Optimal control over the full counting statistics in a non-adiabatic pump

This paper introduces a systematic optimal control procedure to enhance the performance of non-adiabatic Thouless pumps by simultaneously optimizing average transport rates and minimizing noise, thereby enabling independent control over charge and spin currents and their fluctuations in quantum dot systems.

The Gist

Imagine you are trying to move a crowd of people from one room to another using a revolving door. In the world of quantum physics, this "crowd" is made of tiny particles like electrons, and the "revolving door" is a machine called a Thouless pump.

For a long time, scientists knew how to move these particles perfectly, but only if they moved the door very, very slowly. This is called the "adiabatic" limit. If you tried to spin the door faster to get more people through in less time, the system would get chaotic. People would get bumped, the flow would become noisy, and the efficiency would crash.

Recently, researchers tried a "shortcut" method to spin the door faster without the chaos. They added a special counter-force to keep everyone in line. While this worked well for moving the average number of people, it failed to stop the noise. The crowd still jostled and bumped into each other, creating a lot of "static" or fluctuations. This is a problem if you need a perfectly smooth flow, like in high-precision measuring tools.

The New Solution: The "Traffic Controller"

The authors of this paper have introduced a new, smarter way to control this pump using a mathematical tool called Optimal Control Theory. Think of this not just as a speed knob, but as a sophisticated traffic controller that manages the entire flow in real-time.

Here is how their method works, using simple analogies:

1. The "Shadow" System

Usually, scientists only track where the particles are. This new method tracks two things at once:

• The Real Crowd: Where the particles actually are.
• The "Shadow" Crowd: A mathematical ghost version that tracks how much the real crowd is jiggling or fluctuating.

By watching both the real crowd and the shadow crowd simultaneously, the system can adjust the "revolving door" (the pumping rates) to not only move the particles but also to smooth out the bumps and jostles.

2. The Two-Stage Test

The researchers tested this on two different scenarios:

• Scenario A: The Simple Queue (Non-interacting particles)
  Imagine a single line of people where everyone ignores each other. The researchers showed that their new method could spin the door much faster than before.

  • Result: They moved about 20 times more people per cycle than the old "shortcut" method, while cutting the noise (the jostling) in half. It was like turning a chaotic rush hour into a smooth, high-speed conveyor belt.

• Scenario B: The Complex Crowd (Interacting particles with Spin)
  Now, imagine the crowd has two types of people: "Spin-Up" and "Spin-Down" (like wearing red or blue hats). These people interact with each other, making the flow much harder to control.

  • The Goal: The researchers wanted to move only the "Spin-Up" people (creating a "spin current") while leaving the "Spin-Down" people behind, and doing so without creating any noise.
  • Result: They successfully tuned the machine to create a nearly pure stream of "Spin-Up" particles. They suppressed the movement of the "Spin-Down" particles and the overall charge (total number of people) almost to zero. Most importantly, they kept the flow incredibly smooth, improving the "signal-to-noise" ratio by thousands of times.

3. Why This Matters

The paper claims that this method is a "universal remote control" for these quantum systems.

• Independence: You can now control the amount of flow and the smoothness of the flow independently. You can choose to have a lot of flow with low noise, or a specific type of flow (like only spin) with almost no charge.
• Speed: It works even when the system is being driven very fast (non-adiabatic), a regime where previous methods failed or produced unphysical results.
• Versatility: While they tested this on a specific quantum dot model, the math suggests it can be applied to any system where particles move randomly, including heat transfer and other stochastic (random) processes.

In Summary
The authors have built a mathematical "autopilot" for quantum pumps. Instead of just trying to push particles through as fast as possible, this autopilot calculates the perfect, smooth path to move them, ensuring that you get exactly the right amount of flow with the least amount of chaos, even when operating at high speeds. This allows for precise control over both the movement of charge and the movement of spin, which is a significant step forward for future technologies like spintronics.

