[QISCA Journal Club] Drive Crosstalk Cancellation on Surface - 43 chip

[QISCA Journal Club] Drive Crosstalk Cancellation on Surface - 43 chip

🎙 Sangyeon Lee (이상연), SQRT 👥 267 📅 September 11, 2025 ⏱ 40 min 👁 36 📄 original study 🧭 2026-08-15
Available in: English (current) Français

Keywords

drive crosstalkcancellationsurface-43superconducting qubitsrandomized benchmarking

Summary

The presentation by Sangyeon Lee from SQRT, given at the QISCA Journal Club, describes experimental work on drive crosstalk cancellation in a Surface-43 superconducting qubit chip. The speaker spent two months at ETH Zurich in Andreas Wallraff’s group, using the device to implement lattice surgery between distance-3 surface codes. The main focus is on mitigating drive crosstalk, which arises due to the 2D integration of control lines. The method involves applying a compensation drive to the spectator qubit with optimized relative amplitude and phase. The optimization process uses a cost function based on the epsilon_22 element of the quantum channel, which is extracted from Ramsey-like experiments. The speaker found that crosstalk was particularly severe between qubits 23 and 8, likely due to a fixed-frequency coupler crossing the drive line via an airbridge. Verification via randomized benchmarking showed that the error rate per Clifford gate improved from 1.7% to 0.13% with compensation, and leakage was significantly reduced. The talk includes a Q&A session clarifying the definition of crosstalk, the role of geometry, and the hardware-dependent nature of the noise.

178 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information is high as it provides real experimental data on a specific quantum computing challenge, which is often scarce. The argumentation is logical and methodical: the speaker clearly defines the problem, explains the theoretical basis for the compensation method, and presents a systematic optimization procedure. The use of a cost function based on epsilon_22 is well-justified, and the verification via randomized benchmarking strengthens the claims. However, the speaker acknowledges that the optimization was not exhaustive and that the results are preliminary, which tempers the strength of the conclusions.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is moderate: the work is presented as a journal club talk, not a peer-reviewed publication. The speaker does not cite specific references, but the methodology is standard in the field. The title accurately reflects the content. The speaker is transparent about limitations, such as not verifying the root cause of the crosstalk and not fully optimizing the compensation parameters. No comments were provided for analysis.

175 words

Title / Content Match

The title accurately reflects the content, focusing on drive crosstalk cancellation on a Surface-43 chip.

Quality & Reliability

7/10

The presentation is a first-hand account of experimental work on a real device, with detailed methodology and results. However, it is a journal club talk, not peer-reviewed, and the speaker acknowledges limitations and lack of full verification.

Key Moments

Cited Sources

  • No explicit sources cited in the video or description — The presentation did not mention specific references, and the description only includes hashtags.

Concurring Sources

  • No concordant sources provided — No external sources were mentioned in the video.

Dissenting Sources

  • No discordant sources provided — No external sources were mentioned in the video.

Contribution & Novelties

The presentation provides original experimental data on drive crosstalk cancellation in a large-scale superconducting qubit chip, which is valuable for the quantum computing community. The systematic optimization method and the use of randomized benchmarking for verification contribute to the field. The finding that crosstalk is particularly severe due to specific layout features offers insights for future chip design.

Pour aller plus loin :

89 words

Radar Profile

The radar profile shows high scores in quantity and technical level, reflecting the detailed experimental content. Quality and reliability are moderate, consistent with a preliminary presentation. The overall shape suggests a technically strong but not fully verified work.

Reliability 6/10