Characterization of laser cooling in a high-magnetic-field atom trap

We describe cooling and trapping of both ${}^{85}$Rb and ${}^{87}$Rb in a range of magnetic fields up to 2.6 T. Atoms are injected from a low-field pyramidal magneto-optical trap and recaptured in a high-magnetic-field atom trap. The atoms are cooled and trapped by a six-beam optical molasses via the $5S_{1/2}|m_I,m_J=$ 1/2$\rangle$$\to$ $5P_{3/2}|m_I,m_J=$ 3/2$\rangle$ transition ($m_I=$ 5/2 for ${}^{85}$Rb and $m_I=$ 3/2 for ${}^{87}$Rb). We study the trap fluorescence spectra, atom temperatures, density distributions, and lifetimes as a function of magnetic field and detuning parameters. The trap fluorescence spectra are both narrow and asymmetric, as is characteristic for laser cooling of atoms in an external trapping potential. The trap is modeled using a Monte Carlo trajectory simulation technique.

Figure 1

High-magnetic-field atom trapping apparatus. A pyramidal trap (1) injects a stream of cool Rb atoms into a cryogenic magnetic trap (2). The magnetic trap, formed by four superconducting coils, has a field minimum at 2.6 T.

Figure 2

Schematic of the high-magnetic-field trapping apparatus. The high-field trap is surrounded by a four-component electrode package, which allows for ionization of excited atoms and collection of the resultant electron signal upon a microchannel plate (MCP), located outside the high-field region.

Figure 3

Level structure of ${}^{85}$Rb in the strong-magnetic-field regime. The $5S_{1/2}|m_I,1/2\rangle$ ground states are shifted by 14 GHz/T and the $5P_{3/2}|m_I,3/2\rangle$ states by 28 GHz/T. The hyperfine splitting in the Paschen-Back regime is determined by $A_{n\ell }{m}_I{m}_J$. For ${}^{85}$Rb, $A_{5S}$ and $A_{5P}$ are equal to 1012 and 25 MHz, respectively. For ${}^{87}$Rb, $A_{5S}$ and $A_{5P}$ are 3417 and 85 MHz, respectively.

Figure 4

Signal obtained by direct excitation out of the incident atomic beam (circles, $\times 30$) and normalized trap signal versus frequency offset ${\nu }_{\text{ofs}}$ of the cooling laser for a fully optimized trap (squares and triangles). For the atomic beam, the signal is from photoionization of 5$P$ atoms. For the optimized trap we show both ionization and fluorescence signals. The trap shows a narrow linewidth (FWHM less than 4 MHz) and highly asymmetric character. The two insets are fluorescence images, with the trapping light frequency slightly red or blue detuned. Atoms are cooled onto a shell by setting the frequency just above the peak trapping condition.

Figure 5

Measurement of frequency spacing and trapping behavior for two $m_I$ sublevels in ${}^{85}$Rb. The gray curve, taken at high molasses intensity, shows the strongly broadened trapping transition $5S_{1/2}|5/2,1/2\rangle \to 5P_{3/2}|5/2,3/2\rangle$ on the left, as well as the transition $5S_{1/2}|3/2,1/2\rangle \to 5P_{3/2}|3/2,3/2\rangle$ on the right. The inset shows a magnification of the $5P_{3/2}|3/2,3/2\rangle$ signal, peaked at 476 MHz. For comparison, the black line shows the trapping transition under optimal low-intensity trapping conditions (not to scale).

Figure 6

Trap decay and loading behavior, taken by spatial integration of the fluorescence signal recorded with the CCD camera. Time $t=1$ s corresponds to the atomic beam being suddenly blocked or unblocked. Fits to these curves give a $1/e$ lifetime of 18.8 s and a $1/e$ rise time of 16.8 s.

Figure 7

Fluorescence images of the high-magnetic-field trap at 2.6 T for different transverse field conditions. Bottom right: measured trap lifetime as a function of the transverse trap curvature $C_t$, for both photoionization and fluorescence measurements. As $C_t$ is raised the trap becomes increasingly confined toward the central axis and the trap lifetime (and the number of trapped atoms) increases. For clarity, images taken at small values of $C_t$ are rendered using the indicated brightness enhancement factors.

Figure 8

Trap lifetime versus magnetic field $B_0$ for both Rb isotopes. Measurements were taken both with photoionization and fluorescence detection techniques. Errors bars represent the standard deviations of a series of two to seven measurements. For clarity, data taken at identical values of $B_0$ are slightly offset from each other along the $B_0$ axis.

Figure 9

Simulated steady-state number of trapped atoms (squares, left axis), fluorescence signal (circles, arbitrary units), and atom temperature (diamonds and triangles, right axis, logarithmic scale) versus molasses laser frequency offset ${\nu }_{\text{ofs}}$ for a constant input flux of atoms into the trapping region. Trapping is asymmetric in character with respect to frequency. The transverse temperature also attains lower limits than the axial temperature. The displayed simulated fluorescence images are analogous to the ones in Fig. 4 and demonstrate the same characteristic behavior.

Figure 10

Simulated trap position along the field axis versus axial beam asymmetry parameter $\eta =(I_{+z}-I_{-z})/(I_{+z}+I_{-z})$, with the molasses red detuned to one-half linewidth (3 MHz). The bars represent the simulated trap size (standard deviations) in the transverse and axial directions (scales as shown).