Photoprocesses in Derivatives of 1,4- and 1,3-Diazadistyryldibenzenes

Photoprocesses in 1,4-diazadistyrylbenzene (1) and 1,3-diazadistyrylbenzene derivative (2) diperchlorates in MeCN were studied by absorption, luminescence, and kinetic laser spectroscopies. For compound 1, trans-cis-photoisomerization and intersystem crossing to a triplet state are observed. For compound 2, photoelectrocyclization is suggested. Quantum chemical calculations of diazadistyrylbenzene structures in the ground and excited states were carried out. The schemes for photoreactions were proposed.

The present study is focused on the derivatives of 1,4-and 1,3-diazadistyrylbenzenes (1 and 2, respectively, Figure 1). The aim of this work was a comparative study of the photonics of these compounds in MeCN by absorption, fluorescence, and kinetic laser spectroscopies, detecting the triplet state and intermediate products of photoreactions. In parallel with experimental studies, quantum chemical calculations were carried out to elucidate the structure of the chromophore in the course of photoexcitation, as well as to interpret the experimental absorption and fluorescence spectra.
The established ability of isomeric 1,4-and 1,3-derivatives of distyrylbenzenes to undergo various types of photochemical transformations, including those involving triplet states, has not been studied previously and is of considerable interest. This ability makes them promising for use in photoactive materials. The photoprocesses of diazadistyrylbenzenes under study can also be implemented in the construction of photoactive supramolecular systems on their basis with the use of macrocyclic compounds. distyrylbenzenes under study can also be implemented in the construction of photoactive supramolecular systems on their basis with the use of macrocyclic compounds.  Figure 2 shows the absorption spectra, and Table 1 represents the corresponding absorption band maxima and extinction coefficients for compounds 1 and 2. The absorption spectra of 1 and 2 differ both in shape and in the position of the absorption band maxima. A significant blue (40 nm) shift is observed for compound 2 as compared with 1. In this case, the extinction coefficients of both compounds are virtually identical.   Figure 2 shows the absorption spectra, and Table 1 represents the corresponding absorption band maxima and extinction coefficients for compounds 1 and 2. The absorption spectra of 1 and 2 differ both in shape and in the position of the absorption band maxima. A significant blue (40 nm) shift is observed for compound 2 as compared with 1. In this case, the extinction coefficients of both compounds are virtually identical.

Absorption and Fluorescence Spectra
distyrylbenzenes under study can also be implemented in the construction of photoactive supramolecular systems on their basis with the use of macrocyclic compounds.  Figure 2 shows the absorption spectra, and Table 1 represents the corresponding absorption band maxima and extinction coefficients for compounds 1 and 2. The absorption spectra of 1 and 2 differ both in shape and in the position of the absorption band maxima. A significant blue (40 nm) shift is observed for compound 2 as compared with 1. In this case, the extinction coefficients of both compounds are virtually identical.  The calculation shows that the absorption of 1 originates from a single delocalized electronic transition with high oscillator strength, while the absorption of 2 originates from four localized electronic transitions with lower oscillator strength, which sum up to almost the same intensity as the absorption of 1.  Figure 3 shows the fluorescence spectra of compounds 1 and 2, and Table 2 represents their absorption band maxima, fluorescence quantum yields, and fluorescence lifetimes. The calculation shows that the absorption of 1 originates from a single delocalized electronic transition with high oscillator strength, while the absorption of 2 originates from four localized electronic transitions with lower oscillator strength, which sum up to almost the same intensity as the absorption of 1. Figure 3 shows the fluorescence spectra of compounds 1 and 2, and Table 2 represents their absorption band maxima, fluorescence quantum yields, and fluorescence lifetimes.   Δλ is a Stokes shift between the fluorescence emission and excitation maxima; 1 trans isomer; 2 cis isomer.

