Skip to content

Circular near-field mapping of microwave magnetic fields using a projection-based diamond fiber probe

FieldValue
JournalScientific Reports
Published2026-08-07
AuthorsRaynd Huang, Jeson Chen
DOI10.1038/s41598-026-64544-2

A projection-based diamond fiber probe mapped microwave magnetic near fields at 2.4 and 2.87 GHz with approximately 2 mm spatial resolution. Its optically detected magnetic resonance signal reports a transition-selective projection defined by the nitrogen-vacancy quantization axis, not total magnetic-field magnitude. Tests on a double-ridge horn and a helical antenna agreed with HFSS in their main spatial trends.

Projection measured by NV spin transitions

Section titled “Projection measured by NV spin transitions”

The nitrogen-vacancy center in diamond has an (S=1) ground state with a zero-field splitting of 2870 MHz. Its electric spin splitting is approximately 4 MHz, and its gyromagnetic ratio is 2.8 MHz/G.

Under a bias field, the (m_s=\pm1) states split and couple selectively to opposite circular microwave fields. The transverse components are defined as:

[ B_{\sigma^\pm}=\frac{1}{2}(B_{x’}\mp iB_{y’}) ]

The transition amplitudes depend on both transverse field amplitude and the relative phase of the orthogonal components. After power calibration, ODMR contrast tracks these transition-selective projections.

This basis differs from the antenna convention that defines circular polarization relative to propagation. A near field has no unique propagation axis. The NV quantization axis instead provides a physical reference set by the spin Hamiltonian and bias field. Reversing the bias, flipping the diamond, or rotating the probe can alter the apparent helicity without changing the electromagnetic field.

The sensing element was a type Ib CVD diamond slab. A [100] slab measuring 3 × 3 × 0.5 mm was used at 2.87 GHz. A [111] slab measuring 2 × 2 × 0.4 mm served at other frequencies, including 2.4 GHz. Their estimated intrinsic nitrogen concentration was 30 to 180 ppm.

Each slab was fixed to a 3D-printed fiber-optic ring and attached to an SMA fiber ferrule. A 532 nm continuous-wave laser provided optical pumping through a multimode fiber circulator. Back-emitted fluorescence passed through a 550 nm long-pass filter to a CCD spectrometer.

The system integrated fluorescence counts from 600 to 800 nm over 50 ms. ODMR contrast was calculated by dividing the photon count with microwave excitation by the count without it. The microwave source generated 2 to 3 GHz S-band signals and used a 16 W amplifier.

Diamond contains four NV crystallographic classes with axes along [111] and three related orientations. A bias field near one axis separates their Zeeman branches.

For a field aligned with [111], the aligned class approached 2.4 GHz at about 168 G. The other three classes remained near 2.81 GHz and contributed negligibly at 2.4 GHz. This spectral separation selects a well-defined projection axis.

A coaxial ring-shaped neodymium magnet generated the bias field along the SMA axis. Its axial position controlled the Zeeman splitting and set the microwave resonance. At 2.87 GHz, the [100] slab produced contributions from all four NV orientations. That configuration supported stable calibration without magnetic-field alignment.

Quantitative calibration used a straight microstrip transmission line at 2.87 GHz. The laser delivered 20 mW at the fiber output, and ODMR contrast followed a sigmoidal dependence on input power in dBm.

Within ±6 dB around (P_0), the response was approximately linear on a dB scale. Its maximum deviation from a linear fit was below 1% over that 12 dB range. The contrast reached approximately 10% at 2.87 GHz and 1.2% at 2.4 GHz. The reported linewidths were 2 to 4 MHz.

ODMR contrast changed little over a broad range of optical excitation powers. It also showed minimal dependence on the transmission line’s rotation angle relative to the probe. The microwave measurement operated continuously and did not resolve phase.

A scan across a 2 mm-wide line produced an approximately 2 mm response width. This width was comparable to the sensing diamond dimensions, so the diamond size set the current resolution.

The local microwave field was not independently calibrated. The study therefore reports no absolute magnetic-field sensitivity. It treats contrast variation within the quasi-linear region as relative projected-field variation in dB.

Linear-field validation used a broadband double-ridge horn antenna, model LB10180, operating at 2.87 GHz. The probe stayed on a fixed holder while the antenna field was scanned. Every map used microwave power within ±6 dB around (P_0).

HFSS models used the actual antenna dimensions, radiation boundaries, and wave-port excitation normalized to 1 W. The ODMR map agreed well with the simulated magnetic-field magnitude at the same frequency. Both showed a pronounced concentration along the central ridge and gradual decay toward the horn aperture.

