Processing history attached to each value separates conflicting evidence on metastable (Si)GeSn semiconductors

GeSn and SiGeSn with several per cent Sn are metastable direct-gap group-IV semiconductors: anneals, contact sinters or hot measurements can move Sn out of the lattice, so a value describes the film at one point in its processing, not its composition alone. Published values rarely carry that history, and machine-learned simulations add values whose errors are known in kind but not in size; two disagreeing values may reflect an error or two different samples. The hypothesis tested was that recording the processing history with every value separates conflicting evidence that composition and temperature leave unresolved. Each value was stored as one event of a time-resolved evidence hypergraph, whose entries link states before and after, defect species, process step, source and time window on separate clocks for sample processing and campaign decisions; defect reactions must balance. The record held 551 literature values from 163 anneal series, 3028 defect events from molecular dynamics with the machine-learned potential MACE-MP-0, and 159 campaign entries. MACE-MP-0 kept Ge$_{0.88}$Sn$_{0.12}$ crystalline for 0.5 ns only at 500 K; at 600 and 700 K the model lattice disordered, so a 0.13 eV hop activation energy across 500 to 700 K measures disorder, not a barrier. The same potential bound Sn to a vacancy by $0.43 \pm 0.08$ eV. Literature activation energies for Sn diffusion in Ge-rich material spanned 1.13 to 3.10 eV but fell into groups no wider than 0.7 eV once anneal regime and starting state were recorded, and an oxide cap alone decided whether a Si$_{0.08}$Ge$_{0.88}$Sn$_{0.04}$ film degraded at 550 °C. These conflicts follow processing history rather than composition, and a foundation potential’s dynamics hold only within a stated temperature range. Anneals of metastable films between 380 and 650 °C therefore call for the low-temperature film activation energy, not the implanted-bulk value.