Abstract
<title>Abstract</title> <p> Context Metal-free donor–π–acceptor sensitizers allow the frontier levels of a dye-sensitized solar cell to be tuned through the bridge, but in a solid-state cell the photo-oxidized dye must also be reduced by the polymeric hole conductor, so the two components must be designed together. We examined fifteen sensitizers built on seven heteroaryl-substituted thiophene bridges, each as its a- and b-position regioisomer, with seven Schiff-base polythiophene hole conductors. Moving the heteroaryl group to the b-position leaves the HOMO within 0.14 eV but deepens the LUMO by up to 0.43 eV, red-shifting the S <sub>1</sub> band from 361–374 to 393–414 nm at the cost of oscillator strength and of the injection margin, which falls from + 0.14–+0.31 to − 0.08–+0.10 eV. Polymer-limit oxidation potentials span 0.31–1.51 V vs. SCE; only the two most easily oxidized polymers regenerate every dye, leaving 32 of the 105 pairs admissible. The pairs of highest simulated efficiency, including the nominal best at 9.45%, fall outside that set, and a weighted multi-criteria score does not repair the ranking: feasibility must gate the design space, not merely contribute to it. Methods All molecular properties were computed at ωB97X-D/def2-SVP with the SMD continuum model for acetonitrile in Gaussian 16; excitation energies from linear-response TD-DFT, hole–electron analysis in Multiwfn, and ionization potentials by ΔSCF. Regiochemical isomers were prescreened with GFN2-xTB and polymer limits reached by Tanaka 1/n extrapolation with Cardona calibration. Devices were simulated with SCAPS-1D and interpreted with a random-forest model and SHAP attributions. </p>