Optically assigned vector magnetometry with nitrogen-vacancy center ensembles in diamond via polarization anisotropy
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Scientific Reports |
| Published | 2026-04-25 |
| Authors | Suppawit Ngonsamrong, Napoom Thooppanom, Phojet Sudjit, Sorawis Sangtawesin |
| DOI | 10.1038/s41598-026-50385-6 |
Eight polarization measurements can assign four nitrogen-vacancy axes and reconstruct a field vector without a bias field during later measurements. The experiment achieved a sensitivity of η ≈ 17 µT/√Hz and reconstructed a 9.64 ± 0.02 mT test field. The reconstructed direction retained the intrinsic global sign ambiguity.
Why NV ensembles need axis labels
Section titled “Why NV ensembles need axis labels”The negatively charged NV center consists of a substitutional nitrogen atom next to a missing carbon atom. Its ground state is a spin-1 triplet, with ms = 0, ±1 defined along the NV axis. Resonance splitting depends on field magnitude and orientation because nonlinear Zeeman effects combine with zero-field splitting.
NV ensembles measure different projections of one field through four crystallographic axes. For a generic field direction, those projections can produce eight ODMR resonances. A field along [100] gives all four axes the same projection, so only two resonances appear. A field along [111] gives one axis a linear Zeeman splitting and three equal off-axis splittings.
The eight resonances do not identify their own axes. Each transition pair must be assigned to an axis before the vector can be reconstructed. The usual solution applies a small bias field of known strength and orientation to break the degeneracy. This adds field hardware and a switch between field configurations during operation.
Polarization provides the axis label
Section titled “Polarization provides the axis label”Each NV center has two orthogonal optical transition dipoles in the plane normal to its axis. Excitation probability follows the squared projection of the laser electric field onto those dipoles. Oblique illumination therefore gives each crystallographic orientation a distinct polarization response.
The setup used a linearly polarized 532 nm laser. A half-wave plate rotated the linear polarization angle θp within the plane normal to the beam. The excitation model also included a fixed offset θ0 between the polarization basis and each NV dipole.
At an excitation direction of (θk,φk) = (5°,92°), two transitions reached their contrast extrema at the same polarization angle. The overlapping extrema prevented a clear axis assignment. Moving the excitation to (23°,92°) separated the four responses. Their contrast maxima occurred at θp = 117°, 171°, 18°, and 90°. These distinct extrema label the axes without using the field direction.
The experiment used one detector and standard microwave control. The excitation angle also determines whether these maxima remain distinct for a particular diamond facet.
Calibration ties the optical model to the sample
Section titled “Calibration ties the optical model to the sample”The method still requires an initial calibration with a known field. The authors applied |B| ≈ 10 mT at (θ,φ) = (55°,20°). Because the field direction was known, diagonalizing the NV Hamiltonian linked the four crystallographic axes to the four labeled resonance pairs.
The authors then fitted the effective excitation geometry to the measured polarization contrast. This calibration returned θk = 23°, φk = 92°, and θ0 = 129°. Direct geometric calibration would include mirror reflections and refraction into the diamond. The fit avoided those complications and produced a unique solution up to a minus sign. A simple initial observation resolved the sign.
For each later field, the authors measured the polarization response of four ODMR transitions. They calculated modeled contrast for each possible axis permutation. After rescaling the modeled and measured signals, they selected the assignment with the lowest variance-normalized mean squared error.
Eight polarization measurements supplied the extrema needed for unique assignment. Figure 3 demonstrates this procedure for fields directed at (96°,340°), (80°,340°), and (110°,20°). A fixed set of polarization angles can then be used without reapplying a reference field. Calibration can be completed before sample measurement.
Reconstructing field magnitude and direction
Section titled “Reconstructing field magnitude and direction”For each assigned NV axis, the pair of transition frequencies determines the axial field component. The same pair gives the angle ψj between the field and that axis. The calculation uses the zero-field splitting D = 2.87 GHz and the NV gyromagnetic ratio γNV = 28.02 MHz/mT.
At the representative (55°,20°) field, fitting eight Lorentzian peaks gave |B| = 9.64 ± 0.02 mT. The four extracted angles were ψj = 19.9 ± 0.4°, 85.5 ± 0.1°, 52.7 ± 0.1°, and 75.2 ± 0.1°. The reported magnitude uncertainty was consistent with the sensitivity extracted from the experimental parameters.
