A broadband, individually addressing two- and three-dimensional photonic integrated circuit for trapped-ion qubit control

AI summary

Overview

Research area: Quantum computing hardware — integrated photonics for trapped-ion qubit control (quantum physics / nanophotonics / micro-optics).

Technical level: Advanced. The paper assumes familiarity with trapped-ion energy-level structure, waveguide photonics, grating couplers, and numerical optical design tools.

Scope (one sentence): The paper reports a wafer-scale, hybrid two- and three-dimensional photonic integrated circuit that delivers individually addressable laser beams at 405–880 nm to ions trapped above a surface electrode trap, and validates it with trapped Ca and Ba ions.

What This Paper Is About

Trapped-ion processors need many different laser wavelengths (for photoionization, cooling, repumping, shelving, gates and readout) delivered with micron-level precision to every qubit, and today's bench-top optics and on-chip grating couplers both struggle to scale. The authors replace the narrowband grating coupler with a broadband combination of a lithographically defined planar waveguide lens and a 3D-printed micromirror, so that one shared optical element can focus many wavelengths onto individual ions. The goal is an on-chip optical control layer that does not grow in complexity with the number of wavelengths.

Key Contributions

1. A hybrid 2D-3D photonic architecture. A planar waveguide lens (defined in lithography) expands and collimates guided light, and a biconic-aspherical micromirror printed by two-photon polymerization (2PP) redirects and focuses it to a spot approximately 75 µm above the chip surface.

2. Wafer-scale monolithic integration. The full stack — Al₂O₃ waveguide core with SiO₂ cladding, 200 nm Al surface-trap electrodes that double as the mirror reflector, and the 3D-printed optics — is fabricated on 6-inch silicon wafers, with 50 micromirrors printed per wafer. The authors state this is the first demonstration of high-resolution lithography and etch performed on top of 2PP micro-optics.

3. Broadband, individually addressing optics. One shared element addresses three individual ions over 405–880 nm at a 5 µm pitch with −27 dB average intensity crosstalk (−27.6 dB in the reported focal-plane measurement).

4. Real-ion validation. Ca and Ba ions are trapped above the fabricated devices; Ba serves as an optical probe of stray light leakage, and on-chip 866 nm delivery selectively repumps one ion of a 5 µm-spaced Ca pair.

Main Findings

• Focus survives a 500+ nm band: axial beam FWHM grows only from 0.67 µm at 405 nm to 1.46 µm at 880 nm, so a single element covers the cooling, repump, shelving and gate wavelengths for Ca and Ba. Measured values are on average 50% larger than predicted from the geometric NA ≈ 0.34, attributed to non-uniform aperture illumination and spherical aberration.
• Asymmetric spot is a feature, not a bug: radial FWHM stays below 8 µm at all wavelengths and clusters near 5 µm, consistent with the asymmetric aperture set by the printed mirror's topography limit. Tight axial focusing suppresses crosstalk along the ion chain, while the broader radial beam tolerates ion-position uncertainty and beam-pointing error.
• Crosstalk below −27 dB at 5 µm pitch: confirmed by both full-wave simulation and measurement at 729 nm (Ca optical qubit) and 515 nm (Raman gates in Ba). The authors describe this as the first individual addressing at two wavelengths in a monolithic ion-trap PIC using a single shared optical element.
• Three-spot addressing demonstrated: simultaneous three-spot illumination measured at 515 nm and 635 nm with 5 µm pitch; the field of view is limited to approximately 8 by uncorrected coma.
• Loss budget: fiber-to-focal-spot loss is about 10 dB in the NIR, 18.5 dB at 515 nm, 24 dB at 450 nm and 35 dB at 405 nm. The planar-lens/mirror element itself contributes only approximately 2 dB across the whole band, so short-wavelength loss is attributed to the alumina-on-silica waveguide platform rather than the architecture.
• Barium as a leakage probe: with 100 µW of 455 nm light detuned 200 MHz from resonance coupled on chip, the measured intensity 120 µm from the expected peak center is approximately 57 dB below the expected peak intensity — comparable to weak-sidelobe levels near 60 dB resolved tens of microns from focus in precision grating-coupler optics.
• Single-ion repumping in calcium: integrated 866 nm light predominantly excites one ion of a two-ion Ca crystal 5 µm from its neighbor. Contrast is limited by unoptimized axial ion positioning, transition saturation, and imperfect radial alignment.
• Trap performance: Ba ions showed radial secular frequencies repeatedly measured at 2.5 and 2.9 MHz at V_rf = 140 V, with lifetimes of one to two hours adjacent to the mirror; a stable two-ion Ca crystal above the mirror lived approximately one hour.
• Two named failure modes: exposed dielectric at the planar lens develops positive voltage in the Ba experiment, weakening axial confinement and preventing shuttling into the integrated beam (the 413 nm and 531 nm photoionization lasers are suspected contributors); in the Ca experiment, out-of-plane micromotion compensation voltages rose dramatically across the chip until ions could no longer be loaded, hypothesized to be photo-induced conductivity change in the silicon substrate.

Methodology in Plain English

The team built the optics in two stages on a standard 6-inch silicon wafer. First, conventional cleanroom processing defines aluminum oxide waveguides inside silicon dioxide cladding, plus a "planar lens" — a lithographically shaped slab region with a curved facet that spreads the light out from a sub-micron waveguide to a roughly 60 µm aperture and collimates it. Optical facets are cut with a vertical etch of about 9 µm through an amorphous-silicon hard mask, producing sidewalls of approximately 89° to reduce deflection of the exiting beam.

