🔬 optics

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

This paper presents a broadband, monolithic photonic integrated circuit that combines planar waveguide lenses with two-photon polymerized micromirrors to enable individually addressing, low-crosstalk control of multiple trapped-ion qubits across a wide wavelength range, thereby overcoming the scalability and bandwidth limitations of conventional light delivery methods.

Original authors: Daniel Klawson, Yiyang Zhi, Bingran You, Michael Bareian, Elijah Mossman, Chun-Yuan Fan, Arkadev Roy, Ke Sun, Jason Lee, Sung Cheol Yoon, Qiming Wu, Lai Jiang, Wenjun Ke, Weiwei Wu, Sirui Tang, Zachar

Published 2026-07-29

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Klawson, Yiyang Zhi, Bingran You, Michael Bareian, Elijah Mossman, Chun-Yuan Fan, Arkadev Roy, Ke Sun, Jason Lee, Sung Cheol Yoon, Qiming Wu, Lai Jiang, Wenjun Ke, Weiwei Wu, Sirui Tang, Zachary Wall, Jiaxiang Wang, Louis Paul Romero, Sam Vizvary, Steven Diaz, Eric R. Hudson, Wesley C. Campbell, Hartmut Haeffner, Ming C. Wu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to build a super-advanced computer, but instead of using tiny silicon chips like your phone, you are using individual atoms floating in mid-air. These atoms, called ions, are the stars of a new kind of machine known as a quantum computer. To make them do math, scientists have to talk to them using very specific colors of laser light. Think of each atom as a tiny, floating radio that only tunes into one specific station. To control a whole orchestra of these atoms, you need to send different "stations" (colors of light) to each one with incredible precision, without the signal from one atom bleeding into its neighbor's ear.

The problem is that the current way of doing this is like trying to conduct a symphony by holding a giant, wobbly flashlight in a dark room. The equipment is huge, sensitive to the slightest vibration, and gets messy as you try to add more atoms. Scientists have tried shrinking this down onto a chip, but the tools they used were like old-fashioned radio antennas: they only worked for one specific color of light at a time. If you wanted to use three different colors to control your atoms, you needed three different antennas, which took up too much space and made the chip too crowded. This paper tackles that exact bottleneck, asking: "Can we build a single, tiny, smart lens on a chip that can handle a rainbow of colors and point them exactly where they need to go?"

The Paper's Big Idea: A 3D-Printed Magic Mirror

The researchers in this paper have built a new kind of "traffic controller" for light, designed specifically to talk to these floating atoms. They call it a "2D-3D photonic integrated circuit." That's a fancy way of saying they combined flat, printed circuits (2D) with a tiny, 3D-printed mirror (3D) to create a system that can aim different colors of laser light at individual atoms floating just above the chip.

Here is how their invention works, using a simple analogy: Imagine a long hallway (the chip) with a series of doors (waveguides) where light enters. In the old designs, each door had a tiny, flat sticker that scattered the light out, but only if the light was exactly the right color. If you changed the color, the sticker stopped working.

In this new design, the light travels down the hallway and hits a special, flat lens that spreads it out. Then, it hits a tiny, 3D-printed mirror that looks a bit like a curved, bumpy hill. This mirror acts like a high-tech funnel. No matter what color of light comes in (from deep violet to near-infrared), the mirror catches it, bends it, and focuses it into a tight, sharp beam that shoots up into the air. It's like having a single, magical spotlight that can instantly change its color and still hit the bullseye perfectly.

What They Actually Did and Found

The team built this system on a large 6-inch silicon wafer, which is like a giant cookie sheet for making computer chips. They used a process called "two-photon polymerization" to 3D print the tiny mirrors directly onto the chip. This is a bit like using a super-fine 3D printer to draw a microscopic sculpture right on top of a circuit board.

They tested this device with two different types of atoms: Calcium and Barium. Here is what they discovered:

The Color Range: The device successfully handled a huge range of colors, from 405 nanometers (violet) all the way to 880 nanometers (near-infrared). This covers all the specific "stations" needed to control these atoms.
The Precision: They managed to focus the light into a spot that is incredibly small—about 0.67 to 1.46 micrometers wide. To put that in perspective, that's roughly the width of a single bacterium.
The "No-Leak" Test: The most important test was "crosstalk." This is when you try to talk to one atom, but the light accidentally hits its neighbor. The researchers found that their system kept the light focused so tightly that the "noise" or leakage to the neighboring spot was extremely low, measuring at -27 dB. In everyday terms, if the main beam is a shout, the neighbor only hears a whisper.
Real-World Proof: They didn't just simulate this on a computer; they actually trapped real atoms above the chip. They showed that they could use the chip to "repump" (a specific type of laser control) a single Calcium atom while its neighbor, sitting just 5 micrometers away, remained unaffected.

What They Didn't Solve (Yet)

It is important to note that while this is a major step forward, the paper also points out some hurdles they haven't cleared yet. When they tried to trap Barium ions directly over the mirror, they ran into a problem called "charging." The electric fields from the light seemed to build up a static charge on the chip's surface, which pushed the atoms away and made them unstable. They suspect this is because the light interacts with the silicon or the glass-like materials on the chip.

The authors suggest that in the future, they might fix this by adding a special conductive layer (like a transparent metal) or a shield to stop the static charge. They also noted that the Calcium atoms were a bit wobbly in their trap, likely due to similar electrical issues. So, while the "light-bending" part of the machine works beautifully, the "holding the atoms steady" part still needs some fine-tuning.

Why This Matters

This paper shows that we can finally move away from bulky, messy laser setups and toward a clean, chip-based system that can handle multiple colors of light at once. By replacing many different antennas with one smart, 3D-printed mirror, they have opened the door to building quantum computers with hundreds or even thousands of atoms, all controlled by a single, compact device. It's a crucial step toward turning the science fiction of quantum computing into a reality we can hold in our hands.

