Methods for traceable scanning magnetometry using single nitrogen vacancy centers in diamond: determining orientation, distance and localization

Technical Summary: Methods for Traceable Scanning Magnetometry Using Single Nitrogen Vacancy Centers

Problem Statement
Scanning Nitrogen-Vacancy (NV) center magnetometry (SNVM) enables nanoscale, quantitative imaging of magnetic stray fields. However, the quantitative accuracy of these measurements is frequently limited by the imprecise knowledge of two critical geometric parameters: the distance between the NV center and the sample ($d_{NV}$) and the orientation of the NV's high-symmetry axis relative to the laboratory frame (specifically the azimuthal angle $\phi$). While the polar angle $\theta$ is often fixed by the crystal orientation of the diamond nanopillar (e.g., $54.7^\circ$ for a (100) apex), the azimuthal angle $\phi$ is statistically determined by the NV's alignment along one of the four $\langle 111 \rangle$ crystallographic directions. Furthermore, $d_{NV}$ is subject to uncertainty due to ion implantation straggle and variable stand-off distances ($d_{AFM}$) caused by probe contamination, surface roughness, or AFM feedback instability. Traditionally, determining these parameters requires external vector magnets or complex calibration setups, which are not universally available. Additionally, the lateral position of the NV within the nanopillar affects fluorescence collection efficiency and the correlation of magnetic data with topography, yet this is rarely characterized.

Methodology
The authors propose a robust, three-step calibration protocol that utilizes simple, commercially available micro-patterned structures and standard SNVM hardware, eliminating the need for external vector magnets.

1. Distance Calibration ($d_{NV}$): The authors utilize Perpendicular Magnetic Anisotropy (PMA) micro-patterned stripes. By scanning the NV probe across the edge of a semi-infinite magnetic stripe, they measure the stray field profile. The field components ($B_x, B_z$) are modeled analytically as a function of the stand-off distance. By fitting the measured field projection ($B_{NV}$) against these analytic expressions, $d_{NV}$ is extracted as a free parameter alongside the magnetization amplitude. Multiple line scans are averaged to improve statistical reliability.
2. Orientation Calibration ($\phi$): Following the determination of $d_{NV}$, the authors image micron-sized PMA discs. Unlike stripes, discs generate a two-dimensional stray field with both in-plane and out-of-plane components. The azimuthal dependence of the measured field along the disc edge is modeled using a current-loop approximation. The angular position of the field extremum (a "bright arc" on the disc edge) directly reveals the azimuthal angle $\phi$. The magnetization direction ($\sigma$) is simultaneously resolved by comparing the field at the disc center to the background bias field.
3. Lateral Localization and Contamination Detection: The authors employ "inverse AFM," scanning the diamond nanopillar probe over an ultra-sharp, commercially available silicon needle. This process serves two purposes:
   • Contamination Detection: The topographic scan reveals particles attached to the probe apex that would otherwise alter $d_{NV}$.
   • Lateral NV Localization: As the silicon needle scans across the diamond pillar, the NV fluorescence is monitored. The authors observe a distinct dip in fluorescence intensity when the needle passes over the NV center. This dip is attributed to fluorescence quenching and altered coupling to photonic modes caused by the high-refractive-index silicon needle entering the optical near-field of the NV. The position of this dip indicates the lateral ($x-y$) location of the NV within the pillar.

Key Results

• Distance Determination: Using PMA stripes, the authors determined a mean NV-to-sample distance of $d_{NV} = 31.5 \text{ nm}$ with a standard deviation of $1.95 \text{ nm}$ (approx. 6% relative uncertainty). This value significantly exceeds the nominal implantation depth ($10 \text{ nm}$), highlighting the substantial contribution of the stand-off distance ($d_{AFM}$).
• Orientation Determination: The azimuthal angle $\phi$ was determined with a precision of $3^\circ$. For the specific probe used, the measured mean $\phi \approx 97.1^\circ$ aligns closely with the expected $90^\circ$ projection of the $\langle 111 \rangle$ axis onto the (100) plane, confirming the method's validity.
• Lateral Localization: Inverse AFM scans revealed that the lateral position of the NV center varies significantly between probes. In one case, the NV was centered; in another, it was shifted by approximately $80 \text{ nm}$. The fluorescence dip contrast and width (Full Width at Half Maximum) varied depending on the probe geometry and the depth of the NV, with simulations suggesting the dip arises from near-field quenching by the silicon needle.
• Material Characterization: The calibration process also yielded saturation magnetization ($M_s$) values. The authors noted an $\approx 11\%$ discrepancy in $M_s$ between stripe and disc measurements, attributing this to fabrication-induced damage (edge effects) and differences in lithography masks (direct laser writing vs. electron-beam lithography).

Significance and Claims
The paper claims to provide a traceable, self-contained calibration method for SNVM that does not rely on external vector magnets. By using standard PMA patterns and commercially available silicon needles, the authors demonstrate that $d_{NV}$ and $\phi$ can be determined directly within the measurement setup. This allows for the immediate correction of magnetic field reconstructions and magnetization calculations, which are otherwise prone to error due to unknown geometric parameters.

Furthermore, the work highlights the critical role of probe contamination and the variability of NV placement. The "inverse AFM" technique is presented as a practical tool to detect probe contamination (which alters $d_{NV}$) and to map the lateral position of the NV, ensuring that magnetic data can be accurately correlated with topographic features. The authors emphasize that while the calibration is robust, it should be performed before each new set of measurements or upon probe exchange to account for time-dependent changes in the probe's condition. The study concludes that these methods are compatible with both ambient and cryogenic environments, provided the sample chamber has sufficient space for the calibration structures.