Title:Comment on "Using Three-Body Recombination to Extract Electron Temperatures of Ultracold Plasmas"

Abstract: In the recent work Fletcher et al. [Phys. Rev. Lett., v.99, p.145001 (2007)] reported on the novel experimental technique, enabling to measure the temperature of the expanding ultracold plasmas over a considerable time interval (up to 60-70 {\mu}s). It was unexpectedly found that the electron temperature dropped with time as T_e(t) ~ t^{-\alpha} with \alpha = 1.2 +/- 0.1 \approx 1 instead of \alpha = 2, which would be expected for the adiabatic cooling of electrons in the cloud expanding linearly in time. The above-cited authors supposed that 'the difference is likely due to the significant heating effects from 3-body recombination' (i.e. the inelastic processes), but they did not provide sufficient quantitative estimates supporting such a conclusion. The aim of the present comment is to mention that the experimentally revealed t^{-1}-dependence can be explained under quite general assumptions by the purely elastic processes in the ultracold plasma, as it was done a few years ago in our work [Yu.V. Dumin, J. Low Temp. Phys., v.119, p.377 (2000)].
Therefore, t^{-1}-dependence revealed in the above-cited experiment should be primarily a manifestation of "virialization" of the electron velocities in the regime of strong electron-ion correlations, while the heat release by inelastic processes should be of secondary importance (it might be responsible, particularly, for changing the exponent from -1 to -1.2). This is confirmed also by our ab initio molecular-dynamic simulations taking into account the strong electron-ion correlations (i.e. not based on the PIC method, Vlasov approximation for electrons, etc., which are applicable only to small-angular scattering).