Electrically detected magnetic resonance of $^{75}$As magnetic clock transitions in silicon

This paper demonstrates the observation of magnetic clock transitions in near-surface $^{75}$As spins in silicon using low-field continuous-wave electrically detected magnetic resonance (EDMR), establishing the technique as a sensitive method for studying decoherence suppression in silicon-based quantum devices.

The Gist

The Quantum "Sweet Spot": Finding Silence in a Noisy World

Imagine you are trying to listen to a delicate, beautiful violin solo being played in the middle of a roaring construction site. The jackhammers, the shouting, and the heavy machinery are the "noise" (in physics, we call this decoherence). This noise is the enemy of quantum computers; it shakes the tiny quantum particles (qubits) so much that they lose their information, much like the violin music gets drowned out by the construction.

Scientists are currently trying to build quantum computers using tiny impurities in silicon—specifically, an element called Arsenic. These arsenic atoms act like little spinning tops (spins) that can hold quantum information. But there is a problem: the magnetic environment around these atoms is incredibly noisy, making it hard to "hear" the quantum information clearly.

This paper describes a clever way to find a "Sweet Spot" where the noise almost disappears.

---

1. The Magic of the "Clock Transition"

The researchers discovered something called a Magnetic Clock Transition (CT).

Think of it like this: Imagine you are walking on a hilly landscape. Usually, if you take a step in any direction, your altitude changes. If "altitude" represents the frequency of your quantum signal, even a tiny stumble (a magnetic fluctuation) changes your altitude and ruins your measurement.

However, a Clock Transition is like finding the exact, perfectly flat peak of a hill or the very bottom of a valley. If you stand exactly on that peak, you can wobble slightly left or right, but your altitude stays exactly the same. Because the "altitude" (the frequency) doesn't change when the environment wobbles, the signal becomes incredibly stable. It’s a "sweet spot" where the particle becomes temporarily deaf to the magnetic noise around it.

2. How They "Listened" (EDMR)

The researchers didn't use traditional, bulky equipment to listen to these spins. Instead, they used a technique called EDMR (Electrically Detected Magnetic Resonance).

Imagine instead of using a microphone to hear the violin, you built a tiny, sensitive light bulb that only glows when the violin plays a specific note. By measuring the tiny changes in electricity flowing through a silicon chip, they could "hear" the arsenic atoms spinning. This is much more practical for future quantum computers, which will be built on tiny, integrated silicon chips rather than massive laboratory setups.

3. The "Broadening" Paradox

The paper mentions something strange: as they approached this "sweet spot," the signal actually looked messier and wider (this is called linewidth broadening).

This sounds counterintuitive. If the sweet spot is so good, why does it look messy?

Think of it like a camera lens. When you are focusing on a tiny object, if you move the lens just a fraction of a millimeter too far, the image suddenly goes from sharp to a blurry smear. The researchers realized that the "blurriness" they saw wasn't because the quantum information was being destroyed, but because of the way their measuring tool (the electrical signal) translates frequency into magnetic field readings. The "blur" was actually a mathematical side effect of approaching that perfect, flat peak of the hill.

Why does this matter?

To build a scalable quantum computer, we need to move from "laboratory curiosities" to "real-world chips." This paper proves that:

1. We can find these "Sweet Spots" in arsenic-doped silicon.
2. We can detect them electrically, which means we can do it inside a tiny, manufactured device.
3. We can use "qudits" (more complex versions of qubits) to store even more information.

By finding these quiet zones in a noisy world, scientists are one step closer to building a stable, reliable quantum computer that can actually perform complex calculations without being interrupted by the "construction noise" of the universe.

