---
title: Negatively Charged Boron Vacancy (VB-)
url: https://www.emergentmind.com/topics/negatively-charged-boron-vacancy-vb
type: topic
---

# Negatively Charged Boron Vacancy (VB-)

The negatively charged boron vacancy, $V_B^-$, is a point defect in hexagonal boron nitride (hBN) formed by a missing boron atom that traps an extra electron. In hBN it is the only boron-vacancy charge state known to show spin-dependent photoluminescence and therefore functions as the principal spin-photon interface among boron-vacancy configurations. Its combination of a room-temperature spin-triplet ground state, optically detected magnetic resonance (ODMR), broad red emission near $\sim 800\ \text{nm}$, and compatibility with atomically thin van der Waals devices has made it a central defect platform for quantum sensing, integrated photonics, and spin-based quantum control in hBN [2305.05866][2303.06784].

## 1. Defect identity, symmetry, and optical character

Structurally, $V_B^-$ is a boron vacancy surrounded by three neighboring nitrogen atoms. The defect is commonly described as having $D_{3h}$ point-group symmetry, with a ground triplet manifold $^3A_2'$, an excited triplet manifold $^3E''$, and a metastable singlet $^1E'$ in the three-manifold picture used for recent polarization and photodynamics analyses [2406.11953][2509.00912]. The host material, hBN, is a wide-bandgap van der Waals crystal whose atomically thin geometry places defect spins in extreme proximity to external samples, which is why $V_B^-$ has been emphasized for nanoscale sensing of magnetic fields, temperature, pressure, strain, and nearby materials.

The optical response of $V_B^-$ is unusual among solid-state spin defects because the free-space photoluminescence is broad and largely featureless rather than being dominated by a narrow zero-phonon line. This broadening has been attributed to strong electron-phonon coupling and Jahn-Teller mixing of electronic states, and the emission is usually described as peaking around $800\ \text{nm}$ [2202.10980]. Within cavity-enhanced measurements, the zero-phonon line was directly identified at $773\pm2$ nm, in agreement with earlier theoretical estimates reported by Ivady et al., Reimers et al., and Libbi et al.; the dominant transition was assigned to $^3E'' \rightarrow {}^3A_2'$ [2202.10980].

The defect’s basic spin signature is a triplet ground state with zero-field splitting of about $3.5\ \text{GHz}$. This enables optical initialization, microwave manipulation, and optical readout through ODMR. A recurring misconception in the broader boron-vacancy literature is to treat different charge states interchangeably. The available experimental record instead distinguishes them sharply: only the negatively charged state is known to exhibit the spin-dependent photoluminescence used for ODMR-based readout, whereas the neutral state $V_B^0$ does not provide the same operational spin-photon interface [2305.05866].

## 2. Spin Hamiltonians and electronic-state spectroscopy

The ground-state spin Hamiltonian has been formulated in the standard triplet-defect form
$$
H = g\mu_B \mathbf{B}\cdot \mathbf{S} + D\left[S_z^2-\frac{S(S+1)}{3}\right] + \sum_{k=1}^{3} \mathbf{S}\cdot \mathbf{A}_k \cdot \mathbf{I}_k,
$$
with additional nuclear Zeeman and quadrupole terms used in ENDOR treatments [2302.08831]. In zero or weak magnetic field, ODMR resonances are commonly parameterized by
$$
\nu_{2,1}=D/h \pm \sqrt{E^{2}+(g\mu_B B)^2},
$$
where $D$ is the axial zero-field splitting and $E$ is the transverse splitting [2105.12029]. For ion-implanted ensembles at room temperature, two Lorentzian ODMR dips were reported near $\nu_1 \sim 3405$ MHz and $\nu_2 \sim 3548$ MHz, giving $D/h = 3475 \pm 5\ \text{MHz}$ and $E/h = 70 \pm 5\ \text{MHz}$ [2105.12029].

