Cavity Quantum Electrodynamics
The interaction of a single two-level atom with a single quantized radiation mode is the textbook example of quantum electrodynamics. However, it is only in the last two decades that laboratory experiments have been able to mimic this idealized example. We work on such an experiment combining a small, ultra-high finesse optical cavity with an atomic fountain as a source of ultracold atoms. Single atoms can be trapped inside the cavity and probed with an intracavity field that consists of less than one photon on average. In this way, we prepare a system which allows us to study fundamental aspects of matter-light interaction on the quantum level.
Feedback cooling of a single atom
Feedback is one of the most powerful techniques for the control of classical systems. In the quantum world, its implementation is complicated by the inevitable backaction caused by a measurement, as well as by quantum fluctuations perturbing the evolution of an open system. We demonstrate real-time feedback control of an atom trapped in an optical cavity. Information about the atom's position is derived from probe photons transmitted through the cavity. A field-programmable gate array (FPGA) estimates the atomic trajectory from this information and switches the depth of the confining potential in order to counteract its motion. Using this feedback protocol increases the average storage time of an atom in our cavity by a factor of 30 to more than one second.
Three-photon correlations in a strongly driven atom-cavity system
In the strong coupling regime, single excitation quanta are coherently exchanged between the atom and cavity in a process known as vacuum Rabi oscillation. When the system is also strongly driven, it is also possible to see the coherent exchange of energy between the atom-cavity system and the driving laser. We measure this "super Rabi" oscillation via the second-order correlation function of the transmitted light, which gives the probability of detecting two photons separated by a time interval τ.
Investigating also the three-photon correlation function shows a curious interplay between vacuum and super Rabi oscillations. In particular, the output photon stream is asymmetric in time - the probability of detecting two photons separated by τ1 followed by a third after τ2 is not the same as the probability of detecting two photons separated by τ2 and then one after τ1. The occurrence of such time-asymmetry is evidence for the breakdown of the principle of detailed balance in a driven system far from equilibrium.
A high-finesse cavity with variable length
Our optical resonator consists of two highly polished mirrors which reflect more than 99.998% of incident light. The cavity finesse is 195 000, meaning that a photon which enters will bounce back and forth between the two mirrors many thousands of times before leaving the cavity. As a consequence of the ultrahigh finesse, the length of the cavity (which determines its resonance frequency) must be very stable. We actively stabilize our cavity using a piezo-electric tube to a residual rms noise level of less than 0.001 Å. This is more than 1000 times smaller than the diameter of an atom!
In addition, we can vary the length of our cavity between 50 µm and 10 mm without removing it from the vacuum system. This is possible because one mirror is mounted on an inchworm motor, which crawls along a fixed rod in steps of less than 1 nm. Varying the length of the cavity allows us to change the strength of the coupling between atom and cavity as well as the cavity resonance frequencies and field decay rate. This makes the setup much more versatile and suitable for a wide range of experiments.
In our experiment cold rubidium atoms are launched upward with velocities comparable to that of a typical garden fountain. The operating principle of the atomic fountain is quite different, though.
In an atomic fountain, the atoms are launched using light forces. Cold Rb atoms are prepared in a magneto-optical trap (MOT), which confines atoms by light pressure from laser beams illuminating the atom cloud from all directions. We trap and cool typically ten million atoms in a few hundred milliseconds from background vapour.
After trapping, optical molasses cooling is applied to cool the atoms further to about 5 μK. Then the frequency of the lower MOT beams is detuned relative to the upper beams. This still cools the atoms, but now in a moving reference frame, thereby accelerating the atoms in the vertical direction. The lasers are switched off after a few milliseconds, and the atoms continue their ballistic trajectory up towards the optical cavity.
Trapping single atoms
As an atom enters the cavity mode, it tunes the cavity out of resonance with the probe beam. This causes a drop in transmission, which is detected and triggers an increase in the dipole trap depth. The atom is now trapped in the cavity. Its position changes the resonance frequency of the cavity and therefore the transmitted light intensity. Now, feedback can be applied to cool the atom down, increase its coupling to the cavity mode and prevent it from escaping.