Scalable Ion Fluorescence Collection Using a Trap-Integrated Metalens

Hae Gyun Lim, Johannes E. Fröch, Christian M. Pluchar, Arka Majumdar, Sara MouradianarXiv 2025 · DOI · arXiv

Abstract: A scaled trapped-ion quantum computer will require efficient fluorescence collection across a large area. Here we propose and demonstrate a compact monolithically integrated system featuring a metalens fabricated on the backside of a surface ion trap. A 40$\times$100 $\mu$m aperture enables a simulated point-source collection efficiency of 0.91% and a measured point-source detection efficiency of 0.58%. Increasing the aperture area to 40$\times$600 $\mu$m boosts the simulated collection efficiency to 3.17%$-$comparable to that of a conventional objective with a numerical aperture of 0.35. Further improvements are possible by co-optimizing the electrode and aperture geometry. An undercut of the electrode substrate at the aperture ensures a large distance between the ion and dielectric substrate without compromising collection efficiency. The metalens directly collimates the collected fluorescence, eliminating the need for a high numerical aperture objective. An array of such readout zones will offer a compact, scalable solution for high-fidelity parallel readout in next-generation trapped-ion quantum processors.

Introduction

Trapped ions are a leading platform for quantum computing, but scaling to millions of qubits requires overcoming significant engineering hurdles. In architectures like the Quantum Charge Coupled Device (QCCD), ions are shuttled across centimeter-scale chips between various control and readout zones. A major bottleneck in this architecture is the readout phase, which typically relies on bulky, external objective lenses to collect ion fluorescence. Achieving a high numerical aperture (NA) for efficient collection while maintaining a massive field of view (FOV) to cover a large chip is practically impossible with conventional optics.

Conceptual comparison of (a) a traditional bulky objective system and (b) the proposed scalable architecture using an array of trap-integrated metalenses.

Monolithic Integration and Design

To address this scaling bottleneck, we propose a monolithically integrated optical system that builds tiny lenses directly into the trap chip itself. This approach decouples the FOV from the NA by placing a dedicated, highly efficient metalens at every single readout zone.

The system utilizes a standard surface ion trap with a 40$\times$100 $\mu$m aperture etched into the ground electrode to allow fluorescence to pass through. A critical design feature is the "undercut" of the silicon substrate beneath the electrodes. This prevents the bulk material from clipping the optical path and keeps the dielectric substrate far from the ion, mitigating motional heating of the qubit. The metalens, fabricated on the backside of a borosilicate glass substrate, consists of an array of 700 nm tall silicon nitride (SiN) nanopillars precisely sized to collimate 397 nm light.

Experimental Results

The prototype device achieved a measured point-source detection efficiency of 0.58%, which is highly consistent with the simulated 0.91% when accounting for transmittance losses in the glass and metalens. Simulations further show that lengthening the aperture to 40$\times$600 $\mu$m boosts the collection efficiency to 3.17%, matching the performance of a bulky conventional objective with an NA of 0.35.

Comparison of detection efficiency for a conventional objective versus the integrated metalens. (c) demonstrates that the metalens maintains constant efficiency as the readout zone is displaced, whereas the objective's efficiency drops to zero at approximately 1.2 mm. (d) shows the efficiency as a function of ion displacement within the aperture.

The primary advantage of this integrated approach is its off-axis superiority. As shown in Figure 5, a conventional objective's detection efficiency plummets to zero at a displacement of approximately 1.2 mm from the optical axis. In contrast, the integrated metalens maintains its efficiency across the entire chip because the optics are local to each readout zone and move with the trap assembly. This eliminates the need for post-fabrication optical alignment, as the alignment is locked in during the lithographic manufacturing process.

Conclusion

Trap-integrated metalenses provide a compact and scalable path toward high-fidelity parallel readout in next-generation trapped-ion quantum processors. By moving the collection optics onto the chip, we solve the fundamental trade-off between field of view and collection efficiency. Future iterations could feature complex phase profiles to focus light from entire chains of ions simultaneously, further enhancing the throughput of large-scale quantum hardware.

Summary updated

Abstract

A scaled trapped-ion quantum computer will require efficient fluorescence collection across a large area. Here we propose and demonstrate a compact monolithically integrated system featuring a metalens fabricated on the backside of a surface ion trap. A 40$\times$100 $μ$m aperture enables a simulated point-source collection efficiency of 0.91% and a measured point-source detection efficiency of 0.58%. Increasing the aperture area to 40$\times$600 $μ$m boosts the simulated collection efficiency to 3.17%$-$comparable to that of a conventional objective with a numerical aperture of 0.35. Further improvements are possible by co-optimizing the electrode and aperture geometry. An undercut of the electrode substrate at the aperture ensures a large distance between the ion and dielectric substrate without compromising collection efficiency. The metalens directly collimates the collected fluorescence, eliminating the need for a high numerical aperture objective. An array of such readout zones will offer a compact, scalable solution for high-fidelity parallel readout in next-generation trapped-ion quantum processors.
BibTeX
@misc{lim2025scalable,
  author = {Hae Gyun Lim and Johannes E. Fröch and Christian M. Pluchar and Arka Majumdar and Sara Mouradian},
  title = {{Scalable Ion Fluorescence Collection Using a Trap-Integrated Metalens}},
  year = {2025},
  eprint = {2507.23071},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2507.23071},
  howpublished = {arXiv},
}
Figure