🔬 atomic physics

Defect assignment of the clock site in $^{229}\text{Th:CaF}_2$

This paper reassesses the dominant clock-active site in $^{229}\text{Th:CaF}_2$ , using thermodynamic estimates and density functional theory to demonstrate that it is an isolated thorium substitution charge-compensated by fluorine interstitials rather than a thorium dimer, thereby providing a microscopic basis for reducing linewidth broadening in solid-state nuclear clocks.

Original authors: Daniel A. Rehn, Harry W. T. Morgan, Harris E. Mason, H. B. Tran Tan, Ricky Elwell, Igor M. Savukov, Michael J. Martin, Andrei Derevianko, Eric R. Hudson

Published 2026-07-15

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel A. Rehn, Harry W. T. Morgan, Harris E. Mason, H. B. Tran Tan, Ricky Elwell, Igor M. Savukov, Michael J. Martin, Andrei Derevianko, Eric R. Hudson

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

Imagine a super-precise timekeeper hidden inside a crystal. This isn't a regular clock with gears; it's a "nuclear clock" built from a rare atom called Thorium-229, trapped inside a block of Calcium Fluoride (CaF₂). For this clock to tick perfectly, the Thorium atom needs to sit in a very specific, quiet neighborhood within the crystal. If its neighbors are too noisy or arranged wrong, the clock's signal gets blurry, and timekeeping fails.

For a while, scientists thought they knew what this neighborhood looked like. They believed the Thorium atoms were hanging out in pairs, like best friends holding hands, forming a "dimer" (a two-atom cluster). They called this the "D center." It seemed like a good guess because the math looked close to the measurements.

But in this new study, a team of researchers took a second, much closer look. They ran powerful computer simulations and checked the thermodynamics (the rules of energy and heat) to see what's actually happening. Their conclusion? The "best friend" theory is probably wrong. The Thorium atoms aren't pairing up; they are actually living alone, but they have a very specific set of neighbors helping them out.

The Real Neighborhood: The Solo Act with Bodyguards

Instead of two Thorium atoms huddling together, the paper suggests the clock-active Thorium is actually a single atom sitting in a spot meant for a Calcium atom. But here's the catch: Thorium has a different electrical charge than Calcium. It's like a VIP trying to sit in a seat designed for a regular guest; the balance is off.

To fix this, the crystal spontaneously adds two extra Fluorine atoms (the "bodyguards") right next to the Thorium. These two Fluorine atoms arrange themselves in a specific "90-degree" angle, like the corner of a room. This setup neutralizes the charge and creates the perfect, quiet environment the clock needs.

The researchers are quite confident about this. Their computer models show that forming a Thorium pair is actually energetically expensive—it costs about 0.56 eV just to push two Thorium atoms together, and in a dilute crystal, the odds of them randomly bumping into each other are incredibly low. In fact, the energy cost to keep them apart is so high that the "solo act with bodyguards" is the clear winner.

The Evidence: Listening to the Crystal Sing

How do they know this? They listened to the "song" the atoms sing when hit with light. This song is split into different notes by the electric field around the atom. The researchers calculated what the song should sound like for the "90-degree bodyguard" setup.

The result? It matched the real-world measurements almost perfectly.

The main "note" (called the electric field gradient, or EFG) was calculated to be 107.650 V/Å².
Real experiments measured values like 106.3(8) V/Å² and 109.1(7) V/Å².

That's a very close match! The previous "dimer" theory predicted a value of 95 V/Å², which was close, but the new "solo" theory is spot on.

There is a small discrepancy in the "shape" of the song (an asymmetry parameter called η). The simulation gave 0.173, while experiments saw around 0.60. However, the authors explain that this "shape" parameter is very sensitive to tiny details in the computer model, like the size of the crystal block used in the simulation. They argue that the main "note" (the energy splitting) is the most important part, and since that matches, the "solo with bodyguards" theory holds up.

What About the Other Clues?

The paper also addresses a few other things scientists saw:

The "O Center": There's another signal in the crystal that looks like a flat, broad line. The authors think this might be Thorium atoms that are far away from their bodyguards, or maybe the bodyguards moved around too much. It's not the main clock site, so they leave it as a mystery for now.
The "Site 3": There's a weaker signal that matches a slightly different arrangement of the bodyguards (a "1st-2nd obtuse" motif). This is the second most likely candidate, but it's much rarer than the main 90-degree setup.
The Microscope Mistake: Earlier, someone looked at a super-dense crystal under a microscope and saw Thorium atoms clumping together. The authors point out that this crystal was packed with 1.1 at.% Thorium, whereas the clock crystals only have 0.02 at.%. It's like seeing a crowd at a concert and assuming everyone is in a huddle, when in a quiet library (the clock crystal), everyone is actually sitting alone.

Why This Matters

If the Thorium atoms were indeed clumping together in pairs, fixing the clock would mean trying to stop them from huddling. But since they are actually solo atoms with specific Fluorine neighbors, the path forward is different. To make a better clock, scientists need to focus on controlling those Fluorine "bodyguards" and the stress in the crystal lattice.

The paper suggests that by engineering these local defects—making sure the Fluorine atoms sit in that perfect 90-degree corner—we can reduce the "noise" (linewidth) and build a more accurate nuclear clock. It turns out the secret to the perfect timekeeper isn't finding a partner; it's finding the right bodyguards.

