Spectroscopic limits of diamond anvils to 520 GPa and projected bandgap closure
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Nature Communications |
| Published | 2026-02-12 |
| Authors | A. Hilberer, Paul Loubeyre, C. Pépin, F. Occelli, Gunnar Weck |
| DOI | 10.1038/s41467-026-69533-7 |
Diamond anvils lost visible transparency near 500 GPa in broadband absorption measurements reaching 520 GPa. The authors project indirect bandgap closure at a sample pressure of about 555 ± 6 GPa. At 523 ± 13 GPa, the neon sample was observable by infrared imaging but completely obscured by the anvils in visible light. X-ray diffraction still determined its pressure at 523 ± 13 GPa.
Visible transparency collapses above 300 GPa
Section titled “Visible transparency collapses above 300 GPa”Broadband transmission covered 220 to 950 nm through LeToullec membrane diamond anvil cells. The SOLEIL synchrotron supplied ultraviolet light, while a supercontinuum laser covered the visible and near-infrared.
Electronic-grade type IIas synthetic diamond anvils were oriented with the [100] axis along compression. Beveled and toroidal geometries were tested, and each cell contained neon.
Solid neon remains a wide-bandgap insulator into the terapascal range, so its absorption could be neglected. This isolated the pressure-dependent absorption of diamond.
Three runs covered 26 to 139 GPa with 70 µm beveled anvils, 220 to 328 GPa with 20 µm beveled anvils, and 234 to 513 GPa with 24 µm toroidal anvils.
Toroidal anvils were focused-ion-beam machined with a 24 µm flat diameter, 80 µm groove, and 4.6 µm depth. Rhenium gaskets held neon loaded at 140 MPa.
The absorption edge shifted continuously toward lower photon energy from ambient pressure to above 500 GPa. Visible transmission disappeared around 500 GPa. Optical-density thresholds of 1 and 2 followed the same pressure evolution.
The edge shift was reversible under pressure cycling and reproduced across the independent runs.
Raman data reconstruct the stress at the culet
Section titled “Raman data reconstruct the stress at the culet”Under [100] axial loading, backscattering Raman spectra contained only the diamond singlet longitudinal optical mode. Gold samples served as the X-ray pressure standard.
Phonon-edge frequencies measured with beveled and toroidal anvils followed one curve. The result confirmed the Raman edge pressure scale across different anvil geometries.
Hydrostatic compression produced the triply degenerate LO-TO mode and extended the comparison data to 200 GPa. The separation from anvil-tip data reflects anisotropic stress in the loaded diamond.
Combining Raman frequencies with ab initio calculations gave an anvil-tip deviatoric parameter near 0.68. Independent measurements support a nearly constant value of (\delta=0.6 \pm 0.05).
The analysis converts sample pressure (P_s) to average tip pressure as (P=P_s(1+2\delta)/3). Confocal Raman mapping at 77 GPa found a nearly uniform stressed region below the culet.
Pressure then decreased in a quasi-exponential manner with depth. The uniform region’s thickness was governed mainly by the culet diameter. The measurements support a homogeneous, tetragonally distorted diamond slab several microns thick at the sample interface.
Absorption yields the minimum indirect bandgap
Section titled “Absorption yields the minimum indirect bandgap”The standard indirect-gap method plots (A^{1/2}) against photon energy. Absorbance was defined as (A=-\log(I/I_0)), and the linear threshold intercept gives the bandgap.
Diamond’s indirect transitions dominate near the absorption threshold. Their absorption coefficient follows a squared energy offset from the gap, with phonon absorption and emission allowed.
An earlier absorption measurement gave an ambient indirect gap of 5.47 eV. The LO phonon contribution reaches at most 0.25 eV in the 300 GPa range. The authors neglected this contribution because its effect fell within the bandgap uncertainty.
The absorption model combined the axial stress profile, local bandgap evolution and a standard indirect absorption edge. Calculated spectra at 80 GPa and 328 GPa closely matched the observations.
The extraction worked well for beveled anvils. Toroidal thresholds were broader because the maximum-pressure layer was not thick enough to dominate the measured edge.
Despite that broadening, the OD=1 threshold tracked the minimum gap and agreed between geometries. This supported a universal pressure dependence for the minimum diamond bandgap.
Extrapolation places bandgap closure near 555 GPa
Section titled “Extrapolation places bandgap closure near 555 GPa”The extracted minimum bandgap declined linearly with calculated diamond density at the anvil tip. Extrapolation gives a closure density of (5.41 \pm 0.05) g/cm3 and a sample pressure of about (555 \pm 6) GPa.
This is a projected closure, not a directly observed transition. The authors describe metallization of a thin diamond layer as a possibility.
Their effective equation of state combined the IPPS-Ruby2020 hydrostatic diamond equation with the stress model. Its Vinet fit used (K_0=604.5) GPa and (K_0’=3.7).
