Skip to content

Improvement of a focused ion beam fabricated diamond pillar for scanning ensemble nitrogen-vacancy magnetometry probe using an ultrapure diamond

FieldValue
JournalJapanese Journal of Applied Physics
Published2026-06-18
AuthorsDwi Prananto, Yifei Wang, Yuta Kainuma, Kunitaka Hayashi, Masahiko Tomitori
DOI10.35848/1347-4065/ae7f5f

An 800 nm-diameter scanning ensemble nitrogen-vacancy magnetometry probe reached an estimated DC magnetic field sensitivity of 6.7 µT/Hz¹/². It resolved an inferred 300 nm magnetic domain wall in (BiLu)₃Fe₅O₁₂ while preserving the measured spin coherence of a shallow NV ensemble after 30 keV Ga⁺ FIB milling.

The probe had a 1.6 µm-high pillar with an 800 nm tip diameter measured by SEM. It sat at the apex of a 30 µm tungsten rod attached to a quartz tuning fork. The tuning fork supplied an amplitude signal for phase-locked distance control in the integrated AFM system.

PVA and 160 nm Pt/Pd layers protected the shallow NV centers during milling. After cap removal, UV/ozone removed organic adsorbates and oxygen-terminated the surface. The authors used this chemical treatment to avoid an added ICP removal step during fabrication.

The scanning confocal system and the AFM provided optical readout and spatial registration. The quartz tuning fork maintained the tip-sample distance while the sample scanner moved the sample laterally.

The starting material was electronic-grade, (100)-oriented type-IIa diamond measuring 2 × 2 × 0.5 mm³. Measured nitrogen and boron impurities were below 5 ppb and 1 ppb, respectively. The authors polished the piece to 50 µm thickness.

They implanted ¹⁴N⁺ at 30 keV and a dosage of 1 × 10¹² ions/cm². A 900°C anneal for 1 h under Ar formed the NV layer. SRIM Monte Carlo simulation placed its mean depth at 40 nm.

Acid cleaning used a 1:3 HNO₃:H₂SO₄ mixture at 220°C for 30 min, followed by deionized water. Laser cutting produced triangular pieces from the implanted diamond. The pieces were mounted on glass with fluoroelastomer adhesive. Before mounting, acetone-dissolved ABS bound the amorphous carbon residue left by the laser cut.

Ga⁺ milling at 30 keV can create 30 nm-thick amorphous carbon and surface defects. These defects can reduce NV charge stability and coherence time and add background fluorescence. PVA followed by 160 nm Pt/Pd protected the shallow NV centers and was intended to suppress charging during milling.

A 13 µA rectangular milling pattern removed material outside a 7 × 7 × 5 µm³ square island. A doughnut pattern with a 7 µm outer diameter and a 1 µm inner diameter used a 720 pA current. It produced the 1.6 µm-high pillar with an 800 nm diameter.

Hot aqua regia removed the Pt/Pd and PVA. The bath was a 1:3 HCl:HNO₃ mixture held at 60°C for 15 min. A 3 min deionized-water dip followed. UV/ozone then ran at room temperature and 0.1 MPa oxygen pressure. The probe was detached and glued to the tungsten rod with silver paste and epoxy reinforcement.

The scanning confocal system used a 532 nm laser and an objective with NA = 0.7. The laser polarized the NV spins into the mₛ = 0 state. An avalanche photodiode collected 600 to 800 nm photoluminescence through the same objective.

A continuous-wave laser and microwave sweep produced a zero-field ODMR dip at 2.87 GHz. The PL depth profile had a 2.3 µm full width at half maximum. This approached the objective’s 2.2 µm axial diffraction limit and supported the shallow NV distribution. The paper’s figure gave 19 µm for a non-electronic-grade comparison.

The 50 µm platform produced no significant improvement in photon collection compared with an EBL-RIE platform thinner than 3 µm. The authors expect a reduction to a few µm through additional FIB milling to improve collection efficiency.

Spin tests used a 10 mT external field aligned with one axis in the <111> NV crystal-axis family. The Rabi protocol used 4 µs polarization and readout pulses. Microwave power was set to produce a 1.3 MHz Rabi oscillation.

