A magneto-optical trap, or MOT, is the most common class of device used for laser trapping/cooling. The development of the MOT was an important development in atomic physics; its development brought Stephen Chu, Claude Cohen-Tannoudji, and William D. Phillips the 1997 Nobel Prize in Physics. It was an advance over the earlier 'optical molasses' as it allowed for trapping as well as slowing of the atoms. The atoms thus confined are useful for atomic measurements, used for production of Bose-Einstein condensates (itself a Nobel-prizewinning development), and allow high-precision nuclear measurements when used to trap unstable isotopes.
The full MOT depends on a succession of methods which each solve one part of the cooling and trapping problem. The most basic is the ability to push atoms around using laser beams; if this were impossible it would be unlikely that a laser trap would be feasible.
One might expect that a laser beam, carrying a coherent flow of energy, would exert a pressure upon materials in its path. For macroscopic objects, the energy in a laser beam is small and rarely produces a measurable pressure, but free atoms, light as they are, are a different matter. However, atoms are not solid objects but rather have complex structure described by quantum mechanics. Similarly, the behaviour of light on this scale is quantum in nature. Thus, only a quantum mechanical treatment will show how the laser light is able to exert a force on the atom.
Consider an atom with only two internal states, differing in energy by Δ; this is a vast oversimplification, but this is theory so it's allowed. The essential features are the same as for real atoms. If we illuminate the atom with a laser whose frequency is Δ/h, where h is Planck's constant, each photon in the laser beam will also have an energy Δ. Thus, by absorbing one photon from the beam, an atom in the lower (ground) state can transition into the upper (excited) state. Photons carry momentum, so this absorption exerts a force on the atom. The frequency must be tuned, since the atom only interacts with photons that are able to change its state, and the state can only change from photons with just the right energy, since all of these things are quantized. The condition of the photons matching the atom energy is called being "on resonance".
Now the thing about atomic excited states is that they decay by emitting photons. So the atom will, after absorbing a photon, quickly return to the ground state by emitting a second photon with equal energy (and thus the same momentum) to the first. On the face of the matter, this would appear to defeat the goal of exerting a force on an atom with a laser. But, an absorbed photon is always moving in the same direction, while the emitted photon comes out in a random direction. Through many cycles of absorption and emission, the momenta of the emitted photons will cancel out as the momenta of the incident photons add up. Thus, the atom will experience a net force in the direction of the laser.
This effect is not enough to cool down the atoms, which is equivalent to slowing them down. Even if we use two opposing laser beams, the forces will just cancel, leaving the atoms in essentially the same condition as before adding the lasers. To slow down the motion of the atoms, the force exerted by the laser beams must depend on their speed. Fortunately, we can do this using the Doppler effect.
Consider two laser beams, pointing in opposite directions, with a photon energy slightly less than the atomic transition energy Δ. A stationary atom situated in the beams will absorb no photons and thus experience no force, since the laser photons are not on resonance. We want this, since any force on a stationary atom would accelerate the atom, contrary to our purposes. The situation is different for a moving atom. An atom moving along the direction of the beams will observe a Doppler shift in the frequency of the light photons. This increases the frequency of the light from the laser it is moving toward, and decreases the frequency of the other laser light. Frequency is directly related to photon energy, so the one energy increases while the other decreases. If the atom is moving fast enough, the laser light moving in the opposite direction as the atom will shift to have a photon energy of Δ, at which point the photon is on resonance and the atom excites. Thus, the atom will experience a force opposing its motion.
In this case, then, the atoms will experience forces that are selective on their velocity along the direction of the laser beams. For a real system, both the energy of the laser beam and the energy required to excite an atomic transition are 'fuzzy', so the atom will experience a force for a range of velocities rather than a single very well-defined velocity. This arrangement is referred to as 1D optical molasses, slowing the atoms' motion along one axis.
It is a relatively simple matter to combine multiple pairs of lasers along different axes to make multi-dimensional optical molasses. A 2D combination of four laser beams cools atoms in a plane; this arrangement is often called an 'optical funnel'. Six beams can be used to cool atoms in all three directions, making true, 3D optical molasses. This system impedes the motion of atoms in the area covered by all three beam pairs regardless of its direction, acting on the atoms like a viscous fluid does on macroscopic bodies.
Optical molasses have a fatal flaw for trapping atoms; there is no dependence on the position of the atoms, which prevents their confinement to a particular volume. If an atom happens to work its way toward the edge of the molasses region, no force will prevent it from escaping the lasers' influence completely. A result called the Optical Earnshaw Theorem, which I will not prove here, prevents the optical system itself from providing a position-sensitive potential in the manner that is needed. We avoid this and trap the atoms anyways by modifying the structure of the atom itself via the Zeeman effect.
The Operation of the MOT
A MOT cannot be constructed using the two-level atoms used to describe the optical molasses. Consider instead an atom with one ground state and three excited states. These excited states are distinguished from each other by having different projections of the electron's angular momentum along the z-axis: 1, 0, and -1 (in the basic unit of angular momentum, h/2π). Ordinarily, these three states will have exactly the same energy relative to the ground state, Δ.
If we introduce a magnetic field along the z-axis, the energies of the state are modified by the interaction of the electron with the magnetic field. This change in energy is proportional to the product of the angular momentum projection, m, and the magnetic field strength, B, so the states with m=1 and m=-1 change in energy in opposite directions by the same amount which we'll call δ. The other state has m=0, and so does the ground state, so they will not shift in energy. This energy shift is called the Zeeman shift.
