Boron-Vacancy Centers in Boron Nitride

• Boron-vacancy centers are point defects in boron nitride created by a missing boron atom that generates localized, spin-active color centers.
• Stacking order and symmetry engineering enable precise control over their optical transitions and brightness, transforming dim emitters into efficient near-infrared sources.
• Materials engineering through isotopic purification and deformation enhances these defects for applications in nanoscale magnetometry, quantum sensing, and spin-photon interfaces.

Searching arXiv for papers on boron-vacancy centers in boron nitride. Boron-vacancy centers are point defects in boron nitride in which a boron atom is removed from the lattice. In the negatively charged state, denoted $V_\mathrm{B}^-$, the defect forms a localized spin-active color center in a wide-bandgap host and has become a prototypical platform for optically addressable spin defects in two-dimensional materials. Across hexagonal boron nitride (hBN), rhombohedral boron nitride (rBN), and related engineered environments, $V_\mathrm{B}^-$ combines a triplet spin ground state, spin-dependent photodynamics, strong coupling to lattice vibrations, and compatibility with van der Waals heterostructures. Recent work has shown that its properties depend not only on defect identity but also on stacking sequence, isotopic composition, deformation, and local charge environment, making boron-vacancy centers a central case study in symmetry engineering and spin-defect design in low-dimensional hosts (Estaji et al., 22 Mar 2026).

1. Defect identity, host polytypes, and symmetry

The negatively charged boron vacancy $V_\mathrm{B}^-$ is a missing boron atom in boron nitride with an extra electron relative to the neutral vacancy. In hBN and rBN, the defect is surrounded by three neighboring nitrogen atoms whose dangling bonds form localized defect orbitals in the band gap. The ground state is a spin triplet with $S=1$, and the spin density is localized on the three neighboring nitrogen atoms (Estaji et al., 22 Mar 2026).

Layered $sp^2$ boron nitride occurs in more than one stable polytype. Hexagonal BN has $AA'$ stacking, in which boron atoms in one layer sit above nitrogen atoms in the next layer and vice versa, producing an effectively inversion-symmetric crystal. For an in-plane defect in a bulk-like hBN environment, the point group is $D_{3h}$, and the defective layer itself defines a mirror plane perpendicular to the c-axis. Rhombohedral BN has $ABC$ stacking; the layers are laterally shifted so that the crystal is non-centrosymmetric, and the local point group of the defect is reduced to $C_{3v}$ because the mirror symmetry of the defective layer is broken by adjacent layers (Estaji et al., 22 Mar 2026).

This change from $D_{3h}$ to $V_\mathrm{B}^-$0 is not merely crystallographic. In hBN, defect states can be labeled by mirror parity, $V_\mathrm{B}^-$1 or $V_\mathrm{B}^-$2, and this imposes strict optical selection rules. In rBN, these parity labels are no longer good quantum numbers. This suggests that stacking sequence functions as a control parameter for defect photophysics without changing the local chemical identity of the vacancy (Estaji et al., 22 Mar 2026).

The local orbital structure is built from the three nitrogen dangling bonds. In both hBN and rBN they form a fully symmetric $V_\mathrm{B}^-$3-type orbital and a doubly degenerate $V_\mathrm{B}^-$4 pair. In hBN, the corresponding many-body ground state is usually denoted $V_\mathrm{B}^-$5; in rBN it becomes $V_\mathrm{B}^-$6 (Estaji et al., 22 Mar 2026). Earlier work on hBN treated the spin defect as nominally trigonal around the vacancy, with local strain or electric fields producing a finite transverse splitting (Kianinia et al., 2020).

2. Electronic structure and optical transitions

The optical structure of $V_\mathrm{B}^-$7 is strongly host-dependent. In hBN, the relevant electronic levels include a triplet ground state $V_\mathrm{B}^-$8, higher excited triplet $V_\mathrm{B}^-$9, lower excited triplet $V_\mathrm{B}^-$0, and one or more intermediate singlet states participating in intersystem crossing (Clua-Provost et al., 2024). The defect emits a broad near-infrared photoluminescence band, and its optical transition is intrinsically weak, with a radiative rate $V_\mathrm{B}^-$1 (Clua-Provost et al., 2024).

