[Paper Review] Probing charge noise in few electron CMOS quantum dots

Cameron Spence, Bruna Cardoso Paz+15 more|arXiv (Cornell University)|Sep 5, 2022

Quantum and electron transport phenomena4 citations

TL;DR

This study probes low-frequency charge noise in CMOS-fabricated silicon nanowire quantum dots using a novel single-shot tunneling readout method, demonstrating that charge noise can be tuned over two orders of magnitude (1–100 μeV²/Hz) by controlling quantum dot geometry and electron occupancy. The key finding is a strong reduction in charge noise with increasing electron number, indicating that the few-electron regime offers improved noise resilience for spin qubits.

ABSTRACT

Charge noise is one of the main sources of environmental decoherence for spin qubits in silicon, presenting a major obstacle in the path towards highly scalable and reproducible qubit fabrication. Here we demonstrate in-depth characterization of the charge noise environment experienced by a quantum dot in a CMOS-fabricated silicon nanowire. We probe the charge noise for different quantum dot configurations, finding that it is possible to tune the charge noise over two orders of magnitude, ranging from 1 ueV^2 to 100 ueV^2. In particular, we show that the top interface and the reservoirs are the main sources of charge noise and their effect can be mitigated by controlling the quantum dot extension. Additionally, we demonstrate a novel method for the measurement of the charge noise experienced by a quantum dot in the few electron regime. We measure a comparatively higher charge noise value of 40 ueV^2 at the first electron, and demonstrate that the charge noise is highly dependent on the electron occupancy of the quantum dot.

Motivation & Objective

• To characterize low-frequency charge noise in few-electron CMOS quantum dots fabricated on a 300 mm wafer using industrial-compatible processes.
• To identify the dominant sources of charge noise—particularly at interfaces and reservoirs—and evaluate their tunability via gate control.
• To develop and validate a simple, experimentally accessible method for measuring charge noise power spectral density at the single-electron level.
• To investigate the dependence of charge noise on electron occupancy in the quantum dot, especially in the few-electron regime relevant for spin qubits.
• To disentangle chemical potential noise from tunnel coupling fluctuations using time-resolved charge detection.

Proposed method

• Employed a CMOS-fabricated silicon nanowire device with electrostatic gates and a charge detector to monitor tunneling events in real time.
• Used single-shot tunneling measurements to extract time-resolved electron occupancy dynamics, enabling reconstruction of the power spectral density (PSD) of charge fluctuations.
• Applied Fourier transformation to time traces of current fluctuations during Coulomb blockade peaks to obtain the noise PSD at each gate voltage.
• Renormalized the measured PSD using the derivative of average conductance with respect to gate voltage (d⟨N⟩/dV_G) to isolate chemical potential noise.
• Compared results from tunneling rate extraction with those from direct current fluctuation measurements to assess consistency and identify contributions from tunnel coupling fluctuations.
• Simulated the effect of a single trapped charge (TLS) above the interface on the quantum dot’s chemical potential for different electron numbers to support observed trends.

Experimental results

Research questions

• RQ1 How does the charge noise environment in CMOS quantum dots vary with quantum dot geometry and electron occupancy?
• RQ2 What are the dominant sources of charge noise in CMOS-fabricated silicon nanowire devices, particularly in the few-electron regime?
• RQ3 Can a simple, single-shot tunneling measurement method accurately extract charge noise power spectral density at the single-electron level?
• RQ4 How does the charge noise magnitude change as the number of electrons in the quantum dot increases?
• RQ5 To what extent can charge noise be mitigated through gate control of the quantum dot’s extension and tunnel barriers?

Key findings

• Charge noise in the few-electron regime was measured at 40 μeV²/Hz at the first electron, with values ranging from 1 to 100 μeV²/Hz depending on device configuration.
• The noise level can be tuned over two orders of magnitude by adjusting gate voltages, primarily through control of the quantum dot’s spatial extension.
• The top interface and reservoirs were identified as the primary sources of charge noise, with their impact reducible via improved gate control.
• Charge noise power spectral density decreases significantly with increasing electron number—observed as a drop from ~40 μeV²/Hz at N=1 to ~20 μeV²/Hz at N=2 and lower at higher occupancies.
• The observed noise reduction is attributed to increased distance between fluctuators (e.g., interface traps) and the electron wave function, reducing dipolar coupling.
• The single-shot tunneling method successfully disentangled chemical potential noise from tunnel coupling fluctuations, providing a simpler alternative to complex spectroscopic techniques.