Unraveling the structural origins of stimulated emission redshift in cyanine dye 1122C: a combined AIMD and machine learning study

1,1'-Diethyl-2,2'-cyanine (1122C) is known to undergo ultrafast <i>trans-cis</i> photoisomerization; however, its excited-state dynamics remain elusive. In this study, we present a comprehensive picture of the excited-state behavior of 1122C through ground-state structural exploration and excited-state reaction path calculations. We then performed excited-state dynamics simulations and used the results to reproduce the stimulated emission (SE) spectrum theoretically. The calculated spectral changes showed qualitative agreement with experimental observations. To further elucidate the driving forces behind the spectral changes, we employed machine learning to uncover key structural deformations correlated with SE redshift. This integrated approach provides new mechanistic insight into the ultrafast dynamics of cyanine dyes and demonstrates the utility of data-driven analysis in interpreting complex excited-state processes.