Professor Claude Cohen-Tannoudji

Collège de France and Laboratoire Kastler Brossel

École Normale Supérieure, Paris, France

 

Dark states were first observed in optical pumping experiments and interpreted as a quenching of fluorescence due to a destructive interference between two absorption amplitudes connecting two ground state sublevels g1 and g2 to an excited sublevel e. The atom can absorb light from g1 and jump to e. Similarly, it can absorb light from g2 and jump to e. But, if it is in a certain linear superposition of g1 and g2, c1g1+c2g2, the two absorption amplitudes from g1 to e and from g2 to e interfere destructively and cancel out. The fluorescence stops. The state c1g1+c2g2 from which light cannot be absorbed is called a “dark state”.

 

This presentation will start with a description of the first experiments demonstrating the existence of dark states and of the first theoretical interpretations that have been proposed. Dark states appeared to play an essential role in several new physical effects, like Electromagnetically Induced Transparency (EIT), Slow Light, Stimulated Raman Adiabatic Passage (STIRAP), which will be briefly described. A special emphasis will be given to the applications of dark states in the field of ultracold atoms and molecules. First it will be shown how the use of velocity dependent dark states allow atoms to be cooled below the so called “recoil limit”, corresponding to a velocity dispersion of the cooled atoms smaller than the recoil momentum communicated to the atom by the absorption of a single photon. It turned out that a quantitative interpretation of this “sub-recoil cooling” may be given in terms of anomalous random walks and Lévy statistics. Finally, recent applications of dark states to the production of ultracold molecules will be described: production of ultracold polar molecules by the combination of Feshbach resonances and STIRAP; dark states in the photo-association of ultracold atoms allowing a precise determination of the scattering length describing their collisions.