Absorption and Fluorescence Spectra
A weak fluorescence of compounds 1 and 2 was detected. Fluorescence excitation spectra coincide with the corresponding absorption spectra. At the same time, compound 2 demonstrates a large Stokes shift of the fluorescence band (109 nm) and a significant (by 2.5 times) decrease in the fluorescence quantum yield compared with compound 1, which can be explained by a lower degree of conjugation and a significant deformation of  A weak fluorescence of compounds 1 and 2 was detected. Fluorescence excitation spectra coincide with the corresponding absorption spectra. At the same time, compound 2 demonstrates a large Stokes shift of the fluorescence band (109 nm) and a significant (by 2.5 times) decrease in the fluorescence quantum yield compared with compound 1, which can be explained by a lower degree of conjugation and a significant deformation of the structure upon photoexcitation of 2. The measured fluorescence lifetime is 120 ps for 1 and 160 ps for 2. The fluorescence kinetics for compound 2 was found to have a biexponential character, which can confirm the presence of another form of 2 in the excited state with a structure significantly different from the initial one.
The calculated radiative lifetimes are 20-50 times longer than the measured ones, which is consistent with the low fluorescence quantum yield (ϕ = τ/τ r ). In the absence of a quencher and the occurrence of the intersystem crossing to the triplet state, the fluorescence lifetime is determined as τ = 1/(k r + k nr + k ic ). The nonradiative lifetime τ nr for trans isomer 1 is 126 ps, for 2, it is 163 ps, whereas the radiative lifetime is 1420 and 8950 ps, respectively, thus being in agreement with the experimental values.

Kinetic Spectroscopy
Upon laser flash excitation at 355 nm, a change in the absorption spectrum of compound 1 is observed, as illustrated in Figure 4. The corresponding kinetic curves measured in the maxima of the photoinduced absorption bands are shown in Figure 5. ponential character, which can confirm the presence of another form of 2 in the excited state with a structure significantly different from the initial one.
The calculated radiative lifetimes are 20-50 times longer than the measured ones, which is consistent with the low fluorescence quantum yield (φ = τ/τr). In the absence of a quencher and the occurrence of the intersystem crossing to the triplet state, the fluorescence lifetime is determined as τ = 1/(kr + knr + kic). The nonradiative lifetime τnr for trans isomer 1 is 126 ps, for 2, it is 163 ps, whereas the radiative lifetime is 1420 and 8950 ps, respectively, thus being in agreement with the experimental values.

