A high-flux atomic strontium oven with light-driven flux modulation

Abstract

A high-flux source of strontium atoms is required for cold atom quantum technology applications. We present a re-entrant oven design that avoids the need for any vacuum feed-throughs and has an inherent temperature gradient to guard against clogging of the nozzle. The nozzle is fabricated by micro-machining of fused silica using selective laser etching; this specialised technique is capable of making many thousands of fine microchannels and is suitable for batch production. Operating with only electrical heating, using <20W of electrical power, a total flux of $8(1)\times 10^{14}$ atoms/s is achieved at an oven temperature of 475°C, of which we estimate $1.8(2)\times 10^{13}$ atoms/s could be captured. A heated in-vacuum sapphire window grants optical access directly opposite the oven, and can be cleared of metallization without breaking vacuum. We used this optical access to modulate the flux of the atomic beam by direct illumination of the nozzle and the strontium metal with high-power laser light. Heating by laser light increased the useful flux by a factor of up to 16(3) on a timescale of 40s, and a factor of 2.5(5) on a timescale of 1s. This flux modulation serves to increase the operating lifetime of the oven. We report experimental measurements of the performance of the oven in long-term operation over many months.

Figures (13)

• Figure 1: (a) Cross-section of the re-entrant oven, as oriented in the system (see Figure \ref{['fig-fullchamber']}). The top side (front) is in vacuum, with the strontium reservoir (SR) underneath the nozzle (N). The bottom side (back) is in air, with blind holes in a recess where cartridge heaters (CH) and thermocouples are inserted; the numbers mark the locations of thermocouples (see main text). (b) Cross-section of the heated sapphire window component; in this case, the bottom side, with the sapphire window (SW) itself, is in vacuum, while the top side, where the cartridge heaters (CH) are inserted from above, is in air. Grooves in the metal around the SW to allow gases to be pumped out from the inner thin-walled tube, at the top of which is a DN16CF viewport (VP) to seal the vacuum region while preserving optical access. (c) A photograph of the fused silica nozzle (top), and an optical microscope image showing the layout, size, and shape of the microchannels at the edge of the channel array (bottom).
• Figure 2: Cross-section of the vacuum chamber and the experimental setup, with the oven at the bottom, and the heated sapphire window at the top. The diagonal limbs provide optical access for trapping atoms in a 2D magneto-optical trap (MOT); the upper diagonal limbs are extended to ensure the viewports are not in the line-of-sight of the atomic beam. The horizontal limbs allow for a push beam, a spectroscopy probe beam, or a third cooling beam for direct loading into a 3D MOT. The heating laser beam (HLB), atomic beam (AB), probe beam (PB), and photodiode (PD) used for transverse absorption spectroscopy are labelled.
• Figure 3: (a) Cross-section of the oven profile, with a red line highlighting the direction along which the temperature was simulated. The orange rectangles represent the hot ends of the cartridge heaters, which were the only heat sources in the model. (b) Output temperatures from the simulation as a function of position along its length as shown in (a) for a range of heater powers. This was calculated assuming only conductive and radiative losses (no convection), with a fixed effective emissivity of $\varepsilon = 0.2$ for the surfaces involved, as detailed in Appendix \ref{['sec-app-comsol']}.
• Figure 4: Resonant transmission of the probe beam as a function of time. Those plots on the left used a heating laser pulse duration of $\Delta t = \qty{1}{\second}$, while those on the right had $\Delta t = \qty{40}{\second}$. The plots are stacked vertically for different values of $T_\text{oven}$, the values of which are labelled on the far right hand side. The dashed lines mark the start and end times of the heating laser pulse. In all cases shown, the heating laser light had a nominal power of 15.
• Figure 5: Transmission of the probe beam as a function of detuning from the 461$^1$S$_0 \rightarrow ^1$P$_1$ transition in $^{88}$Sr for a range of values of $T_\text{oven}$. The plot on the left shows features taken when there was no heating laser light present, while that on the right shows equivalent features taken from the end of a 1 heating laser pulse. In both cases, the circular markers are experimental data points, while the solid lines are fits to the data using the theoretical lineshape for free molecular flow, given by equation \ref{['eq-fit-function']} in Appendix \ref{['sec-app-lineshape-flux']}. The heating laser light had a nominal power of 15. The values of $T_\text{oven}$ in the legend on the left axes apply to the data in both plots. Also labelled are the fit vapour temperatures, $T_\mathrm{fit}$, for the theoretical lineshapes.