Direct experimental access to the excited-state spin structure was obtained through continuous-wave ODMR and then validated by pulsed ODMR. In continuous-wave ODMR, the known ground-state resonances near $3.48\ \text{GHz}$ were accompanied by a broad additional resonance near $2.1\ \text{GHz}$; in pulsed ODMR, those broad $\sim 2.1$ GHz features disappeared when the microwave acted only while the laser was off, confirming their excited-state origin [2112.02912]. The excited-state spin was modeled as
$$
H = D_{es}(S_Z^2 - S(S+1)) + \mu_B g_{es}\,\vec{B}\cdot \vec{S} + E_{es}(S_X^2 - S_Y^2) - A_{es}\,\vec{S}\cdot \vec{I}.
$$

From field-dependent spectroscopy, the room-temperature excited-state parameters were extracted as $D_{es} = 2.06 \pm 0.04\ \text{GHz}$, $g_{es} = 2.04 \pm 0.05$, and $E_{es} = 93.1 \pm 23.3\ \text{MHz}$ [2112.02912]. At cryogenic temperature, a closely related excited-state zero-field splitting of $D_{\mathrm{ES}} \approx 2160~\text{MHz}$ was reported, together with excited-state ODMR contrast of $12\%$ at $7$ K; the excited-state $g$-factor was found to be similar to the ground-state value [2112.06420].

| Manifold | Quantity | Reported value |
|---|---|---|
| Ground state | Zero-field splitting | about $3.5\ \text{GHz}$ |
| Excited state | Zero-field splitting | $2.06 \pm 0.04\ \text{GHz}$ at room temperature |
| Excited state | Zero-field splitting | $\approx 2160$ MHz at $7$ K |
| Excited state | $g$-factor | $2.04 \pm 0.05$ |
| Excited state | Transverse anisotropy | $93.1 \pm 23.3\ \text{MHz}$ |
| Excited state | Hyperfine scales | $\sim 90$ MHz and $\sim 18.8$ MHz |

Level anti-crossings are a central part of the $V_B^-$ spectroscopic phenomenology. Using the spin Hamiltonian, the excited-state LAC was predicted near $750$ G and the ground-state LAC near $1240$ G at room temperature, while low-temperature measurements placed the excited-state LAC near $800$ G and the ground-state LAC near $1307$ G [2112.02912][2112.06420]. Near these crossings, strong spin mixing reduces the selectivity of optical pumping and lowers ODMR contrast; photoluminescence anomalies, angle dependence, and residual contrast suppression at nominal alignment were attributed to coherent spin precession and anisotropic relaxation, including residual mixing from hyperfine-induced effective tilted fields [2112.02912]. A noteworthy contrast with the NV center in diamond is that in $V_B^-$ the excited-state LAC remains prominent at cryogenic temperature and the excited-state ODMR contrast becomes larger rather than disappearing [2112.06420].

## 3. Optical cycle, lifetimes, and spin polarization

Time-resolved photoluminescence established a spin-dependent optical cycle in which the excited-state decay rates differ strongly between the $m_s=0$ and $m_s=\pm1$ manifolds. A seven-level description with ground and excited triplets plus an effective metastable singlet was used to analyze the dynamics. In this framework, optical pumping occurs at rate $k_p$, spin-conserving excited-state decay at rate $k_r$, intersystem crossing from the excited triplet to the singlet at rates $\gamma_0$ and $\gamma_1$, and singlet return to the ground triplet at rates $\kappa_0$ and $\kappa_1$ [2404.14155].

Using an all-optical protocol, room-temperature averages over 18 flakes yielded $\tau_0 = 1.22(9)\ \mathrm{ns}$, $\tau_1 = 0.53(6)\ \mathrm{ns}$, and spin polarization $\mathcal{P} = 75(7)\%$. At $4$ K, the same analysis gave $\tau_0 = 4.46(7)\ \mathrm{ns}$, $\tau_1 = 1.50(5)\ \mathrm{ns}$, and $\mathcal{P} = 71(1)\%$. From power-dependent transient fits, the metastable lifetime was inferred as $\tau_m = 18(3)\ \mathrm{ns}$ at room temperature and $\tau_m = 30(3)\ \mathrm{ns}$ at $4$ K [2404.14155]. These data showed that the absolute lifetimes lengthen strongly on cooling, whereas the spin selectivity of the cycle remains substantial.

