Loading of a surface-electrode ion trap from a remote, precooled source

We demonstrate loading of ions into a surface-electrode trap (SET) from a remote, laser-cooled source of neutral atoms. We first cool and load $\sim$${10}^6$ neutral ${}^{88}$Sr atoms into a magneto-optical trap from an oven that has no line of sight with the SET. The cold atoms are then pushed with a resonant laser into the trap region where they are subsequently photoionized and trapped in an SET operated at a cryogenic temperature of 4.6 K. We present studies of the loading process and show that our technique achieves ion loading into a shallow (15 meV depth) trap at rates as high as 125 ions/s while drastically reducing the amount of metal deposition on the trap surface as compared with direct loading from a hot vapor. Furthermore, we note that due to multiple stages of isotopic filtering in our loading process, this technique has the potential for enhanced isotopic selectivity over other loading methods. Rapid loading from a clean, isotopically pure, and precooled source may enable scalable quantum-information processing with trapped ions in large, low-depth surface-trap arrays that are not amenable to loading from a hot atomic beam.

Figure 1

Cryogenic apparatus and atomic systems. (a) Diagram depicting the cryogenic UHV magneto-optical trap (MOT) and ion-trap system (not to scale). (b) Detail of cold head and trap stage (not to scale). Control signals are routed from the lead connector, along the filter boards (through single-stage low-pass filters), onto the interposer, and onto the trap chip, with gold wire bonds used at each junction. The radio-frequency (rf) signal is routed similarly, excepting the filter components. Macor rods thermally isolate the trap stage from the cold head, allowing the former to be heated separately from the cold stage of the cryocooler. (c) Photograph of the surface-electrode ion-trap chip (the chip is 1 cm on a side). The trap is niobium on a sapphire substrate, and the patterned metal electrodes form a segmented linear Paul trap 50 $\mu$m from the chip surface. The two rf electrodes are fed from the central lead on the left, and all other electrodes are dc control electrodes. The inset shows a fluorescence image of two ions in this trap (the interion distance is 6 $\mu$m). (d) Relevant electronic level structures of neutral and singly ionized strontium. Solid arrows represent transitions driven directly in this work and dotted arrows represent other relevant excitation or decay paths.

Figure 2

Number of atoms trapped in the MOT as a function of MOT loading time. These data were taken at an oven temperature of $425{}^{\circ }\mathrm{C}$ and a total MOT laser beam intensity of 40 mW/cm${}^2$. The line is a fit to a simple exponential (see text) with a characteristic time of $\tau =10.5$ ms and a steady-state atom number of $N_{\text{ss}}=1.4\times {10}^6$.

Figure 3

Ion loading probability and rate as a function of MOT loading time. The upper set of points, the (blue) circles, are for an oven temperature $T_{\mathrm{oven}}=425{}^{\circ }$C and a total MOT beam intensity of 40 mW/cm${}^2$. The lower three sets of points are for varying $T_{\mathrm{oven}}$ at a lower MOT beam intensity of 23 mW/cm${}^2$: the (blue) triangles are for $T_{\mathrm{oven}}=425{}^{\circ }$C, the (green) squares are for $T_{\mathrm{oven}}=400{}^{\circ }$C, and the (red) diamonds are for $T_{\mathrm{oven}}=375{}^{\circ }$C. The inset shows the ion loading rate as a function of MOT loading time obtained from the upper set of points in the main figure. The data are taken with the push and photoionization (PI) lasers on for 3 ms after the MOT beams are turned off. All lines are guides for the eye.

Figure 4

Ion loading in saturated and unsaturated regimes. (a) Ion loading probability (circles, left axis) and MOT atom number (triangles, right axis) as a function of MOT loading time, showing the ion loading probability (ILP) saturating before the MOT atom number. Here the photoionization (PI) lasers are on for 3 ms after the MOT beams are turned off. (b) As (a), except the first-step PI laser is pulsed on for only 10 $\mu$s for the ion loading data (circles, left axis). In this case, the ILP is not saturated and is proportional to the MOT atom number (triangles, right axis). Lines in (a) and (b) are guides to the eye. (c) Unsaturated ILP as a function of delay of the first-step PI laser pulse after the MOT beams are shut off, with a fixed pulse width of 25 $\mu$s and a fixed MOT loading time of 10 ms. The ILP is here proportional to the temporal density profile of the pulsed cold atomic beam. A background loading rate due to first-step PI excitation from the push beam (which is on continuously for 5 ms after the MOT beams are shut off) has been subtracted. All data are taken with $T_{\mathrm{oven}}=425{}^{\circ }$C and a MOT beam intensity of 40 mW/cm${}^2$.