Rapid all-optical loading of trapped ions using a miniaturised atom source

TL;DR

This work tackles rapid, low-heat loading of trapped ions by introducing a microfabricated optically heated calcium oven with an integrated collimator. It demonstrates Ca+ loading at rates up to $24(3)\, s^{-1}$ with heating powers below $85\, mW$ and loads a single ion in under $30\, s$ with $41.4(4)\, mW$, using a two-photon ionisation scheme with a measured ionisation probability of $q = 1.5(5)\times10^{-5}$. A simple thermal model shows radiative losses dominate the oven’s performance, enabling extrapolations to higher flux with modest power increases or alternate wavelengths. The results imply substantial heat load reduction and on-demand reloading capabilities for scalable ion-trap quantum processors, with potential applicability to a broad range of metals used in ion traps.

Abstract

We characterise an efficient optically-heated neutral atom source for ion trapping. We observe loading rates of up to $24(3)\,\mathrm{s}^{-1}$ with heating powers below $85\,\mathrm{mW}$, and demonstrate loading of a single ion in under $30\,\mathrm{s}$ with $41.4(4)\,\mathrm{mW}$ of optical power in a room-temperature ion trap system with an ionisation probability of $1.50(5)\times10^{-5}$. We calibrate a thermal model for the source's internal temperature by imaging the fluorescence of a collimated flux of neutral calcium that effuses from the oven at various optical heating powers. We show that the thermal performance of this oven is mainly limited by radiative losses. We explore the effect of second-stage photo-ionisation laser power on the loading rate, and identify a path beyond the loading rates reported in this study. We predict that this source is also well-suited to a wide range of metals used in ion-trapping.

Figures (5)

• Figure 1: Simplified experimental setup (a) Construction of the oven. Shown schematically are: i, the aperture for input heating laser light; ii, the crucible for containing the atom source material; iii, the integrated collimator; and iv, the outer wall which is thermally isolated from the hot crucible. An outer jacket, used for mechanical stability, is omitted for clarity. (b) A photograph of the ion trap and oven mount installed in the chamber, the oven used in this work is highlighted. The electrodes are 0.5mm in diameter and are spaced by 0.65mm from the RF null. (c) An illustration depicting the midplane of the ion trap system. Simplified diagram of the imaging system; the central region is approximately to scale. Further, the coordinate axes used to analyse the atomic beam are rotated by 30 to the trap axis and are consistent with (b).
• Figure 2: A false-colour image of resonant fluorescence from neutral $^{40}$Ca taken using the imaging system shown in \ref{['fig:neutral-setup']}. The axes are scaled to the object plane by the known magnification of $M=5.95(2)$. The dashed contours show one and two standard deviations ($\sigma$) and the point, marked by $\times$, shows the peak location. Both the mean and standard deviations for the two axes are extracted from a fit to a two-dimensional Gaussian.
• Figure 3: Determination of the oven's thermal performance from fluorescence measurements. (a) Peak photon count $C_\mathrm{max}$ extracted from Gaussian fits of fluorescence images as shown in \ref{['fig:neutral-image']}. Datasets are shown for two excitation laser powers, 1.6(2)µW (yellow triangles) and 5.2(5)µW (green circles). (b) The peak atomic density is inferred from the camera data in (a). The dashed blue line shows $n_{\mathrm{peak}}$ as inferred by the thermal model. (c) Synchronous camera and PMT readings are combined to quantify the total atomic flux and deduce the crucible temperature. The red line is the best fit of the thermal model to the temperature data. The inset shows the distribution of the model parameters obtained by bootstrapping, each point corresponds to a unique set of parameters. The conductive loss coefficient $\alpha_\mathrm{c}/\varepsilon = 4(3)e-5W\per K$ could not be well established due to the thermal response being dominated by radiative losses in this regime. The radiative loss coefficient is fitted to be $\alpha_{\mathrm{r}}/\varepsilon = 1.2(1)e-5m\squared$. The thermal model obtained from data shown in (c) is used to predict $n_\mathrm{peak}$ in (b). Shaded areas represent $1\sigma$ confidence intervals.
• Figure 4: Loading rate comparison to number density and pulse scheme. (a) Loading rate of $^{40}\mathrm{Ca}^+$ ions (black circles, right axis) versus varying heating power with modelled central number density (dashed line, left axis). Dashed vertical lines indicate the region used to determine the relationship between the loading rates and number densities. The thermal model is reproduced from \ref{['fig:neutral-density']}. The shaded area represents one standard deviation for the thermal model. The ion loading rate data follows the same trend as the modelled number density. (b) Pulse sequence for the loading rate measurement. The repump (866nm wavelength) laser is briefly turned off at the beginning to measure background counts. The excitation (423nm wavelength) laser pulse is scanned in duration followed by a cooling and probe cycle using 80MHz, 30MHz and 15MHz red detuned pulses of cooling (397nm wavelength) laser light. Detection is carried out during the 15MHz red detuned cooling laser pulse in order to check if ions have been loaded.
• Figure 5: Diagrammatic representation of the atomic flux coming from the oven. The flux out of the oven can be inferred from the known density and the velocity distribution at a plane set at a distance $d$ from the crucible. Given that $d\gg50\ \mathrm{\upmu m}$ (collimator diameter), the oven is modelled as a point source.