In hBN, the mirror symmetry of $V_\mathrm{B}^-$2 makes the lowest triplet transition dipole-forbidden. The ground state is $V_\mathrm{B}^-$3, while the lowest triplet excited states are $V_\mathrm{B}^-$4 and $V_\mathrm{B}^-$5. Electric-dipole transitions between opposite mirror parity are forbidden, so the lowest-energy emission is phonon-assisted through Herzberg–Teller activation. As a result, coherent zero-phonon-line emission is essentially absent, the spectrum is dominated by phonon sidebands, radiative lifetimes are $V_\mathrm{B}^-$6, and quantum efficiency is $V_\mathrm{B}^-$7 (Estaji et al., 22 Mar 2026).

Cavity-enhanced spectroscopy in hBN identified the room-temperature zero-phonon line of $V_\mathrm{B}^-$8 at $V_\mathrm{B}^-$9, providing a direct spectroscopic anchor for the otherwise weak purely electronic transition (Qian et al., 2022). The same work interpreted the broad emission as arising from strong electron-phonon coupling and Jahn–Teller mixing, with a Huang–Rhys factor $S=1$0 from earlier theoretical modeling (Qian et al., 2022).

In rBN, first-principles GW+BSE calculations found that the lowest triplet excitation is a $S=1$1 state in the near infrared and that the transition from the $S=1$2 ground state is optically allowed because the mirror-symmetry selection rule no longer applies. The first triplet excited state lies at about $S=1$3 eV, corresponding to a zero-phonon line near $S=1$4 nm, and has a transition dipole moment $S=1$5 (Estaji et al., 22 Mar 2026). The radiative lifetime is then estimated as $S=1$6, with the authors stating “an approximately two-orders-of-magnitude increase in the radiative decay rate” relative to hBN (Estaji et al., 22 Mar 2026).

The optical brightening in rBN is therefore attributed to reduced crystal-field symmetry. A plausible implication is that rBN converts a defect that is optically dim in hBN into a bright near-infrared emitter while preserving the same fundamental defect chemistry (Estaji et al., 22 Mar 2026).

3. Vibronic structure, photodynamics, and charge-state behavior

The photoluminescence of boron-vacancy centers is governed by strong vibronic coupling. In rBN, Huang–Rhys theory yields a total Huang–Rhys factor

$S=1$7

with

$S=1$8

showing that the phonon sideband is dominated by Jahn–Teller-active $S=1$9 modes (Estaji et al., 22 Mar 2026). At $sp^2$0 K the calculated photoluminescence is a broad band centered near $sp^2$1 eV, whereas at $sp^2$2 K a pronounced zero-phonon line emerges with a structured phonon sideband (Estaji et al., 22 Mar 2026).

In hBN, time-resolved studies established a seven-level model consisting of a triplet ground state, a lower triplet excited state reached after internal conversion from $sp^2$3, and an effective metastable singlet manifold (Clua-Provost et al., 2024). Optical pumping is spin-conserving, radiative decay is weak and phonon-assisted, and intersystem crossing is spin-dependent. The rates $sp^2$4 and $sp^2$5 polarize population into the bright $sp^2$6 state and generate ODMR contrast (Clua-Provost et al., 2024).

Room-temperature excited-state lifetimes in hBN were extracted as

$sp^2$7

with optical spin polarization

$sp^2$8

At $sp^2$9 K, the corresponding values become

$AA'$0

(Clua-Provost et al., 2024). The metastable lifetime is short, about $AA'$1 at room temperature and $AA'$2 at $AA'$3 K, which helps sustain measurable photoluminescence despite low radiative efficiency (Clua-Provost et al., 2024).

A later study directly measured the singlet lifetime in neutron-irradiated hBN flakes using a pulse-recovery protocol and obtained

$AA'$4

across 16 sub-micron flakes (Escalante et al., 7 Apr 2025). That work also found that a simple seven-level model fails to reproduce all power-dependent photoluminescence transients and introduced a nine-level model with an additional two-level manifold, interpreted as another electronic or charge state, possibly $AA'$5 (Escalante et al., 7 Apr 2025). This suggests that charge conversion is part of the photodynamics, especially under stronger optical excitation.

Evidence for charge-state complexity also appears in implantation studies. Focused He-ion irradiation produces many boron vacancies, but only a small fraction are in the optically active $AA'$6 state. By combining ODMR splitting with a microscopic charge model and molecular-dynamics estimates of total vacancy density, one study reported a lower bound for the fraction of all vacancies in the negatively charged, optically active state of $AA'$7 (Carbone et al., 30 Jan 2025). A related density-dependent study found that at high implantation dosage only a small portion of created boron vacancy defects are in the desired negatively charged state (Gong et al., 2022).

