Measuring temperature and observing graphitization of heavy-ion-heated diamond
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Scientific Reports |
| Published | 2026-07-06 |
| Authors | J. Lütgert, P. Hesselbach, A. Bergermann, A.J. Roy, V. Bagnoud |
| DOI | 10.1038/s41598-026-59428-4 |
Heavy-ion heating produced diamond temperatures above 2000 K, and the x-ray diagnostic agreed with stopping-power simulations up to about 2000 K. At the highest ion numbers, elastic scattering reached 50% of inelastic scattering, above the 0.27 expected for pristine diamond even at 5000 K. Modified diffraction peaks and changed optical properties make thermal graphitization plausible, but direct phase evidence is still lacking.
X-ray temperature measurement
Section titled “X-ray temperature measurement”The experiment used spectrally resolved x-ray Thomson scattering (XRTS) to infer the bulk lattice temperature. Heating increases atomic motion, reduces coherent Laue interference through the Debye-Waller effect, and raises diffuse elastic scattering. The measured elastic-to-inelastic ratio was linked to the static ion-ion structure factor. DFT-MD calculations for pristine diamond then converted that ratio into a temperature.
Each spectrum was fitted as elastic and inelastic contributions. PrismSpect modeled the titanium-plasma source, and the instrument response was convolved with the source spectrum. The inelastic contribution came from an unirradiated sample. Its room-temperature elastic contribution was about 1.7%, and this value was subtracted from heated shots.
A cylindrically bent HOPG crystal in von Hámos geometry collected XRTS at 130°. Placing the x-ray diagnostics upstream of the target relative to the ion beam, with added shielding, raised the signal-to-noise ratio by more than 3 times. The earlier campaign reached about 1300 K but was limited by signal noise and heating strength.
Heavy-ion heating setup
Section titled “Heavy-ion heating setup”The target was a commercial, optical-quality monocrystalline CVD diamond disc with a 3 mm diameter and 100 µm thickness. Its cylinder faces were oriented along [100], and the sample was rotated to place the (111) Laue peak on the XRD detector. The manufacturer reported nitrogen impurities below 1 ppm.
Uranium-238 ions had charge state 73+ and an energy of 400 MeV per ion. Each bunch contained up to 4.4 × 10^9 ions, with pulse lengths varied between 125 and 300 ns FWHM. The PHELIX laser generated 4.75 keV He α x-rays from a titanium backlighter. It delivered 160 to 180 J at 527 nm in a 2 ns pulse, focused to a 30 to 40 µm spot.
A glassy carbon degrader about 5 cm upstream slowed the ions while retaining accelerator focusing. With no degrader, the ion-energy Bragg peak did not fall within the x-ray probe. The 12 mm short setting put the peak inside the diamond but outside the probed area. The 14 mm long setting placed it in the probe and raised the peak temperature, while reducing heating homogeneity.
SRIM and ray tracing modeled energy deposition and used measured beam focus areas up to 0.87 × 0.48 mm². FLUKA predicted about 2.4% projectile fragmentation per mm of glassy carbon. The degraded-beam focus correction reduced the calculated long-degrader temperature by 200 K at high ion numbers and was applied throughout.
Temperature agreement and its limit
Section titled “Temperature agreement and its limit”The XRTS ratios and stopping-power predictions agreed while the target retained its diamond structure. XRTS temperatures in the table assume pristine diamond.
| Configuration | Measured elastic-to-inelastic ratio | XRTS temperature | SRIM temperature |
|---|---|---|---|
| No degrader, maximum ion numbers | 0.064 | 1200 ± 120 K | 1480 ± 135 K |
| 12 mm short degrader | 0.11 | 2000 ± 200 K | 2000 ± 190 K |
| 14 mm long degrader, highest ion numbers | up to 0.50 | Not consistent with pristine diamond | Not consistent with pristine diamond |
The no-degrader case reproduced conditions from the earlier campaign. The short degrader raised the measured temperature to 2000 ± 200 K, consistent with the SRIM value of 2000 ± 190 K.
In the long-degrader configuration, the methods coincided below 1900 K. Above that range, the measured ratio rose beyond predictions for both average and peak temperatures. Pristine diamond would reach only 0.27 at 5000 K. The authors therefore concluded that the pristine-diamond assumption had failed. Graphitization or amorphization remained possible.
XRD and optical imaging
Section titled “XRD and optical imaging”Cold XRD lines split into multiple segments as the delay between heating and probing increased. Driven shots with and without a degrader showed fracture, including shots with different elastic scattering ratios. Fracture can add lattice disorder, but its presence at lower ratios showed that it did not explain the steep high-ratio increase.