For this predominantly linear field, the two circular projections contribute symmetrically. The measured response is therefore magnitude-like and supports direct comparison of spatial features. It remains a transition-selective projection rather than a full vector magnitude, so quantitative differences from simulation remain expected.

At 2.4 GHz, the authors mapped a helical antenna that produces circularly polarized far-field radiation under axial-mode operation. The base-plate map showed strong response beside the helix.

Its spatial pattern agreed qualitatively with one simulated circular component defined relative to the antenna symmetry axis. A slight left-right asymmetry appeared in the measurement. The authors attribute this to local field orientation, steep gradients near the metal baseplate, and antenna differences not captured by the ideal HFSS model. These differences include fabrication tolerances, feed asymmetry, solder joints, and connector imperfections.

Along the helix side, the measured signal was strongest near metal and reduced in air gaps. It tracked the simulated magnetic-field intensity in those regions.

The simulation showed no clear separation between opposite circular components when the field was analyzed transversely. Their similar spatial distributions made the measured response resemble field intensity rather than a specific rotation sense.

Measurements on opposite sides of the helix used different effective projection axes. They therefore did not directly compare opposite circular components in one common frame. The measured and simulated observables also remained different. HFSS used an antenna-axis decomposition, while the probe used its local NV axis.

Comparison with prior probes and engineering use

Section titled “Comparison with prior probes and engineering use”

The paper positions this probe between microscope-based NV systems and conventional metallic near-field probes. NV microscopes provide about micrometer-scale resolution. This probe, the cited 2018 fiber probe, and EMC probes are listed at millimeter scale. The present resolution is about 2 mm.

The fiber-coupled geometry permits scanning around antenna structures that fit poorly within a microscope field of view. Unlike conventional probes, which mainly detect linear field components, NV selection rules provide sensitivity to circular handedness in the NV frame. The ODMR contrast and linewidth were comparable to prior NV microwave sensing platforms.

Previous Rabi-calibrated work reported sensitivities in the μT/√Hz range. The present study cannot make that absolute comparison because its local field lacked independent calibration. For antenna diagnostics, relative spatial maps may be sufficient when absolute field metrology is not required.

Users must record the bias direction, diamond orientation, probe orientation, and local geometry when interpreting circular fields. Otherwise, apparent helicity may reflect a changed measurement basis rather than a changed antenna field.

Absolute amplitudes would require a calibrated Rabi protocol, which the paper does not implement. The prototype covers 2.4 to 2.87 GHz, and alternate bias configurations may extend that range. Frequency modulation, lock-in detection, and smaller sensing diamonds were proposed as routes to improved sensitivity or resolution. These are suggested developments, not demonstrated results.

The supplied paper is an unedited accepted manuscript and warns that errors may remain before final publication.

The paper’s custom probe needs a small, nitrogen-containing, crystallographically oriented CVD diamond sensing plate for fiber-coupled ODMR readout. 6CCVD’s diamond materials catalog includes single-crystal diamond grown by MPCVD, providing a matching plate material family, while the required [100] and [111] orientations, type Ib nitrogen content, and exact slab dimensions would need confirmation for this device. Engineering support can be used to request a custom sensing-element specification against the probe’s 2.4 to 2.87 GHz requirements.

View Original Abstract

Abstract Accurate characterization of microwave (MW) near fields is essential for antenna diagnostics and electromagnetic engineering, yet conventional probes primarily access linear components and provide limited polarization information. Here, we demonstrate a fiber-integrated diamond probe based on nitrogen-vacancy (NV⁻) centers for projection-based mapping of MW magnetic fields with circular selectivity. The measurement exploits transition-selective coupling between Zeeman-resolved spin transitions and circular components of the MW magnetic field defined with respect to the NV quantization axis, such that the observed optically detected magnetic resonance (ODMR) contrast reflects a projection onto a spin-defined circular basis rather than the field magnitude. This framework enables a physically consistent description of polarization-dependent response in near-field environments without a unique propagation direction. Operating in a continuous-wave regime, the probe provides robust measurements without absolute calibration and yields spatial distributions directly comparable with simulations. Measurements of a double-ridge horn antenna show agreement with linear field distributions, while mapping of a helical antenna reveals circularly selective response consistent with local measurement geometry. These results are performed using practical antenna structures relevant to wireless communication systems, establishing a compact, quantum-enabled platform for polarization-sensitive microwave diagnostics in realistic electromagnetic environments.