🔬 mesoscale physics

Large spin shuttling oscillations enabling high-fidelity single qubit gates

This paper demonstrates that mobile electron architectures utilizing large spin shuttling oscillations can achieve high-fidelity single-qubit gates with superior Rabi frequencies and reduced charge noise compared to static methods, provided that spin-valley physics bottlenecks are mitigated through quantum optimal control.

Original authors: Akshay Menon Pazhedath, Alessandro David, Max Oberländer, Matthias M. Müller, Tommaso Calarco, Hendrik Bluhm, Felix Motzoi

Published 2026-09-16

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Akshay Menon Pazhedath, Alessandro David, Max Oberländer, Matthias M. Müller, Tommaso Calarco, Hendrik Bluhm, Felix Motzoi

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quest to build a quantum computer, scientists are racing to create tiny, reliable switches that can hold and process information in ways classical machines cannot. One promising path involves using the spin of a single electron trapped inside a semiconductor chip, much like a tiny magnet that can point up or down to represent a zero or a one. These electron spins are incredibly stable, holding their state for long periods, but getting them to perform calculations requires precise control. The challenge is that as we try to connect millions of these tiny switches to build a powerful machine, the wiring becomes a nightmare. If every switch needs its own dedicated wire, the chip becomes too crowded to function. To solve this, researchers have proposed a "conveyor belt" approach, where the electron itself is physically moved across the chip to different locations to perform its tasks, rather than trying to reach it with a static wire.

However, moving an electron introduces new problems. As the electron travels, it encounters a complex landscape of energy levels that can confuse its magnetic orientation, causing the information to scramble and disappear. This is particularly true in silicon-based chips, where the electron can get stuck in a "valley" state—a specific energy configuration that interferes with the intended operation. A new study by Akshay Menon Pazhedath and colleagues at research institutes in Germany and Italy explores whether moving the electron quickly and over larger distances can actually help solve these problems, or if it makes them worse. They investigated a method where the electron is shuttled back and forth under a tiny magnet to make it spin, a process known as electric-dipole spin resonance. While previous attempts kept the electron moving only a tiny fraction of a nanometer, this team simulated moving it much further, up to twenty nanometers, to see if the increased motion could be turned into an advantage.

The researchers found that moving the electron over these larger distances allows them to use much weaker magnetic fields to control the spin, which in turn reduces the noise that usually destroys the quantum information. In their simulations, they modeled two different ways the atomic structure of the silicon chip might be imperfect: one where stray atoms of germanium are scattered inside the silicon, and another where the layers of the chip have tiny, jagged steps from the manufacturing process. Both of these imperfections create "valleys" in the energy landscape that can trap the electron and ruin the calculation. The team discovered that while these traps are indeed a major obstacle, they are not a dead end. By using advanced mathematical techniques to carefully shape the path the electron takes—speeding it up, slowing it down, or changing its rhythm at just the right moments—they could guide the electron through these tricky zones without losing the information.

The results of these simulations were striking. When the electron was moved using standard, unadjusted pulses, the error rate in the calculation was high, especially when the electron passed through areas where the energy valleys were very low. However, when the researchers applied their optimized control pulses, the error rate dropped dramatically, becoming more than a thousand times smaller. This improvement held true even when they simulated one thousand different versions of the chip, each with a unique pattern of imperfections. The study suggests that the "conveyor belt" method is not only viable but potentially superior to static methods, provided the movement is carefully choreographed. The team showed that even in the worst-case scenarios, where the energy traps are deepest and the interference is strongest, the optimized movement could still achieve gate fidelities high enough for practical use.

This work does not claim to have built a working machine yet; it is a detailed theoretical demonstration of how such a machine could work. The authors emphasize that the success of this approach relies on the ability to precisely control the electron's trajectory in real-time. They also noted that while their simulations focused on specific types of atomic imperfections, the method is robust enough to handle a wide variety of unknown defects that might exist in a real device. By proving that the electron's motion can be used to suppress errors rather than create them, the study offers a clear path forward for scaling up quantum computers. It suggests that the future of silicon-based quantum computing may depend not on keeping the electron still, but on teaching it how to move with precision, turning the very act of transportation into a tool for maintaining perfect control.

Technical Summary: Large Spin Shuttling Oscillations Enabling High-Fidelity Single Qubit Gates

Problem Statement
Semiconductor quantum dots, particularly in Si/SiGe heterostructures, have achieved high single- and two-qubit gate fidelities. However, scalability is hindered by the short range of exchange interactions, which necessitates complex wiring and leads to crosstalk. Mobile electron architectures, such as the "SpinBus," propose shuttling qubits between static storage zones and dedicated manipulation zones to solve this. In these manipulation zones, single-qubit gates are typically performed via Electric-Dipole Spin Resonance (EDSR) using a micro-magnet to create a magnetic gradient.

