Armadillo Labs · QuantArm
The problem

Nearly every superconductor ever discovered was found by accident.

A superconductor carries electric current with zero resistance — no waste heat, no loss. The catch: every one we know of only works far below room temperature, locked inside cryogenic jackets. A material that superconducts at ordinary temperatures would change how the world moves power.

The blocker is discovery itself. No one can reliably predict which materials will superconduct before making them — so the field advances by trial and error, one painstaking sample at a time.

Backdrop: illustrative rendering of a materials search space. The lit points stand for the small set of known superconductors; everything else is unexplored.

The observation

The crystal lattice carries the tell.

These are real crystal structures, and the atomic motion you see is a real vibrational mode — computed from the quantum mechanics of each material, not an artist's loop. Cool the crystal and the thermal shaking dies away, exposing the ordered vibrations underneath.

In the copper-oxide superconductors, neutron-scattering experiments (measured — Reznik et al., inelastic neutron scattering) show a specific copper–oxygen stretching vibration softening anomalously as these materials approach superconductivity. The lattice is telling us something about where pairing lives.

Displacement patterns: density-functional lattice dynamics (Quantum ESPRESSO). Amplitudes scale with temperature for visibility. Cool below the measured Tc and a dissipationless supercurrent switches on — a wavelike, phase-coherent condensate in the conducting planes (schematic, not a simulation).

The signature

Superconductors live in a particular lattice regime — and it is computable.

Plotted here is the full vibrational spectrum of the material, computed from first principles. Known superconductors share a pattern: vibrational modes poised near — but not past — an instability, electronic states piled up near the Fermi level, and the electron–lattice coupling concentrated in a few specific modes rather than spread evenly.

When those signatures co-locate, the lattice regime is favorable for superconductivity. That co-location is a computable signature — the thing QuantArm screens for.

The screen identifies where superconductivity is favorable; it does not claim to settle why materials pair. For a discovery engine, that is the right question — and the honest one.

The engine

QuantArm: interrogating a material in days, not months.

The standard workhorse for lattice and coupling calculations is accurate but painfully slow for exactly the soft, correlated materials that matter here. QuantArm computes the same physical observables through hundreds of small, independent calculations that run in parallel — and every result is gated against the reference method before it counts.

0.06%
BaBiO₃ vibrational frequencies, mean deviation vs reference
1.05%
MgB₂ zone-center — 1.02% at zone boundary
0.16%
Ba₁₋ₓKₓBiO₃ vs certified reference
4–5%
electron–lattice coupling strength vs reference; mode ranking exact

Backdrop: the method itself — one crystal, many independently displaced copies, each a small self-contained calculation. Validation figures from the QuantArm v3 gate ledger, July 2026.

The path

Screen many. Interrogate few. Predict. Synthesize.

The discovery funnel: thousands of candidate materials screened cheaply for the lattice regime; survivors get the full coupling interrogation; the strongest emerge as doping-resolved predictions — which composition, and where its transition-temperature dome should peak — sealed before synthesis.

Superconductors have historically been found by accident; the rare predicted-first successes — the high-pressure hydrides — exist only at millions of atmospheres. The product here is prediction at ambient conditions: naming the material before it exists on a bench.

Funnel stages reflect the QuantArm discovery pipeline. Candidate counts are design targets for the screening tiers, not results.