Rabi contrast was defined by the difference between the maximum and minimum of a damped oscillation fit. A pulse sequence without microwaves supplied the normalization reference. Hahn echo used π/2, t, π, t, π/2 pulses. The final pulse alternated between π/2 and 3π/2 to read different spin projections, and subtraction of the two readouts canceled common noise.

Spin measurementProbe resultCondition or reference
Rabi oscillation contrastAbout 2%Preserved during UV/ozone exposure; about 1.5 times the acid-treated control after treatment
Hahn echo T₁(30.4 ± 3.6) µsUnchanged before and after FIB fabrication and UV/ozone treatment
Longitudinal relaxation T₂(0.72 ± 0.015) ms before UV/ozone and (1.83 ± 0.52) ms after 6 hAcid-treated control was (1.81 ± 0.31) ms

A separate electronic-grade diamond implanted under the same conditions served as the acid-treated spin control. The authors associate the T₂ gain with oxygen termination, mainly C-O-H bonds, and a lower surface ratio of sp² to sp³ carbon. This may indicate fewer FIB-induced damage states and less paramagnetic noise near the pillar rim.

The authors assign T₁’s dominance to bulk P1 nitrogen centers created during implantation. T₁ is therefore less sensitive to surface defects than T₂.

Magnetic domain imaging and DC sensitivity

Section titled “Magnetic domain imaging and DC sensitivity”

The test sample was BiLuIG, or (BiLu)₃Fe₅O₁₂. It has in-plane magnetization and maze-like domains with Néel walls at zero field. The probe scanned a 17.5 × 17.5 µm² area at 256 × 256 pixels while monitoring PL.

Transverse stray fields above 10 mT mixed the NV ground and excited states. This increased non-radiative transitions, reduced spin polarization, and quenched PL. Dark narrow lines in the PL map indicated strong off-axis fields associated with the domain walls.

A line cut gave an inferred wall width of about 300 nm. A separate domain-wall estimate using δ = π√(A/Kᵤ) gave 400 nm. That calculation used an exchange coupling constant of 3.7 pJ/m and cubic anisotropy of 180 J/m³.

DC sensitivity was estimated from the steepest ODMR slope and a photon count rate of 350 kcps. High transverse fields prevented ODMR measurement at contact distances of a few nm, so the probe was lifted a few µm. The resulting 6.7 µT/Hz¹/² was better than the prior FIB-fabricated probe’s 12.3 µT/Hz¹/². It remained within the same order of magnitude as EBL-RIE probes.

The ensemble design sets a direct trade-off between spatial resolution and DC sensitivity. The present 800 nm pillar is estimated to contain 25 NVs. A smaller pillar should improve spatial resolution, but fewer NVs would reduce sensitivity.

Ga⁺ milling below 30 keV could suppress amorphous carbon formation. The authors project probes below 200 nm with about 75 nm spatial resolution. That estimate assumes spatial resolution scales linearly with pillar diameter.

The surface-recovery mechanism was inferred from spin changes and prior XPS work. The paper did not report direct surface spectroscopy of the finished pillar. A proposed extension combines ICP removal of FIB damage, control of ensemble depth, and UV/ozone treatment.

For this SNVM probe, the relevant material family is MPCVD-grown single-crystal diamond. The paper requires a 50 µm, (100)-oriented type-IIa platform with a 160 nm Pt/Pd cap for Ga⁺ FIB milling. 6CCVD’s diamond materials catalog lists single-crystal MPCVD diamond from 0.1 to 500 µm thick, which includes the 50 µm platform, while engineering support provides the route for checking orientation and cap specifications before fabrication.

View Original Abstract

Abstract Scanning diamond nitrogen-vacancy probe microscopy (SNVM) is an important tool for studying nanoscale condensed-matter phenomena. Ga + -ion-focused-ion-beam (FIB) milling has been introduced as a method for fabricating SNVM, while the probe diameter is limited to a few micrometers due to Ga + -induced damage. We report a method for improving the quality of FIB-fabricated SNVM probes by polyvinyl alcohol and Pt/Pd capping, followed by post-fabrication UV/ozone exposure. The effectiveness of the method is confirmed by the fabrication of an 800 nm diameter probe with shallow NV centers, demonstrating an imaging of maze-like magnetic domain structure with a resolution of a few hundred nanometers and an improved sensitivity of 6.7 μT/Hz −1/2 , while preserving spin coherence properties.