Technical Summary: Broadband, Individually Addressing 2D-3D Photonic Integrated Circuit for Trapped-Ion Qubit Control

Problem Statement
Trapped-ion quantum processors currently face a significant scalability bottleneck in optical engineering. While trapped ions offer high-fidelity gate operations, scaling beyond ~100 physical qubits requires delivering multiple distinct laser wavelengths (spanning the near-ultraviolet to near-infrared) to individual ions with micron-level precision. Conventional free-space optical assemblies do not scale favorably due to footprint, vibration sensitivity, and alignment complexity. Conversely, existing on-chip solutions relying on diffractive grating couplers suffer from narrow operational bandwidth. Because grating emission properties are wavelength-dependent, achieving high efficiency and low crosstalk typically requires multiple gratings per ion or limits the number of wavelengths that can be delivered simultaneously. This constraint increases the circuit footprint and optical interfacing complexity, limiting the density of qubits that can be controlled within a single interaction zone.

Methodology and Device Architecture
The authors present a hybrid, monolithic photonic integrated circuit (PIC) that combines guided-wave photonics with 3D-printed free-space optics to achieve broadband, individually addressable beam delivery. The device is fabricated on 6-inch silicon wafers using a process that integrates conventional waveguides with additive manufacturing.

Photonic Stack: The core consists of an amorphous aluminum oxide (Al $_2$ O $_3$ ) waveguide symmetrically clad with silicon dioxide (SiO $_2$ ). This material platform supports broadband routing from 375 nm to 650 nm with propagation losses better than 3 dB/cm.
Optical Path: Light is routed through strip waveguides and conditioned by a lithographically defined planar waveguide lens. Upon exiting the waveguide, the beam diverges and is captured by a 3D biconic-aspherical micromirror ( $\mu$ -mirror).
Additive Manufacturing: The $\mu$ -mirrors are fabricated using two-photon polymerization (2PP) at the wafer scale. They are printed flush within an etched trench in the center DC electrode of a 5-wire surface Paul trap.
Integration: A 200 nm aluminum layer is deposited over the printed optics to serve dual purposes: acting as a reflective coating for the $\mu$ -mirror and forming the surface trap electrodes. This process includes a novel post-2PP step where high-resolution lithography and etching are performed directly on the printed micro-optics to define electrode gaps and clear optical facets.
Alignment: The system achieves alignment accuracy of $\le$ 1 $\mu$ m between the substrate and the printed mirror, with the mirror held at the same potential as the center DC electrode to minimize electric field distortion.

Key Results
The device was characterized across a wavelength range of 405 nm to 880 nm, covering the key transitions for $^{40}$ Ca $^+$ and $^{138}$ Ba $^+$ ions.

Broadband Focusing: The system successfully delivers individually addressable optical beams to a height of 75 $\mu$ m above the chip surface. The focal spot retains a tight "bowtie" shape across the entire bandwidth.
Spot Size and Crosstalk: The axial full-width at half-maximum (FWHM) ranges from 0.67 $\mu$ m at 405 nm to 1.46 $\mu$ m at 880 nm. The radial FWHM remains below 8 $\mu$ m. Crucially, the system achieves an average intensity crosstalk of -27.6 dB at a 5 $\mu$ m pitch (the typical spacing between ions in a linear trap). This performance was verified at both 515 nm and 729 nm, demonstrating individual addressing at two distinct wavelengths using a single shared optical element.
Loss Characterization: End-to-end fiber-to-ion losses were measured, ranging from ~10 dB in the near-infrared to 35 dB at 405 nm. The optical contribution of the planar lens and $\mu$ -mirror itself is approximately 2 dB across the band, indicating that higher losses at short wavelengths are due to the waveguide platform rather than the 3D optical architecture.
Ion Trapping and Control:
- $^{138}$ Ba $^+$ : Ions were successfully trapped above the device using free-space lasers. Optical crosstalk was characterized by measuring the excitation rate of the S $_{1/2} \to$ P $_{3/2}$ transition. At a distance of 120 $\mu$ m from the beam center, the intensity was suppressed by approximately 57 dB, comparable to precision grating couplers.
- $^{40}$ Ca $^+$ : A stable two-ion crystal was trapped directly above the $\mu$ -mirror. The integrated 866 nm beam was used to demonstrate selective repumping of a single ion in a pair separated by 5 $\mu$ m, confirming individual addressing capability with real ions.

Significance and Claims
The paper claims to demonstrate the first monolithic ion-trap PIC capable of individually addressing multiple ions at two different wavelengths using a single shared optical element. By replacing wavelength-specific grating couplers with a broadband planar lens and 3D-printed mirror, the architecture decouples the number of optical delivery elements from the number of required wavelengths.

The authors posit that this approach offers a scalable path for trapped-ion quantum hardware. Specifically, it allows for a higher density of qubits within a single interaction zone without increasing the photonic footprint, thereby reducing the number of ion transport operations required per clock cycle in a Quantum Charge-Coupled Device (QCCD) architecture. The work also highlights the potential of integrating additive manufacturing (2PP) into quantum devices to expand the design space for novel quantum architectures.

Limitations and Future Work
The authors note that dielectric charging on the exposed planar lens facets and photo-induced conductivity changes in the silicon substrate currently limit trap stability and coherence. They propose future mitigations, such as depositing a transparent conductive oxide (e.g., ITO) on the lens facets or adding a ground plane to the material stack to shield ions from substrate charging effects. These improvements are expected to enable extended trap operation times.

Technical Summary: Broadband, Individually Addressing 2D-3D Photonic Integrated Circuit for Trapped-Ion Qubit Control

More like this