4. Spin structure, hyperfine coupling, and coherence

The ground state of $AA'$8 is an $AA'$9 triplet with a zero-field splitting near $D_{3h}$0 GHz. In rBN, ab initio calculations gave

$D_{3h}$1

in excellent agreement with experiment,

$D_{3h}$2

(Estaji et al., 22 Mar 2026). In hBN, room-temperature ODMR typically yields

$D_{3h}$3

with a low-temperature value

$D_{3h}$4

in the photodynamics work (Clua-Provost et al., 2024). An earlier engineering study extracted

$D_{3h}$5

from two zero-field ODMR lines at $D_{3h}$6 and $D_{3h}$7 (Kianinia et al., 2020).

The hyperfine interaction is dominated by the three nearest nitrogen nuclei. In rBN, calculations yielded a dominant component

$D_{3h}$8

with a characteristic seven-line ODMR pattern for three equivalent $D_{3h}$9N nuclei (Estaji et al., 22 Mar 2026). In isotopically purified $ABC$0N hosts, the hyperfine structure simplifies substantially. For h$ABC$1B$ABC$2N the splitting is

$ABC$3

whereas for h$ABC$4B$ABC$5N it becomes

$ABC$6

and the multiplet reduces from seven lines to four because $ABC$7 has $ABC$8 rather than $ABC$9 (Clua-Provost et al., 2023). Boron isotope engineering also narrows ESR linewidths: $C_{3v}$0 versus $C_{3v}$1 (Clua-Provost et al., 2023).

Coherent manipulation has been demonstrated in dense $C_{3v}$2 ensembles. Echo-limited coherence times of about $C_{3v}$3 ns were extended by more than fivefold using advanced dynamical decoupling, and comparison between XY-8 and DROID sequences showed that dipolar interactions within the $C_{3v}$4 ensemble substantially contribute to decoherence (Gong et al., 2022). That work also used the many-body dynamics to estimate $C_{3v}$5 density and extracted a transverse electric-field susceptibility

$C_{3v}$6

(Gong et al., 2022).

Recent theoretical work on spin-lattice relaxation across BN polytypes found that the room-temperature $C_{3v}$7 of $C_{3v}$8 in monolayer BN and hBN are nearly identical, while rBN unexpectedly has a longer $C_{3v}$9 despite reduced symmetry (Estaji et al., 31 Mar 2025). The calculated hBN value is on the order of

$D_{3h}$0

consistent with experiment, and the dominant mechanism is a two-phonon Raman process with

$D_{3h}$1

over the 150–300 K range (Estaji et al., 31 Mar 2025). The longer $D_{3h}$2 in rBN was attributed to stiffer out-of-plane phonon modes (Estaji et al., 31 Mar 2025).

5. Defect creation, isotopic control, and materials engineering

Creation of boron-vacancy centers in hBN has been demonstrated with several irradiation methods. Focused ion beams of xenon, argon, and nitrogen at $D_{3h}$3 keV produce ensembles of $D_{3h}$4 in exfoliated hBN with nanoscale spatial precision (Kianinia et al., 2020). Xe implantation yields vacancy profiles peaking at about $D_{3h}$5 nm depth, Ar at about $D_{3h}$6 nm, and N at about $D_{3h}$7 nm, according to SRIM simulations (Kianinia et al., 2020). Nitrogen implantation at equal fluence gives higher photoluminescence than Xe and Ar, attributed to a more homogeneous vacancy distribution and reduced sputtering (Kianinia et al., 2020).

Focused He-ion irradiation at $D_{3h}$8 keV also systematically generates $D_{3h}$9 ensembles and was used to quantify the yield of optically active centers (Carbone et al., 30 Jan 2025). At moderate doses, the defect-related photoluminescence increases monotonically; at the highest doses, a strong background and a defect-activated Raman line near $V_\mathrm{B}^-$00 appear, indicating substantial lattice damage (Carbone et al., 30 Jan 2025).

Isotopic purification provides another axis of materials control. Enrichment in $V_\mathrm{B}^-$01N simplifies the hyperfine spectrum and removes quadrupolar terms from the nearest-neighbor nuclei, while enrichment in $V_\mathrm{B}^-$02B narrows ESR lines by reducing second-neighbor hyperfine broadening (Clua-Provost et al., 2023). The combination h$V_\mathrm{B}^-$03B$V_\mathrm{B}^-$04N was identified as the optimal host for $V_\mathrm{B}^-$05 spin defects in quantum technologies (Clua-Provost et al., 2023).

