⚛️ quantum physics

Semiclassical spin-bath calculation of the nitrogen-isotope effect on NV $^-$ ensemble coherence in diamond

This paper demonstrates through semiclassical spin-bath simulations that the nitrogen nuclear isotope alone reduces the $T_2/T_2^*$ coherence ratio in NV $^-$ ensembles by approximately 12% due to a higher fraction of hyperfine-degenerate P1 pairs in $^{15}$ N compared to $^{14}$ N, though this effect only partially explains the larger reduction observed in recent experiments.

Original authors: Chikara Shinei (University of Tsukuba, J-FAST, CNRS / University Grenoble Alpes / Grenoble INP / University of Tsukuba)

Published 2026-10-01

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Chikara Shinei (University of Tsukuba, J-FAST, CNRS / University Grenoble Alpes / Grenoble INP / University of Tsukuba)

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

Deep within the crystal lattice of a diamond, tiny defects known as nitrogen-vacancy centers act as incredibly sensitive magnetic sensors. These defects are essentially single atoms missing from the diamond's grid, replaced by a nitrogen atom and an empty space, creating a quantum system that can detect magnetic fields with extraordinary precision. The performance of these sensors depends on how long the quantum information inside them remains stable before it gets scrambled by the surrounding environment. Scientists measure this stability in two ways: one that captures the immediate, static noise of the surroundings, and another that measures how long the system can survive a specific type of pulse sequence designed to cancel out that noise. In diamonds rich with nitrogen, the main source of this scrambling is a sea of other nitrogen atoms, each carrying an electron that acts like a tiny, fluctuating magnet.

For years, researchers have known that the concentration of these nitrogen atoms dictates how quickly the sensor loses its coherence. However, a recent puzzle emerged when scientists compared diamonds made with two different types of nitrogen atoms. One type, called nitrogen-14, is the most common form found in nature. The other, nitrogen-15, is a rarer isotope with a slightly different nuclear structure. When researchers built a sensor using a layer of diamond doped with nitrogen-15, they found that its stability was significantly worse than what was observed in natural diamonds containing nitrogen-14, even when the total number of nitrogen atoms was the same. This was surprising because the two isotopes are chemically identical; the only difference lies in the nucleus of the atom. The question became whether this tiny nuclear difference alone could be responsible for such a dramatic drop in performance, or if other factors were at play.

To solve this, a researcher at the University of Tsukuba and the Japanese-French Laboratory for Semiconductor Physics and Technology constructed a detailed computer model of the diamond's interior. The goal was to isolate the effect of the nitrogen nucleus from everything else. The model treated the diamond as a grid where nitrogen atoms were placed randomly, and it simulated how the electron spins of these atoms interacted with one another. Crucially, the simulation accounted for the fact that the nitrogen nucleus can exist in different energy states, which changes the magnetic field felt by the nearby electron. The model focused on a specific mechanism: how often two nitrogen atoms could swap their spin states, a process known as a flip-flop. These swaps are the primary way the environment creates noise that disrupts the sensor.

The simulation revealed a clear, purely mathematical reason for the difference. The ability of two nitrogen atoms to swap spins depends on whether their internal energy levels match up perfectly. In the natural nitrogen-14 isotope, the nucleus has three possible energy states, while the nitrogen-15 isotope has only two. The model calculated that because of this difference in the number of available states, the nitrogen-15 environment offers more opportunities for these disruptive swaps to occur. Specifically, the simulation showed that in a nitrogen-15 bath, the fraction of atom pairs that can freely swap spins is higher than in a nitrogen-14 bath. This means the environment around a nitrogen-15 sensor is inherently more active and chaotic, causing the sensor to lose its coherence faster.

The results of the simulation were precise and consistent. The model predicted that the sensor's ability to withstand noise over time would be about 12 percent shorter in the nitrogen-15 sample compared to the nitrogen-14 sample, regardless of how many nitrogen atoms were present. This prediction held true across a wide range of nitrogen concentrations, from very sparse to very dense. Importantly, the model showed that the initial, static noise level remained exactly the same for both isotopes, confirming that the difference lay entirely in the dynamic, time-dependent fluctuations of the environment. This finding aligns with the direction of the experimental data, which showed a drop in performance for nitrogen-15, but the simulation only accounted for about half of the total difference observed in the lab.

The researchers concluded that while the nuclear isotope alone is sufficient to cause a measurable reduction in stability, it is not the whole story. The fact that the simulation, which relied only on counting the possible energy states, reproduced the trend but not the full magnitude suggests that more complex interactions between the atoms are also involved. The study establishes that the nitrogen isotope acts as a fundamental control knob for the sensor's performance, with the more common nitrogen-14 offering a quieter, more stable environment than its nitrogen-15 counterpart. This insight provides a solid benchmark for future, more complex calculations and helps explain why the choice of isotope matters for the next generation of ultra-sensitive diamond-based magnetometers.

Technical Summary: Semiclassical Spin-Bath Calculation of the Nitrogen-Isotope Effect on NV⁻ Ensemble Coherence

Problem Statement
Nitrogen-vacancy (NV⁻) ensembles in diamond serve as solid-state magnetometers, with sensitivity governed by the ensemble dephasing time ( $T_2^*$ ) for broadband DC sensing and the Hahn-echo coherence time ( $T_2$ ) for narrowband AC sensing. In nitrogen-rich diamond, both times are limited by the bath of substitutional nitrogen (P1) electron spins, scaling inversely with nitrogen concentration $[N]$ . While the ratio $T_2/T_2^*$ is observed to be approximately 16 in natural isotopic abundance diamond (dominated by $^{14}\text{N}$ , nuclear spin $I=1$ ), a recent study of a deliberately $^{15}\text{N}$ -doped ( $I=1/2$ ) ensemble magnetometer reported a significantly lower ratio of $8.05 \pm 0.18$ in the single-quantum convention. The central question addressed is whether the nitrogen nuclear isotope alone can account for this reduction in coherence, specifically investigating if the nuclear spin multiplicity alters the dynamics of the P1 electron bath.

