Computational Discovery of a V-Ti-Sn-Ga High-Entropy Superconductor Candidate

We report the computational identification of a high-entropy alloy (HEA) superconductor candidate usinga physics-informed simulated annealing optimiser (GEM) applied to the thirteen-element set reported asanomalous transmutation products in an ultrasound-mercury experiment (Cardone et al., 2017). Acrossmore than 44,000 independent optimisation runs spanning nine systematic passes, the optimiserconverged reproducibly on the composition V0.796Ti0.103Sn0.067Ga0.034 — which we designate theSTAVING alloy, an anagram of its constituent elements (V, Ti, Sn, Ga). The alloy has a valenceelectron count (VEC) of 4.762 and an Allen-Dynes modified McMillan estimated Tc of approximately 13K. This composition has not been reported in the HEA superconductor literature, which is dominated byTa-Nb-Hf-Zr-Ti systems. The convergence of the attractor is independent of the resonance hypothesisused during optimisation, as demonstrated by a full frequency sweep (5-50 kHz) producing identicalcompositions regardless of drive frequency. We present the computational methodology, theprogressive element-set refinement that identified the STAVING alloy, the physical basis for thepredicted Tc, and five falsifiable experimental predictions. The origin of the element set in the Cardoneexperiment is described and assessed honestly; the computational prediction stands independently ofwhether that experiment is valid. (Code files have .txt extension due to upload constraints; rename to .py before use.)