Low-Crosstalk, Silicon-Fabricated Optical Waveguides for Laser Delivery to Matter Qubits

This paper reports the development and demonstration of CMOS-fabricated silicon nitride optical waveguides that achieve crosstalk reduction exceeding 50 dB, enabling the precise, scalable delivery of laser fields to address chains of trapped barium ions for quantum information processing.

The Gist

The Big Picture: A "Light Traffic Controller" for Quantum Computers

Imagine you are trying to talk to a row of eight people sitting in a dark room. You want to whisper a secret to just one specific person without the person sitting next to them hearing it. If you use a giant flashlight, the light spills over, and everyone hears the whisper. This "spillover" is called crosstalk, and in the world of quantum computers (which use tiny particles called qubits to store information), even a tiny bit of spillover ruins the calculation.

This paper describes a new, high-tech "flashlight" made of silicon that solves this problem. It is a microscopic chip that takes one beam of laser light and splits it into eight separate beams, aiming each one perfectly at a specific ion (a charged atom) without the beams leaking into their neighbors.

The Problem: The "Messy Room" of Light

In the past, scientists used big, bulky mirrors and lenses to aim lasers at these atoms. It was like trying to direct traffic in a crowded city using a single person with a megaphone. It was hard to scale up, hard to keep precise, and light often leaked where it shouldn't.

The researchers wanted to build a chip that could do this job automatically, like a pre-programmed traffic light system, but for light.

The Solution: A Silicon "Highway" for Light

The team built a chip using Silicon Nitride (a type of glass-like material). Think of this chip as a tiny, invisible highway system for light.

1. The Highway (Waveguides): Instead of light flying through the air, it travels inside tiny, narrow tunnels (waveguides) carved into the chip. This keeps the light contained, just like a train stays on its tracks.
2. The Exit Ramps: The chip splits the light into eight different "exit ramps." The tricky part is that the atoms they are trying to hit aren't sitting in a perfect straight line; they are spaced out unevenly. The chip was designed to match this messy spacing perfectly.
3. The "Moat" (Trenches): This is the paper's biggest innovation. To stop light from leaking from one ramp to the next, the engineers dug deep "moats" (trenches) between the exit ramps.
   • The Analogy: Imagine two houses next to each other. If you want to stop sound from traveling from one house to the other, you might dig a deep ditch between them. If the sound wave hits the ditch, it falls in and dies out instead of crossing over. These "moats" on the chip catch stray light and stop it from bothering the neighbor.

The Results: Silence is Golden

The team tested this chip with different colors of laser light (blue, yellow, and red).

• The Test: They shone light into the chip and measured how much "spillover" happened between the exit ramps.
• The Score: They found that the light leaking to the neighbor was reduced by more than 50 decibels.
  • The Analogy: That is like the difference between a jet engine roaring right next to your ear and a library being completely silent. It is a massive reduction in noise.
• The Proof: They used this chip to cool down a chain of eight barium atoms (ions). When the light hit the atoms, they glowed (fluoresced). When the light missed the atoms (because the chip was moved slightly), the glow stopped. This proved the chip could hit the targets precisely without blinding the neighbors.

Why This Matters (According to the Paper)

The paper claims this is a major step forward because:

1. It's Mass-Producible: They didn't build this in a messy lab with hand-held tools. They used a standard computer chip factory (a "foundry"). This means they can make thousands of these identical chips, just like making computer processors.
2. It's Scalable: Because it's a small chip, you can put many of them together to control hundreds or thousands of qubits, which is necessary for building a powerful quantum computer.
3. It's Precise: It can handle atoms that are spaced irregularly, which is a common problem in real-world quantum traps.

Summary

The researchers built a tiny silicon chip that acts like a precision laser pointer array. By digging deep trenches between the light paths, they stopped the light from leaking, ensuring that each quantum bit gets its own private message without interference. They proved it works by using it to control and cool a chain of atoms, showing that this technology is ready to help build the next generation of quantum computers.

Technical Summary

1. Problem Statement

Controlling matter-based quantum bits (qubits), such as trapped ions, requires precise delivery of laser fields to individual qubits. As quantum systems scale to larger arrays, a significant challenge arises: spatial cross-talk.

• The Challenge: Adjacent qubits are often closely spaced and irregularly positioned. Delivering laser light to specific qubits without leaking energy into neighboring qubits degrades gate fidelity and causes decoherence.
• Limitations of Current Tech: Traditional bulk optics are not scalable. Existing integrated solutions (like laser-written glass waveguides) often lack the high index contrast and fabrication precision required for the visible wavelengths (493 nm, 585 nm, 650 nm) used in ion trapping, or they suffer from insufficient cross-talk suppression.
• Goal: Develop a scalable, CMOS-compatible photonic chip that can deliver laser light to an array of unequally spaced qubits with ultra-low cross-talk, enabling high-fidelity quantum control.

