Skip to content

Simultaneous determination of local magnetic fields and sensor orientation with nitrogen-vacancy centers in nanodiamond

FieldValue
JournalPhysical Review Applied
Published2026-09-03
AuthorsYizhou Wang, Haochen Shen, Zhongyuan Liu, Yue Jin Yu, Shengwang Du
DOI10.1103/zh44-55ms

Four non-coplanar bias magnetic fields were sufficient to recover both an unknown local vector magnetic field and the crystallographic orientation of nitrogen-vacancy (NV) centers in a nanodiamond. In the controlled bulk-diamond test, mean axis deviation was below 0.5° and reconstructed field consistency was below 2 μT. For a single nanodiamond, axis consistency was about 1° and local-field standard deviation was about 50 μT.

Four fields resolve orientation and field together

Section titled “Four fields resolve orientation and field together”

Most orientation-tracking methods assume that the external field at the sensor is already known. Magnetometry then needs the NV orientation, while orientation reconstruction relies on a known field.

The authors remove this dependency by measuring the total field under several known bias configurations. Each measurement constrains the unknown local field to a sphere centered opposite the applied bias. One field leaves the full sphere, two leave a ring, and three leave two mirror-symmetric points. A fourth field with a non-coplanar sphere center selects one point and removes the symmetry.

Four is therefore the minimum for unambiguous joint reconstruction without a prior local-field magnitude. This matches the four-field count of the prior orientation-only method, so vector-field recovery does not raise the minimum measurement count.

Each tested single-crystal particle hosted an ensemble of negatively charged NV centers. The four possible axes follow tetrahedral directions in the diamond lattice. The total field is the sum of the applied bias and local field, and each axis measures its projection to leading order.

Excitation used a 532 nm laser. NV fluorescence included a 637 nm zero-phonon line and a phonon sideband extending to about 800 nm. ODMR records two microwave transitions from the m_s = 0 state to the m_s = +1 and m_s = -1 states.

The spin-1 Hamiltonian used Dgs = 2π × 2.87 GHz and γe = 2π × 28.0 MHz/mT. The model treated the magnetic field as dominant and neglected small local strain and electric-field mixing between the m_s = ±1 levels.

Three-dimensional electromagnetic coils supplied the static bias fields. A Monte Carlo multistart fit extracted each ODMR transition splitting. MATLAB Multistart then solved a least-squares problem over three orientation angles and three local-field components.

Before the joint solve, tetrahedral symmetry allowed each total-field magnitude to be calculated from four projections without knowing the axes. The supplement recommends bias directions with pairwise unit-vector dot products near -1/3.

The controlled proof-of-concept used a [100]-cut bulk plate with known crystallographic axes. The authors applied 24 bias fields of approximately 1 mT and ensured they were not coplanar.

Across 50 random four-field subsets, the mean great-circle axis deviation was below 0.5°. The reconstructed axes agreed with the known crystal orientation. The local field matched the geomagnetic field in the U.S. Midwest, with sub-2 μT consistency across reconstructions.

The bulk ODMR linewidth was about 0.6 MHz. The bulk diamond was isotopically purified to carbon-13 concentrations below 10^-4, which reduced nuclear spin noise. The paper attributes the lower field spread partly to this purification and the bulk sample’s lower surface-to-volume ratio.

The sub-degree axis result is comparable to earlier biological orientation-tracking results. The bulk experiment also showed that vector-field reconstruction can operate without a separate calibration of the local field.

Individual single-crystal nanodiamonds tested the unknown-axis case. Figure 1 shows particles about 750 nm across, while the sample preparation describes particles around 1 μm. Each particle had a random crystallographic orientation and an ensemble of NV centers.

The authors applied 34 fields of about 5 to 10 mT to separate the eight resonances. The ODMR linewidth was about 10 MHz, compared with 0.6 MHz in bulk diamond. This broadening raised the noise floor and reduced field precision.

