron absorbs a photon, the atom slows down. We’ve arranged the lasers and determined the correct frequency to ensure this occurs. But when the electron emits a photon later, it does so in a random direction. Sometimes that speeds up the atom, and other times that slows it down. Averaged over all the possible directions, the net result is that close to nothing happens. So even though the atom experiences random kicks with each emission, those kicks balance each other out in the long run.
This process we’ve described sounds perfect, so why haven’t we been able to cool things to whatever arbitrary temperature we desire? That has to do with the electrons’ emissions of photons. The effect on the average velocity cancels out, but the average squared velocity is different. Consider 101^2=10,201 and 99^2=9,801. In both examples, the velocities (101 and 99) are “off” by the same amount; both numbers are 1 away from 100. If the "off" velocities are equally likely, that difference averages out over time, and average velocity is unchanged. But for velocity squared, the two velocities are different distances from 100^2=10,000; the first is larger by 201, but the second is only smaller by 199. So when these get averaged, the squared velocity increases. And since temperature, or average kinetic energy, depends on velocity squared, the temperature is increased by the random emissions. As we go to lower temperatures, we discover that the cooling effect of absorbing photons is balanced by the temperature increase from the random emissions. So there is a lower bound on the temperature we can achieve. That lower bound is the Doppler cooling limit. However, Doppler cooling is still capable of reaching tens of microkelvin - tantalizingly close to zero kelvin.
So why do we care? Are we going to use super cold ice cubes for our sodas? Certain superconducting magnets, like those in magnetic resonance imaging (MRI) machines, must be kept at temperatures below 10 K. Ultracold atoms are also being investigated for use in quantum computers, which may require atoms in the millikelvin temperature range. Quantum computers take advantage of the fact that quantum bits can store 1s and 0s simultaneously in a superposition of both values. But if quantum bits interact with their environment, they will snap to one of the two states, becoming useless. Using ultracold atoms would reduce their random vibrations, which would help prevent such interactions. For that and more, Doppler cooling will remain relevant as the easiest way to create ultracold atoms.