LIF versus magnetic field value for the F_g=2longrightarrow F_e=3 transition of <sup>87</sup>Rb for different values of the laser power density <em>I</em>: (a) 0.14 mW cm<sup>−2</sup>, (b) 1 mW cm<sup>−2</sup>, (c) 10 mW cm<sup>−2</sup>

<p><strong>Figure 4.</strong> LIF versus magnetic field value for the F_g=2\longrightarrow F_e=3 transition of <sup>87</sup>Rb for different values of the laser power density <em>I</em>: (a) 0.14 mW cm<sup>−2</sup>, (b) 1 mW cm<sup>−2</sup>, (c) 10 mW cm<sup>−2</sup>. The bottom right panel shows the contrast of the central minimum as a function of the laser power density. Filled circles correspond to experimentally measured values, whereas the solid line shows the result of a calculation. Note the different scales in (a), (b) and (c).</p> <p><strong>Abstract</strong></p> <p>We present the results of an investigation of the different physical processes that influence the shape of nonlinear magneto-optical signals both at small magnetic field values (~100 mG) and at large magnetic field values (several tens of Gauss). We used a theoretical model that provided an accurate description of experimental signals for a wide range of experimental parameters. By turning various effects 'on' or 'off' inside this model, we investigated the origin of different features of the measured signals. We confirmed that the narrowest structures, with widths of the order of 100 mG, are related mostly to coherences among ground-state magnetic sublevels. The shape of the curves at other scales could be explained by taking into account the different velocity groups of atoms that come into and out of resonance with the exciting laser field. Coherent effects in the excited state can also play a role, although they mostly affect the polarization components of the fluorescence. The results of theoretical calculations are compared with experimental measurements of laser-induced fluorescence from the <em>D</em><sub>2</sub> line of atomic rubidium as a function of the magnetic field.</p>