Researchers Achieve Lasing From a Single Trapped Atom

Conventional laser cavities contain huge numbers of atoms and photons, each contributing to the radiation field by stimulated emission. The effect of an individual photon on the dynamics of such a system is negligible. But if reduced to its bare essentials, a laser would become a pumped single atom that interacts with a single electromagnetic mode in a cavity. In so minimal a system, quantum fluctuations on the scale of single photons can become dominant and the usually familiar behavior of a laser can become very unfamiliar.

In the mid-1980s, groups led by Herbert Walther (Max Planck Institute for Quantum Optics) and Serge Haroche (Ecole Normale Supérieure) approached that conceptual limit when they built single-atom micromasers based respectively, on one-and two-photon transitions. ^1,1 The devices worked using a dilute beam of Rydberg atoms that transit a superconducting cavity a few centimeters in length. In a cavity tuned to the frequency of an atomic transition, each excited atom exchanges energy with the cavity to produce a photon. That photon intensifies an interaction with the next atom that enters and makes the atom’s transition to the ground state more likely. The effect of the interactions is to build up an equilibrium intracavity field that sustains the micromaser.

In 1994, Michael Feld, Kyungwon An, and their MIT colleagues extended the achievement to optical frequencies by developing a single-atom laser that used a similar stream of atoms introduced one by one into a cavity. ³ (See Physics Today, February 1995, page 20 .) The much shorter optical wavelengths required mirrors with ultrahigh reflectivities. ⁴

The use of a dilute beam provides a way for single atoms to interact with the cavity. But their introduction into the cavity is itself a source of intensity fluctuations, because each atom enters the cavity at random intervals. Moreover, the microlasers achieve steady state through incremental contributions of many atoms that transit the cavity.

Now, Jeffrey Kimble and colleagues at Caltech have removed this final source of fluctuations by eliminating the beam entirely. ⁵ Their single-atom laser device, shown below, operates by using “one and the same atom,” as Kimble describes it. The atom is confined within the resonant cavity using a laser trapping beam, while external pump beams force it to emit repeatedly.

Cavity QED and laser cooling

Strong coupling of an atom to a single cavity mode accounts for the lasing from a one-atom system. Such coupling arises simply from an atom’s presence within a cavity: The vacuum field fluctuates and polarizes the atom, and in return the induced atomic dipole fluctuates and polarizes the cavity field. (See the article by Yoshihisa Yamamoto and Richart Slusher in Physics Today, June 1993, page 66 .)

The coherent interaction prompts an exchange of a photon between the atom and cavity field at the single-photon Rabi frequency 2g, with an interaction energy given by ℏg. To ensure that the coherent rate g is large enough to exceed the cavity decay rate and the rate of atomic spontaneous emission, researchers work with tiny cavities—the electric field per photon scales as √ℏω/V_m, where V_m is the mode volume—and with very highquality mirrors. Kimble and his colleagues Jason McKeever, Andreea Boca, David Boozer, and Joseph Buck used mirrors spaced just 42 µm apart and chose to work with cesium atoms, whose large dipole moments strongly couple to an electric field.

To transform a cavity into a laser system comprising a localized single atom, Kimble’s group had to combine the techniques of laser cooling and trapping with those of cavity quantum electrodynamics. Much of the work over many years involved increasing the lifetime of atoms within the cavity beyond that possible with thermal atomic beams. A significant experiment in 1996 demonstrated that cold atoms could be observed in real time as they fell one-by-one through an optical cavity. Using cold atoms increased the transit time to roughly 100 µs (about 1000 times longer than conventional beams). ⁶

That work set the stage for atom trapping. By monitoring the dramatic change in transmission from the mirrors, the group could tell precisely when trapping occurred and how long it took stray heating to boil off the atom. In 2003, Kimble’s group achieved a major milestone: the development of a novel type of far-off resonance dipole-force trap (FORT). The FORT sets up a standing wave in the cavity from a laser tuned far from the atomic transition frequency and weakly polarizes the atom, prompting it to move to regions of highest light intensity. Because an atom’s center-of-mass motion is only weakly dependent on the atom’s internal state, the FORT could keep an atom trapped and strongly coupled to the cavity for lifetimes up to 3 s. ⁷

To load the cavity, roughly 10⁴ atoms are dropped from a magneto-optical trap that sits a few millimeters above the cavity. The FORT is then sprung on the few atoms that intersect with the cavity mode volume. At the same time, the researchers focus another laser, again detuned from the atom–cavity resonance, to remove the few-mK energies gained by the atoms during their fall.