A complementary semiclassical analysis combined excited-state spectroscopy, power-dependent photoluminescence traces, and spin-resolved polarization dynamics to extract $\Gamma_\mathrm{ISC} = 1286~\text{MHz}$, $\Gamma_E = 849~\text{MHz}$, $\Gamma_s = 22.3~\text{MHz}$, $r \lesssim 0.04$, and $k = 0.21$ [2406.11953]. On that basis, off-resonant optical pumping was predicted to produce electronic spin polarization $p \gtrsim 95\%$ under ambient conditions. This suggests that the intrinsic polarization of $V_B^-$ may be substantially higher than earlier rate models had implied, and it helps rationalize previously reported unusually large ODMR contrasts.

The metastable singlet lifetime was later measured directly in neutron-irradiated sub-micron flakes through pulse-pair recovery measurements. Averaging over 16 flakes gave $\tau_s = 15(3)$ ns at room temperature [2504.05289]. In the same study, a conventional 7-level fit yielded $\tau_s = 23(1)$ ns and failed to reproduce high-power quenching and the full thermal photoluminescence peak amplitude, whereas a 9-level model coupled to an additional 2-level subsystem produced $\tau_s = 17.6(3)$ ns and matched the direct lifetime more closely. The additional subsystem was interpreted as another electronic state or possibly a charge-converted manifold, and larger flakes exhibited behavior consistent with optically induced conversion of $V_B^-$ to another state, possibly $V_B^0$ [2504.05289]. This is one of the clearest current model-refinement issues in the field: the seven-level picture is robust for many datasets, but it is not universally sufficient at higher excitation power or in all sample geometries.

## 4. Hyperfine structure, quadrupole couplings, and the nuclear-spin environment

Electron-nuclear couplings in $V_B^-$ are dominated by the three nearest nitrogen atoms around the vacancy. Conventional ESR and high-frequency ENDOR established that the nearest-neighbor hyperfine interaction is axially symmetric, with principal values
$$
A_\perp = 45.5 \pm 0.9~\text{MHz}, \qquad A_\parallel = 87 \pm 0.5~\text{MHz},
$$
and that the nuclear quadrupole interaction is characterized by
$$
C_q = 1.96 \pm 0.05~\text{MHz}, \qquad n = -0.070 \pm 0.005.
$$
The hyperfine principal axis aligns with the nitrogen dangling-bond direction rather than the crystallographic $c$-axis, whereas the quadrupole tensor is nearly axially symmetric about the same local axis [2302.08831].

A major consequence of those measurements is the spatial picture of the defect wavefunction. Using a linear-combination-of-atomic-orbitals analysis together with the measured hyperfine values, the spin density on a single nearest nitrogen was estimated as about $28\%$, implying that about $84\%$ of the electronic spin density is localized on the three nearest nitrogen atoms [2302.08831]. This established experimentally that $V_B^-$ is not a broadly delocalized spin texture but a strongly localized defect state confined mainly to the first coordination shell within a single BN layer.

Excited-state spectroscopy revealed a related but distinct hyperfine pattern. Fine scans of the excited-state ODMR line showed a seven-peak structure analogous to the ground-state pattern of three equivalent nitrogen nuclei, but with a much larger splitting of about $90$ MHz; an additional finer splitting of about $18.8$ MHz was tentatively attributed to a boron nucleus farther from the vacancy [2112.02912]. The stronger nitrogen-related excited-state splitting was interpreted as evidence of greater electron density near the nitrogen nuclei in the excited state.

ENDOR has also resolved more remote nuclei. High-field measurements detected $^{14}\mathrm{N}$ spins in the third nitrogen shell, N(3), approximately $0.4$ nm from the vacancy, with
$A_{xx} \approx 4.41 \pm 0.02$ MHz,
$A_{yy} \approx 4.43 \pm 0.02$ MHz,
$A_{zz} \approx 6.86 \pm 0.02$ MHz,
$C_q \approx 0.37 \pm 0.03$ MHz, and
$\eta \approx 0.55 \pm 0.01$ [2504.06798]. Density-functional calculations reproduced these values closely and confirmed the shell assignment. This extended the role of $V_B^-$ from a defect whose own hyperfine structure can be resolved to a local probe of remote nuclear magnetic moments in the hBN host.

