Copper, graphene-like layers and battery particles break simple models
A charge-storing nickel electrode, a graphene-like insulator lasting to 110 kelvin, copper that avoids a predicted collapse and three lithium-motion regimes reveal where simple material models fail.
Science··Evening
A nickel electrode banks oxidising power
Chih-Jung Chen's group at National Taiwan University separated nitrile synthesis from hydrogen production. A NiOOH/Ni(OH)2 electrode is first charged electrochemically, then converts benzylamine to benzonitrile in an organic solvent such as hexane with no applied voltage; the reduced electrode is recharged while hydrogen evolves at the cathode. The peer-reviewed work appears in Angewandte Chemie International Edition. The usual electrochemical route runs benzylamine oxidation in strongly alkaline water, where the benzylimine intermediate and benzonitrile degrade with water and hydroxide — a problem that worsens as concentration rises. Decoupling the half-reactions lets the synthesis step move into a solvent that does not hydrolyse the product. Hexane performed best, and high yield, selectivity and Faradaic efficiency held even above 100 millimolar benzylamine. Several benzylamine derivatives also converted. This is laboratory electrochemistry, not industrial scale or continuous-reactor evidence.[1]
A 110-kelvin insulator in a slowed graphene-like stack
Strong electron repulsion is predicted to open an insulating gap at graphene's Dirac point, but the Mott state has never been found in pristine graphene because the Fermi velocity swamps interactions. A team built a honeycomb moiré superlattice in twisted MoSe2 homobilayers, cut the electron speed by about a hundredfold, and observed the gap up to 110 kelvin. The peer-reviewed Nature Communications paper of 14 August 2026 — an unedited early-access manuscript — folds slow bands from the valence-band maximum at the Γ valley, with negligible spin-orbit coupling and a Fermi velocity lower by nearly two orders of magnitude. Rydberg-exciton sensing read with moiré exciton-polarons revealed a Mott gap at the Dirac point persisting to 110 kelvin; the team also found correlated states at filling ν = -1 with weak ferromagnetic coupling and further states at fractional fillings. This is low-temperature condensed-matter experiment, not a room-temperature electronics claim. The simple graphene model fails here because the interaction regime is opened by slowing the carriers.[2]
Copper did not collapse; lithium took three paths
The MeV-UED femtosecond electron camera at SLAC National Accelerator Laboratory followed a thin copper film through the solid-to-liquid transition. In the Nature Communications early-access paper of 6 August 2026, at absorbed energy densities 2 to 4 times the melting threshold, melting began at the surface slightly below the nominal melting point of 1,085 degrees Celsius and then spread rapidly through the volume. Neither the measurements nor accompanying molecular-dynamics simulations showed the sudden lattice collapse predicted at the superheating limit near 1,424 degrees Celsius — about 1.25 times the melting temperature. The authors read this as evidence that a metal's own atomic dynamics limit how fast it can lose order; copper alloys are candidates for fusion-chamber heat sinks, so the transient behaviour matters for design. Separately, a University of Chicago team's peer-reviewed Nature Communications study tracked lithium-6/lithium-7 isotope exchange and lithium-for-sodium swaps inside lithium iron phosphate particles about 20 nanometres across. Three transport modes separated: ordinary Fickian diffusion, a slower subdiffusive regime and a faster superdiffusive one. Kinetic Monte Carlo simulations attribute the departures to one-dimensional single-file channels, nanoscale confinement, lattice softening and ion-electron coupled reactions. The standard diffusion equation is incomplete for this material. Together the four results show simple material models breaking on a nickel redox reservoir, slow Dirac bands, superheated copper and single-file lithium channels.[3], [4], [1], [2]
