Skip to content

High-pressure magnetic transition in iron observed via diamond quantum sensing

FieldValue
JournalHigh Pressure Research
Published2026-04-24
AuthorsKouki Yamamoto, Kenshin URIU, Ryotaro Suda, Misaki Sasaki, Mari Einaga
DOI10.1080/08957959.2026.2661257

NV centers fabricated on the culet of a nearly (111)-oriented diamond anvil imaged iron’s stray field up to 30.0 GPa. The field weakened during compression, vanished at 30.0 GPa, and returned during decompression. The pressure path showed hysteresis in the ferromagnetic alpha to paramagnetic epsilon transition at room temperature.

Sensor design places NV centers in the sample chamber

Section titled “Sensor design places NV centers in the sample chamber”

NV centers were fabricated directly on the upper culet, so the sensor sat above the iron in the high-pressure chamber. Helium-ion implantation created vacancies. Annealing allowed the vacancies to diffuse and combine with isolated nitrogen in the type-Ib anvil. The resulting NV layer was about 1 µm thick.

In a (111)-oriented anvil, one NV axis is perpendicular to the culet. This axis retained sufficient signal contrast at 30.0 GPa. The selected anvil was several degrees off (111) because polishing prevented perfect orientation.

The experiment used a DAC designed for X-ray diffraction. Its upper anvil had a 300 µm culet. The cell used a nonmagnetic rhenium gasket and an 80 µm chamber.

Pressure cell elementSpecificationPurpose
Upper anvilType-Ib diamond, nearly (111), 300 µm culetNV sensing surface
GasketRhenium, 250 µm initial thickness, preindented to 42 µmSupports the chamber
Chamber80 µm diameterHolds iron, ruby, and NaCl
Iron4N purity, about 35 µm × 30 µm × 10 µmPlate-like magnetic specimen
Pressure mediumNaClHydrostatic pressure transmission
CalibrationRuby spherePressure determined from fluorescence

NaCl was repeatedly loaded until the chamber became optically transparent. Iron and ruby were then surrounded by additional NaCl. The pressure was determined using the ruby fluorescence method.

The optical setup used a 520 nm laser, a dichroic mirror, and a 50× objective with NA 0.55. Fluorescence passed 550 nm and 600 nm long-pass filters, then a 200 mm imaging lens, before reaching a CMOS camera. A beamsplitter diverted ruby fluorescence to a spectrometer.

Optically detected magnetic resonance spectra were acquired for each camera pixel. This produced a wide-field magnetic map. A microwave antenna with a 2.5 mm central opening delivered the excitation field.

The NV spin levels have a zero-field splitting of 2870 MHz at room temperature and ambient pressure. Under a magnetic field, the splitting is 2γB, with γ = 28 MHz/mT, for the field component along the NV axis.

The bias field was a few millitesla. Its value was adjusted at each pressure to separate the relevant resonance dips. The adjustment also had to account for pressure-dependent shifts and the pressure gradient across the culet.

At each pressure, the authors applied an external field of about 37 mT to saturate the iron magnetization. This field was applied only temporarily. Only the bias field remained during ODMR measurements.

Magnetization-reversal tests at 6.6 GPa reversed the field pattern when the external field changed direction. Reapplying the original direction restored the pattern. These tests confirmed that the mapped field came from the iron.

Pressure pathPressureBias fieldObserved iron stray field
Compression6.6 GPa2.5 mTStrong rightward field pattern
Compression13.2 GPa1.7 mTReduced but substantial rightward field
Compression21.3 GPa2.5 mTField largely suppressed, with a faint rightward remnant
Compression30.0 GPa1.0 mTNo iron stray field detected
Decompression20.2 GPa0.5 mTNo iron stray field detected
Decompression13.5 GPa0.6 mTClear recovery of the iron field
Decompression7.8 GPa1.0 mTMagnitude comparable with 6.6 GPa

At 30.0 GPa, the remaining weak map came from bias-field inhomogeneity rather than iron. At 20.2 GPa during decompression, the visible weak distribution had the same origin.

The decompressed field at 7.8 GPa had a magnitude comparable with the original field at 6.6 GPa. Its spatial distribution differed, which the authors attribute likely to a compression-induced change in the iron shape.

Hysteresis and comparison with earlier sensors

Section titled “Hysteresis and comparison with earlier sensors”

The compression data retained a faint rightward field at 21.3 GPa but no iron field at 30.0 GPa. During decompression, no iron field appeared at 20.2 GPa, while recovery was clear at 13.5 GPa. The different paths show hysteresis, consistent with earlier iron studies cited by the authors.

The sampled points place full field loss between 21.3 and 30.0 GPa during compression. They place recovery between 20.2 and 13.5 GPa during decompression. The measurements therefore capture both sides of the magnetic transition without assigning a sharper transition pressure than the pressure steps allow.

The (111)-oriented anvil retained usable contrast over the whole field of view through 30.0 GPa. Earlier (100)-oriented studies reported reduced contrast and could not analyze the weak signal directly above iron. Cited work also states that (100)-oriented sensors become unusable above approximately 50 GPa.

The authors expect the other three NV orientations to vanish above approximately 50 GPa. They did not test that limit in this experiment.

The type-Ib anvil contained many NV centers before implantation. These internal centers responded only to the bias field and produced a background signal. When the iron stray field was weak, internal and surface dips overlapped. A single-dip fit was then shifted, so an accurate field magnitude was unavailable. The presence or absence of a stray field remained reliable.

Below about 50 GPa, all four NV orientations remained visible. Their different zero-field splitting shifts, together with the culet pressure gradient, complicated dip assignment and bias tuning.

The bias field also changed between pressure steps. The authors argue that comparisons remain meaningful because the bias was normal to the sample plane and only a few millitesla. Magnetization reversal required fields of several tens of millitesla.

The authors identify two ways to reduce the internal background. They suggest quantifying internal NV density before surface fabrication and choosing a low-density anvil. Eliminating or deactivating internal NV centers remains an open challenge.

Implications for sensor material selection

Section titled “Implications for sensor material selection”

The measured result favors a nearly (111)-oriented anvil when a sensor must read weak stray fields under pressure. Surface defect processing must coexist with low internal NV background, which creates a material selection tradeoff. Type-Ib diamond provides isolated nitrogen for creating surface centers, but its internal centers reduce spectral separation.

Below about 50 GPa, instrument design must include bias-field tuning, position-dependent frequency calibration, and separate fits for surface and internal signals. Weak-field operation should rely on field presence until those signals can be separated. The authors expect imaging above 100 GPa to become feasible, but this paper provides measurements only to 30.0 GPa.

For DAC-based NV quantum magnetometry, the relevant material family is MPCVD-grown single-crystal diamond. The 6CCVD diamond materials catalog lists single-crystal plates and wafers up to 125 mm, without specifying (111) orientation or finished anvil sensors. The paper requires a nearly (111) type-Ib anvil, a 300 µm culet, and a helium-implanted, annealed NV layer about 1 µm deep, so those requirements should be confirmed through engineering support.

View Original Abstract

Diamond quantum sensors offer high precision and spatial resolution as magnetic probes, making them promising for a wide range of applications. While diamond anvil cells (DACs) can generate extremely high pressures, techniques for magnetometry under such conditions remain limited. By fabricating an ensemble of NV centers directly on the anvil diamond surface, we enable precise magnetic measurements under high pressure. In this work, we employ this NV ensemble to image the stray magnetic field of iron up to 30 GPa, enabling the observation of the magnetic transition (α−ε transition) in iron.