Univerza v Ljubljani

Fakulteta za Matematiko in Fiziko

Oddelek za Fiziko

Seminar za 4. letnik

L A S E R C O O L I N G

Jurij Pahor

Mentor: prof. Martin Čopič

Ljubljana, maj 2002

1. Introduction

This paper will present a brief overview of methods of laser cooling and physical models behind it. The methods of laser cooling exploit forces which arise from interaction of laser light with neutral atoms. Essentially, there are two basic mechanisms for deccelerating a beam of neutral atoms.

The first one relies on radiative force as a result of light pressure. Atoms in a beam lose their kinetic energy by scattering light of a counterpropagating laser beam. The decceleration occurs because atoms gain momentum in the direction opposite to their motion and lose it in random directions by spontaneous emissions. This process provides for a dissipative net force acting on neutral atoms in the direction of propagating light.

The second mechanism embodies spatial variation of light field, resulting in a gradient of atomic energy. The force, driven by this energy gradient is called the dipole force. It can be thought as the force acting on an optically induced electric dipole moment of the atom in an inhomogeneous light field.

Another very large field of atom manipulation with light is trapping of neutral atoms. This is most readily accomplished by combining optical forces with magnetic field gradients, although experiments based on purely optical forces have been carried out.

2. The Radiative Force

The radiative or radiation-pressure force originates from an atom absorbing a photon, thus gaining a momentum in the direction of light propagation. The atom is left in the excited state and eventually returns to the ground state by spontaneous emission. The photon is emitted with random polarisation in random direction and after many scattering processes its average contribution to the atomic momentum equals zero. The resulting average force is then , where is the scattering rate.

We shall calculate using a simple two-level atom model in semiclassical approximation with the light field as a classical electric field. The result is illustrative and to some extent practically applicable although real atoms have multiple energy levels. Let us begin the discussion with first defining:

laser light frequency

atomic transition frequency

detuning of laser frequency from atomic transition frequency

natural radiative decay rate of the excited state , with as lifetime of the state

intensity and saturation intensity of laser light

the Rabi frequency at which the atomic state oscillates between and for zero detuning.

The time-dependent Schödinger equation is

, / (1)

with the field free hamiltonian and the atom-light interaction hamiltonian . We write the solution as an expansion in terms of eigenfunctions of

. / (2)

After inserting expansion (2) into equation (1), multiplying from left with and integrating over spatial coordinates we get the equation relating expansion coefficients

, / (3)

where and . In our approximation there are only two coefficients in expantion (2), namely cg and ce , for which eq. (3) becomes

. / (4)

With the interaction hamiltonian and the light field we write the diagonal elements as

, / (5)

where is the Rabi frequency. Here we neglected the spatial variation of of order of several hundred nanometers in the region containing atomic wavefunction, which is of order < 1nm. Next, by differentiating one of the equations (4) and substituing it in the other, we obtain two uncoupled, second order differential equations for cg and ce. The initial conditions for this set of equations are and .

Let us write the solution for time dependence of the excited state coefficient:

. / (6)

The probability of finding an atom in the excited state oscillates with the frequency and is equal to the Rabi frequency as defined above for zero detuning of the laser light frequency from the atomic resonance frequency.

The result (6) doesn't allow any net force but we haven't accounted for spontaneous emission yet. Using the Schrödinger equation while taking into account all possible states the photon can be emitted to, would be a rather difficult task. Instead we introduce the density operator with the matrix elements

. / (7)

With this formalism it is possible to describe the evolution of a pure state into , which is not a pure state any more, because the information about the emitted photon is disregarded. The probability for the excited state of an atom is now . The time evolution of the density operator is described with

. / (8)

Spontaneous emission is taken into account with an exponential decay of the coefficient with a constant rate , thus . Using , along with eqs. (4) we obtain a set of equations for density matrix coefficients

, / (9)

where and . The terms including take into account the spontaneous emission effect. The set of equations (9) can be solved by applying two general relationships between the density matrix coefficients; the conservation of the population and the identity .

The solution for is

/ (10)

and with this result we can finally write the scattering rate as

. / (11)

From expression (11) follows that the scattering rate saturates to for high laser intensities with being proportional to the optical field amplitude , thus being proportional to field intensity. The maximum force based on spontaneous emission is then

. / (12)


The discussion until now included the force exerted on atoms by a traveling wave but this is not a true cooling force since it is acts in one direction only. A dissipative force that vanishes for an atom at rest is required for cooling. Such force arises in a standing wave, detuned below resonance, as a consequence of the Doppler shift. The atoms traveling with velocity see the frequency of the light to be (Figure 1), so that becomes .





Figure 1: A moving atom sees the light traveling in the opposite direction with higher frequency and the light traveling in the same direction with lower frequency than an atom at rest.

