Theory of the charge response of the boron-vacancy spin in monolayer hexagonal boron nitride

Dongyeon, Kim, Theodore Scharlotta, Daniel Hashemi

Nov 1, 2026arXiv:2611.02137v1

cond-mat.mtrl-sci

Scorecard· 16/16

6.5/10 impact

Careful, pre-registered DFT study giving a new and interpretable direction-dependence result for V_B- charge response, but limited by PBE-dependent magnitudes, small sample, monolayer-only scope and no experimental validation.

Abstract

The boron vacancy $V_{\mathrm{B}}^{-}$ in hexagonal boron nitride is a leading two-dimensional spin sensor, but the microscopic connection between nearby charge positions and its spin Hamiltonian remains poorly understood. Calculations for carbon and oxygen donors 3.85--8.08 $\mathring{\mathrm{A}}$ away show that direction matters, not only distance. In the frozen lattice, reversing a carbon donor's azimuth at 4.35 $\mathring{\mathrm{A}}$ (C1 to C1m) more than halves $∣ E ∣$ and turns its axis across the mirror line. The ratio is 0.467 with the Perdew--Burke--Ernzerhof (PBE) functional, against 1 for distance alone. A field-plus-gradient model fitted to six donor configurations, including C1 but excluding C1m, predicted a ratio of 0.443 before C1m was calculated. Heyd--Scuseria--Ernzerhof hybrid statics give 0.432 and PBE relaxation 0.600. Exploratory analysis suggests why direction matters: a charge's potential differs at the three vacancy nitrogens and moves the spin of the two $e^{'}$ holes among them, more when the charge faces one nitrogen than the gap between two. $E$ follows this spin imbalance through one constant, $c\approx8$ GHz per $μ_{\mathrm{B}}$ , across donors and applied fields, whereas in PBE $D$ is nearly insensitive to it. A three-site model with one fitted screening constant reproduces the frozen-lattice $E$ to 83 MHz rms. Results are for one charge state in a free-standing monolayer; absolute $∣ E ∣$ varies by tens of percent across approximations.

AI Impact Assessments

(1 model)

Assessment: Theory of the charge response of the boron-vacancy spin in monolayer hBN

Core Contribution

This is a DFT study of how a nearby ionized donor (C_B or O_N) changes the spin Hamiltonian of V_B⁻ in monolayer hBN. It makes three linked claims.

Direction dependence. At fixed distance, the donor's azimuth matters, not only its distance. Reversing a carbon donor's azimuth at 4.35 Å (C1 to C1m) cuts the frozen-lattice |E| from 988 to 462 MHz (ratio 0.467) and switches the E axis from along to across the mirror line. Field-only point-charge models, which the NV-center literature has used, predict a ratio of 1.

Mechanism. The transverse parameter E tracks a "spin imbalance" among the three vacancy nitrogens through a single constant, c ≈ 8 GHz/μB. This holds across donors, relaxed and frozen lattices, applied fields and HSE06. A hole-pair dipolar derivation shows that redistributing spin around the triangle leaves D unchanged but changes E as a second angular harmonic. A three-site tight-binding model with one screening constant (ε ≈ 3.6) reproduces the frozen-lattice E to 83 MHz rms.

Practical output. A map of |E| and θ_E against charge position, an inversion analysis, estimates of the charge densities consistent with a 50 MHz E, and a simulated-spectra analysis of how well configurations can be told apart.

The conceptual advance is an interpretable microscopic link between charge position and the E tensor, plus the point that the second angular harmonic of the potential, not only the field magnitude, controls E.

Methodological Rigor

The paper is unusually careful about epistemic status.

Pre-registration. The field-plus-gradient model's predictions were timestamped and archived before C1m, C3m, C3, O3 and Om were computed. The paper separates "tested" results from "exploratory" ones and marks post hoc checks. The 0.443 prediction against 0.467 in DFT is a real out-of-sample test, although C1 was in the fit.

Uncertainty accounting. It separates numerical convergence, finite-size effects (|E| changes by 7–35% from 8×8 to 10×10) and functional dependence. HSE06 raises |E| by about 55%, and spin decontamination lowers it by 37–53%. The authors state that absolute |E| is only a PBE value. The C1m/C1 ratio survives these tests (0.43–0.48 in PBE and HSE06, 0.346 after decontamination), but its magnitude is functional-dependent.

State checks. Electronic-state checks and restart tests support the high-spin ground-state assignment.

Weaknesses:

The sample is small, with eleven placements and a calibration set of about six independent constraints.

The spin-imbalance relation uses outputs of the same calculation on both sides, so it is descriptive rather than causal. The 36 tensors used to fit c are not independent, and c depends on the PAW partition and on the ideal azimuths (7.28 versus 7.95 with relaxed azimuths).

The gradient mechanism is not uniquely identified. The nonlinear and field-only variants also give the right C1m branch.

The model has worse leave-one-out |E| error than a distance power law (94 versus 42 MHz) and does not describe ΔD.

Free-standing monolayer, a single charge state, Γ-point supercells and an uncorrected PBE |E| scale all limit applicability. The PBE D is 20% below experiment.

There is no experimental validation, and the paper offers only testable predictions.

Spin-decontamination-corrected ratios are only indicative.

Potential Impact

The audience is mainly the hBN spin-defect and 2D quantum sensing community, plus the broader point-defect theory community. The practical lessons are useful.

Report the E axis along with |E|.

Use ratios at fixed distance rather than absolute |E|.

The charge-noise and density estimates (about 7.6 MHz per 10¹² cm⁻²) offer order-of-magnitude guidance for interpreting ODMR linewidths and E splittings.

The inversion analysis shows that |E| alone cannot localize a charge (15 candidate sites), though the axis narrows this to one.

Direct application to experiments is limited. Real samples have strain, multiple charges, thicker flakes and optical-readout charge dynamics, and the authors acknowledge this. The three-site picture could extend to other three-site-doublet defects (NV-like), but this is untested.

Timeliness

V_B⁻ in hBN is a very active platform, with a push toward nanoscale quantum sensing in few-layer flakes where charge environments matter. Understanding E-parameter and charge-noise physics is a real bottleneck, and the paper arrives as shallow-vacancy experiments (Zhou et al.) appear. The area has few such microscopic analyses.

Strengths

A clear, falsifiable prediction (direction dependence) backed by pre-registered tests.

An interpretable compact model (Eqs. 2–5) with a closed-form derivation that D is invariant to spin redistribution while E is not.

Honest uncertainty reporting, including functional sensitivity.

Open data deposit on Zenodo.

Limitations

Dense, jargon-heavy writing with heavy reliance on the supplement. The main text is difficult to parse because many sections and numbers are compressed into it.

Heavy reliance on PBE for a quantity (|E|) that is highly functional-sensitive.

Small sample and a phenomenological model.

Limited to one charge state and the monolayer.

Complex outcomes that depend on the partition of spin density.

Overall

This is a solid, careful, specialist theory paper with a genuinely new qualitative insight and an unusually transparent validation protocol. Its impact is likely moderate and confined to the defect-theory and hBN-sensing communities. Without experimental confirmation and with strong method-dependence of magnitudes, it is unlikely to be a field-defining reference, but it could be cited as the explanation of angular effects on E in charge-sensing proposals.

Rating:6.5/ 10

Significance 6Rigor 7.5Novelty 6.5Clarity 5

Generated Nov 3, 2026

Comparison History (0)

No comparisons yet.

Scorecard· 16/16

6.5/10 impact