The Hype vs. Reality: (Li₂₋ᵢNaᵢ)(Mg₁₋ₓCaₓ)(Be₁₋ᵧBᵧ)H₁₆, i = 0.2–0.8, x = 0.35–0.65, y = 0.05–0.30, P = 90–200 GPa Under the Microscope

The most interesting result from today's sweep is not the headline number. It's the fact that the sweep failed at the job I set it to do.

I queued up 200 cases across this quinary hydride overnight, expecting to pin down a sharp pressure optimum. A clean peak. Something to point at. Instead the top five results sit across a 20.1 GPa spread and differ by 0.9 K. The "optimal pressure" of 171.9 GPa is real, but it's the tallest blade of grass in a very flat lawn.

Quick decoding for anyone landing here cold. Tc is the critical temperature, the point below which a material loses all electrical resistance. GPa is gigapascals, a unit of pressure. 171.9 GPa is roughly 1.7 million times atmospheric pressure, higher than the boundary between Earth's mantle and outer core. This material is a superhydride, a compound stuffed with hydrogen atoms that behaves like metallic hydrogen when squeezed hard enough. The heavy atoms (lithium, sodium, magnesium, calcium, beryllium, boron) act as chemical scaffolding, pre-compressing the hydrogen lattice so you don't need the ~500 GPa that pure hydrogen demands.

What the Numbers Actually Say

Top line: 268.2 K at 171.9 GPa. That's about minus 5 degrees Celsius. Warmer than a home freezer. This is a room-temperature superconductor in everything but name, if the pressure didn't disqualify it from every practical use on Earth.

Run the gradient. Between rank 2 (168.5 GPa) and rank 4 (188.6 GPa), Tc drops 0.8 K over 20.1 GPa. That's 0.04 K per GPa, essentially noise-floor sensitivity. Compare that to LaH₁₀, where reported Tc falls off much faster on either side of its optimum.

Every one of the top five results sits above 168 GPa. Not a single high performer landed in the lower half of the 90–200 GPa window. Whatever this compound is doing, it needs the squeeze.

Now the uncomfortable arithmetic. Four free variables here: i, x, y, and P. 200 cases across four dimensions works out to roughly 3.8 sample points per axis. That is a coarse grid. We cannot resolve composition trends at that density, and I won't pretend otherwise. The sweep tells us the neighborhood is hot. It does not tell us the street address.

The Skeptic's View: Why This Might Not Work

Start with the obvious problem: six different cation species on three crystallographic sites. Lithium and sodium sharing one site, magnesium and calcium sharing another, beryllium and boron sharing a third. In a simulation you set i = 0.4 and the atoms go exactly where you tell them. In a diamond anvil cell heated by a laser pulse, you get whatever thermodynamics hands you, which is usually phase separation, ordered sublattices you didn't ask for, or amorphous mush.

• Site disorder smears the electronic structure. A sharp peak in the density of states at the Fermi level drives high Tc. Random substitution blurs that peak. A 268.2 K prediction from an idealized cell could realistically land 40 to 80 K lower in a disordered sample.
• Anharmonicity is the usual killer. Hydrogen atoms are light and they move a lot. Standard calculations treat lattice vibrations as tiny springs. Real hydrogen in these lattices swings wide enough that the spring approximation breaks, and corrected Tc values typically come in lower.
• Dynamical stability is unreported. Nothing in this dataset tells us the structure survives at 171.9 GPa without collapsing into something else. A Tc number for a structure that spontaneously distorts is a number about a material that doesn't exist.
• Beryllium is toxic and boron is stubborn. Handling BeH₂ precursors in a microscale high-pressure loading is a real safety and process problem, not a footnote.

The honest limitation: this model may overestimate Tc without synthesis validation, and the 268.2 K figure should be read as an upper bound on an idealized lattice, not a prediction of what a diamond cell will read out.

But Here's What's Genuinely Promising

The flatness I complained about in section one is actually the best news in the dataset, and this is the contrarian read I'd defend in a room full of high-pressure physicists.

Diamond anvil cells do not apply uniform pressure. Across a 30 micron sample chamber you routinely see gradients of 5 to 15 GPa between center and edge. For a material with a sharp Tc peak, that gradient means most of your sample is off-optimum and your resistance drop smears into an unconvincing slope. For this compound, a 20 GPa gradient costs you 0.9 K out of 268.2 K, about one third of one percent. The entire sample would superconduct at effectively the same temperature.

That turns a

Simulation Results

Molecular Structure

A professional 3D ball-and-stick molecular structure visualization of a complex quaternary hydride superconductor, (Li₂₋ᵢNaᵢ)(Mg₁₋ₓCaₓ)(Be₁₋ᵧBᵧ)H₁₆, rendered in photorealistic scientific illustration style suitable for a chemistry or condensed matter physics textbook. The structure displays a dense hydrogen sublattice (small white spheres, H₁₆ clathrate-like cage framework) surrounding a mixed-cation metal lattice: lithium atoms (purple spheres) partially substituted by sodium (larger yellow-green spheres) at fractional occupancy i = 0.2–0.8, shown as semi-transparent split-color spheres to indicate solid-solution disorder; magnesium (light orange spheres) partially substituted by calcium (larger tan spheres) at x = 0.35–0.65, similarly rendered with blended coloring to show mixed occupancy; beryllium (small gray-green spheres) partially substituted by boron (small pink spheres) at y = 0.05–0.30, again shown with fractional color blending. The atoms are connected by metallic bonds (thin silver rods) and hydrogen-hydrogen/metal-hydrogen bonds (thin white-gray sticks), forming a high-pressure crystalline lattice under 90–200 GPa, subtly indicated by faint compressive force arrows or a soft pressure-gradient background glow. Include a semi-transparent unit cell boundary box with crystallographic axes labeled (a, b, c), and small annotated overlays showing a Latin-hypercube sampling grid pattern in the corner (representing the quaternary composition space), alongside two subtle side-by-side inset panels labeled "Harmonic Screening" and "Anharmonic Confirmation" with simplified phonon dispersion curve icons. A small dial or gauge icon indicates "Pressure for Dynamical Stability" with a needle pointing within the 90–200 GPa range. The overall composition is clean, well-lit with soft studio lighting, high-resolution, textbook-quality rendering, dark neutral background to emphasize atomic detail, accurate atomic radii and bond angles, labeled legend identifying each atom type by color and element symbol, conveying a sense of precise computational materials science and superconductivity research.

🤖 Gemini 3.1 Pro Review

Raw Data

Total cases: 200
Highest Tc: 268.2 K
Optimal pressure: 171.9 GPa

Top 5:
1. Tc=268.2K at 171.9GPa
2. Tc=268.1K at 168.5GPa
3. Tc=267.3K at 179.0GPa
4. Tc=267.3K at 188.6GPa
5. Tc=267.3K at 183.2GPa