Landau-Level-Resolved Mode Mixing and Shot Noise in Gate-Defined Graphene Quantum Point Contacts

This paper establishes that shot noise measurements in gate-defined graphene quantum point contacts reveal a distinct Landau-level-resolved crossover between a universal chaotic-cavity Fano factor of $F \simeq 1/4$ for higher levels and a unique $F = 1/3$ for the zeroth level, a signature arising from sublattice-polarized single-channel transport that cannot be detected by conductance alone.

The Gist

The Big Picture: A Quantum Highway with a Traffic Jam

Imagine a super-highway made of electricity, where tiny particles called electrons zoom along the edges of a piece of graphene (a material as thin as a sheet of paper but incredibly strong).

Usually, when scientists study these highways, they just count how many cars (electrons) get from point A to point B. This is called conductance. It's like counting the total number of cars passing a toll booth.

But this paper asks a deeper question: How are those cars actually moving? Are they driving smoothly in their own lanes, or are they crashing into each other, changing lanes, and getting mixed up? To find out, the researchers didn't just count the cars; they measured the noise (the "static" or "hiss") created by the traffic.

Think of it this way:

• Conductance is like counting how many people enter a concert hall.
• Shot Noise is like listening to the crowd. If everyone walks in perfectly in a line, it's quiet. If people are shoving, tripping, and bumping into each other at the door, it's loud and chaotic. That "loudness" tells you exactly how the crowd is behaving.

The Setup: The "Split Gate" Funnel

The researchers built a special device called a Quantum Point Contact (QPC). Imagine a wide highway that suddenly narrows down into a tiny, single-lane tunnel.

They used magnets to force the electrons into specific "lanes" called Landau Levels.

• Higher Landau Levels: Think of these as a busy highway with many lanes (like 4 or 8 lanes).
• The Zeroth Landau Level: This is a special, unique lane that only exists in graphene. It's like a "ghost lane" that behaves differently from all the others.

They also added a "split gate" (a tiny metal fence) over the tunnel to control how many lanes are open.

The Discovery: Two Different Types of Chaos

The team used a mix of computer simulations and real experiments to see what happens when electrons try to squeeze through this tunnel. They found two completely different "universes" of behavior, depending on which lane the electrons were in.

1. The "Chaotic Ballroom" (Higher Landau Levels)

When electrons are in the higher lanes (the busy multi-lane highway), they hit the tunnel and get thrown into a chaotic dance.

• The Analogy: Imagine a huge ballroom with 100 people trying to get through a single door. They push, shove, and mix with everyone else. No one knows who will get through and who will be pushed back.
• The Result: Because everyone is mixing perfectly, the "noise" (the Fano factor) settles at a specific, universal number: 1/4 (0.25).
• Why it matters: This is the standard "chaotic" behavior you expect when many things mix together.

2. The "Solo Dancer" (The Zeroth Landau Level)

Here is the magic part. When the electrons are in the zeroth lane (the special ghost lane), something weird happens. Even though there is a chaotic environment, the electrons act like they are in a single-lane tunnel.

• The Analogy: Imagine the same ballroom, but this time, the "ghost lane" electrons are wearing special shoes that make them invisible to the crowd. They don't mix with the other 99 people. They glide through the chaos as a single, unified stream.
• The Result: Because they are effectively just one single channel mixing chaotically with itself, the noise settles at a different number: 1/3 (0.33).
• Why it matters: This is a unique signature of graphene. It proves that this special lane is fundamentally different from all the others.

The "Aha!" Moment: Why Noise is Better Than Counting

If you just looked at the number of electrons getting through (conductance), you might think both situations look the same. You'd see a mix of electrons going through and bouncing back.

But by listening to the noise, the researchers could hear the difference:

• Noise Level 0.25 = "We have a huge crowd mixing together." (Multi-lane chaos)
• Noise Level 0.33 = "We have a single, special stream moving through the chaos." (Single-lane chaos)

This is like being able to tell the difference between a herd of cows and a single horse just by listening to the sound of their footsteps, even if they are walking at the same speed.

Why Should You Care?

1. It's a New Way to Test Materials: This gives scientists a new "ruler" to measure how well graphene works. If you see the noise jump from 0.25 to 0.33, you know you've successfully isolated that special zeroth lane.
2. It's Unique to Graphene: You can't do this with regular silicon chips. Graphene's special physics (relativistic electrons) creates this unique "ghost lane" that behaves differently than anything else in the world.
3. Future Tech: Understanding how electrons mix (or don't mix) is crucial for building future quantum computers. If we can control this "noise," we might be able to build faster, more reliable quantum devices.

Summary

The paper shows that in graphene, electrons in the "zeroth" lane act like a solo dancer in a chaotic ballroom, while electrons in higher lanes act like a massive crowd. By measuring the "noise" of the traffic, scientists can tell these two groups apart, revealing a hidden layer of physics that simple counting would miss.

Technical Summary

1. Problem Statement

Graphene Quantum Point Contacts (QPCs) in the quantum Hall regime exhibit complex transport phenomena involving chiral edge propagation, valley degeneracy, and gate-induced mode mixing. While conductance measurements reveal the average transmission of edge channels, they fail to capture the microscopic statistics of transmission eigenvalues or the specific mechanisms of mode partitioning (mixing vs. filtering).

• The Gap: It remains unclear how the unique sublattice polarization of the zeroth Landau level ($N_L=0$) in graphene affects noise statistics compared to higher Landau levels ($N_L > 0$). Specifically, does the $N_L=0$ state realize a distinct transport universality class within the same device?
• The Challenge: Existing models often rely on simplified potentials or average transmission. A framework is needed that combines realistic device electrostatics with statistical universality to predict shot noise signatures across different Landau level regimes.