Kinetic Spectroscopy
Upon laser flash excitation at 355 nm, a change in the absorption spectrum of compound 1 is observed, as illustrated in Figure 4. The corresponding kinetic curves measured in the maxima of the photoinduced absorption bands are shown in Figure 5.  For compound 2, the difference in the photoinduced absorption in the visible spectral range is negligible and more than an order of magnitude smaller than for compound 1. Figure 6 illustrates the different spectrums of photoinduced absorption of compound 2 under the accumulation of absorption curves. The photoinduced absorbance change at 350-450 nm can be attributed to trans-cisisomerization of 1 to give a relatively long-lived cis-isomer ( Figure 5a). In this case, the rate of decrease in absorption at 400 nm does not depend on the ambient oxygen content in the solution and characterizes the process of cis-trans isomerization. The absorption band in the region of 500-650 nm is observed only for the deoxygenated solution and can be associated with the T-T absorption. The triplet state lifetime of 1 is 30 µs (Figure 5b).
The absence of bleaching in the principal absorption band of compound 1 can be attributed to the similarity of the absorption spectra of the trans and cis isomers of the compound.
For compound 2, the difference in the photoinduced absorption in the visible spectral range is negligible and more than an order of magnitude smaller than for compound 1. Figure 6 illustrates the different spectrums of photoinduced absorption of compound 2 under the accumulation of absorption curves.
For compound 2, the difference in the photoinduced absorption in the visible spectral range is negligible and more than an order of magnitude smaller than for compound 1. Figure 6 illustrates the different spectrums of photoinduced absorption of compound 2 under the accumulation of absorption curves. The observed photoinduced changes in the absorbance of compound 2 can be explained by assuming that these changes occur mainly in the short-wavelength region of the spectrum, which made it impossible to detect them by kinetic laser spectroscopy. In addition to spectral measurements performed upon laser excitation, the absorption spectra of compounds 1 and 2 were measured upon stationary UV irradiation. Figure 7 shows the obtained data. The observed photoinduced changes in the absorbance of compound 2 can be explained by assuming that these changes occur mainly in the short-wavelength region of the spectrum, which made it impossible to detect them by kinetic laser spectroscopy. In addition to spectral measurements performed upon laser excitation, the absorption spectra of compounds 1 and 2 were measured upon stationary UV irradiation. Figure 7 shows the obtained data. Upon UV irradiation, an increase in optical density of the solution of compound 1 was observed, which corresponds to a photoinduced change in absorption measured upon laser excitation ( Figure 4) and attributed to the formation of a long-lived cis isomer of 1, which has an absorption spectrum similar to that of the trans isomer and a significant extinction coefficient.
On the contrary, UV irradiation of compound 2 results in a decrease in absorption of the principal absorption band and an increase in absorption in the short-wavelength region of the spectrum, which is consistent with the experimental data obtained upon laser excitation. It can be assumed that for compound 2, photoelectrocyclization takes place to give the final product, which absorbs in the short-wavelength region of the spectrum. This conclusion is supported by the literature data confirming the participation of 1,3-distyrylbenzenes in the electrocyclization reaction [1][2][3]8]. Photocyclization of 1,3-diazadistyrylbenzene seems to occur in the singlet excited cis isomer, which has a short lifetime. The possible scheme of electrocyclization of compound 2 is depicted in Figure 8. Upon UV irradiation, an increase in optical density of the solution of compound 1 was observed, which corresponds to a photoinduced change in absorption measured upon laser excitation ( Figure 4) and attributed to the formation of a long-lived cis isomer of 1, which has an absorption spectrum similar to that of the trans isomer and a significant extinction coefficient.
On the contrary, UV irradiation of compound 2 results in a decrease in absorption of the principal absorption band and an increase in absorption in the short-wavelength region of the spectrum, which is consistent with the experimental data obtained upon laser excitation. It can be assumed that for compound 2, photoelectrocyclization takes place to give the final product, which absorbs in the short-wavelength region of the spectrum. This conclusion is supported by the literature data confirming the participation of 1,3-distyrylbenzenes in the electrocyclization reaction [1][2][3]8]. Photocyclization of 1,3-diazadistyrylbenzene seems to occur in the singlet excited cis isomer, which has a short lifetime. The possible scheme of electrocyclization of compound 2 is depicted in Figure 8.
excitation. It can be assumed that for compound 2, photoelectrocyclization takes place to give the final product, which absorbs in the short-wavelength region of the spectrum. This conclusion is supported by the literature data confirming the participation of 1,3-distyrylbenzenes in the electrocyclization reaction [1][2][3]8]. Photocyclization of 1,3-diazadistyrylbenzene seems to occur in the singlet excited cis isomer, which has a short lifetime. The possible scheme of electrocyclization of compound 2 is depicted in Figure 8.

Model
In general, the processes of laser flash photolysis experiments can be viewed as follows. The reactions of trans-cis and cis-trans isomerization proceed via a conical intersection (CI), and the minimum energy paths from the local trans or cis minima on the S 1 potential energy surface to the CI include saddle points, which limit the photoisomerization rate. The energy of the transition state relative to the trans or cis local minimum can be considered as the activation energy of trans-cis photoisomerization.
A laser pulse (355 nm) gives the molecule (trans-1 or trans-2) excessive energy, which may be sufficient to overcome the barriers of trans-cis photoisomerization. After that, the molecules can easily reach the region of the conical intersection and relax nonradiatively either to the trans or to the cis form in the ground state. This explains why the trans-cis isomerization takes place in the nanosecond timescale under the laser flash photolysis conditions. When photoinduced absorption spectra are measured (after at least 5 µs after the laser pulse), the excess energy is already dissipated either radiatively or nonradiatively, and the mixture consists of the initial molecule in the ground state and its primary photoproducts in their lowest states (S 0 for singlet products and T 1 for triplet products). In order to measure photoinduced absorption, the molecules in the mixture are excited directly to their absorption bands without pumping excessive energy. In this case, the barriers for the excited-state reactions can be overcome only through thermal activation, and the reaction rates can be estimated using the Arrhenius formula, given that the corresponding excited state is sufficiently long-lived. Therefore, after the laser pulse, the mixture consists of both trans and cis of 1 or 2 in their ground state and, where possible, their triplet forms. Upon measuring the photoinduced absorption spectra, the triplet forms absorb at~500-600 nm, and the singlet forms absorb at~350-400 nm, undergo thermally activated or barrierless phototransformations, or emit light (fluoresce).