Each angle defines a cone of allowed field directions around one NV axis. The four cone curves were expressed on a unit sphere. Their common intersection determines the field direction. If the curves do not intersect exactly, the authors minimize the RMS residual between predicted axial projections and measured cos ψj values.
The reconstruction gave (θ,φ) = (53.5 ± 0.1°, 21.2 ± 0.1°). The opposite solution was (126.5 ± 0.1°, -158.8 ± 0.1°). The reported intervals are 95% confidence intervals.
Comparison with earlier bias-free methods
Section titled “Comparison with earlier bias-free methods”The paper compares the method with three earlier bias-free approaches. Chen et al. used an optical vortex beam, but their method applied only to single NV centers in (111)-oriented diamond. Münzhuber et al. required polarization-selective microwave excitation for a commonly used (100)-oriented diamond. Magaletti et al. controlled polarization during both excitation and detection and relied on specific diamond facet geometries.
The present method breaks the symmetry at the excitation stage by using oblique incidence. It uses one detector and standard microwave control. The authors state that it remains compatible with different diamond facets when the excitation angle is engineered correctly.
They also state that the assignment protocol can apply to NV magnetometry platforms with different coherence properties. The reported sensitivity remains limited by the ensemble linewidth rather than the axis assignment step. The method can support rapid acquisition when all eight ODMR transitions are well resolved.
Operating limits
Section titled “Operating limits”The demonstrated protocol requires all eight ODMR transitions to be well resolved. The spectrometer linewidth was approximately 30 MHz. The authors therefore required a transition separation of at least 120 MHz, corresponding to |B| ≈ 4.3 mT in this setup. These values describe the experiment rather than a general threshold for other NV sensors.
At that field range, the authors neglected strain-induced splitting, hyperfine structure, and temperature drift. They treat degenerate or overlapping transitions as separate edge cases in the Supplementary Information. The main demonstration therefore covers the resolvable-resonance regime.
The ±B ambiguity remains after axis assignment. It is intrinsic to NV ensemble vector magnetometry. Prior field constraints or auxiliary sensors would be needed when the global sign matters.
Oblique excitation can also complicate a miniaturized sensor head. Calibration still needs one known field, although later measurements do not require bias-field switching. These limits define where the current protocol is most reliable.
Engineering consequences for diamond selection
Section titled “Engineering consequences for diamond selection”Diamond selection must account for crystallographic orientation and optical access. The experiment used a (100)-oriented diamond with NV ensembles along four ⟨111⟩ directions. The excitation beam must meet that sample at an oblique angle that produces distinct polarization extrema. A half-wave plate then provides the controlled excitation polarization required for assignment.
Narrow ODMR linewidth also controls the practical field floor and sensitivity of this implementation. The method removes the need for a bias field during later measurements, although one known field is needed for initial calibration. The authors see this geometry as compatible with compact portable magnetometers and real-time magnetic field imaging.
A compact implementation still needs a stable oblique optical path, calibrated polarization control, and enough spectral resolution to separate each transition pair. Those requirements define the near-term engineering target for this reference-free vector magnetometry approach.
View Original Abstract
The nitrogen-vacancy (NV) center in diamond is a promising platform for quantum magnetometry due to its spin-dependent optical readout and ability to operate under ambient conditions. For vector magnetometry, an ensemble of NV centers aligned along four different crystallographic orientations can be used to determine both the magnitude and direction of a magnetic field via optically detected magnetic resonance (ODMR) spectra. However, due to the symmetry and equivalent crystallographic planes of the diamond lattice, full reconstruction of the magnetic field vector typically requires an external bias field to lift the degeneracy between NV orientations. Here, we demonstrate that this degeneracy can be lifted by illuminating the diamond at an oblique angle and analyzing the resulting polarization-dependent ODMR spectra. The unique dipole orientation of the NV centers enables the unambiguous assignment of ODMR transitions to specific NV axes, and we show that as few as eight measurements are sufficient to reconstruct the full magnetic field vector. Our method enables rapid vector magnetometry of arbitrary magnetic fields without the need to apply or vary a reference field, especially when all the ODMR transitions are well-resolved. This approach opens the door to more compact and portable magnetometers and may facilitate real-time magnetic field imaging in materials science and biological systems.
Single Crystal Cvd Diamond Plate
View Specs →
Boron Doped Single Crystal Diamond Plate
View Specs →