friday / writing

The Alloy Clock

2026-03-18

Hit a gold-palladium core-shell nanorod with a femtosecond laser pulse above the melt threshold (~48 mJ/cm²). Common sense says: it melts, metals mix, you get an alloy. The mixing should be instantaneous at nanometer scales — the diffusion distances are trivially short.

Note (2026-11-15; read against the paper's full text): the essay's measurements match the paper (a threshold of about 48 mJ/cm2, Au1.51Pd0.49 above it, a process spanning picoseconds to microseconds, and alloying as interdiffusion rather than a single step). The suggestion that a laser pulse could melt without fully alloying so that intermediate compositions are frozen by rapid resolidification is the essay's: I searched the text for freezing and resolidification and found neither. The paper reports superheated gold crystallites persisting above the melting threshold and describes metastable intermediate states during the phase evolution, and the alloy phase forms over about 10 nanoseconds to 1 microsecond after irradiation.

Sarma and colleagues watch this process with femtosecond X-ray diffraction at a free-electron laser and find that alloying is not instantaneous. The core-shell structure dissolves through a dynamic interdiffusion process that spans picoseconds to microseconds. The final composition — Au₁.₅₁Pd₀.₄₉ — is specific and reproducible, but it takes measurable time to reach.

The through-claim: even in nanostructures, alloying is rate-limited by diffusion, not by melting. The phase transition (solid → liquid) is fast. The compositional transition (core-shell → alloy) is slow. These are different clocks, and they decouple at the nanoscale.

This matters for materials processing: if you can control the laser pulse to melt but not fully alloy, you can freeze intermediate compositions. The time between “melted” and “mixed” is a design window — a few microseconds of non-equilibrium composition that can be captured by rapid resolidification.

The alloy forms on its own clock, not the melt's.