The quintet-to-singlet back-switching process in {[Fe(tpy)2]2+} (tpy = 2,2′:6′,2″-terpyridine), {[Fe(dftpy)2]2+} (dftpy = 6,6″-difluoro-2,2′:6′,2″-terpyridine) and {[Fe(dctpy)2]2+} (dctpy = 6,6″-dichloro-2,2′:6′,2″-terpyridine), together with the photorelaxation dynamics of the {[Fe(dftpy)2]2+} spin-crossover and the {[Fe(dctpy)2]2+} high-spin complexes with exceptionally long-lived 5/7MLCT states were investigated by full-dimensional trajectory surface hopping (TSH) dynamics.
Following photoexcitation, the initial population of the 5Eg state decays quickly via a branching mechanism of direct (5Eg → 5T2g) and 3MC-crossed (5Eg → 3MC → 5T2g) processes, with the latter dominating the dynamics.
The quintet-to-singlet conversion, the key step of the back-switching process, remains inefficient (<10%). This low efficiency arises because the singlet–triplet–quintet crossing region is readily accessible only during the initial sub-picosecond dynamics (<500 fs). Shifting the crossing region to improve its accessibility could substantially enhance the quintet-to-singlet conversion efficiency, suggesting that more extensive chemical modification could significantly optimize the back-switching process.
Upon excitation to the quintet MLCT manifold, the population decays within <200 fs to the dark 5Eg (5MC) state (followed by the relaxation to the 5T2g (5MC) in less than 1 ps); only 1–2% of the initial population remains in the ligand-centered state in all complexes. Because the previously performed experimental investigations were on picosecond timescales and used probe wavelengths at which the 5MLCT is bright but the 5MC is dark, this small MLCT population could be detected as a long-lifetime component.
We identify a mismatch in comparison to previous transient absorption spectroscopy assignments (analyzing picosecond timescales and not detecting dark 5MC states), which calls for further fs-resolved investigations.