Pumping the system

To function as a laser, the atom–cavity system simply required pumping to initiate and sustain emission (see figure 2). The standing-wave FORT established along the cavity axis traps the Cs atom within the cavity, while 2 external laser beams pump and recycle the ground and excited hyperfine states of the Cs 6S _1/2 and 6P _3/2 levels. Strong coupling of the cesium atom with a single mode enforces the single-atom lasing transition between the two states in the 6P excited and 6S ground manifolds.

The 3-s trapping lifetime turned out to be critical for configuring the device as a laser. Dan Stamper-Kurn, a former postdoc with Kimble, explains that the trapped atom suffers momentum kicks and recoil heating each time it spits out a photon and the cavity fills and empties with light. “That [recoil] subtly shifts the energy of the atom, which is oscillating around in the trap—it’s a recipe for disaster.” Still, with pump beams running continuously, the atom typically remains in the cavity for 50–100 ms, far longer than the photon cavity lifetime of 20 ns.

Because one atom is doing the work normally shared by many, the output flux from the atom is just 100 000 photons/second (about 20 femtowatts) of 852-nm light, small compared with conventional lasers. Two photodetectors register arrival times of each photon to determine the mean counting rates and the statistical properties of the light.

The nature of those photon statistics reveals a key difference with many-atom systems. Conventional lasers involve coherent emission from the collective polarization of many atoms and, far above threshold, produce light that exhibits Poissonian statistics. That is, no correlation exists between the arrival time of one photon and the next. In the case of the single-atom laser, though, a bottleneck in the recycling process produces a distinct anticorrelation between successive photons. After the atom emits a photon, a characteristic time is required for the atom to be “reloaded” and emit another. Photons end up spaced more evenly apart in time, with only a small likelihood of arriving together.

This antibunching of the photons produces sub-Poissonian (nonclassical) light. In that respect, the system resembles one-atom fluorescence, the reemission of light scattered off an atom into free space. The difference is that, within a cavity tuned to the frequency of the atomic transition, strong coupling entrains each photon into a single cavity mode, rather than in all directions.

Another key difference with conventional lasers is the “thresholdless” nature of the emission from the one-atom laser. Conventional lasers require a gradual buildup of the radiation field to stimulate lasing. But in the single-atom case, owing to the large single-photon Rabi frequency (which dominates the cavity and atomic decay rates), the vacuum field itself is sufficient to prompt the emission to turn on immediately without requiring other photons in the cavity. Indeed, the mean number of photons in the cavity is 0.005 in the Caltech experiment, making the cavity dark most of the time. By comparison, the mean photon number in Feld’s system is about 10 at lasing threshold. Still, the saturation photon number in the one-atom system is just 0.013, so even numbers as small as 0.005 have an important influence on the dynamics of the atom–cavity interaction.

Howard Carmichael of the University of Auckland in New Zealand comments that so low a mean photon number suggests that stimulated emission is not really dominating the Caltech system; rather, the strong coupling simply forces spontaneous emission in one direction and ensures that photons are pulled into the proper cavity mode. That makes it more akin to engineering the boundary conditions to control the light flux than what is usually understood as a laser.

In that sense, the single-atom laser is an example of the growing field of what Haroche calls “state engineering,” by which researchers prepare particular light states for particular ends. Other groups, for example, are hoping to confine quantum dots, or clusters of atoms within a cavity, to produce light with desired quantum statistical properties. The fluctuations in light from the one-atom laser are minimized by the conceptual simplicity of the system—pumping, emission, and recycling of a single excited state—and the photons arrive at their destination in an orderly fashion.

Kimble’s group is currently working on a follow-up experiment in hopes of making a deterministic stream of photons. In that scheme, the continuous driving fields (Ω_P and Ω_R in figure 2) would be pulsed and the technique would be one among several current approaches designed to obtain photons on demand. Kimble commented that such sources and their extensions might be employed for diverse tasks in quantum information science, including quantum cryptography. Ultimately, he and his colleagues are “attempting to develop tools for the implementation of new protocols for quantum computation and communication, including those needed to develop quantum networks.”