Technical Summary

Technical Summary: Optimal Control over the Full Counting Statistics in a Non-Adiabatic Pump

Problem Statement
Thouless pumping enables minimal dissipation charge and spin transfer in mesoscopic systems by cyclically modulating system-reservoir couplings. However, a critical trade-off exists between pumping speed and adiabaticity. Operating beyond the adiabatic limit (non-adiabatic regime) leads to deviations from the stationary state, causing a sharp decline in the geometric pumping rate. While Shortcut-to-Adiabaticity (STA) techniques, such as counterdiabatic driving, have been applied to maintain average particle transfer at high frequencies, they suffer from significant limitations:

1. They often require controlling all transition rates, which is challenging when only a subset is accessible.
2. At sufficiently high frequencies, the required time-dependent rates for counterdiabatic drives can become negative (unphysical).
3. Crucially, while STA preserves the average number of transferred particles, it fails to control the fluctuations (noise) associated with the transfer. This lack of control over Full Counting Statistics (FCS) limits applications in metrology and spintronics.

Methodology
The authors introduce a systematic procedure based on Optimal Control Theory (OCT), specifically utilizing the Pontryagin Maximum Principle (PMP), to simultaneously optimize the average flux and the counting statistics of particle transfer.

• FCS Formalism: The system is modeled as a stochastic system coupled to reservoirs, described by a master equation. The dynamics are extended to include a "counting field" $\chi$ to track particle number moments. The characteristic function $\phi(\chi, t)$ is derived from a modified Lagrangian $L(\chi, t)$, where tunneling rates to a specific reservoir are weighted by $e^{\pm i\chi}$.
• Effective State Space: To control both the average current and the noise (variance), the authors define an effective state vector $|p, \dot{p}\rangle$ that combines the system's probability vector $|p(t)\rangle$ and its derivative with respect to the counting field. This expands the dynamics into a higher-dimensional space governed by a block-matrix operator $L_4$ (for a two-state system) or $L_9$ (for an interacting spin system).
• Optimization Framework: The control problem is formulated to minimize a cost function $C[u(t)] = \int_0^T (w_1 i(t) + w_2 s(t)) dt$, where $i(t)$ is the current and $s(t)$ is the noise current. The weights $w_1$ and $w_2$ allow for independent tuning of the direction of current flow and the suppression of fluctuations.
• Implementation: The optimal control vector $u(t)$ (the time-dependent tunneling rates or physical parameters like voltages and magnetic fields) is found numerically using an iterative gradient descent routine on a discrete time grid. The method ensures the positivity and periodicity of the rates by parameterizing them appropriately (e.g., using squared functions or sinusoidal increments).

Key Contributions and Results

1. Non-Adiabatic Two-State System (Vanishing Coulomb Interaction):

   • The method was applied to a single-resonance Quantum Dot (QD) where spin-up and spin-down electrons are treated independently.
   • At a high non-adiabatic frequency ($\Omega = 10\Gamma_0$), the optimized protocol achieved an average particle transfer $\langle N \rangle_T = 0.22$ and reduced the variance $\Delta N^2_T$ to $0.23$.
   • Compared to standard STA techniques, the optimized protocol increased the average transfer by over an order of magnitude while reducing noise by approximately 50%.

2. Interacting Regime (Coulomb Blockade and Spin Transport):

   • The approach was extended to a QD in the Coulomb blockade regime, where transition rates depend on physical parameters like gate voltages ($V_L, V_R$) and Zeeman splitting ($\epsilon_{kZ}$).
   • The authors demonstrated the ability to generate a nearly pure spin current while suppressing charge current and minimizing spin fluctuations.
   • Results: The optimized cycle yielded a dimensionless spin transfer $\langle \tilde{S} \rangle_T \approx 1.55$ with negligible charge transfer ($\langle N \rangle_T \approx 0.02$). The signal-to-noise ratio for spin transfer improved by several orders of magnitude compared to the initial unoptimized cycle.
   • The method successfully decoupled spin and charge transfer statistics, allowing for the independent tuning of spin and charge fluctuations.

Significance and Claims
The paper claims that this systematic procedure offers unprecedented independent control over charge and spin fluctuations in the non-adiabatic regime. Unlike STA methods, which are limited in frequency range and fail to address noise, this OCT-based approach:

• Maintains high-fidelity quantized currents over a wider frequency range.
• Actively mitigates current fluctuations, a crucial requirement for metrological applications.
• Is applicable to a broad class of stochastic systems where only specific physical parameters (rather than all abstract transition rates) are controllable.

The authors conclude that this method is particularly relevant for high-accuracy pumping and can be extended to control higher-order moments of the distribution. They suggest its potential applicability to various mesoscopic phenomena, including charge transport in nanostructures, stochastic thermodynamics, heat transfer, and Brownian motors, without inventing specific experimental proposals beyond the QD models analyzed.