Impact of Absorption due to Zero-Field Splitting on Loss in Dielectrics: A Case Study in Sapphire

This paper proposes that transitions between zero-field-split states of paramagnetic impurities (such as Cr, Fe, and V) in sapphire generate magnetic absorption that contributes a loss tangent of $10^{-9}$–$10^{-8}$ at 4.5 GHz, suggesting this mechanism is a significant limiting factor for the coherence times of superconducting qubits.

The Gist

The Big Picture: Why Quantum Computers Are "Shaky"

Imagine you are trying to balance a spinning top on a table. If the table is perfectly smooth and still, the top spins for a long time. But if the table is wobbly, or if there are tiny vibrations shaking the surface, the top falls over quickly.

In the world of quantum computing, the "spinning top" is a qubit (the basic unit of information). Scientists want these qubits to spin (stay in a specific state) for as long as possible so they can do complex calculations. This duration is called coherence time.

The problem? The "table" (the material the qubit sits on) isn't perfect. It has tiny imperfections that absorb energy and make the qubit fall over. For years, scientists thought the main culprit was "Two-Level Systems" (TLS)—tiny defects where atoms get stuck in a weird, jiggly state, like a door that won't quite latch.

This paper suggests we've been looking at the wrong door. The authors propose that the real troublemaker might be magnetic impurities (tiny bits of metal like Chromium, Iron, or Vanadium) hidden inside the crystal, acting like tiny, invisible magnets that steal energy from the qubit.

---

The Analogy: The Silent Thief vs. The Loud Thief

To understand the difference between what we knew before and what this paper found, let's use an analogy of a house (the quantum computer) and two types of thieves:

1. The Loud Thief (Electric Dipole Loss): This is the "Two-Level System" we already knew about. Imagine a thief who breaks a window to get in. It's loud, obvious, and creates a lot of noise (energy loss). Scientists have been working hard to fix the windows (improving surfaces and materials) to stop this thief.
2. The Silent Thief (Magnetic Dipole Loss): This is what this paper is about. Imagine a thief who doesn't break a window but instead uses a tiny, invisible magnet to pull a coin out of your pocket without you noticing. This thief is much quieter and harder to see.

For a long time, scientists thought the "Loud Thief" was the only one stealing energy. They thought the "Silent Thief" (magnetic loss) was so weak it didn't matter. This paper says: "Wait a minute, that Silent Thief is actually quite strong!"

The Mechanism: The "Zero-Field Splitting" (The Invisible Spin)

How does this Silent Thief work?

Inside the crystal (usually Sapphire, which is used as the floor for these quantum computers), there are tiny impurities like Chromium or Iron. These atoms have electrons that are spinning.

• Normal Spin: Usually, if you have a spinning electron, its energy depends on how you turn a magnet near it.
• Zero-Field Splitting (ZFS): The authors focus on a special trick these electrons play. Even if you have no magnet nearby (Zero Field), the electron's spin is already split into different energy levels. It's like a spinning top that naturally wants to wobble in two different directions even when the table is perfectly still.

When a quantum computer sends out a microwave signal (the "hum" of the computer), it accidentally hits the exact frequency that makes these hidden magnetic spins jump from one level to another. When they jump, they absorb energy from the computer, just like a thief stealing a coin.

The Calculation: How Much Energy is Stolen?

The authors did the math to see how much energy these "Silent Thieves" steal.

• They looked at real-world sapphire crystals used in quantum computers.
• They found that even "high-quality" crystals have tiny amounts of Chromium, Iron, and Vanadium (about 1 in every 100 billion atoms).
• They calculated that these tiny amounts are enough to cause a loss of energy that matches exactly what scientists have been measuring in experiments.

The Result: The "Silent Thief" (magnetic loss) is stealing about as much energy as the "Loud Thief" (electric loss). This means we can't just ignore magnetic impurities anymore.

