Science1 distinct publisher3 min readUpdated
LLNL says its Nature Physics measurement reconciles diamond melting data with quantum simulations, and that applying it to inertial-confinement fusion could triple energy gain.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
Researchers at Lawrence Livermore National Laboratory have measured how diamond melts at pressures roughly three times those at Earth's core and published the result in Nature Physics [1][2]. The measurement matters because the material in question is not a curiosity: diamond is used to make the capsules that hold fuel in inertial-confinement fusion experiments [3], and the lab says the numbers had been wrong by about 20% for two decades [6].
The problem starts with work done at LLNL about 20 years ago, when Jon Eggert and colleagues ran pioneering high-pressure diamond melting experiments [4]. They found something unusual: diamond became denser when it melted [5]. LLNL scientist Marius Millot compares it to water, where liquid is denser than ice, so ice cubes float; by the same logic, he says, diamond would float in liquid carbon at high pressure [11]. The measurement was an advance and a puzzle at once. Melting temperatures from the laboratory differed from theoretical predictions by roughly 20%, and according to Millot no simulation technique, however advanced, could reproduce the experiments [6][7].
A second open question came from Sandia National Laboratories, where researchers used the Z machine's magnetic fields to shock-compress diamond samples and saw signals suggesting the material passes through another crystalline structure before becoming liquid carbon [8]. Simulations backed that reading, but nobody had directly observed the atomic structure of the compressed material, so the intermediate phase stayed unconfirmed [9].
The new experiments were laser-driven, run at the Omega Laser Facility at the University of Rochester's Laboratory for Laser Energetics [10]. Intense laser energy vaporizes the outer layer of a small sample, which launches a shockwave through the diamond inside [12]. The extreme conditions last about a billionth of a second, and within that window the team had to capture several properties at once, including X-ray diffraction to read the atomic arrangement [13]. Millot describes the campaign as shock-compressing tiny diamond samples to temperatures hotter than the surface of the sun and pressures higher than the centre of Neptune and Uranus while still measuring atomic structure, temperature, density and optical reflectivity [14]. He also calls it the first time shock-compressed diamond has been probed with X-ray diffraction to these conditions [15].
LLNL says the results settle two long-standing discrepancies and bring the measurements into close agreement with simulations based on quantum mechanics [16]. The consequential part of the announcement is the fusion claim: applying the findings to inertial-confinement fusion may allow researchers to triple energy gain, and the same data should improve planetary interior models [17]. That is a single-source statement from the lab's own release, which gives no baseline, no mechanism and no timeline for the factor of three.
Two things to watch. First, whether the paper explains which way the Sandia intermediate-phase question resolved; the release states that two discrepancies are settled without saying the phase was confirmed or excluded [16][9]. Second, whether the revised melting curve actually moves capsule design at the National Ignition Facility, or only tightens an equation-of-state table that designers were already working around [3][17].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
The new study was published in Nature Physics; the release is dated August 20, 2026 and sourced to Lawrence Livermore National Laboratory.
The melting temperatures measured in the laboratory differed by roughly 20% from temperatures predicted by theoretical models.
Millot: "No matter what the theorists did -- even with the most advanced computer simulation techniques -- they could not reproduce the experiments."
At Sandia National Laboratories, researchers used the magnetic fields of the Z machine to shock compress tiny diamond samples, and their measurements produced signals suggesting diamond might pass through another crystalline structure before melting completely into liquid carbon.
The LLNL team performed laser-driven dynamic compression experiments at the Omega Laser Facility, part of the University of Rochester's Laboratory for Laser Energetics (LLE).
At the Omega facility, intense laser energy vaporized the outer layer of a tiny sample, launching a powerful squeezing shockwave through the diamond inside.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Detailed but single-source institutional release
The account is specific and internally coherent: named researchers, a named facility and diagnostic approach, a quantified prior discrepancy (~20%, later characterised as >1,000 degrees), and an explicit negative result on the intermediate phase. It is also tied to a peer-reviewed Nature Physics paper. Against that, the cluster contains exactly one item, which is the laboratory's own release republished by ScienceDaily, with no journal citation, no paper text, no independent expert comment, and a body that is truncated mid-sentence.
No adoption signal supplied
Nothing in the cluster records uptake of the corrected diamond melt data: no fusion campaign adopting it, no code or equation-of-state release, no capsule fabrication change, no third-party replication, and no usage disclosure. Adoption is therefore not assessable from the supplied material rather than low.
Physics well-specified, fusion payoff overstated
The measurement claims are proportionate to what is described, and the release even reports a negative result and an admission that earlier LLNL temperatures were wrong, which pulls against inflation. The gap comes from the framing layer: 'may allow researchers to triple energy gain' is presented in the dek and body as a practical consequence with no baseline, mechanism, design change or timeline, and the 'first ever' X-ray diffraction claim is self-asserted with no adoption or independent confirmation anywhere in the cluster.
Lab-authored release touching its own fusion programme
The only item is sourced to Lawrence Livermore National Laboratory and republished essentially verbatim, and its forward-looking claim concerns inertial confinement fusion energy gain — the programme area that most directly benefits from favourable framing. The corrective element (Eggert conceding the original measurements were off by more than 1,000 degrees, and the intermediate-phase hypothesis being rejected) shows the release is not purely promotional, which caps the score below the high range.
Moderate-low
Confidence is limited by structure, not by internal inconsistency: a single publisher, that publisher reprinting the interested party's release, no access to the underlying Nature Physics paper, no independent or Sandia comment, no adoption evidence, and a truncated body. The descriptive physics claims can be reported with reasonable confidence as the release's account; the fusion gain projection cannot be relied on at all.
product
LLNL closes a 20 percent gap in diamond melting, and stakes a fusion gain claim on it1 distinct publisher
product
Inertia's fusion claim is a factory claim: fuel pellets in hours, not a week1 distinct publisher
build
Nuclear AI program adds an AI security vendor, and the $60M is not the company's1 distinct publisher
science
A timescale, not a diffraction peak: how ultrafast XUV makes exciton condensation testable1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 20, 2026