We have demonstrated the slowing of a rubidium atomic beam by isotropic monochromatic light. The results agree with a model calculation, thus allowing its use for designing a practical isotropic light slower. The large hyperfine splittings of rubidium lead to natural multifrequency slowing, which is also included in our model.
We have modeled and experimentally investigated the ``Zeeman tuned slowing'' of rubidium using ${\mathrm{\ensuremath{\sigma}}}^{\mathrm{\ensuremath{-}}}$ polarized laser light. Previous studies have all used ${\mathrm{\ensuremath{\sigma}}}^{+}$ light. We find that important details of the slowing process enable this new method to produce significantly slower and more intense atomic beams than for the ${\mathrm{\ensuremath{\sigma}}}^{+}$ case. An additional advantage: ${\mathrm{\ensuremath{\sigma}}}^{\mathrm{\ensuremath{-}}}$ slowing exhibits much weaker dependence on laser wavelength and intensity.
We have developed a rubidium atomic-beam light source that emits Rb lines with a Doppler width of only 0.0015 cm−1. When the light from this source was measured with a scanning Fabry–Perot interferometer and the data were reduced by digital computer utilizing numerical deconvolution techniques, we were able to determine a hyperfine level splitting in natural rubidium as small as 0.0014 cm−1 with an uncertainty of 0.0001 cm−1. We have measured the relative intensities and positions of 19 of the 20 hfs components of the two resonance lines, 7800 and 7974 Å, of natural Rb. From these data, we calculate isotope shifts σ87−σ85 = 0.00268 cm−1 for 52P32 and σ87′−σ85′ = 0.00259 cm−1 for 52P12.
Short pulses of narrow-line low-intensity dye-laser light nearly resonant with the Zeeman-split 2P21 resonance line (7948 \AA{}) of rubidium were observed to propagate through dilute rubidium vapor as slowly as (141)c. These slow pulse velocities showed that most of the energy in the propagating wave was contained by the vapor as coherent atomic excitation. The observed pulse velocities vp are in good agreement with the equation ${v}_{p}=\frac{d\ensuremath{\omega}}{\mathrm{dk}}$ for the group velocity obtained from linear-dispersion theory. Also, the experimental results are quantitatively explained by adiabatic following, in which the pseudomoments of the atoms remain aligned along the effective field of the laser light. The adiabatic-following model allows for a direct comparison of our results with the work on self-induced transparency. For high-intensity light, adiabatic following predicts a nonlinear pulse velocity and the possibility of observing self-steepening.
We have laser cooled an atomic beam of rubidium with diode lasers and have used these slow atoms to load a magneto-optical trap. The cooling is performed with the frequency-chirp method,1,2 in which the laser's frequency is ramped linearly (by the injection currant) to maintain resonance with the decelerating atoms. Two synchronously chirped lasers are required to prevent optical hyperfine pumping of the ground state. The chirp typically covers 500 MHz in 10 ms. At the end of each ramp, the resulting velocity distribution is measured with a pulsed probe laser. By varying the end point of the chirp, we are able to obtain slow atoms with average velocities ranging from 200 to −50 m/s. The widths of the velocity distributions are measured to be approximately 40 m/s. This relatively large width results from the inherent linewidth (~30 MHz) of the free-running diode lasers used for both cooling and probing.
A combined experimental and theoretical investigation of time- and alignment-dependent propagation of light in an ultracold atomic gas of atomic 85Rb is reported. Coherences among the scattering amplitudes for light scattering off excited hyperfine levels produce strong variations of the light polarization in the vicinity of atomic resonance. Measurements are in excellent agreement with Monte Carlo simulations of the multiple scattering process.