Why This Matters: A New Clue for the Mystery

For years, scientists have been trying to figure out why quantum computers lose their "memory" (coherence). They've been polishing surfaces and cleaning materials, but the problem persists.

This paper suggests a new theory:

• Maybe the "Two-Level Systems" (TLS) that everyone talks about aren't just random jiggling atoms.
• Maybe many of them are actually these magnetic impurities doing their "Zero-Field Splitting" dance.

If this is true, it changes how we build quantum computers. Instead of just polishing the surface, we might need to:

1. Grow crystals that are even purer (removing the Chromium and Iron).
2. Design the computer to avoid the specific frequencies where these magnetic thieves are most active.

The Bottom Line

Think of the quantum computer as a high-stakes game of Jenga.

• Old View: The tower is falling because the blocks on the outside are loose (surface defects).
• New View (This Paper): The tower is also falling because there are invisible, magnetic magnets inside the blocks themselves, pulling the tower apart from the inside.

The authors have shown that these internal magnetic magnets are strong enough to be a major reason why quantum computers aren't as stable as we'd like them to be. It's a "lightbulb moment" that could help engineers build better, longer-lasting quantum computers in the future.

Technical Summary

1. Problem Statement

The coherence times of superconducting qubits are currently limited by dielectric loss in the materials used for their substrates and packaging. While surface losses and Two-Level Systems (TLSs) in amorphous materials have been extensively studied, the microscopic origins of bulk dielectric loss in high-quality crystalline substrates (like sapphire) remain elusive.

• The Gap: Standard models attribute loss to electric dipole transitions (TLSs). However, recent experiments suggest that bulk loss in crystalline sapphire is higher than predicted by electric loss mechanisms alone.
• The Hypothesis: The authors propose that magnetic dipole transitions arising from paramagnetic impurities (specifically transition metals like Cr, Fe, and V) within the crystal lattice contribute significantly to bulk loss. These transitions occur between spin states split by Zero-Field Splitting (ZFS), even in the absence of an external magnetic field.

2. Methodology

The authors developed a theoretical framework to quantify magnetic loss and applied it to a specific case study.

A. Theoretical Derivation

• Absorption Cross-Section: They derived the absorption cross-section ($\sigma_{MD}$) for a magnetic dipole transition. Starting from Fermi's Golden Rule and the interaction Hamiltonian for electromagnetic radiation, they expanded the vector potential to the first order (magnetic dipole term) and second order (electric quadrupole term).
• Key Equation: The derived cross-section is:
  $$\sigma_{MD}(\omega) = n_r \pi^2 \alpha^3 a_0^2 |\xi_B \cdot M_{IF}|^2 \omega \delta(\omega - \Omega_{IF})$$
  Where:
  • $n_r$: Refractive index.
  • $\alpha$: Fine-structure constant.
  • $a_0$: Bohr radius.
  • $M_{IF}$: Transition matrix element involving orbital ($L$) and spin ($S$) angular momentum operators.
  • $\Omega_{IF}$: Transition frequency (ZFS energy).
• Broadening: They accounted for homogeneous broadening (due to finite state lifetimes) by replacing the Dirac delta function with a Lorentzian profile. They noted that while inhomogeneous broadening (strain) exists, the Lorentzian dominates far off-resonance.
• Loss Tangent Calculation: They linked the absorption coefficient to the bulk loss tangent ($\tan \delta$) using the relation:
  $$\tan(\delta) = \frac{c}{n_r \omega} N_{def} \sigma(\omega)$$
  Where $N_{def}$ is the defect density.

