⚛️ quantum physics

Ion trap on borosilicate substrate with integrated femtosecond-laser-written waveguide

This paper presents a scalable ion-trap platform on a borosilicate substrate featuring an integrated femtosecond-laser-written waveguide for on-chip light delivery, which was successfully validated in a cryogenic system with $^{40}$ Ca $^+$ ions to demonstrate coherent operations and low-loss guidance at both 729 nm and 405 nm.

Original authors: Jakob Wahl, Alexander Zesar, Philipp Hurdax, Marco Schmauser, Victoria Schwab, Michael Pasquini, Marco Valentini, Clemens Rössler, Thomas Monz, Bernhard Lamprecht, Klemens Schüppert, Philipp Schindler

Published 2026-08-14

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Jakob Wahl, Alexander Zesar, Philipp Hurdax, Marco Schmauser, Victoria Schwab, Michael Pasquini, Marco Valentini, Clemens Rössler, Thomas Monz, Bernhard Lamprecht, Klemens Schüppert, Philipp Schindler

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 a world where computers don't just calculate numbers but solve mysteries that would take today's supercomputers a million years to crack. This is the promise of quantum computing, and one of the most promising ways to build these machines is by trapping tiny, charged atoms called ions in mid-air using invisible electric fields. Think of these ions as the "qubits," the basic building blocks of this new kind of computer. To make them work, scientists need to talk to them using very specific beams of laser light, like shining a flashlight on a single firefly in a dark forest to tell it what to do.

However, as scientists try to build bigger quantum computers with thousands of these fireflies, a major problem arises: how do you shine a laser on each one individually without the beams getting in each other's way? Currently, they use giant, clunky lenses and mirrors floating above the chips, which is like trying to juggle a thousand flashlights with your hands while wearing oven mitts. It's messy, hard to scale, and prone to shaking. The goal is to shrink this whole setup down, integrating the "flashlights" directly onto the chip itself, creating a sleek, stable system where the light travels through tiny tunnels built right into the machine.

This is exactly what a team of researchers has achieved in a new study. They have built a special "trap" for ions using a type of glass called borosilicate (the same kind used in sturdy kitchenware) and, for the first time, have successfully built a microscopic light tunnel, or "waveguide," directly into a separate piece of glass that sits right on top of the trap. Instead of shining a laser from the ceiling, they guide the light through this glass tunnel so it pops out right next to the trapped ion.

The team didn't just build it; they tested it rigorously. They created a system where they could trap a calcium ion, move it around like a tiny marble on a track, and then stop it right in front of the glass tunnel. When they sent laser light through the tunnel, it successfully hit the ion and made it perform complex quantum "dances" (known as Rabi oscillations), proving the light delivery works perfectly. They also checked if the extra glass block sitting on the trap would mess things up. They found that while the glass did create a tiny bit of static electricity (like a balloon rubbed on hair), it was incredibly stable and didn't heat up the ion or cause it to jitter.

The researchers also figured out how to tune the size of the light beam coming out of the tunnel. By changing how they wrote the tunnel with a powerful laser, they could make the beam wider or narrower, even down to a size that works for blue light, which is crucial for cooling the ions. They discovered that they could bend these light tunnels quite sharply—down to a curve with a radius of just 6 millimeters—without losing much light, which is a big deal for packing everything into a small space.

In short, this paper shows a new, robust way to build the "wiring" for future quantum computers. By separating the optical part (the light tunnels) from the electrical part (the trap electrodes) and bonding them together, they created a platform that is ready for mass production and can be scaled up to handle thousands of ions. It's a significant step toward turning the dream of a massive, stable quantum computer into a reality, proving that you can deliver light to a trapped atom with precision, stability, and without the need for a room full of bulky mirrors.

Technical Summary: Ion Trap on Borosilicate Substrate with Integrated Femtosecond-Laser-Written Waveguide

Problem Statement
Scaling trapped-ion quantum computers to thousands of qubits requires overcoming critical engineering bottlenecks in optical light delivery. Current systems rely on free-space optics, which necessitate dedicated objectives for each gate zone and become geometrically unfeasible as qubit numbers increase. While integrated photonic circuits offer a solution, state-of-the-art approaches typically place waveguides beneath trap electrodes, requiring complex out-of-plane coupling elements (e.g., grating couplers or micro-mirrors) that introduce fabrication challenges and bandwidth limitations. Furthermore, integrating optics and electrodes on a single substrate is complicated by material incompatibilities and the need for high-precision alignment.

