Assembling a Ring-Shaped Crystal in a Microfabricated Surface Ion Trap

We report on experiments with a microfabricated surface trap designed for confining a chain of ions in a ring. Uniform ion separation over most of the ring is achieved with a rotationally symmetric design and by measuring and suppressing undesired electric fields. After reducing stray fields, the ions are confined primarily by a radio-frequency pseudopotential and their mutual Coulomb repulsion. Approximately 400 ${{}^{40}\mathrm{Ca}}^{+}$ ions with an average separation of $9\mu \mathrm{m}$ comprise the ion crystal.

Figure 1

Device layout and labeling. Of the 89 available control electrodes, 86 performed as expected (black labels and dashed arcs), while $e22$, $e67$, and $e89$ (red labels) are shorted to ground. The trapping volume is above the central control electrode $e89$, and is designated by the green-dashed arcs. The irregularities at the loading hole and the diametrically opposite location result from reduced control over the ion motion at these locations. The trapping sites are labeled “$g\mathrm{XX}$” at the gaps in the control electrodes, with the loading hole at $g00$. An $R\text{−}T\text{−}Z$ reference frame is shown at $g25$. “GND” designates electrical ground.

Figure 2

Cross-sectional SEM of the trap, showing the four metal levels, interconnects, and insulators. The trap is fabricated using aluminum-(Al-)1/2% copper (Cu) metallization, planarized silicon dioxide (${\mathrm{SiO}}_2$) interlevel dielectric, and vertical electrical connections using tungsten plug technology.

Figure 3

Tangential-field suppression before and after correction. The error bars (too small on this scale to be readily visible) are propagated from secular frequency and ion-position uncertainties. The field at $g28$ and $g39$ is not calculated after correction due to difficulty in measuring the secular frequency.

Figure 4

(a) A composite image of a full ring after correction. The stitched image comprises separate images of eight octants (labeled by the numbers shown). Without correcting the background fields, ions would primarily fall in an unwanted potential well (marked UW) in proximity to the shorted electrodes $e22$ and $e67$ (see Fig. 1). With the correction, ions occupy the trapping volume except for above the loading hole (marked LH). Uneven illumination in octants 1, 2, 3, 5, 6, and 7 is due to the Gaussian profile of the cooling and repump beams, which are elliptically focused and cover the 1.3-mm diameter of the trap. In octant 4, uneven illumination is primarily a consequence of having limited control over electrodes $e22$ and $e67$, thus preventing field measurement and subsequent correction. The chain interruption at the loading hole is due to deficiencies in the model used to calculate the correction. (b) Scaling the correction from 100% (i) to 0% (v) in 25% steps. As the corrective voltage decreases, the irregularities in the imaged segment increase.

Figure 5

Ion separation in three neighboring octants for which all ions are resolved. Overlap is inferred from the composite image of the full ring (Fig. 4). Imaging-system and cooling-beam imperfections prevent resolving all ions in the remaining five octants. The error bars are based on the uncertainty of the total magnification, but the positions are not corrected for aberrations at the edge of the collection optic.