A different form of engineering is stacking control. The rBN study demonstrated that changing only the stacking sequence from $V_\mathrm{B}^-$06 to $V_\mathrm{B}^-$07 can increase brightness by at least one order of magnitude while preserving comparable or improved spin properties (Estaji et al., 22 Mar 2026). Another route is deformation engineering: suspended, locally deformed regions of hBN show photoluminescence enhancement of up to $V_\mathrm{B}^-$08 times relative to supported regions while preserving ODMR contrast, linewidth, and spin lifetime (Geng et al., 31 Aug 2025). In that work, zero-field ODMR, Raman spectroscopy, and Kelvin probe force microscopy all indicated that local deformation breaks the symmetry and activates otherwise forbidden or weak optical transitions (Geng et al., 31 Aug 2025). This suggests that symmetry lowering, whether by stacking or deformation, is a recurring strategy for brightening $V_\mathrm{B}^-$09.

6. Applications and broader significance

Boron-vacancy centers are primarily studied as quantum sensors and spin-photon interfaces in two-dimensional hosts. Their $V_\mathrm{B}^-$10 ground state, room-temperature ODMR, near-surface location, and compatibility with van der Waals heterostructures make them attractive for nanoscale magnetometry, electric-field sensing, strain sensing, and thermometry (Clua-Provost et al., 2024). The 2D geometry allows defects to reside within nanometers of a target sample, which is a distinctive advantage over defects in bulk hosts (Escalante et al., 7 Apr 2025).

Temperature-dependent spin-phonon studies showed that the zero-field splitting of $V_\mathrm{B}^-$11 changes by more than $V_\mathrm{B}^-$12 MHz between $V_\mathrm{B}^-$13 and $V_\mathrm{B}^-$14 K, with a room-temperature susceptibility of about

$V_\mathrm{B}^-$15

in isotopically purified h$V_\mathrm{B}^-$16B$V_\mathrm{B}^-$17N and

$V_\mathrm{B}^-$18

in natural hBN (Liu et al., 2024). This large temperature response makes $V_\mathrm{B}^-$19 a nanoscale thermometer with estimated sensitivity

$V_\mathrm{B}^-$20

at room temperature (Liu et al., 2024).

High-field relaxometry extends the sensing range further. Over temperatures of $V_\mathrm{B}^-$21–$V_\mathrm{B}^-$22 K and magnetic fields up to $V_\mathrm{B}^-$23 T, the longitudinal relaxation of $V_\mathrm{B}^-$24 ensembles in hBN exhibits distinct regimes: low-temperature spin-spin-interaction-driven dynamics with stretched exponentials, a universal

$V_\mathrm{B}^-$25

high-temperature regime, and a direct single-phonon contribution at high fields scaling as

$V_\mathrm{B}^-$26

(Solanki et al., 22 Jul 2025). This establishes $V_\mathrm{B}^-$27 as a candidate for high-field, sub-terahertz quantum sensing (Solanki et al., 22 Jul 2025).

Nuclear-spin functionality has also been developed around the defect. Optical pumping in h$V_\mathrm{B}^-$28B$V_\mathrm{B}^-$29N produces $V_\mathrm{B}^-$30N nuclear polarization of about $V_\mathrm{B}^-$31 near the ground-state level anticrossing (Clua-Provost et al., 2023). At the control-theory level, a recent proposal showed how a single $V_\mathrm{B}^-$32 electron spin could implement synchronous three-qubit $V_\mathrm{B}^-$33, $V_\mathrm{B}^-$34, and Hadamard gates on the three nearest nuclear spins and prepare nuclear GHZ states with fidelity $V_\mathrm{B}^-$35 (Sakuldee et al., 2024). This suggests that the intrinsic nuclear environment of the boron vacancy can function not only as a decoherence source but also as a quantum register.

A persistent misconception is that boron-vacancy centers in BN are defined solely by the vacancy itself. The cumulative literature shows that their observable properties depend strongly on host stacking, isotopic composition, implantation history, deformation, and charge environment. A plausible implication is that the defect identity sets the basic spin and orbital structure, while the host symmetry and local environment determine whether the center is dim or bright, narrow or broad, and limited by spin-bath noise or by spin-phonon relaxation (Estaji et al., 22 Mar 2026).

In that broader sense, boron-vacancy centers are exemplary of defect engineering in two-dimensional materials: the same chemically simple vacancy can serve as a weak ensemble emitter in hBN, a bright near-infrared center in rBN, a mechanically activated emitter in deformed membranes, a high-field relaxometric probe, or the central spin of a nuclear-spin processor, depending on how the surrounding boron nitride lattice is designed (Estaji et al., 22 Mar 2026).