B. Case Study: Sapphire Substrates

• Materials: They focused on Sapphire ($Al_2O_3$), the standard substrate for superconducting qubits.
• Impurities: They modeled three common paramagnetic impurities found in Heat-Exchanger Method (HEMEX) sapphire:
  • Chromium (Cr$^{3+}$): Ground state spin $S=3/2$, ZFS $\approx 11.45$ GHz.
  • Iron (Fe$^{3+}$): Ground state spin $S=5/2$, ZFS $\approx 12.03$ GHz.
  • Vanadium (V$^{2+}$): Ground state spin $S=3/2$, ZFS split by hyperfine coupling (multiple transitions between 8.68–10.40 GHz).
• Parameters: They utilized experimental values for ZFS parameters ($D$), electron g-factors ($g_e$), and defect concentrations ($N_{def} \approx 10^{16} - 10^{17} \text{ cm}^{-3}$). They assumed the orbital wavefunction remains unchanged during the transition, simplifying the matrix element calculation to depend primarily on spin operators.

3. Key Contributions

1. Quantification of Magnetic Loss: This is the first work to explicitly derive and quantify the contribution of magnetic dipole absorption from ZFS-split states to the bulk dielectric loss tangent in superconducting qubit substrates.
2. Re-evaluation of TLS Models: The authors challenge the assumption that all bulk loss is electric in nature. They demonstrate that magnetic transitions, often overlooked due to their weaker coupling compared to electric dipoles, can be significant in the GHz regime.
3. Matrix Element Analysis: They analyzed the magnitude of magnetic dipole matrix elements ($|M_{IF}|$), showing that for pure spin transitions (common in these impurities), the matrix element is of order 1, making the absorption non-negligible despite the weak magnetic coupling.
4. Temperature and Power Dependence: They discussed how ZFS-induced loss behaves under varying temperature and power, noting that while it mimics TLS behavior, off-resonant contributions may dominate at typical qubit operating frequencies, potentially explaining discrepancies in power saturation observations.

4. Results

• Calculated Loss Tangents: For a typical HEMEX sapphire with impurity concentrations of $10^{17} \text{ cm}^{-3}$, the calculated bulk loss tangent at 4.5 GHz (a standard qubit frequency) is in the range of $10^{-9}$ to $10^{-8}$.
  • Specifically, for Cr, $\tan(\delta) \approx 9.0 \times 10^{-9}$.
• Comparison with Experiment: These calculated values are comparable to the experimentally measured bulk loss tangents in high-quality sapphire (e.g., Read et al. reported $19 \times 10^{-9}$).
• Resonance vs. Off-Resonance: While the loss peaks sharply at the ZFS resonance frequencies (e.g., 11.45 GHz for Cr), the "tails" of the Lorentzian broadening extend down to the 4.5 GHz operating range, contributing significant loss even when the qubit is not directly on resonance with the impurity transition.
• Impurity Impact: Vanadium, due to hyperfine splitting, creates a broader distribution of transition frequencies, potentially contributing to loss across a wider band.

5. Significance and Implications

• Limiting Factor for Coherence: The study suggests that magnetic loss from transition-metal impurities is a limiting factor for the coherence times of superconducting qubits. It implies that simply removing surface TLSs is insufficient if bulk magnetic impurities are present.
• Material Purity Requirements: The results highlight the critical need for ultra-high purity sapphire substrates with minimized transition-metal impurities (Cr, Fe, V) to achieve lower loss tangents.
• Reinterpretation of TLS Data: The authors propose that the "TLS" behavior observed in bulk crystals might not solely be electric dipole tunneling but could include magnetic dipole transitions. This challenges the interpretation of power saturation and temperature dependence data, suggesting that off-resonant magnetic absorption could explain loss mechanisms previously attributed to other sources.
• General Applicability: While focused on sapphire, the derived equations apply to any dielectric material containing paramagnetic defects, offering a new tool for analyzing loss in quantum materials.

In conclusion, Turiansky and Van de Walle provide a rigorous theoretical and quantitative argument that magnetic dipole absorption due to Zero-Field Splitting is a major, previously underestimated source of bulk dielectric loss in superconducting qubit substrates, necessitating a re-evaluation of material specifications and loss models in quantum hardware design.