Methodology
The authors employ a semiclassical spin-bath model, extending the framework established by Bauch et al. [2] with Jahn-Teller-resolved hyperfine tensors derived from Park et al. [5]. The approach is intentionally non-many-body, relying on a transparent, parameter-free counting argument to serve as a benchmark for future cluster-correlation-expansion (CCE) treatments.

Key methodological elements include:

Model Setup: An NV⁻ center is placed at the origin of a diamond lattice, with P1 centers randomly assigned to substitutional sites based on probability $[N]$ .
Hyperfine Resolution: The model incorporates the Jahn-Teller (JT) distortion axis of P1 centers. The effective hyperfine splitting ( $A_{\text{eff}}$ ) depends on the angle between the JT axis and the magnetic field, yielding distinct values for on-axis and off-axis orientations.
Flip-Flop Dynamics: P1–P1 spin exchange (flip-flops) is treated as a two-level problem. The population transfer rate is sharply peaked when the hyperfine energy difference (detuning $\delta$ ) between two P1 centers is zero.
Resonant Channel Counting: The bath correlation time $\tau_c$ $τ_{c}$ is governed by the fraction of P1 pairs that are hyperfine-degenerate ( $f_{\text{res}}$ $f_{res}$ ). The authors calculate this fraction analytically based on nuclear spin multiplicity ( $2I+1$ $2 I + 1$ ) and JT degeneracy:
- For $^{14}\text{N}$ ( $I=1$ ): $f_{\text{res}} = 1/4$ .
- For $^{15}\text{N}$ ( $I=1/2$ ): $f_{\text{res}} = 5/16$ .
Simulation: The model simulates ensemble coherence times over a nitrogen concentration range of $[N] = 0.1\text{--}100$ ppm. It averages single-NV decays over distributions of the dipolar field strength ( $\Delta_{\text{single}}$ ) and correlation time ( $\tau_{c,\text{single}}$ ) to extract ensemble $T_2^*$ and $T_2$ .

Key Contributions and Results

Isotope Independence of $T_2^*$ : The simulations confirm that $T_2^*$ is identical for both isotopes. This is because $T_2^*$ is determined solely by the quasi-static dipolar field of the P1 bath (Eq. 3), which depends on P1 positions and is blind to the nuclear spin state. The simulated slope ( $11.21 \pm 0.08$ $\mu\text{s}\cdot\text{ppm}$ ) agrees with experimental values within $\sim 17\%$ .
Isotope Dependence of $T_2$ : The model predicts that $T_2$ is shorter for $^{15}\text{N}$ than for $^{14}\text{N}$ at every concentration. The increased number of resonant flip-flop channels in $^{15}\text{N}$ (5/16 vs. 1/4) leads to a faster bath correlation time.
Quantitative Ratio: The model yields a concentration-independent ratio of $T_2(^{14}\text{N})/T_2(^{15}\text{N}) = 1.118 \pm 0.011$ . This value is $11\sigma$ above unity, confirming that $^{15}\text{N}$ enrichment intrinsically shortens $T_2$ at a fixed nitrogen concentration.
Discrepancy with Experiment: While the model reproduces the direction and concentration independence of the isotope effect, it underestimates the magnitude. The experimental ratio (derived from combining Ref. [2] and Ref. [9]) is $2.07 \pm 0.25$ , which is $3.8\sigma$ larger than the simulated value of $1.118$. The model accounts for only $\sim 17\%$ of the observed contrast on a logarithmic scale.

Significance and Claims
The paper claims to establish a lower bound on the isotope contribution to coherence reduction. The primary significance lies in demonstrating that the nitrogen nuclear isotope controls $T_2$ (but not $T_2^*$ ) through a purely combinatorial mechanism: the fraction of hyperfine-degenerate P1 pairs.

The authors emphasize three specific findings:

Structural Constant: The isotope ratio is a genuine material constant, independent of nitrogen concentration, arising from nuclear spin multiplicity and JT degeneracy rather than sample-specific parameters.
Mechanism Distinction: Unlike dilute nuclear baths where higher nuclear spin typically reduces coherence, the P1 bath behaves inversely. Here, the nucleus acts as a detuner of electron flip-flops; a larger spin multiplicity ( $^{14}\text{N}$ ) spreads transition frequencies over more values, suppressing flip-flops and slowing the bath, whereas $^{15}\text{N}$ allows more resonant channels, accelerating the bath.
Benchmark for Many-Body Physics: The discrepancy between the semiclassical counting argument (1.12) and the experimental observation (2.07) suggests that many-body effects are significant. The authors posit that a full CCE treatment, which accounts for near-degenerate channels and correlated evolution beyond simple rate equations, is the necessary next step to close this gap.

The paper concludes that while the semiclassical model successfully predicts the ordering and concentration independence of the effect, the full magnitude of the observed contrast requires physics beyond the mean-field stochastic treatment, specifically the inclusion of correlated many-body dynamics.

Technical Summary: Semiclassical Spin-Bath Calculation of the Nitrogen-Isotope Effect on NV⁻ Ensemble Coherence

More like this