If we set up the magnetic field for our MOT such that it increases proportional to the distance from a centre, where the field is zero, the Zeeman shift will also increase proportionally with this distance. If the magnetic field switches direction at the centre, one state will be shifted down at one side of the minimum and the other will be shifted down at the other side. This allows us to have the desired position sensitivity.
For this to work, we need the beam from one side to only excite the m=1 state, and the beam from the other side to only excite the m=-1 state. This can be arranged by the use of the correct beam polarization. There exists a polarization state that only excites to the m=1 state, called σ+, and one that only excites to the m=-1 state, σ-. If we put σ+ light in from the side where the m=1 state shifts down, and σ- light in from the other side, and fix its frequency at less than Δ/h, then the atoms will experience a restoring force toward the centre, having essentially the same form as a simple harmonic oscillator.
Moreover, the laser setup remains an optical molasses, so this oscillator is damped. Thus the atoms oscillate around the centre from the Zeeman force, and have their energy drained away by the Doppler force. This very quickly produces a cloud of slow, trapped atoms around the centre of the magnetic field.
The Construction of the MOT
All of these careful preparations are spoiled if air molecules are present along with the atoms to bump our trapped atoms away from the centre of the trap. Thus, a MOT must be made within a high-vacuum chamber, which can be a glass bulb or a metal chamber with optical windows. A perfect vacuum would mean that there are no atoms available to trap, though, so the vacuum chamber must be filled with a diffuse vapour of the element to be trapped. The MOT is strong enough to capture some atoms directly from the vapour.
The magnetic field is most conveniently provided by a quadrupole arrangement, realized with a pair of coils situated at the top and bottom of the chamber. If the currents in the coils travel in opposite directions, then the magnetic field generated is as desired. (This arrangement is opposite that of Helmholtz coils.) The polarizations σ+ and σ- are oppositely-oriented circular polarizations, making them relatively easy to realize. In fact, since the orientation of the circular polarization is defined relative to the direction of the light, the two beams can come from the same source and become oppositely oriented by virtue of being oppositely directed.
The Limitations of the MOT
The most significant limitation of the MOT technique is that it requires that the atoms to be trapped have a relatively simple atomic structure. Specifically, it requires that the atom have a 'cycling transition', where if the atom is excited there is a reasonable chance it will return to the state it was excited from. This is difficult to achieve if there are multiple valence electrons available to excite and de-excite, as they will tend to "get in each other's way" and prevent trapping on a cycling transition.
Thus, the most common atoms used in MOTs are alkali metal atoms; most commonly rubidium and caesium, but potassium MOTs are also common, and sodium and lithium MOTs not unheard of. Francium is fairly easy to trap, but is very radioactive and so is difficult to obtain. Hydrogen is, in theory, easy to trap, but the (far-ultraviolet) laser frequency required is so high that producing and manipulating a laser beam at that frequency is presently infeasible. (Most optical materials are opaque to far-ultraviolet light) Present research interest towards hydrogen traps is directed with an eye towards containing anti-hydrogen atoms.
This is not to say that alkali atoms are the only ones that can be trapped using MOTs. The alkaline earth metals are being worked on, and magnesium, calcium, and strontium traps have been achieved. Noble gases have effectively no valence electrons, but if one electron is excited to a higher state it remains there long enough to trap with. This has already been done for all noble gases except radon, and radon will be trapped someday as well.
Applications of the MOT
As mentioned above, the MOT is a valuable tool in many fields of atomic and subatomic physics. Its applications include precision spectroscopy, production of Bose-Einstein condensates, and precision decay measurements. All of these applications leverage the MOT as a source of slow, cold atoms. Spectroscopy is improved by the use of a MOT because the motion of atoms in a warm gas smears out the atomic spectrum through the Doppler effect. The MOT, along with other laser cooling methods such as Sisyphus cooling enabled the production of Bose-Einstein condensates, which can only form at very low temperatures.
In my own research, we use the MOT to study the beta decay of short-lived alkali isotopes produced with the ISAC radioactive beam facility at TRIUMF. The Standard Model of particle physics gives very precise predictions of the distribution of decay products from a nucleus undergoing beta decay. That distribution is difficult to measure because one of the decay products, the neutrino, is essentially undetectable. If the energy and momentum of the beta particle and the daughter nucleus can be measured, though, conservation of energy and momentum can be used to determine the neutrino momentum, if the initial momentum of the nucleus is known. This is where the MOT comes in; if the nucleus decays while in a MOT, its initial momentum will be known to be very small. In addition, the MOT is well-defined in position, so the detection of the decay products can use methods that depend on knowing the initial position of the atom. These advantages combine to produce a more precise measurement than other methods of studying beta decay.
H. Metcalf and P. van der Straten's book Laser Cooling and Trapping, published 1999.
Lectures at the 2005 TRIUMF Summer Institute, notes available at http://triumf.ca/tsi/
Experience as a graduate student with the TRINAT group at TRIUMF over the past year.
This writeup is copyright 2005 D.G. Roberge and is released under the Creative Commons Attribution-NoDerivs-NonCommercial licence. Details can be found at http://creativecommons.org/licenses/by-nd-nc/2.0/ .