The diamond diffraction signal increased in the regime where XRTS and SRIM agreed, despite the expected Debye-Waller reduction. That reduction reaches only 5% when diamond is heated to 2000 K. A grain-size decrease from 20 to 1 µm could increase reflectivity by about 7 times through primary extinction.
At high scattering ratios and long delays, the diamond diffraction signal decreased and the lines blurred. Fractured pieces could rotate their diffraction signal away from the detector, while a phase change offers another explanation. Thermal expansion and temperature gradients also broaden the peaks.
Optical imaging around 3 µs after the ion-bang onset showed both fracture and an opaque region. The authors linked the opaque region to the ion-energy Bragg peak, where the diamond’s optical properties had changed.
Graphitization models
Section titled “Graphitization models”Thermal graphitization begins at about 1800 K and accelerates with temperature. Classical MD simulations used AIREBO-M, REBO-II, and EDIP interatomic potentials. All three placed graphitization at the surface under ambient pressure and predicted faster propagation along [111].
At 2000 K along [111], the graphitization velocities were 0.2 ± 0.2 ms⁻¹ for AIREBO-M, 1.43 ± 0.27 ms⁻¹ for REBO-II, and 5.57 ± 0.05 ms⁻¹ for EDIP. At 3000 K, REBO-II estimated that 80% of the target would graphitize along [111], while EDIP calculated a complete transition.
The [100] direction was more stable. Only EDIP predicted graphitization there at 2000 K, with a velocity of 1.82 ± 0.14 ms⁻¹. Probe times of 500 ns or less would leave at least 95% of the diamond intact even at the highest [111] velocity near 2000 K. Stress-induced fracture could create smaller splinters and more surface area, which might increase the transformation rate.
Remaining uncertainty and engineering implications
Section titled “Remaining uncertainty and engineering implications”Radiation damage remained an alternative explanation. An amorphization threshold of 3 × 10²² vacancies per cm³ has been reported, with other studies differing by up to three times. SRIM predicted vacancy densities more than three orders of magnitude below this threshold. However, those calculations assumed a cold target and omitted dynamic effects. Tracks from uranium had not been detected in ambient diamond.
The direct observation needed to identify the phase was still missing. Large temperature gradients made the XRD data inconclusive beyond early-time assessment of macroscopic integrity. Electron microscopy or Raman spectroscopy after irradiation could provide complementary evidence, although recovering the sample would require a setup change. A higher ion flux could reach 1800 to 2000 K without a degrader and would separate bulk heating from individual-particle energy deposition.
Diamond fracture has direct consequences for ion-heated anvil cells. The authors state that reducing the highest energy deposition rate may be needed to preserve anvil stability. The paper does not establish an allowable fluence or maximum operating temperature for detector and anvil applications. Material qualification must therefore consider thermal graphitization, mechanical fracture, and accumulated radiation damage under the intended gradients.
Materials and support
Section titled “Materials and support”The paper used a 3 mm diameter, 100 µm thick monocrystalline CVD diamond disc as its heavy-ion diagnostic target, and it identifies diamond anvil cells as an application concern. 6CCVD offers single-crystal and polycrystalline diamond grown by MPCVD, with plates and wafers up to 125 mm and thicknesses from 0.1 to 500 µm; available formats are listed in the diamond materials catalog. Exact geometry and custom fabrication requirements for a comparable target or anvil-cell component can be reviewed through the supplier’s engineering support.
View Original Abstract
The graphitization of diamond is of significant interest for nanodiamond synthesis from laser-shocked hydrocarbons, for its use as a detector material, and for the application in diamond anvil cells. The transition can be triggered either locally through sufficiently rapid energy deposition, or thermally at temperatures above 1800 K. We report on an experiment in which a heavy-ion beam was used to volumetrically heat monocrystalline diamond, investigating the intermediate regime between these two competing mechanisms. Our sample is probed using x-ray radiation generated from a laser-driven titanium plasma. The ratio of elastic to inelastic scattering in backward direction allows for measuring the samples’ bulk temperature. We observe good agreement between the x-ray based technique and stopping power simulations up to ∼ 2000 K. When increasing the heating further by fully stopping the ions in the target, we reach conditions where graphitization is predicted to occur. In this regime, elastic x-ray scattering is notably stronger than expected for pristine, heated diamond. Simultaneous x-ray diffraction measurements show a modification of the previously sharp diamond peaks in this range, which is also accompanied by changes in the optical properties of the sample. A thermally induced transition to graphite provides the most plausible explanation of the observations.
Single Crystal Cvd Diamond Plate
View Specs →