Problem Statement
The performance of solid-state $^{229}$ Th nuclear clocks is critically dependent on the microscopic environment of the thorium nucleus within the host crystal. While $^{229}$ Th:CaF $_2$ is a leading candidate material due to its high VUV transparency and sufficient doping density, the specific defect configuration responsible for the dominant clock-active, quadrupole-split spectral site has been a subject of debate. Recent spectroscopic work assigned this site to a "D center," interpreted as a thorium dimer (two adjacent Th $^{4+}$ ions) without local charge compensation. However, the composition and structure of this site are crucial for clock development: if the transition is hosted by Th–Th aggregates, linewidths and frequency shifts would be tied to concentration-dependent many-dopant complexes. Conversely, if the spectrum arises from a locally charge-compensated single-Th site, optimization efforts should focus on controlling nearby charge-compensation defects and strain.

Methodology
The authors reassess the dominant defect configuration using a combination of thermodynamic estimates, density functional theory (DFT) calculations, and electric-field-gradient (EFG) comparisons.

Thermodynamics: The study evaluates the free energy of formation for both the proposed Th–Th dimer and various locally charge-compensated single-Th motifs. This includes calculating association energies and accounting for configurational entropy at relevant temperatures (300 K and near the CaF $_2$ melting point).
DFT Calculations: The authors performed DFT calculations using the VASP code with the PBE functional and projector-augmented-wave (PAW) method. They utilized a 5 $\times$ 5 $\times$ 5 supercell (377 atoms) to model neutral, same-stoichiometry two-fluorine-interstitial (2F $_i$ ) motifs around a substitutional Th $^{4+}$ site. Spin-orbit coupling (SOC) was included for all structures.
Spectral Comparison: The study computes the principal EFG component ( $V_{ZZ}$ ) and the asymmetry parameter ( $\eta$ ) for various relaxed defect geometries. These calculated values are compared against experimental transition frequencies and EFG parameters reported in recent literature (e.g., Zhang et al., Morawetz et al., Hiraki et al.).

Key Contributions and Results

Rejection of the Bare Dimer Hypothesis: Thermodynamic analysis indicates that the formation of a bare Th–Th dimer is energetically unfavorable. The calculated energy to form the dimer from separated defects is +0.56 eV, reflecting repulsion between nearby Th $^{4+}$ substitutions. Furthermore, configurational entropy strongly disfavors dimerization in the dilute crystals used for clocks (concentration $\sim$ 0.02 at.%), adding a significant free-energy cost. The observation of Th clustering in previous studies is attributed to much higher doping concentrations (1.1 at.%) than those used in clock crystals.
Identification of the Dominant Site: The calculations identify a locally charge-compensated single-Th defect as the thermodynamically preferred configuration. Specifically, a relaxed 90 $^\circ$ motif consisting of an isolated Th $^{4+}$ on a Ca $^{2+}$ site compensated by two nearby fluorine interstitials (Th $^{\bullet\bullet}_{Ca} + 2F'_{i}$ ) is the lowest-energy member of the two-interstitial family.
Spectral Consistency: The 90 $^\circ$ two-F $_i$ motif yields a calculated $V_{ZZ}$ of 107.650 V Å $^{-2}$ , which is in excellent agreement with experimental values ranging from 106.3 to 109.1 V Å $^{-2}$ . This confirms that the dominant quadrupole-split site is consistent with a locally compensated single-Th center rather than a dimer.
Minor Site Assignment: The study identifies a "1st-2nd obtuse" mixed-shell interstitial motif as the next most likely configuration. This structure provides a plausible assignment for a weaker, higher-energy spectral site (Site 3 in Ref. [8]), with a calculated $V_{ZZ}$ of -309.049 V Å $^{-2}$ matching experimental observations.
Resolution of Previous Discrepancies: The authors note that earlier calculations predicting a high $V_{ZZ}$ for the 90 $^\circ$ motif (223 V Å $^{-2}$ ) were likely artifacts of using smaller supercells with relaxed lattice vectors. The current larger, fixed-lattice calculations recover the observed branch scale, indicating the disagreement was a convergence issue rather than evidence against fluoride compensation.

Significance and Claims
The paper claims that the dominant clock-active quadrupole-split site in $^{229}$ Th:CaF $_2$ is controlled by local fluoride compensation rather than unavoidable thorium aggregation. This reassignment has direct implications for the development of solid-state nuclear clocks:

Linewidth Control: The observed concentration-dependent linewidth is consistent with a picture of neutral, locally compensated defect complexes acting as elastic sources (Kanzaki forces) that induce inhomogeneous broadening. This contrasts with the expectation for uncompensated dimers, which would generate Coulomb-like long-range interactions.
Optimization Strategy: Improving the useful quadrupole-split site fraction and reducing linewidths should focus on engineering the defect chemistry of local fluoride compensation, managing strain, and mitigating radiation-induced defect conversion. It suggests that promoting thorium aggregation is not the path forward.
Microscopic Basis: The findings provide a microscopic basis for understanding broadening mechanisms and suggest that alternative hosts (such as ThF $_4$ ) where $^{229}$ Th incorporation does not require compensating defects might offer further improvements.

The authors remain modest regarding the "O center" (zero-EFG site), noting that its origin remains a separate problem potentially involving distant compensation, dynamic averaging, or larger defect complexes, and that the assignment of the minor Site 4 remains unresolved within the current static model set.

More like this