BEC Ends Globally but Starts Locally

Bose Einstein condensations (BEC), essentially dilute gas clouds in which millions of atoms enter into a single, corporate coherent object, have proven to be a versatile testbed for numerous quantum effects. But having attained the critical conditions necessary for making BEC in the first place, physicists have not paid much attention to the collapse process itself. Now an experiment conducted by scientists from the FOM Institute for Atomic and Molecular Physics (Netherlands) and the Kurchatov Institute (Russia) look at the collapse more closely and find something surprising while analyzing cigar shaped samples. In their experiment atoms enter the BEC state through the use of "shock cooling," in which radio-frequency waves used to cool atoms are provided in a single one millisecond burst rather than in a sustained way as in conventional evaporative cooling. The work shows that BEC is a local effect with local coherence (atoms acting in concert) and that coherence over the whole of a condensate occurs only later. In other words, the condensation has happened so fast that not all atoms are in the ground state; that is, the atoms are not all in equilibrium. Instead, the cloud is much elongated, with warmer atoms near the center and cooler atoms toward the ends of a cigar shaped condensate. While coming to eventual equilibrium, the condensate undergoes oscillations in its shape. This is observed by absorption imaging after switching-off the trap (a figure will posted soon at www.aip.org/mgr/png ). Usually this release gives rise to a cloud expanding in all directions. But in this case oscillating condensates released at the proper moment contract axially while expanding radially. The axial size reaches a minimum value as the sample drops under the influence of gravity. This is equivalent to focusing of a cavity dumped atom laser. The size of the focus is determined by the distribution of axial momenta among the condensate atoms and therefore contains valuable information on the phase fluctuation in the condensate at the moment of release. (Shvarchuck et al., Physical Review Letters, 30 December 2002; contact Jook Walraven, walraven@amolf.nl, 31-20-608-1234; text at www.aip.org/physnews/select ; website at www.amolf.nl/)

CORRECTION. In last week's Update (619), the stability or uncertainty in several frequency measurements was incorrectly reported because of a stray negative sign in the exponent. Thus, for example, the stability of the Mossbauer radiation emission line at a wavelength of 0.086 nm is at the level of one part in 10¹¹, not 10^-11.