The Photoisomerization Pathway(s) of Push–Pull Phenylazoheteroarenes**

Abstract

Azoheteroarenes are the most recent derivatives targetedt of urtheri mprovet he properties of azo-based photoswitches. Their light-induced mechanism for trans-cis isomerization is assumed to be very similart ot hat of the parent azobenzene. As such, they inherited the controversy about the dominant isomerization pathway (rotationv s. inversion)d epending on the excited state (np*v s. pp*). Althought he controversy seems settledi na zobenzene, the extent to which the same conclusions applyt ot he more structurally diverse family of azoheteroarenes is unclear. Here, by means of non-adiabatic molecular dynamics, the photoisomerization mechanism of three prototypicalp henylazoheteroarenes with increasing push-pull characteri su nraveled. The evolutiono fthe rotational and inversion conical intersection energies, the preferred pathway,a nd the associated kinetics upon both np*a nd pp*e xcitationsc an be linked directlyw ith the push-pull substitution effects. Overall, the working conditions of this family of azo-dyes is clarified and ap ossibility to exploit push-pull substituents to tune their photoisomerizationm echanism is identified, with potentialimpact on their quantum yield. [**] Ap revious versiono ft his manuscript has been deposited on ap reprint server (

Tuning the Thermal Stability and Photoisomerization of Azoheteroarenes through Macrocycle Strain**

“…Comparison of the photoisomerization kinetics of compounds 2113 and 1127,w ith as horta nd al ong alkyl bridge, respectively, and the unsubstituted compounds 2 and 1,s tudied in ref. [62].I ts hows the (top) ratio of trajectories reachingaC oIn before the time limit (1 ps), and (below) the characteristic times describedi nt he main text.S ee Ta ble S5.3 (in the Supporting Information)f or an assessment of thee rror associated with these values. Figure 3.…”

“…[62] That is mainlya ssessed from the time required to reach aC oIn in our trajectories (t CoIn ), anda lso with t S1 and t S2 ,d escribed as the time spent in S 1 and S 2 states,r espectively.I nu nsubstituted azoheteroarenes,t he S 2 /S 1 decay is ultrafast (100-200 fs), whereas the overall time to reacha nS 1 /S 0 CoIn after irradiation is approximately 500 fs, with small differences depending on the compound and on the excitation energy. [62] In the bridged heteroarenes 2113 and 1127,t he photoisomerization is significantly faster.I nt he case of 2113,a ll 25 trajectories initiated at S 1 andS 2 reach an S 1 /S 0 CoIn at an average t CoIn of 27 and 165 fs, respectively.I nb oth cases, the S 1 PES leads very efficiently towards aC oIn, with very short residence times in S 1 (t S1 )o f2 7a nd 37 fs. Thus, the longer t CoIn associated with the S 2 excitation is exclusively due to the extra S 2 !S 1 relaxation step, the timescale of which (t S2 ¼ 128 fs) is indeedv ery similar to the non-bridge compound (t S2 ¼ 113 fs).…”

“…[46,83] Unsubstituted compounds 1 and 2 feature as imilar distribution of CoIn A -a nd CoIn B -like structures within the crossing seam, which indicates that both rotationa nd inversion pathways are accessible. [62] Such scenario is completely different in bridgeda zoheteroarenes;w eo bserve in Figure 3apronouncedp reference of both 1127 and 2113 towards the rotationp athway.O ne reason is that the E-isomer minimao fb oth compounds already showa non-planar CNNC angle, so the initial structures are indeed closer to the CoIn A -like region of the crossing seam than in acyclic heteroarenes.T his has been previously identified as a drivingf orce towards af ast photoisomerization in bridged AB. [26] Another reason is that the opening of the CNN angles associated with the inversion pathway is energetically too unfavorable, presumably owing to the steric constrains: the trajectories of 2113 barely explore CNN angles above 1408 (see FigureS5.2 in the Supporting Information), which is where the inversion-like CoIna ppear in 1 and 2.…”

“…To evaluate the impact of the bridgeo nt he photochemistry of the azoheteroarene derivatives, we have studied compounds 2113 and 1127.T hese are examples of short-and long-bridge derivatives, respectively.W eh ave selected these specific compounds because both feature the E-isomer as the most stable one. That is an advantage for practical reasons and in long-bridge derivatives, the E-isomerh as brighter transitions at Franck-Condon (FC) than the Z-isomer (see Ta blesS5.1 and S5.2 in the Supporting Information for np*a nd pp*s tate energies and oscillator strength),a nd also for the sake of comparison, because recent literature has dealt with the E-to-Z photoisomerization of 1 and 2, [62] providing an excellent opportunity to compare the photochemistry of bridged versus non-bridged azoheteroarenes.M oreover,u sing ac ompound with as horter bridget han 2113 would have implied the study of the Z-to-E isomerization, which is expected to be less affected by ab ridge. [26] To study the photoisomerization of 2113 and 1127,w ehave used the same computational protocol as in ref.…”

“…[26] To study the photoisomerization of 2113 and 1127,w ehave used the same computational protocol as in ref. [62].T hati s, we performed non-adiabatic molecular dynamics simulations (NAMD)b ased on the fewest-switches surface hopping method. [79] Swarms of 25 trajectories were initiat-ed at both the S 1 (np*) and S 2 (pp*) states for 1127 and 2113, which are located at 433/268nma nd 449/300nm, respectively, at the Franck-Condon geometry (the evolution of the np*a nd pp*b ands with the bridge length can be found in sectionS5 in the Supporting Information, revealing as ignificant increase of band separation for short-bridged derivatives).T he trajectories are then propagated until an S 1 -S 0 energy gap below 0.1 eV is reached( see Computational Details).…”