2. Methodology

The authors employ a hybrid theoretical framework combining microscopic device simulation with statistical random matrix theory (RMT).

• Experimental Validation:

  • Device: hBN-encapsulated graphene Hall bars with a split-gate QPC (100 nm width) and a graphite back gate.
  • Technique: Low-frequency differential lock-in measurements to map conductance, transmission ($N$), and reflection ($M-N$) as functions of back-gate ($V_{bg}$) and top-gate ($V_{tg}$) voltages.
  • Analysis: Extraction of mode numbers using the Landauer-Büttiker formalism from multi-terminal voltage measurements.

• Microscopic Modeling (Tight-Binding):

  • Simulation: Used the kwant package to simulate gate-defined graphene QPCs.
  • Electrostatics: Mapped realistic gate potentials (top and back) to a spatially varying onsite potential $V(x,y)$ based on carrier density and Dirac dispersion, rather than using idealized rectangular or saddle-point potentials.
  • Physics: Included a perpendicular magnetic field via Peierls substitution, spin/valley degeneracy, and phenomenological disorder (random onsite potential) to model charge puddles and substrate inhomogeneity.
  • Output: Computed transmission eigenvalues, shot noise, and current density distributions.

• Statistical Analysis (Random Matrix Theory - RMT):

  • Used RMT to interpret the numerical transmission eigenvalue distributions.
  • Modeled the QPC as a chaotic cavity coupled to leads, analyzing the universality classes for single-channel ($N=1$) vs. multi-channel ($N \gg 1$) mixing.

3. Key Contributions

1. Identification of Three Transport Regimes: The study maps the QPC behavior into three distinct microscopic regimes based on the interplay between bulk filling ($\nu_{bg}$) and local filling ($\nu_{QPC}$):

   • Adiabatic Propagation: When $\nu_{bg} \approx \nu_{QPC}$, edge channels propagate without scattering.
   • Sharp Mode Filtering: When $\nu_{bg} > \nu_{QPC}$, the QPC acts as a barrier, transmitting a fixed number of modes and reflecting the rest.
   • Multi-Mode Mixing: When $\nu_{QPC} > \nu_{bg}$ (particularly in the unipolar regime), localized states beneath the split gate induce strong mixing between propagating edge channels and localized states, leading to partial transmission/reflection.

2. Landau-Level-Resolved Noise Signatures: The paper establishes that shot noise (specifically the Fano factor) distinguishes between transport universality classes that are indistinguishable in conductance.

3. Microscopic Origin of $N_L=0$ Anomaly: The authors demonstrate that the unique $F=1/3$ limit for the zeroth Landau level arises not from diffusive transport (as in zero-field graphene), but from sublattice polarization. This polarization suppresses coupling to mixed-sublattice localized states, effectively confining transport to a single chaotic channel ($N=1$).

4. Key Results

• Conductance Maps: The simulations successfully reproduced experimental conductance maps, showing quantized plateaus and transitions. The maps revealed that in the bipolar regime (opposite gate polarities), transport is governed by resonant tunneling via localized states, while the unipolar regime with high local density leads to significant mode mixing.
• Fano Factor Crossover:
  • Higher Landau Levels ($N_L > 0$): As the top-gate voltage increases (inducing mixing), the Fano factor ($F$) saturates at $F \simeq 1/4$. This corresponds to the universal limit for a chaotic cavity with many channels ($N \gg 1$), where the transmission eigenvalue distribution is bimodal (peaked at $T \to 0$ and $T \to 1$).
  • Zeroth Landau Level ($N_L = 0$): The Fano factor saturates at $F \simeq 1/3$. This value is derived from a single-channel ($N=1$) chaotic mixing scenario where the transmission eigenvalue distribution is flat ($\rho(T)=1$).
• Mechanism Distinction: The $F=1/3$ in this quantum Hall regime is mechanistically distinct from the pseudo-diffusive $F=1/3$ seen in zero-field graphene. The former is caused by sublattice polarization confining transport to one channel; the latter is caused by evanescent modes in a diffusive bulk.
• Current Density Visualization: Simulations of current density confirmed that in the mixing regime, edge trajectories branch and intertwine with localized states, whereas in the filtering regime, paths remain spatially separated.

5. Significance

• New Diagnostic Tool: The crossover from $F=1/3$ to $F=1/4$ serves as a direct, quantitative discriminator between single-channel and multi-channel chaotic transport in graphene QPCs. This signature is absent in conductance measurements.
• Universality Classes: It demonstrates that a single mesoscopic device can host two distinct transport universality classes simultaneously, selectable simply by tuning the Landau level index. This is a unique feature of graphene's relativistic band structure.
• Experimental Benchmark: The paper provides robust theoretical benchmarks ($F \approx 0.33$ vs. $F \approx 0.25$) for future noise spectroscopy experiments. Deviations from these values would indicate incomplete mixing, residual coherence, or inelastic relaxation.
• General Framework: The hybrid approach (realistic electrostatics + RMT) offers a general strategy for analyzing noise in gate-defined 2D conductors, applicable to bilayer graphene, moiré systems, and topological edge channels.

In conclusion, this work resolves the microscopic origin of shot noise in graphene QPCs, revealing that the sublattice structure of the zeroth Landau level fundamentally alters the statistical nature of transport, creating a unique noise signature that distinguishes it from higher Landau levels.