## 5. Generation, charge-state control, and optical-brightness engineering

Ion irradiation is the dominant route to creating $V_B^-$, but the relation between vacancy production and usable optically active centers is nontrivial. Large-area ion implantation of exfoliated hBN produced room-temperature ODMR with $D/h = 3475 \pm 5\ \text{MHz}$, $E/h = 70 \pm 5\ \text{MHz}$, ODMR contrast up to $22\%$, and a longest measured $T_1$ of about $17\ \mu\text{s}$. In that study, nitrogen implantation at $30$ keV and $1 \times 10^{14}\ \text{ions/cm}^2$ generated a strong photoluminescence band from about $700$ to $1000$ nm centered near $820$ nm [2105.12029]. Fluence, energy, and ion species were all found to matter, but not in the same way: fluence controlled defect density and disorder, energy mainly affected brightness within the thin-flake regime studied, and heavier ions increased strain-related transverse splitting and reduced $T_1$.

A later focused-ion-beam framework emphasized that robust fabrication in thin flakes is limited at least as much by impurity recoil implantation as by nominal vacancy creation. In flakes thinner than about $10\ \text{nm}$, surface and interface effects dominate. In a capped-versus-uncapped comparison using $10$ keV H$^+$ at $5 \times 10^{14}\ \text{cm}^{-2}$, a region capped by $7$ nm multilayer graphene before irradiation showed about a $50\%$ reduction in $V_B^-$ emission, whereas a control region covered with $22$ nm multilayer graphene only after irradiation showed about a $17\%$ reduction attributable to optical absorption; the difference was assigned to recoil implantation of carbon into hBN [2303.06784]. The same work identified an optimal fluence in a thick-flake example irradiated with $15$ keV H$^+$, where the photoluminescence intensity peaked at about $5 \times 10^{15}\ \text{cm}^{-2}$.

Quantifying the yield of optically active $V_B^-$ remains a separate issue from counting all vacancies. By comparing fluence-dependent ODMR splittings with a microscopic charge model and molecular-dynamics vacancy counts for $30$ keV He irradiation, a lower bound of about $0.2\%$ was obtained for the fraction of all vacancies that enter the optically active negatively charged state [2501.18481]. This result strongly indicates that charge-state stabilization and environmental control are as important as vacancy production itself.

Charge-state control is therefore fundamental. Reversible switching between $V_B^-$ and $V_B^0$,
$$
\mathrm{V_B^- \rightleftharpoons V_B^0 + e^-},
$$
was demonstrated under $5$ keV electron-beam irradiation and $532$ nm laser excitation [2305.05866]. Electron-beam exposure quenched the $V_B^-$ photoluminescence by about $10\%$ at low flux and by more than $30\%$ around $\sim 3\times10^{16}\,\mathrm{e^- / cm^2 / s}$, whereas the laser drove recovery toward the negative state. In an FLG/hBN/FLG heterostructure, the balance could be tuned electrically: at $-6$ V the quenching was about $30\%$, at $+6$ V it increased to about $57\%$, and without the electron beam the bias had no significant effect on photoluminescence [2305.05866]. The physical interpretation was that holes promote $V_B^0$ formation and electrons promote recovery of $V_B^-$.

Because the intrinsic photoluminescence of $V_B^-$ is weak, substantial effort has gone into brightness enhancement without destroying spin functionality. Several strategies have now been demonstrated. Low-loss plasmonic nano-patch antennas produced overall intensity enhancement up to $\sim 250\times$, corresponding to an estimated actual enhancement of $\sim 1685\times$ after correcting for laser spot size, while preserving ODMR contrast [2207.08357]. In suspended hBN, implantation-induced local deformation correlated with photoluminescence enhancement of up to $30\times$ relative to supported regions, while the bright spot retained linewidth $165(3)$ MHz, ODMR contrast $2.18(3)\%$, and $T_1 = 10.3(3)\ \mu\text{s}$, comparable to darker supported regions [2509.00912]. In PbI$_2$/hBN heterostructures, donor-assisted energy transfer increased $V_B^-$ photoluminescence by $5$–$45\times$ and improved continuous-wave ODMR sensitivity by $190\%$ at $488$ nm and $325\%$ at $555$ nm, with a best sensitivity of about $76\ \mu\text{T Hz}^{-1/2}$ at $160\ \mu\text{W}$ [2602.02256]. Taken together, these results show that weak emission is not a fixed property of the defect alone; it can be strongly modified by photonic environment, strain symmetry breaking, and heterostructure-mediated energy transfer.

