🔬 mesoscale physics

Correlating Superconducting Qubit Performance Losses to Sidewall Near-Field Scattering via Terahertz Nanophotonics

This paper demonstrates that noninvasive terahertz nano-imaging and spectroscopy can effectively correlate sidewall near-field scattering and dielectric responses with superconducting qubit coherence, offering a high-throughput alternative to destructive characterization methods for optimizing quantum circuit performance.

Original authors: Richard H. J. Kim, Samuel J. Haeuser, Joong-Mok Park, Randall K. Chan, Jin-Su Oh, Thomas Koschny, Lin Zhou, Matthew J. Kramer, Akshay A. Murthy, Mustafa Bal, Francesco Crisa, Sabrina Garattoni, Shaoji

Published 2026-02-19

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Richard H. J. Kim, Samuel J. Haeuser, Joong-Mok Park, Randall K. Chan, Jin-Su Oh, Thomas Koschny, Lin Zhou, Matthew J. Kramer, Akshay A. Murthy, Mustafa Bal, Francesco Crisa, Sabrina Garattoni, Shaojiang Zhu, Andrei Lunin, David Olaya, Peter Hopkins, Alex Romanenko, Anna Grassellino, Jigang Wang

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-fast, super-quiet library where books (quantum information) can be read without any noise disturbing them. In the world of quantum computing, these "books" are stored in tiny circuits called superconducting qubits. The goal is to keep the information in these qubits for as long as possible before it gets scrambled by noise. This duration is called coherence time (or $T_1$ ).

However, right now, these qubits are like fragile glass houses. Even the tiniest imperfection in the walls or the floor can cause the glass to shatter (the information is lost). Scientists have been trying to figure out exactly where these cracks are hiding.

The Problem: The "Invisible" Cracks

Traditionally, to find these cracks, scientists had to use two methods, both of which were problematic:

The "X-Ray" Method (Electron Microscopy): This is like taking a sledgehammer to the library to see the bricks. It destroys the qubit, so you can't use it again.
The "Freeze-Test" (Cryogenic Measurement): This involves putting the qubit in a freezer colder than outer space and waiting hours or days to see if it works. It's slow, expensive, and you can't test many of them quickly.

The researchers in this paper wanted a way to check the qubits without breaking them and without freezing them.

The Solution: The "Terahertz Flashlight"

The team developed a new tool using Terahertz (THz) light. Think of this as a super-powered, ultra-precise flashlight that can see things smaller than a human hair, but without touching them.

They combined this light with an Atomic Force Microscope (AFM). Imagine a tiny, invisible needle (the tip of the microscope) hovering just above the surface of the qubit. When they shine the Terahertz light on this needle, it acts like a tiny antenna.

If the surface underneath is smooth and perfect, the light bounces off nicely.
If there is a crack, a rough edge, or a weird chemical layer, the light scatters wildly.

This allows them to create a "heat map" of the qubit's surface, showing exactly where the energy is getting lost, all while the qubit sits at room temperature.

The Big Discovery: The "Sidewall" Secret

The qubits they tested were made of Niobium (a special metal) and covered with a protective layer (Gold-Palladium) to keep them from rusting (oxidizing).

The Analogy: Imagine building a wall out of bricks (Niobium) and then painting the top of the wall to protect it. The researchers thought, "Great, the top is safe!" But they discovered that the sides of the wall (the sidewalls) were left unpainted and exposed to the air.

The Finding: The Terahertz flashlight revealed that the rough, oxidized edges on the sides of the metal were acting like tiny magnets for energy loss.
The Correlation: They found a direct link: The rougher and more "scattered" the light was at the sidewalls, the shorter the qubit's life ( $T_1$ ) was. Conversely, the smoother the sidewalls looked to the Terahertz light, the longer the qubit lasted.

It's like realizing that while you painted the roof of your house, you left the gutters full of holes. The rain (energy loss) was coming in through the gutters, not the roof.

Bonus: Finding a "Ghost" in the Machine

The team also used this tool to look at the Josephson Junction, which is the tiny heart of the qubit (like the engine of a car). They found a tiny, invisible defect—a "dent" in the metal—that was only a few nanometers wide.

Normally, a dent would make the surface look lower.
But with this Terahertz tool, they saw that the electrical properties changed drastically at that dent, even though the physical shape didn't change much. It was like finding a ghost in the machine that only appeared when you looked at it with this specific "ghost-hunting" light.

Why This Matters

This research is a game-changer for two reasons:

Speed: Instead of waiting days to freeze-test a qubit, engineers can now scan it in minutes at room temperature. It's like going from a slow mail-in test to an instant blood test.
Guidance: It tells manufacturers exactly what to fix. "Don't just worry about the top of the metal; polish the sides!" This helps them build better, longer-lasting quantum computers faster.

In summary: The scientists built a magical, non-invasive "Terahertz X-ray" that can see the tiny, invisible flaws on the sides of quantum computer chips. They discovered that these side-flaws are the main reason the computers lose their memory, and now they have a fast way to fix them.

1. Problem Statement

Superconducting qubits, particularly those based on niobium (Nb), face a critical bottleneck in scalability due to limited coherence times ( $T_1$ ). The primary source of microwave loss is attributed to Two-Level Systems (TLS) residing in surface oxides and structural defects.

