A Descriptor-Constrained Target-Response Closure Assessment for a Frozen ECSM Electron-Like Packet in Measured Molecular Scattering

This paper reports V41 of the ECSM electron-like packet measured-scattering programme. V38–V40 established that a frozen ECSM electron-like packet does not reproduce measured molecular scattering unaided, but that a target-response layer can substantially reduce held-out residuals across sevoflurane and methane. V40 showed that the ECSM response form survives generic baseline competition but does not win it outright. V41 performs the next closure test: whether the target-response layer can be constrained by simple molecular descriptors rather than treated as a separate molecule-by-molecule fit. Two measured elastic electron-scattering targets are used: sevoflurane and methane. Both are evaluated over the measured 25–125 degree angular domain, with 50, 150, and 250 eV used for training and 100, 200, and 300 eV held out for testing. The frozen ECSM electron-like packet is unchanged. The V41 assessment compares bare ECSM, target-specific upper benchmark fits, a shared-global response, and descriptor-scaled constrained responses using electron count, heavy-atom count, atom count, and mass-per-electron proxies. The best constrained model is the atom-count-scaled response. It gives a two-target held-out mean relative residual of 0. 1711613305747449, compared with 0. 18230255059554087 for the target-specific upper benchmark and 1. 0214466733961964 for bare ECSM. The final V41 label is PASSV41CONSTRAINEDRESPONSECLOSESTARGETLAYER. This result suggests that the target-response layer can be compressed into a descriptor-scaled structure across the two available molecules. However, because only two targets are used, V41 is not claimed as a validated universal molecular descriptor law. It is a closure and identifiability result that motivates a third/fourth-target validation before any universal response claim is made.