Skip to content

Localized nitrogen and hydrogen defect heterogeneities at mineral inclusion interfaces in eclogitic diamonds

FieldValue
JournalDiamond and Related Materials
Published2026-08-08
AuthorsAntonio Angellotti, Lorenzo Barni, Marta Morana, G. Cinque, Y. Lu
DOI10.1016/j.diamond.2026.114060
Citations1

Mineral inclusions were surrounded by localized hydrogen-related FTIR absorption extending several to tens of micrometers into four eclogitic diamonds from Yakutia. In LGM12, the 3107 cm−1 peak height reached 0.06 to 0.08 over a graphite inclusion, compared with 0.01 to 0.03 in the surrounding diamond. A sulfide aggregate in DIA4 produced a rise in total nitrogen from about 50 to about 250 ppm along the inclusion.

The study examined four lithospheric eclogitic diamonds from the Udachnaya kimberlite pipe in the Daldyn-Alakit field, Yakutia Province. The samples were DIA4, DIA5, LGM12, and LGM15. All were originally octahedral and were double-polished parallel to the (110) planes for transmission spectroscopy.

Visible inclusions included clinopyroxene, garnet, graphite, and sulfides. Raman spectroscopy examined selected inclusion areas using a 532 nm laser and an approximately 1.5 µm focal spot. The system was calibrated against the 520.7 cm−1 silicon Raman peak.

Synchrotron micro-FTIR data were collected in transmission mode at the MIRIAM beamline of Diamond Light Source. Spectra covered 4000 to 500 cm−1 at 2 cm−1 resolution, with 256 scans per spectrum. Maps covered areas of approximately 100 by 100 µm².

The maps used a 10 by 10 µm² aperture and a 7 µm stage step, producing 3 µm point overlap. Line scans contained 30 spots, with a 10 µm step and the same aperture size. Spectra affected by saturation, inclusion interference, or surface irregularities were excluded.

Asymmetric least squares baseline correction was applied to the measured spectra. Nitrogen concentrations and residence temperatures were calculated with QUIDDIT software. Hydrogen absorption was extracted separately at 3107, 1405, and 3236 cm−1.

All four diamonds were classified as Type IaAB. The nitrogen analysis separated A-center and B-center concentrations and calculated B% as the B-center concentration divided by total nitrogen.

DiamondTotal N, ppmA-center / B-center, ppmB%Estimated residence temperature, °C
DIA4122 ± 6358 ± 30 / 57 ± 3651 ± 101180 ± 14
DIA5144 ± 5390 ± 34 / 50 ± 2435 ± 101155 ± 12
LGM12114 ± 45104 ± 39 / 9 ± 78 ± 41112 ± 14
LGM15330 ± 101169 ± 50 / 159 ± 5650 ± 21149 ± 8

The statistics use the central 85% of observations and exclude data tails. Across all four diamonds, mean total nitrogen was 137 ± 60 ppm. Mean A-center and B-center concentrations were 91 ± 38 ppm and 41 ± 32 ppm, respectively.

The residence-temperature calculation used formation ages of 2.00 and 2.733 Gy because the age of the Udachnaya diamonds is not uniquely established. The two models differed by 7.5°C.

The global eclogitic reference dataset reports that about 32% of diamonds contain 100 ppm nitrogen or less. Only 2.8% exceed 1000 ppm, and about 25% have B% at or below 10%. The Russian subset has an average nitrogen concentration of about 600 ppm. The four-sample mean of 137 ± 60 ppm was below that subset average.

The 3107 and 1405 cm−1 bands were assigned to VN3H or N3VH0 defects. The 3236 cm−1 band was assigned to VN4H defects. A weak feature near 2786 cm−1 was interpreted as an overtone of the VN3H0 defect. None of the samples was classified as H-rich because its 3107 cm−1 absorption remained below the general lattice absorption near 2158 cm−1.

The spatial measurements separated the responses of different inclusion types. Sulfides produced the clearest nitrogen response, while graphite produced strong hydrogen enrichment without a consistent nitrogen increase.

MapInclusionMeasured nitrogen and hydrogen response
DIA5 Map2Covellite, about 150 by 100 µm²Over the inclusion rim, A-center nitrogen was 100 to 120 ppm and B-center nitrogen was 45 to 60 ppm. Nearby values were 40 to 60 ppm and 10 to 20 ppm. The 3107 cm−1 peak increased from 0.02 to 0.06 to 0.14 to 0.18.
DIA4 Map1Covellite and chalcopyrite aggregate, about 150 by 150 µm²Total nitrogen rose from about 50 to about 250 ppm along the inclusion. The 3107 cm−1 peak increased from 0.02 to 0.03 to 0.05 to 0.06.
LGM12 Map1Sulfide inferred from metallic luster, about 50 by 50 µm²Total nitrogen increased from 100 to 200 ppm in the surrounding diamond to 300 to 350 ppm over the inclusion. The 3107 cm−1 peak increased from 0.08 to 0.15 to 0.20 to 0.25.
LGM12 Map2Graphite, about 20 by 20 µm²Total nitrogen was 60 to 80 ppm in the inclusion-associated area and 110 to 130 ppm in graphite-free regions. The 3107 cm−1 peak was 0.06 to 0.08 over the inclusion and 0.01 to 0.03 around it.