SampleApplied bias fieldsFields per reconstructionOrientation resultLocal field result
Bulk [100]-cut diamond24 fields at about 1 mT4, across 50 subsetsMean deviation below 0.5°Sub-2 μT consistency
Single nanodiamond34 fields at about 5 to 10 mT4 for axes, 10 for fieldAbout 1°Standard deviation about 50 μT

Four-field runs produced axis consistency near 1°. Ten fields gave local-field components of -7, -33, and -7 μT along x, y, and z. Their respective uncertainties were 36, 11, and 24 μT.

Two nanodiamonds on the same substrate provided a cross-particle check. ND1 gave components of 11, 10, and 46 μT, with uncertainties of 11, 14, and 27 μT. ND2 gave -20, 2, and 68 μT, with uncertainties of 28, 26, and 41 μT. The authors report that these estimates agree within experimental uncertainty.

Synthetic bulk spectra started from a local field of about 50 μT toward -z and contained eight peaks. Tests added Gaussian noise from 0.1 to 0.5 MHz and used 4 to 10 of the 24 bulk fields. Nanodiamond simulations used noise from 1 to 5 MHz and 4 to 20 of the 34 measured fields.

Both field error and axis error increased approximately linearly with injected noise. Adding fields improved accuracy, but the nanodiamond benefit saturated at about ten fields per reconstruction.

At 0.3 MHz noise, the bulk simulation gave a mean field deviation below 10 μT and a mean axis deviation below 0.1°. Its component standard deviations stayed below 10 μT, and its axis standard deviation stayed below 0.5°.

Scaling all bias fields by factors of 2 to 10 improved axis determination but gave only marginal local-field gains. The authors explain that axis precision depends on relative bias directions. Local-field precision depends on the intersection of total-field vectors, which remains noise-limited once the bias field is much larger than the local field.

The reconstruction assumes quasi-static conditions. Nanodiamond orientation and local field must remain approximately constant across the spectra used for one reconstruction.

The Hamiltonian also neglects small strain and electric-field mixing between the m_s = ±1 levels. The local field may change with the applied bias. For ferritin-like moments, a low-field estimate suggests bias-induced alignment is small at room temperature and millitesla fields. The authors leave systematic quantification for other magnetic species and operating conditions unresolved.

The paper identifies the carbon-13 nuclear spin bath, strain variation, and structural inhomogeneity as the main nanodiamond noise sources. P1 centers and other paramagnetic defects may also contribute to the spin bath. Nanodiamonds also have a higher surface-to-volume ratio than bulk diamond, increasing surface-spin contributions.

Peak overlap is a separate measurement limit. Random orientations can produce six or fewer visible peaks instead of eight, and overlapping resonances become hard to separate.

For sensor selection, narrow ODMR linewidth and low spin-bath noise matter more for field precision than particle orientation alone. For calibration, four non-coplanar fields are the minimum, while ten fields often reduced nanodiamond field error. The authors expect the combined readout to support ferritin-related magnetic measurements in living systems and nanoscale magnetism measurements in quantum materials.

The paper’s sensing element is a single-crystal nanodiamond particle, about 0.75 to 1 μm in the reported samples, with an ensemble of negatively charged NV centers. 6CCVD’s diamond materials catalog lists single-crystal diamond grown by MPCVD, which is the base material family for this application. Use engineering support to confirm particle format, NV content, and surface condition before substituting a catalog material for the paper’s sensing particles.

View Original Abstract

Nitrogen-vacancy (NV) centers in nanodiamonds have emerged as a promising quantum sensing platform for biomedical imaging applications, yet random orientations of individual particles present significant challenges in large-scale sensor calibration. In this study, we demonstrate a novel approach to simultaneously determine each particle’s crystallographic axes and the surrounding local vector magnetic field. Specifically, a minimum of four distinct bias fields is required to unambiguously extract both the orientation and the local field. We validate our method experimentally using NV centers in two scenarios: (1) in a bulk diamond with known crystal orientation as a proof of concept, and (2) on various single nanodiamonds to mimic real-world applications. Our work represents a crucial step towards unlocking the full potential of nanodiamonds for advanced applications such as in-situ biomedical imaging and nanoscale sensing in complex environments.