2. Methodology

The authors designed, fabricated, and characterized a silicon nitride ($Si_3N_4$) photonic chip using a commercial foundry process.

A. Fabrication Platform

• Material: 150 nm thick $Si_3N_4$ waveguides deposited on Silicon-on-Insulator (SOI) wafers via Low-Pressure Chemical Vapor Deposition (LPCVD).
• Why $Si_3N_4$? Unlike standard silicon waveguides which absorb visible light (bandgap ~1.1 eV), $Si_3N_4$ is transparent in the near-UV to visible spectrum, making it ideal for ion transitions (493 nm, 585 nm, 650 nm).
• Process: Fabricated at AIM Photonics (a commercial CMOS foundry) using 193 nm deep-UV lithography. This ensures high reproducibility and scalability.

B. Design Architecture & Crosstalk Mitigation

To minimize stray light and cross-talk, the design incorporates several key features:

1. Strategic Bending: The waveguides utilize $90^\circ$ radial bends. This ensures that stray light (slab modes) excited at the input or splitter interfaces is directed away from the output facet, preventing it from exiting the chip alongside the intended signal.
2. Deep-Etched Air Trenches: Air trenches are etched between output waveguides, extending through the cladding, buried oxide (BOX), and partially into the substrate. This physically breaks the waveguide path for any stray light, forcing it to diverge rapidly rather than coupling to adjacent channels.
3. Spot Size Converters (SSCs): Inverse-tapered SSCs expand the highly confined waveguide mode (500 nm width) to a larger mode field diameter (100 nm width at the tip) at the output facet. This allows for efficient coupling to high-magnification imaging optics and diffraction-limited focusing on individual ions.
4. Irregular Spacing: The output spacing is patterned to match the specific, unequal spacing of ions in a Paul trap, scaled to work with the imaging system's magnification.

C. Characterization Techniques

• Scanning Slit Method: A 5 $\mu$m slit scanned across the output image plane (magnified 50x) to measure the transversely integrated intensity profile. This measures the "conservative" cross-talk, including stray light.
• Direct Fiber Scan: A single-mode fiber scanned across the output to measure the 2D intensity profile and background noise floor, isolating intrinsic chip performance from imaging aberrations.
• Ion Demonstration: The chip was used to drive the repumping transition ($D_{3/2} \leftrightarrow P_{1/2}$) of a chain of eight trapped $^{138}Ba^+$ ions to perform Doppler cooling.

3. Key Results

A. Optical Performance

• Cross-talk Suppression:
  • Using the scanning slit method, the measured cross-talk between adjacent waveguide outputs at 650 nm was $-50.8(1.3)$ dB.
  • Devices designed for 493 nm and 585 nm showed similar performance.
  • The direct fiber scan revealed a peak-to-background ratio of $-60.6(2.5)$ dB, indicating extremely low intrinsic leakage.
• Impact of Trenching: Comparisons showed that devices without trenches had a much slower intensity falloff between channels. The trenches were proven to be the critical factor in achieving the sharp cutoff and high cross-talk suppression.
• Loss: The waveguides exhibited a propagation loss of 1.7 dB/cm, with bend losses simulated to be below 0.1 dB per $90^\circ$ bend.

B. Ion Trapping Demonstration

• The chip successfully delivered 650 nm light to a chain of eight unequally spaced Barium ions.
• Doppler Cooling: The chip provided the necessary repumping light to maintain the ions in a crystallized, cooled state.
• Spatial Resolution: When the chip outputs were translated relative to the ion chain, individual ions flickered in and out of fluorescence as they moved between the focused beam spots and the dark regions (high extinction). This confirmed that the light was tightly focused on individual ions and that the cross-talk was low enough to prevent unintended excitation of neighbors.

4. Significance and Impact

• Scalability: This work demonstrates a CMOS-foundry-compatible method for producing complex photonic circuits. This moves quantum control hardware from custom, hand-built optics to reproducible, mass-producible chips.
• Performance Benchmark: The achieved cross-talk of ~51 dB is an order of magnitude better than previously reported low-cross-talk literature for similar applications, setting a new standard for integrated quantum photonics.
• Versatility: The platform supports multiple visible wavelengths (UV to Red) and can be tailored to different qubit spacings and trap geometries.
• Future Applications: While demonstrated here as a passive device for cooling, the architecture supports the integration of active components (modulators, switches) for individual qubit addressing, mid-circuit measurements, and quantum networking. It provides a robust interface between optical fibers and matter qubits, a critical step toward modular, large-scale quantum computers.

Conclusion

The paper presents a breakthrough in integrated quantum photonics by successfully fabricating a low-loss, ultra-low cross-talk silicon nitride waveguide array. By leveraging foundry processes and innovative design features like deep-etched trenches and strategic bending, the authors achieved a device capable of precisely addressing irregularly spaced trapped ions. This technology solves a critical scaling bottleneck for matter-based quantum computing, offering a path toward high-fidelity, modular quantum systems.