Two maps from regions without visible inclusions provided controls. DIA5 Map3 had nitrogen concentrations of 220 to 240 ppm, except along the upper and lower margins. Its 3107 cm−1 peak height did not exceed about 0.03. LGM15 Map1 had nitrogen concentrations of about 300 to 350 ppm and a maximum 3107 cm−1 peak height of about 0.03.

The LGM15 signal was extremely low, so apparent spatial variation was probably driven by measurement noise. The contrast with inclusion-bearing maps supports the association between inclusions and elevated hydrogen-related absorption.

A garnet and graphite map in DIA4 showed no consistent nitrogen trend. Hydrogen absorption increased where both inclusion types were present. Their overlapping positions prevented the authors from assigning the response to either mineral.

Distance dependence and mineral-specific behavior

Section titled “Distance dependence and mineral-specific behavior”

Three-dimensional maps expressed the 3107 cm−1 absorbance against distance from each inclusion center or rim. In LGM12 Map2, the absorbance decreased almost radially away from the small graphite inclusion. The gradient was visible because the inclusion was only about 20 µm across.

The sulfide in DIA5 Map2 was about 150 µm across. Its size prevented clear detection of an external hydrogen gradient. Internal absorbance variation instead followed the inclusion’s polycrystalline morphology. The authors relate that structure to destabilization of a monosulfide solid solution during exhumation.

Graphite and sulfide inclusions produced different nitrogen responses. Graphite was associated mainly with hydrogen enrichment. Sulfides were associated with coupled nitrogen and hydrogen enrichment in several maps. Garnet produced a less systematic effect.

The platelet band at 1365 cm−1 followed another pattern. Its absorbance varied from 0.625 to 0.425 in DIA5 Map1, 0.375 to 0.175 in DIA5 Map2, and 0.26 to 0.16 in LGM12 Map2. These spatial patterns did not align with the nitrogen or hydrogen maps, suggesting a separate distribution mechanism.

The authors propose that residual growth-medium fluids, localized hydrogen diffusion, and chemical interaction with graphite or sulfides produced the observed enrichment. Graphite has a strong affinity for hydrogen and can retain adsorbed or chemisorbed species. Sulfides can also incorporate hydrogen and nitrogen.

Methane-bearing fluids are a possible source of diamond carbon and hydrogen. The authors also note that nitrogen could occur as NH4+, N3−, or Fe-nitride under mantle oxygen fugacity conditions. They therefore interpret the sulfide maps as evidence for possible nitrogen and hydrogen incorporation into the host sulfide.

The data do not distinguish fluid-mediated chemical control from stress-assisted defect stabilization. Local stress and strain around inclusions may favor hydrogen-defect formation. The repeated spatial agreement between inclusions and hydrogen absorption led the authors to regard fluid control as the dominant factor, with stress acting as a secondary contribution.

Post-entrapment diffusion from the diamond matrix into a sulfide remains possible. The measurements therefore establish spatial association, not a unique reaction pathway.

The study used four samples from one locality. Although their nitrogen concentrations and aggregation states fall within global eclogitic ranges, the authors state that broader application requires samples from other cratonic settings and parageneses. Independent geochemical and structural constraints are also needed to resolve the controlling mechanism.

Some uncertainty comes from spectral normalization. Raw and normalized absorbance at 3107 cm−1 had a correlation factor of about 0.70. Part of the scatter in normalized maps may therefore reflect normalization error.

Spectral attenuation caused a lower apparent nitrogen concentration over a thick part of the DIA5 sulfide. The value fell to 40 ppm even though the spectra remained above the authors’ quality threshold. The 3236 cm−1 band was also weak or absent in LGM12 Map2, making it unsuitable for a spatial trend there.

For natural diamond selection, bulk nitrogen concentration and calculated residence temperature do not describe local hydrogen heterogeneity near inclusions. Micro-FTIR mapping can reveal whether the 3107, 1405, and 3236 cm−1 bands vary at the scale relevant to an application. The paper does not establish a device-performance threshold, but it shows that inclusion-centered hydrogen heterogeneity can extend several to tens of micrometers while nitrogen aggregation and residence temperature remain spatially uniform.

View Original Abstract

Abstract Diamonds in nature are known to have formed by redox-assisted precipitation of carbon from oxidized fluids containing also hydrogen and nitrogen, both of which are hosted in the diamond lattice. Minerals entrapped during the growth of diamonds might have played a role in nitrogen and hydrogen incorporation and abundance although, so far, no evidence has been reported. We used synchrotron micro-FTIR mapping to investigate the interplay between the distribution of nitrogen and hydrogen and mineral inclusions (graphite, sulfide and garnet) in four eclogitic diamonds from the Yakutian kimberlite field, Siberian craton. Combined mapping of nitrogen and hydrogen-related infrared bands reveals heterogeneities near minerals. In particular, a marked absorbance of N- and H-related peaks observed in correspondence with sulfide inclusions supports their incorporation in the host mineral, while near graphite and garnet inclusions, only the H peak absorbance of peaks at 3107 cm − 1 , 1405 cm − 1, and 3236 cm − 1 increases. These zones can extend to tens of micrometers from the inclusion-diamond interfaces and provide insight into the potential control that entrapped minerals exert on the distribution and abundance of N and H in the diamond lattice during and after its formation at mantle depths.