Toroidal cells have reported sample pressures near 600 GPa without an observed loss of diamond strength. The authors note that metallization may have occurred in a thin layer at the sample interface.
An advanced calculation cited by the paper places the atomic metallic hydrogen transition at 577 GPa. That value lies close to the projected diamond bandgap closure.
Spectroscopy has pressure-dependent operating windows
Section titled “Spectroscopy has pressure-dependent operating windows”The measurements define three regimes. Optical measurements remain unaffected below the indirect gap, become partially perturbed above it, and become impossible after reaching OD=2.
| Diagnostic | Pressure or wavelength | Result | Operational limit |
|---|---|---|---|
| Visible microscopy | 523 ± 13 GPa | No resolved sample | Anvils are fully opaque |
| Raman edge | 660 nm excitation above 480 GPa | Edge cannot be determined | Excitation reaches OD=2 threshold |
| Infrared microscopy | 1650 to 2150 nm at 523 ± 13 GPa | Sample remains observable | Transparency is reduced |
| X-ray diffraction | 523 ± 13 GPa | Ne and Re diffraction determines pressure | Gap closure does not remove X-ray access |
Reflection measurements follow the same classification because the light path crosses one anvil twice. The paper notes that indirect-gap closure does not affect X-ray transparency.
Hard X-ray measurements above 15 keV remain usable in the terapascal range. The highest-pressure experiment used 30.2 keV X-rays, corresponding to a wavelength of 0.411 Å.
X-ray diffraction of hydrogen remains difficult. The cited maximum achieved pressure for hydrogen diffraction is about 250 GPa, well below the projected metallization region.
Visible evidence does not resolve atomic metallic hydrogen
Section titled “Visible evidence does not resolve atomic metallic hydrogen”Infrared absorption reports of probable molecular metallic hydrogen at 425 GPa and deuterium at 460 GPa fall within the transparent-anvil regime. The authors state that those measurements do not require reevaluation.
The reported atomic metallic hydrogen observation at 495 GPa relied on visual inspection and reflectance measurements at four wavelengths. Three wavelengths were visible, while one was near-infrared.
At 495 GPa, the visible range lies deep within the opaque-anvil regime. Visual observations cannot provide unconstrained access to the sample, and visible reflectance is strongly compromised.
Only the near-infrared reflectance point may remain accessible, although anvil absorption still affects it. The authors raise serious doubts about the claimed observation and leave atomic metallic hydrogen unresolved.
Near-infrared reflectance is the only viable optical method for the expected transition. Its absorption may increase around 560 GPa, but the authors expect a reflectivity jump to remain detectable. If atomic metallic hydrogen shifts beyond 600 GPa, its observation in a cell would likely require X-ray diagnostics.
Engineering choices for extreme-pressure measurements
Section titled “Engineering choices for extreme-pressure measurements”Anvil geometry did not change the OD=1 pressure trend, but it changed threshold sharpness. Quantitative bandgap extraction was more reliable for beveled anvils than for the broadened toroidal response.
Instruments must therefore match the optical window of the anvil. Visible microscopy and 660 nm Raman become unsuitable above 480 GPa, while infrared and X-ray access remain available.
Pressure calibration should retain the Raman edge as a local indicator and cross-check it with an X-ray pressure standard. The authors used gold and the IPPS-Ruby2020 scale to refine the calibration.
The closure estimate depends on several assumptions. Transverse stress is held at the tip value, the model uses a semi-infinite anvil, and only indirect-gap absorption is included.
The phonon contribution is neglected, although the authors judge its effect smaller than the present uncertainty. Toroidal threshold broadening also shows where the simplified stress model needs geometry-specific treatment.
At 520 GPa, infrared or X-ray diagnostics retain sample access, while visible imaging and 660 nm Raman do not.
View Original Abstract
Diamond anvils serve as optical windows in static ultrahigh-pressure experiments, now reaching the terapascal regime. However, they exhibit poorly understood changes in their optical properties under multimegabar pressure. Here, we present broadband absorption measurements (ultraviolet to infrared wavelengths) up to 520 GPa, revealing a pronounced loss of transparency with pressure. Diamond Raman scattering is used to infer the stress profile along the anvils’ axis under the assumption of tetragonal distortion, and crucially at the sample interface. This enables a quantitative analysis of absorption spectra, showing an indirect bandgap narrowing towards the infrared, with metallization projected near 560 GPa sample pressure within our stress model. A universal optical behavior is observed across different anvil geometries, which is consistent with the universality of the Raman edge pressure scale, here refined. These findings help define the spectroscopic operational limits of diamond anvil cells under extreme pressure with important implications for recent claims of hydrogen metallization. Diamond anvils are widely used in high-pressure research to investigate matter under extreme conditions. Here, broadband spectroscopy is used to measure pressure-driven opacity of diamond anvils to 520 GPa, revealing bandgap narrowing and optical behavior that redefines the limits of high-pressure spectroscopy.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
Free Standing Polycrystalline Cvd Diamond Membrane
View Specs →