Enhanced shot noise in graphene quantum point contacts with electrostatic reconstruction

M.o. Garg, Olivier Maillet, Noah L. Samuelson, Tianle Wang, Juanjuan Feng, Liam Cohen, Aifei Zhang, Kenji Watanabe, Takashi Taniguchi, P. Roulleau, Maura Sassetti, Michael P. Zaletel, Andrea F. Young, Dario Ferraro, P. Roche, François ParmentierarXiv 2025 · DOI · arXiv

TL;DR

Researchers have discovered that graphene quantum point contacts (QPCs) exhibit significantly higher shot noise than standard theory predicts. This "noise boost" is caused by the self-assembly of a quantum dot at the device's narrowest point, which creates "memory effects" where one electron's passage makes it easier for the next one to follow.

What Problem Does This Paper Address?

In the world of mesoscopic physics, Quantum Point Contacts (QPCs) are the ultimate "valves" for electrons. By narrowing a conducting channel to the scale of an electron's wavelength, physicists can study fundamental phenomena like the quantization of conductance and the fractional charges of anyons.

Ideally, a QPC acts as a simple partitioner: an electron arriving at the "valve" is either transmitted or reflected. This randomness creates "shot noise." Standard theory (the Schottky formula) provides a clear prediction for this noise based on the transmission probability. However, graphene QPCs—despite being cleaner than their gallium arsenide predecessors—often show much higher noise levels than expected. This paper investigates why this happens, pointing to a phenomenon called electrostatic reconstruction.

Methods

The team used a high-quality van der Waals heterostructure: a graphene flake encapsulated in hexagonal boron nitride (hBN) with few-layer graphite gates. This setup minimizes disorder and strain.

  • Experimental Regime: They operated the device in the Integer Quantum Hall (QH) regime at a high magnetic field ($B = 13$ T) and ultra-low temperatures ($T = 100$ mK).
  • Measurement: They measured "excess shot noise"—the current fluctuations beyond thermal noise—as a function of the DC current flowing through the QPC.
  • Modeling: To explain the data, they used Thomas-Fermi calculations to simulate the electron density at the QPC's "saddle point" and developed a semiclassical model to account for correlated tunneling events. They also applied a plasmon model to describe how energy is stored and released within the system.

Datasets

As an experimental physics paper, it does not use standard machine learning datasets. Instead, the "dataset" consists of:

  • Differential conductance maps ($G_{QPC}$ vs. gate voltages and bias).
  • Shot noise spectra measured across two independent amplification lines (auto- and cross-correlations).
  • Calibration data using thermal noise (Johnson-Nyquist noise) to precisely determine the gain of their cryogenic amplifiers.

Key Insights and Ideas

  • The Accidental Quantum Dot: At high magnetic fields, the competition between the gate's confinement and electron-electron interactions causes the electron liquid to "reconstruct." Instead of a smooth transition, a small island—a quantum dot—spontaneously forms at the QPC's narrowest point.
  • Correlated Tunneling (The "Memory Effect"): In a standard QPC, electron transfers are independent (Markovian). Here, the quantum dot introduces a memory. When an electron tunnels through, it can leave the dot in an excited "plasmonic" state.
  • Neutral Mode Excitation: Specifically, the authors suggest that an electron passage excites a "neutral mode" (a dipolar plasmon). If this excitation lives long enough, it can be absorbed by the next electron, significantly increasing its probability of tunneling.
  • Super-Poissonian Noise: This "bunching" of electrons leads to noise levels that are 50% to 100% higher than the standard partition noise, characterized by an effective Fano factor ($\alpha$) between 1.5 and 2.

Key Results

  • Enhanced Fano Factor: The measured shot noise consistently exceeded the standard prediction ($F = \tau(1-\tau)$). By introducing a scaling prefactor $\alpha$, they found that $\alpha \approx 1.5$ to $2$ provided the best fit across different edge channels.
  • Coulomb Diamonds: Conductance maps showed clear "Coulomb diamond" features, a hallmark of quantum dots, confirming that the QPC was not a simple barrier but a resonant tunneling system.
  • Fractional Island Formation: Thomas-Fermi simulations confirmed that the local filling factor at the saddle point likely jumps to a fractional value (like $\nu = 1 + 2/3$), creating the circulating edge channels necessary for these correlations to occur.

Limitations and Future Work

The authors acknowledge that their semiclassical model is a simplification; it captures the "bunching" physics but ignores the specific microscopic details of the edge channel structure within the dot.

The presence of this "parasitic" quantum dot is a double-edged sword. While it provides a fascinating look at many-body physics, it can be "detrimental" to experiments aiming to measure the fundamental properties of anyons, as the dot's internal dynamics can mask the intrinsic noise signatures of the fractional quantum Hall effect. Future work will likely focus on how to either suppress this reconstruction or use it as a more controlled probe of edge excitations.

Summary updated

Abstract

Shot noise measurements in quantum point contacts are a powerful tool to investigate charge transport in the integer and fractional quantum Hall regime, in particular to unveil the charge, quantum statistics and tunneling dynamics of edge excitations. In this letter, we describe shot noise measurements in a graphene quantum point contact in the quantum Hall regime. At large magnetic field, the competition between confinement and electronic interactions gives rise to a quantum dot located at the saddle point of the quantum point contact. We show that the presence of this quantum dot leads to a $50-100~\%$ increase in the shot noise, which we attribute to correlated charge tunneling. Our results highlight the role played by the electrostatic environment in those graphene devices.
BibTeX
@misc{garg2025enhanced,
  author = {M.O. Garg and Olivier Maillet and Noah L. Samuelson and Tianle Wang and Juanjuan Feng and Liam Cohen and Aifei Zhang and Kenji Watanabe and Takashi Taniguchi and P. Roulleau and Maura Sassetti and Michael P. Zaletel and Andrea F. Young and Dario Ferraro and P. Roche and François Parmentier},
  title = {{Enhanced shot noise in graphene quantum point contacts with electrostatic reconstruction}},
  year = {2025},
  eprint = {2503.17209},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2503.17209},
}

Loading related content...

Figure