Methodology
The authors propose and validate a hybrid platform where the optical delivery layer is physically separated from the trap electrode substrate. The architecture consists of:

Substrate Stack: A surface-electrode ion trap patterned on a borosilicate glass layer atop a silicon wafer. A second, structured borosilicate glass block is anodically bonded directly onto the trap surface.
Waveguide Fabrication: Femtosecond-laser-written (FLW) waveguides are inscribed within the bonded optics block. These waveguides run parallel to the trap surface at a height ( $170\ \mu\text{m}$ ) matching the ion's trapping position, eliminating the need for out-of-plane couplers.
Process Engineering: The fabrication utilizes industrial semiconductor processes (Infineon Technologies) for the trap and direct laser writing (JOANNEUM RESEARCH) for the optics. The process includes active fiber alignment and bonding to the chip.
Waveguide Characterization: The authors systematically engineered the waveguide mode-field diameter (MFD) by tuning laser writing parameters, specifically pulse energy, repetition rate, and lateral scan offsets. They characterized transmission, bending loss, and far-field divergence at relevant wavelengths ( $729\ \text{nm}$ and $405\ \text{nm}$ ).
Experimental Validation: The platform was tested in a cryogenic ( $\approx 8\ \text{K}$ ) system using $^{40}\text{Ca}^+$ ions. Performance metrics included heating rates, stray electric field stability, and coherent qubit operations driven by light delivered through the integrated waveguide.

Key Contributions and Results

Integrated Light Delivery: The team successfully demonstrated a platform where light is delivered directly to the ion via an in-plane waveguide, bypassing the need for complex out-of-plane coupling structures.
Waveguide Engineering:
- Achieved single-mode, low-loss guidance at $729\ \text{nm}$ with tunable MFDs ranging from $\approx 3\ \mu\text{m}$ to $8\ \mu\text{m}$ by adjusting pulse energy and lateral scan offsets.
- Demonstrated low-loss curved waveguides with radii of curvature down to $6\ \text{mm}$ , enabled by strong mode confinement.
- Extended single-mode operation to the blue spectral range ( $405\ \text{nm}$ ), crucial for cooling and state preparation.
Trapping Performance:
- Heating Rates: Measured heating rates between $0.5$ and $1.5$ quanta/s at an axial frequency of $1\ \text{MHz}$ . The presence of the dielectric block at a $1\ \text{mm}$ distance did not measurably increase heating rates.
- Stray Fields: Characterized stray electric fields, finding them remarkably stable with an average drift of $(-4.48 \pm 0.16) \times 10^{-4}\ \text{V/mm/hour}$ . The spatial dependence of these fields was consistent with a uniform surface charge density of $\approx 2\ e/\mu\text{m}^2$ on the glass block.
Coherent Operations: Successfully trapped a $^{40}\text{Ca}^+$ ion, shuttled it to an addressing zone in front of the waveguide, and drove coherent qubit dynamics (Rabi oscillations) on the $4^2S_{1/2} \leftrightarrow 3^2D_{5/2}$ transition using $729\ \text{nm}$ light delivered through the integrated waveguide.

Significance and Claims
The paper claims to provide a "robust, scalable route to integrated light delivery" for trapped-ion devices. By physically separating the optical and electrical layers, the architecture maintains compatibility with silicon-based integration (ASICs) and standard microfabrication processes while avoiding the fabrication complexities of out-of-plane couplers.

The authors emphasize that the platform is compatible with hybrid micro-optics (e.g., pick-and-place lenses) to achieve single-ion addressing, though the current demonstration utilized a global beam. They note that while the current waveguide output is a global beam with a radius of $\approx 100\ \mu\text{m}$ at the ion, the system validates the end-to-end functionality of the platform. The work establishes that inscribing waveguides parallel to the trap surface is a viable strategy for scaling, offering a path to circumvent the geometrical constraints of free-space optics without introducing significant dielectric heating or instability.

More like this