Abstract
The performance of superconducting qubits is degraded by a poorly characterized set of energy sources breaking the Cooper pairs responsible for superconductivity, creating a condition often called "quasiparticle poisoning". Both superconducting qubits and low threshold dark matter calorimeters have observed excess bursts of quasiparticles or phonons that decrease in rate with time. Here, we show that a silicon crystal glued to its holder exhibits a rate of low-energy phonon events that is more than two orders of magnitude larger than in a functionally identical crystal suspended from its holder in a low-stress state. The excess phonon event rate in the glued crystal decreases with time since cooldown, consistent with a source of phonon bursts which contributes to quasiparticle poisoning in quantum circuits and the low-energy events observed in cryogenic calorimeters. We argue that relaxation of thermally induced stress between the glue and crystal is the source of these events.
Introduction
Superconducting qubits rely on the formation of Cooper pairs—electrons bound together at cryogenic temperatures. However, these pairs are frequently broken by unknown energy sources, creating "quasiparticles" that cause decoherence and the loss of quantum information. This phenomenon, known as quasiparticle poisoning, has been a persistent challenge in the development of scalable quantum computing. Similarly, low-threshold dark matter experiments observe a mysterious "excess" of low-energy phonon events that limit their sensitivity. While researchers have previously considered infrared radiation or environmental radioactivity as culprits, these do not fully explain the burst-like nature of the noise or its temporal decay. This study investigates whether mechanical stress from mounting hardware is the hidden source of these disturbances.
Experimental Design: High Stress vs. Low Stress
To test the stress hypothesis, researchers conducted a head-to-head comparison between two functionally identical silicon crystals equipped with Quasiparticle-trap-assisted Electrothermal Feedback Transition Edge Sensors (QETs). These sensors detect athermal phonons with eV-scale resolution.
- High Stress (HS): The crystal was glued to a copper mount using GE varnish. Due to mismatched thermal contraction between the glue and silicon at millikelvin temperatures, this creates significant mechanical stress.
- Low Stress (LS): The crystal was suspended using 50 \mu m diameter aluminum wire bonds, allowing the materials to contract independently and minimizing stress.
The Mechanism of Phonon Bursts
The study proposes that as the glue-silicon interface "settles" after cooling, it undergoes microscopic fractures. Each tiny snap releases a burst of athermal phonons into the crystal. These phonons travel to the superconducting films and break Cooper pairs, creating the quasiparticles that disrupt qubit coherence or trigger spurious detector events. This mechanism explains the "bursty" nature of the noise observed in many cryogenic experiments.
Results: A 100-fold Reduction in Noise
Data collected over 80 hours revealed a dramatic difference between the two mounting methods. In the primary energy range of interest (3–38 eV), the glued HS crystal exhibited an event rate more than two orders of magnitude higher than the suspended LS crystal.
Key findings include:
- Quantifiable Decay: The event rate in the HS crystal followed a clear exponential decay curve with a time constant of 6–10 days, matching the expected behavior of stress relaxation. In contrast, the LS crystal's rate remained low and stable.
- Quasiparticle Density Impact: Simulations estimated that the stress-induced phonons in the HS setup produce a quasiparticle density ($x_{qp}$) of $\approx 5.0 \times 10^{-8}$, which is high enough to be a dominant error source in many current superconducting qubit architectures.
Discussion and Future Outlook
While this work identifies mechanical stress relaxation as a major source of noise, some questions remain. Even in the low-stress setup, a residual noise floor exists, possibly due to stress within the thin metal films deposited on the silicon. Future work will explore different substrates, alternative adhesives, and the effects of repeated thermal cycling.
Nevertheless, this research offers a clear engineering path to more stable quantum computers and more sensitive dark matter detectors. Transitioning to low-stress mounting techniques can reduce phonon-induced errors by a factor of 100, potentially overcoming a significant bottleneck in the scaling of superconducting quantum processors.
Summary updated