Now the atoms see the light of the counterpropagating beam Doppler shifted upward, closer to the resonance, wheras they see the light of the other beam shifted further away from the resonance (Figure 2). Consequently, they scatter more light from the beam propagating opposite to their motion than from the other beam, which results in an average force slowing the atoms down.

Figure 2: Schematic dependence of the scattering rate R from laser light detuning in units of natural width parameter . The cooling process takes place when te atoms traveling through a standing wave of laser ligth, detuned below resonance (), see one beam Doppler shifted towards the resonance and the other beam shifted away from the resonance.

The light must be originally detuned below resonance, as can be seen from the following.The magnitude of the force in one direction is

, / (13)

where the '+' sign stands for the direction of and '-' sign stands for the opposite direction. The total force acting on the atom is the sum , so

. / (14)

The force is deccelerating for negative detuning but is accelerative for positive detuning. Also, for low velocites, the terms in the denominator can be neglected towards , thus making the force proportional to the velocity. This is then a true damping force, thus it compresses the velocity distribution of the atoms, which is the essence of the cooling process. Because of the viscous damping the technique is referred to as optical molasses. To make an example, the term amounts to ~10 MHz for and , whereas the detuning can be of the order of , tipically 100 MHz.

Although it may seem that atoms can be cooled to an arbitrary low temperature, it is not so in reality. The scattering force is oriented towards the direction opposite to the atoms motion only on average. The individual momentum kicks of spontaneously emitted photons have random directions resulting in heating the sample. Obtainable temperature is then limited by the competition between damping and heating, both being equal in a steady state. The lower temperature limit is called the Doppler limit. A rough estimate for this limit is the kinetic energy, the atom gains, after emitting a photon. The velocity and the kinetic energy of the atom, after considering the conservation of the momentum and the energy of the atom-photon system, is

. / (15)

The typical value for is 1 eV and for the is several Gev which amounts to a few 10 cm/s for atomic velocity and several tens μK. This estimate is rather low, since it accounts only for a single photon emission. The lower temperature limit is in reality larger for a magnitude, that is of the order of several hundred μK.

3. Cooling Below the Doppler Limit – The Dipole Force

In the previous discussion we derived a force, which relies on the process of the spontaneous emission and can cool a sample of atoms down to the Doppler limit. A different type of force, which is conservative instead of dissipative, can be produced. The origin of the dipole force lies in a gradient of the atomic energy in the light field. In an appropriate setup, the potential can be harmonic. An ansamble of atoms, trapped in the potential, is being cooled by evaporation of the atoms out of the trap with simultaneously lowering the depth of the trap.

We shall begin the discussion with calculating the energy shifts of a two-level atom states under the influence of the light field. Because of their origin they are called light shifts. The light shifts are most readily found by eliminating the explicit time dependence in eqs. (4).

This is done by the rotating frame transformation, as in eqs. (9), that is

. / (16)

Inserting the new coefficients (16) into eqs. (4) gives

. / (17)

This set of equations is basically the Schrödinger time-dependent equation for the coefficients (16) with the hamiltonian

. / (18)

The shifted energies are given as the eigenvalues of the hamiltonian (18)

. / (19)

For the sake of simplicity we shall now concentrate on the case of , so that the energy of the ground state shifts to (without the light field it is set to zero), while the energy of the excited state remains unchanged . In a standing wave, formed by two light beams of equal polarisation, the intensity changes from zero at the nodes to maximum value at the antinodes. The Rabi frequency is proportional to the field amplitude, thus enabling a gradient force to arise

. / (20)

The dipole force is, in contrast to the radiative force, position dependent and is thus appropriate for trapping the atoms. Effective optical traps can be build with first cooling a sample of atoms with the radiative force and then trapping it with the dipole force. The atoms in the trap gets in thermal equilibrium by collisions. The atoms with higher velocities evaporate from the trap which results in lower temperature of the remaining sample. Eventually, lower temeratures can be achieved by lowering the potential well of the trap.

4. Conclusion

We have considered the interaction of the neutral atoms with light, accounting for two diffrerent types of forces. The radiative force is a dissipative, velocity-dependent force, which cools the atoms. The mechanism producing the radiative force is absorption, folowed by spontaneous emission of the light. The reachable temperature is of order of several hundred μK. The dipole force is on the other hand conservative, since it originates from a gradient of atomic energy in the light field. The energy gradient emerges from the light shifts of the atom. The dipole force potential is appropriate for trapping the atoms, whereas the cooling process includes evaporation of the atoms out of the trap. With this technique, lower temeratures are achievable, of the order of μK.

References

H. Metcalf, P. van der Straten, Laser Cooling and Trapping, Springer-Verlag, New York (1999)

W. D. Phillips, Optical manipulatoin of atoms at the nanoscale, Proc. of the international school of physics »Enrico Fermi«, Società Italiana di Fisica (2000)

Y. Shimizu, H. Sasada, Mechanical force in laser cooling and trapping, Am. J. Phys. 66 (11), Nov. 1998

1