Reference-state transferability limits in IR140: working-solution spectroscopy, bleachable optical depth, and apparent femtosecond saturation.

Aggregation-sensitive nonlinear absorbers are often characterised using dilute spectroscopic references, although their highly loaded nonlinear-optical (NLO) formulations may occupy different states. Here, IR140 is used to test whether reference-state assignments remain transferable to 800 nm femtosecond measurements. UV-visible-near-infrared absorption, steady-state photoluminescence, absolute photoluminescence quantum yield (PLQY), streak-camera time-resolved photoluminescence, and open-aperture Z-scan/I-scan measurements were combined with a unified finite-depth slow-absorber model. The dilute methanol-stock series develops a broad blue-shifted branch near 700-800 nm, whereas aqueous NLO solutions prepared from a dimethyl sulfoxide (DMSO) stock display a distinct red-shifted branch near 1000 nm. Thus, nominal solvent identity and a dilute spectrum do not uniquely define the formulation used for nonlinear measurement. Unified analysis yields apparent saturation fluences of 86.5-312.8 μJ cm-2 in methanol, 168.4 μJ cm-2 in DMSO, and 225.5-444.0 μJ cm-2 in aqueous formulations; the overlapping ranges show that solvent ranking alone is insufficient. These values are model-conditioned estimates rather than independent-preparation averages. Along the methanol-stock reference branch, PLQY decreases from 14.55% to 0.08%, while the monitored near-infrared emissive channel shortens markedly. The measured picosecond emissive relaxation is too slow to control bleaching within a 35 fs pulse, although unobserved ultrafast non-emissive channels cannot be excluded. We therefore establish an IR140-specific reference-to-working-state audit: when spectra are not equivalent, the low-fluence transmittance T0 must be fixed from the formulation-matched working solution, and comparisons should be restricted to equivalent formulations or explicitly model the mismatch.