## 6. Relaxation, coherence, and quantum-technology roles

The longitudinal and transverse spin dynamics of $V_B^-$ are strongly environment-dependent. At room temperature, single- and double-quantum relaxation measurements gave $\Omega = 33.26 \pm 1.83~\text{kHz}$ and $\gamma = 81.60 \pm 9.12~\text{kHz}$ at $293$ K, with both rates increasing from $293$ to $393$ K and the double-quantum channel growing more rapidly [2506.14358]. Above $400$ K, the double-quantum rate was reported to become much greater than the single-quantum rate and may dominate the decoherence channel. A second-order spin-phonon model using effective phonon modes at $23.48$, $77.39$, and $165.75$ meV reproduced the temperature dependence and implicated higher-energy phonons as especially important [2506.14358].

Field- and temperature-dependent relaxometry over a broader regime revealed three distinct relaxation regimes. Between $15$ and $250$ K and up to $7$ T, low-temperature low-field behavior was dominated by spin-spin interactions and disorder-induced stretched exponential relaxation, high fields activated a first-order direct single-phonon process, and at $T \gtrsim 125$–$150\ \mathrm{K}$ a Raman-like two-phonon process with approximately $T^2$ scaling dominated [2507.16786]. In the high-field regime the relaxation rate scaled approximately as
$$
\Gamma \sim T\,H_{\mathrm{ext}}^{1.6},
$$
and the spin transition frequency could be pushed from a few GHz to roughly $200\ \text{GHz}$ by magnetic field alone. This suggests a route toward broadband and sub-terahertz relaxometry using atomically thin hBN sensors.

Direct coherence measurements with broadband microwave control showed that short coherence remains a major limitation in many current samples. Using an isotopically enriched $\mathrm{h}^{10}\mathrm{B}^{15}\mathrm{N}$ thin film and sub-GHz Rabi driving, Ramsey interference gave $T_2^* = 13.8 \pm 0.5$ ns with Gaussian-like decay, while Hahn echo gave $T_2 = 108.7 \pm 5.5$ ns and stretch factor $\alpha = 1.25 \pm 0.11$ [2604.27843]. The same work argued that strong, broadband pulses are necessary for reliable coherence extraction because the hyperfine-broadened ODMR spectrum otherwise produces substantial pulse errors.

First-principles many-body simulations indicate that the dominant decoherence mechanism changes with magnetic field in the dense nuclear-spin bath of hBN. A transition boundary was predicted at about $5020$ G for h-$^{10}$B$^{14}$N and $2050$ G for h-$^{11}$B$^{14}$N. Below that boundary, decoherence is governed by independent nuclear-spin dynamics and occurs within submicrosecond timescales; above it, pair-wise flip-flop transitions dominate and Hahn-echo $T_2$ reaches tens of microseconds, with reported values at $1$ T of $T_2 \approx 74\,\mu\mathrm{s}$ for h-$^{10}$B$^{14}$N and $T_2 \approx 23\,\mu\mathrm{s}$ for h-$^{11}$B$^{14}$N [2505.03306]. This is a theoretical result, but it provides a concrete operating principle: isotope engineering and sufficiently large magnetic fields should move $V_B^-$ into a more favorable coherence regime.

These relaxation and coherence properties frame the defect’s present and prospective applications. High optical polarization, including the model-based estimate $p \gtrsim 95\%$, directly benefits sensing and initialization protocols [2406.11953]. Excited-state level anti-crossings and hyperfine couplings are central to dynamic nuclear polarization, and the experimentally resolved remote $^{14}$N shells show that the defect can serve as a local NMR-like probe inside hBN itself [2112.02912][2504.06798]. On the quantum-information side, proposals for synchronous nuclear-spin control around the defect use the electron spin as a mediator for collective $X$, $Z$, and Hadamard gates on the three nearest nuclei, with GHZ-state preparation reported at fidelity $0.99$ in numerical analyses that include decoherence effects [2411.02828]. A plausible implication is that the long-term significance of $V_B^-$ will depend not on any single metric—brightness, polarization, or coherence—but on the degree to which fabrication, charge-state stabilization, nuclear-spin control, and photonic engineering can be made mutually compatible in realistic hBN devices.

Source: https://www.emergentmind.com/topics/negatively-charged-boron-vacancy-vb