Technical Summary

Technical Summary: Electrically Detected Magnetic Resonance of $^{75}$As Magnetic Clock Transitions in Silicon

Problem: Decoherence in Silicon-Based Quantum Systems

A primary obstacle in developing scalable silicon-based quantum information processors is decoherence—the loss of quantum information due to environmental noise. In donor-spin systems, this noise typically arises from magnetic field fluctuations (from the $^{29}$Si nuclear spin bath or interface defects) and electric field fluctuations. While isotopic enrichment ($^{28}$Si) and dynamical decoupling are standard mitigation strategies, a more intrinsic approach involves operating at Clock Transitions (CTs). CTs are specific magnetic field values where the transition frequency is first-order insensitive to magnetic field fluctuations ($df/dB_0 = 0$), theoretically providing a "protected" operating point for qubits. While CTs have been studied via Electron Spin Resonance (ESR), their behavior in Electrically Detected Magnetic Resonance (EDMR)—which is more compatible with integrated semiconductor devices—remains poorly understood, especially in near-surface environments.

Methodology

The researchers investigated near-surface $^{75}$As (arsenic) donor spins in an isotopically enriched $^{28}$Si substrate.

1. Device Fabrication: The team fabricated an EDMR device using a combination of thermal oxide patterning, $^{75}$As drive-in diffusion from a spin-on glass (SOG) source, and $^{31}$P ion implantation. The $^{75}$As donors were intentionally introduced via lateral diffusion from the source-drain contacts into the active region.
2. Detection Mechanism: The study utilized Spin-Dependent Recombination (SDR). Under 633 nm laser illumination, electron-hole pairs are generated. The recombination rate between $^{75}$As donor electrons and paramagnetic $P_b^0$ interface defects is spin-selective: parallel spin pairs are long-lived, while antiparallel pairs recombine rapidly. Resonant RF excitation flips spins, altering the recombination rate and thus the measurable source-drain current.
3. Spectroscopy: Low-field ($< 10$ mT) continuous-wave EDMR was performed at 5 K. The researchers used RF Frequency Modulation (RF-FM) lock-in detection, which maps frequency modulation into an effective magnetic field modulation.
4. Modeling: The transition frequencies and strengths were modeled using the substitutional $^{75}$As Hamiltonian, accounting for the electron Zeeman interaction, nuclear Zeeman interaction, and isotropic hyperfine coupling ($A \approx 198.4$ MHz).

Key Contributions

• Observation of CTs via EDMR: The paper provides the first observation of magnetic CTs for $^{75}$As donors using EDMR in a near-surface ensemble.
• Linewidth Broadening Model: The authors developed and applied a phenomenological model to explain the anomalous linewidth behavior observed near the CT points, accounting for magnetic ($\Delta B_0$), hyperfine ($\Delta A$), and modulation-induced ($\Delta B_{\text{mod}}$) broadening.
• Characterization of Near-Surface Environments: The study successfully linked spectral broadening to local environmental factors, such as strain and magnetic disorder, inherent to the Si/SiO$_2$ interface.

Results

• Identification of CTs: Two closely spaced magnetic CTs (labeled C1 and C2) were identified at approximately 3.8 mT with a resonant frequency of $\sim$383 MHz.
• Phase Inversion: Consistent with the $df/dB_0 = 0$ condition, the researchers observed a characteristic phase inversion in the lock-in signal as the magnetic field swept through the CT turning point.
• Linewidth Divergence: As the CT condition is approached, the peak-to-peak linewidth ($\Delta B_{pp}$) increases significantly. This is not a loss of coherence, but rather a mathematical consequence of the mapping between frequency and magnetic field: as the slope $df/dB_0$ approaches zero, small frequency variations (due to hyperfine disorder) translate into large apparent magnetic field widths.
• Quantification of Disorder: By fitting the linewidth data, the authors extracted:
  • Hyperfine variation ($\Delta A$): $(0.26 \pm 0.05)$ MHz, corresponding to a local strain of $\epsilon \sim 7 \times 10^{-5}$.
  • Magnetic disorder ($\Delta B_0$): $(0.10 \pm 0.02)$ mT.

Significance

This work establishes low-field EDMR as a powerful and sensitive probe for studying clock transitions in near-surface donor systems. Because EDMR can be performed on fully fabricated devices, these findings are directly relevant to the engineering of silicon-based quantum architectures. The ability to identify and characterize these "protected" operating points in a realistic, device-ready environment provides a roadmap for implementing high-coherence qudits (using the higher-dimensional Hilbert space of $^{75}$As) in scalable quantum technologies.