The primary challenge addressed in this work is the trade-off between gate speed and coherence. Standard static EDSR implementations use small oscillation amplitudes ( $\sim 1$ nm) to minimize charge noise effects, requiring large magnetic gradients that induce spin dephasing. Conversely, shuttling architectures allow for much larger oscillation amplitudes ( $\sim 10\text{--}20$ nm), which could reduce gradient requirements and improve coherence times. However, large-amplitude shuttling introduces new decoherence mechanisms:

Charge Noise: While reduced gradients help, the larger spatial excursion increases sensitivity to position-dependent noise.
Spin-Valley Physics: In Silicon, the conduction band has low-lying valley states. As the electron shuttles, it encounters spatially varying valley splittings and spin-valley couplings. This leads to position-dependent energy shifts and phase randomization (decoherence), particularly at "Low Valley Splitting Points" (LVSPs) where the valley splitting is small ( $<15$ $\mu$ eV). These LVSPs can severely degrade gate fidelity, potentially rendering manipulation zones unusable.

Methodology
The authors model the shuttling-based EDSR operation using a combined spin-valley Hamiltonian within an open quantum system framework.

Hamiltonian: The system includes Zeeman terms, Rabi driving (via the micro-magnet gradient), valley dynamics (modeled as a two-level system), and a position-dependent spin-valley interaction ( $H_{spin-valley}$ ).
Valley Models: Two distinct microscopic models are employed to simulate realistic valley landscapes along the device:
1. Ge-diffusion model: Accounts for the random distribution of Ge atoms in the Si well.
2. Step model: Accounts for fabrication miscuts at the heterostructure interface.
Dynamics: The system evolution is governed by a master equation incorporating valley relaxation ( $T_{1,v}$ ) and spin dephasing ( $T_{2,s}$ ) due to charge noise.
Optimization: The authors utilize Quantum Optimal Control (specifically the GRAPE algorithm implemented with JAX for auto-differentiation) to shape the electron trajectory $x_{qd}(t)$ . Unlike standard sinusoidal drives, the control pulse is optimized to minimize the average gate infidelity ( $1 - \bar{F}$ ) while tracing out the valley degree of freedom. The target gate is a $\pi$ -rotation about the y-axis.

Key Contributions

Demonstration of Shuttling Advantage: The paper theoretically demonstrates that large-amplitude shuttling ( $x_0 \approx 10\text{--}20$ nm) allows for an order-of-magnitude reduction in the required magnetic gradient compared to static EDSR. This reduction significantly lowers the spin dephasing rate caused by charge noise, even when operating at larger spatial scales.
Optimal Control for Valley Physics: The authors show that while spin-valley coupling and LVSPs create a bottleneck for standard analytical Rabi pulses, Quantum Optimal Control can mitigate these effects. By shaping the electron trajectory, the control sequence can effectively navigate through regions of low valley splitting and strong spin-valley coupling.
Robustness Across Models: The study validates the approach across two competing physical models (Ge-diffusion and Step) and a statistical ensemble of 1000 sampled devices, proving that the benefits are not specific to a single idealized device configuration.

Results

Static vs. Shuttling: Simulations confirm that for a given valley splitting, the larger amplitudes of shuttling allow for lower magnetic gradients, resulting in longer $T_2^*$ times compared to static dots operating at high gradients.
Fidelity Improvement:
- Analytical Pulses: Standard sinusoidal drives result in high infidelity (often $>10^{-2}$ ) when encountering LVSPs or strong spin-valley coupling.
- Parameter Optimization: Tuning the driving frequency ( $\omega$ ) and pulse length ( $T_g$ ) improves fidelity for many cases but fails to overcome the worst-case LVSPs.
- Trajectory Optimization: Using optimal control to shape the trajectory consistently achieves average gate infidelities below $10^{-3}$ (fidelities $>99.9\%$ ). This holds true even for devices with valley splittings as low as a few $\mu$ eV and for the full ensemble of 1000 sampled devices.
Scaling: The optimal control approach improves gate fidelity by at least two to three orders of magnitude compared to analytical formulas, effectively neutralizing the detrimental effects of LVSPs.

Significance and Claims
The paper claims that shuttling-based EDSR is a viable and preferable method for performing single-qubit operations in semiconductor platforms, provided that optimal control techniques are employed.

Scalability: The approach supports the scalability of quantum bus architectures by enabling high-fidelity operations without the need for prohibitively large magnetic gradients.
Resilience: The work asserts that the "drastic measure" of avoiding LVSPs is unnecessary; instead, control theory can render these difficult regions usable.
Practicality: The authors suggest that while the study is model-based, the conclusions likely extend to in-situ closed-loop control, offering a pathway to real-time calibration and high-fidelity gate execution even in the presence of unknown or fluctuating valley landscapes.

The paper concludes that while valley physics and charge noise present significant challenges, they can be overcome through trajectory shaping, making the SpinBus architecture a strong candidate for fault-tolerant quantum computation.

Technical Summary: Large Spin Shuttling Oscillations Enabling High-Fidelity Single Qubit Gates

More like this