The wafers then go to a commercial 2PP foundry, where a polymer micromirror is printed into a trench at the center DC electrode, aligned to etched marks with accuracy within 1 µm in all directions. The mirrors have approximately 13 nm RMS surface roughness and part-per-thousand shape deviation from design. Back in the cleanroom, 200 nm of aluminum is evaporated with a tilted, rotating geometry so it coats the polymer mirror conformally and simultaneously forms the surface-trap electrodes; the metal is then patterned and wet-etched, and the mirrors are given the same potential as the center DC electrode to limit field distortion.

For optical testing, the authors image the emitted beam with a piezo-actuated confocal profiler stepping vertically in 1.25 µm increments, using six fiber-pigtailed diode lasers at 405, 450, 515, 635, 730 and 880 nm injecting 3–30 mW. Crosstalk is measured by exciting one entrance waveguide and watching leakage at neighboring spot positions. Ion tests use a 5-wire surface trap: Ba is loaded by laser ablation and photoionization at 413 nm and 531 nm with cooling at 493 nm and repumping at 650 nm, while Ca is loaded from a resistively heated oven using 375 nm and 422 nm light with 397 nm Doppler cooling. In the Ba experiment, the ion itself becomes the detector — the 455 nm transition rate is extracted from quantum-jump fluorescence traces and converted to local intensity, letting the team map stray light versus ion position.

Why This Matters

Today's trapped-ion processors are at roughly 100 physical qubits, and the optics are a recognized scaling bottleneck: grating couplers offer either high efficiency or wide bandwidth, not both, so a multi-wavelength machine needs many gratings, and coupler footprint ultimately caps how many qubits a zone can hold.

Real-world applications (as framed by this work):

• Scalable quantum computing: a single broadband mirror replaces many gratings, so the number of optical delivery elements scales with the number of ion registers rather than the number of wavelengths, reducing ion transport operations per clock cycle in a QCCD machine.
• Precision optical addressing hardware: the planar-lens-plus-printed-mirror design is a general recipe for delivering many wavelengths to small, densely packed targets, relevant to atomic clocks and neutral-atom arrays.
• Additive manufacturing in semiconductor process flows: the post-2PP lithography and etch step shows 3D-printed micro-optics can survive and be integrated into standard fab processing.
• Quantum sensing and metrology instruments: the broadband, low-crosstalk addressing scheme suits any experiment requiring wavelength-multiplexed, spatially separated optical channels.

Industry relevance: the work is explicitly wafer-scale (6-inch, 50 mirrors per wafer, stealth dicing, commercial 2PP and fiber-packaging partners), which is the language of manufacturability rather than lab demonstration. It also identifies concrete integration fixes — conducting oxides, 3D-printed metallized shields, and ground planes — that map onto existing surface-trap process know-how.

Future Directions

• Eliminate charging: deposit a transparent conducting oxide such as indium tin oxide on the planar lens facet, and/or print a metallized overhanging shield atop the lens, building on prior ITO-based trap work.
• Add a ground plane: place one above the silicon handle layer or above the optically active dielectrics to terminate RF field lines and screen photo-generated carriers, so traps can run for extended periods.
• Fix off-axis aberrations: introduce additional optical elements to correct the coma that currently limits the field of view to approximately 8, enabling larger addressed ion chains without sacrificing per-emitter bandwidth.
• Improve short-wavelength loss: the authors attribute the 35 dB at 405 nm and 24 dB at 450 nm to the alumina-on-silica platform rather than the architecture, and point to established process refinements as the remedy.
• Resolve the barium trapping question: trapping Ba directly above the micromirror is described as still under active investigation, and Ca operation at other wavelengths could not be tested after the trap became unstable.

Target Audience

Quantum-hardware engineers and applied physicists working on trapped-ion systems, integrated photonics researchers interested in broadband on-chip beam delivery, and fabrication specialists evaluating two-photon polymerization as a front-end integration step. It is also useful for program managers and technical leads assessing manufacturable paths to scale ion-trap machines past roughly 100 qubits, and for graduate students seeking a worked example of end-to-end device design, fab, optical characterization and ion testing in a single paper.

Authors’ abstract

Trapped ions provide a high-fidelity platform for quantum information processing, yet delivery of multiple, distinct wavelengths across large networks of interaction zones remains a bottleneck. Conventional free-space light delivery lacks scalability, while on-chip grating couplers suffer from narrow operational bandwidth that increases circuit footprint and optical interfacing complexity. Here we show a broadband photonic integrated circuit capable of addressing individual ions. The circuit combines a planar waveguide lens with a micromirror fabricated using two-photon polymerization at wafer scale. This implementation can address three individual ions from $λ$ = 405 - 880 nm with -27 dB average intensity crosstalk at $5\,μ\mathrm{m}$ pitch. We trap $^{40}\mathrm{Ca}^{+}$ and $^{138}\mathrm{Ba}^{+}$ ions above such devices, characterize optical crosstalk with barium ions, and demonstrate individual repumping of calcium ions. This monolithic photonic architecture brings broadband addressing in an on-chip modality to trapped-ion technology. More generally, integrating additive manufacturing into quantum devices is poised to unlock expanded design space for implementing novel quantum architectures.