Current Limitations: Existing diagnostic methods rely on destructive electron microscopy (TEM/SEM) or low-throughput cryogenic quantum measurements (millikelvin $T_1$ characterization).
The Gap: While encapsulating Nb surfaces with capping layers (e.g., AuPd) has shown promise in reducing oxide formation and improving $T_1$ , the specific impact of sidewall defects and structural heterogeneity introduced during fabrication remains poorly understood. There is a lack of non-invasive, high-throughput tools to correlate nanoscale structural features with qubit performance at room temperature.

2. Methodology

The authors employed Terahertz Scattering-type Scanning Near-field Optical Microscopy (THz-sSNOM) combined with Atomic Force Microscopy (AFM) to characterize encapsulated Nb transmon qubits.

Experimental Setup:
- Source: Broadband THz pulses generated by a Yb-laser (20 $\mu$ J, 1 MHz, 91 fs pulse width, 1030 nm center wavelength).
- Probe: A tapping-mode AFM tip (nominal radius ~20 nm) scatters the THz light.
- Detection: Electro-optic sampling of the backscattered near-field signal. Signals are demodulated at higher harmonics ( $n=1, 2, 3, 4$ ) of the tip-tapping frequency to isolate the near-field component ( $s_n$ ) from far-field background.
- Samples: Nb transmon qubits fabricated on HEMEX sapphire substrates, featuring a 155-nm Nb layer capped with 6-nm AuPd.
Measurement Strategy:
- Room-Temperature Imaging: Raster scans of Nb sidewalls and Josephson junction (JJ) areas to map local conductivity and field enhancements.
- Line Scans: 16 line profiles were taken across Nb sidewalls on four different qubits to extract near-field amplitude peaks ( $s_{peak}$ ) and substrate baselines ( $s_{sub}$ ).
- Correlation: The extracted near-field metrics were compared against cryogenic $T_1$ measurements and Quality Factors ( $Q$ ) obtained from microwave characterization.
- Validation: Cross-sectional Transmission Electron Microscopy (TEM) was used to correlate THz signals with physical structural defects (oxide thickness, trench depth).
- Nanospectroscopy: Time-domain THz spectroscopy was performed across a Josephson junction to extract complex dielectric constants ( $\epsilon_1, \epsilon_2$ ).

3. Key Contributions

Non-Destructive Proxy Characterization: Demonstrated that THz-sSNOM can serve as a rapid, room-temperature screening tool to predict qubit coherence, eliminating the need for time-consuming cryogenic testing for initial material selection.
Sidewall Correlation: Identified a direct, linear correlation between sidewall near-field scattering amplitudes and qubit coherence times ( $T_1$ ), specifically linking edge defects to performance loss.
Dielectric Mapping: Successfully extracted effective dielectric functions of Al/AlOx/Al Josephson junctions with ~20 nm spatial resolution, revealing sub-wavelength structural features that behave like metamaterials.
Defect Identification: Visualized specific structural anomalies (e.g., incomplete capping layers, trench depth variations) that are invisible to standard AFM topography but detectable via THz near-field scattering.

4. Key Results

Correlation with Coherence ( $T_1$ and $Q$ ):
- A strong linear trend was observed between the normalized near-field scattering signal at Nb edges, defined as $\frac{s_{peak} - s_{sub}}{s_{sub}}$ (specifically the 2nd harmonic $s_2$ ), and the qubit's Quality Factor ( $Q$ ).
- Inverse Relationship: Higher near-field scattering amplitudes (indicating stronger field concentration at defects) correlated with longer $T_1$ times and higher $Q$ in the encapsulated samples.
- Note: This correlation was absent in non-encapsulated qubits, suggesting the encapsulation process creates a specific defect profile where sidewall quality is the dominant loss factor.
Structural Insights (TEM vs. THz):
- TEM revealed that the AuPd capping layer often stopped short of the Nb edge, leaving a ~100 nm exposed Nb top surface prone to oxidation.
- Trench Depth: A relationship was found between trench depth (etched substrate area) and $Q$ . Deeper trenches correlated with higher $Q$ , consistent with simulations showing reduced electric field participation in the lossy substrate.
- The THz near-field signal acts as a comprehensive proxy for these combined geometric and chemical factors (oxide presence, trench depth, edge curvature).
Josephson Junction Analysis:
- THz imaging detected a "dark spot" (defect) in an Al JJ arm, showing a ~30–40% drop in scattered amplitude and a slight topographical indentation (<5 nm).
- Spectroscopy revealed that the defect region exhibited suppressed conductivity at lower frequencies, fitting a Drude-Smith model indicative of carrier confinement, distinct from bulk material behavior.
- While the specific JJ defect did not immediately correlate with a drop in $T_1$ for that specific device, the ability to resolve local electrodynamics at the nanoscale was proven.

5. Significance and Future Outlook

Optimization of Fabrication: The findings suggest that optimizing the sidewall geometry and ensuring complete capping layer coverage are critical for minimizing TLS losses in Nb-based qubits.
High-Throughput Screening: THz-sSNOM offers a fast, non-invasive alternative to destructive TEM or slow cryogenic testing, enabling rapid feedback loops for material selection and process optimization.
Future Directions:
- Extending THz-sSNOM to cryogenic temperatures (below the Al critical temperature) to directly probe quasiparticle dynamics and Cooper pair distributions.
- Using these techniques to study collective modes and topological phases in quantum materials.
- Developing quantitative simulations linking THz electrodynamics directly to GHz qubit performance metrics.

In summary, this work establishes THz near-field scattering as a powerful, non-destructive diagnostic tool that bridges the gap between nanoscale structural defects and macroscopic quantum coherence, providing a clear pathway to improving the performance of superconducting quantum circuits.