Figure 1

A schematic of the experimental layout: A 4.5 cm long hollow-core photonic band-gap fiber is mounted inside a steel chamber with glass windows connected to an ultrahigh-vacuum system. A 795 nm diode laser beam is scanned mode-hop free over 10 GHz and split into the pump and signal beams, which counterpropagate through the HC-PBGF. An 805 nm light-induced atomic desorption beam is mixed with the pump beam at a polarizing beam cube and is coupled into the fiber with $10\times$ microscope objectives. An acousto-optic modulator (AOM) allows for the pulsing of the LIAD beam in order to regulate the optical depth, if needed. The coupling is monitored using a charge-coupled device (CCD) camera. The probe beam can alternately be passed through the cell and can be detected by a photodiode to monitor the ambient cell Rb density. The coupled probe light is detected with photomultiplier tube (PMT). lens (L); photodiode (PD); polarizer (Pol); polarizing beam cube (Pol cube); half-wave plate ($\lambda /2$).Reuse & Permissions

Figure 2

Saturable-absorption spectroscopy of the ${}^{87}\mathrm{Rb}$ $F=1\to F^{\prime }=2$ transition in the presence and the absence of a 20 mW desorption beam. In the presence of LIAD, a significant ac Stark shift is observed, and the line shape is seen to broaden and to distort. The arrow indicates the ${}^{87}\mathrm{Rb}$ $F=1\to F^{\prime }=1$ transition. Solid lines are fits to a Lorentzian peak inside a Gaussian.Reuse & Permissions

Figure 3

The effect of guided mode nonuniformity on the ac Stark shift. (a) Radial mode profiles for 10 mW flat-top and Gaussian-core modes. (b) Simulated ac Stark shifts of a Lorentzian saturated-absorption peak by flat-top and Gaussian-core modes. The vertical dashed line represents the maximum shift experienced by the most intense component of the core mode. The nonuniformity of the core mode is also seen to broaden and to steepen the line toward higher frequencies.Reuse & Permissions

Figure 4

ac Stark shifts as functions of desorption beam power. 30-nW and 150-nW powers were used for the counterpropagating signal and the pump beams, respectively. The experimental shifts are larger than those expected for a flat-top mode (dashed line) and are well fit by the simulated shift of a nonuniform Gaussian-guided mode (solid line).Reuse & Permissions

Figure 5

Measured power broadening of Rb atoms confined to the core of the band-gap fiber and predicted linewidth broadening as a function of pump power measured via saturable-absorption spectroscopy on the ${}^{87}\mathrm{Rb}$ $F=1\to F^{\prime }=2$ transition in our system. The intercept represents the minimum linewidth available limited by transit-time broadening.Reuse & Permissions

Figure 6

Passive slow light in the center of the Rb $D_1$ line. Pulse delay in HC-PBGF as a function of optical depth for a 4.6 ns, 795 nm pulse centered between the ${}^{85}\mathrm{Rb}$ $F=2\to F^{\prime }=2$ and $F=3\to F^{\prime }=3$ transitions. The solid gray line is a fit to an exponential function that estimates the experimentally accessed homogeneous linewidth (see text).Reuse & Permissions