Results
The structures of 1,4-diazadistyrylbenzene (1) and its electrocyclization product calculated by PBE0/6-31+G(d,p)/PCM(MeCN) are shown in Figure 9. The trans isomer is planar both in the ground and in the excited states. The cis isomer is nonplanar; the C-C=C-C torsion angle is~9 • in the ground state and~23 • in the excited state. These torsions are close to the same angles in the DHP (~8 and 18 • in the ground and excited states, respectively). The ground-state DHP is a local minimum, while the excited-state DHP is a conical intersection (CI). Trans-cis photoisomerization proceeds via another conical intersection at 90 • C-C=C-C torsion angle. The minimum energy path from both trans and cis minima to the CI includes two saddle points at~121 and 66 • C-C=C-C torsion angle.
The structures of 1,4-diazadistyrylbenzene (1) and its electrocyclization product calculated by PBE0/6-31+G(d,p)/PCM(MeCN) are shown in Figure 9. The trans isomer is planar both in the ground and in the excited states. The cis isomer is nonplanar; the C-C=C-C torsion angle is ~9° in the ground state and ~23° in the excited state. These torsions are close to the same angles in the DHP (~8 and 18° in the ground and excited states, respectively). The ground-state DHP is a local minimum, while the excited-state DHP is a conical intersection (CI). Trans-cis photoisomerization proceeds via another conical intersection at ~90° C-C=C-C torsion angle. The minimum energy path from both trans and cis minima to the CI includes two saddle points at ~121 and 66° C-C=C-C torsion angle.      -1 and (b) cis-1 and (c) its electrocyclization product dihydrophenanthrene DHP-1. Figure 10 shows the energy level diagram of diazadistyrylbenzene 1 (a), its possible phototransformations (b), and the absorption maxima of the phototransformation products (c). On the basis of the calculated data, we may infer the following picture of the photoprocesses in 1. As mentioned above, after the laser pulse, the mixture consists of trans-1, cis-1, and, possibly, 3 trans-1. When photoinduced absorption is measured, trans-1 absorbs at 420 nm (390 nm in the experiment) with further relaxation to the minimum of the S1 state ((trans-1)*). From the minimum, the molecule can emit light at 455 nm (470 nm in the experiment), convert to the triplet state ( 3 trans-1) through S1-T2 intersystem crossing with further phosphorescence and nonradiative relaxation, or undergo a trans-cis isomerization. Without excess excitation energy provided by the laser pulse, this process proceeds in the microsecond timescale and, therefore, is sufficiently slow. The reverse cis-trans isomerization in the excited state proceeds much faster, namely, in the nanosecond As mentioned above, after the laser pulse, the mixture consists of trans-1, cis-1, and, possibly, 3 trans-1. When photoinduced absorption is measured, trans-1 absorbs at 420 nm (390 nm in the experiment) with further relaxation to the minimum of the S 1 state ((trans-1)*). From the minimum, the molecule can emit light at 455 nm (470 nm in the experiment), convert to the triplet state ( 3 trans-1) through S 1 -T 2 intersystem crossing with further phos-phorescence and nonradiative relaxation, or undergo a trans-cis isomerization. Without excess excitation energy provided by the laser pulse, this process proceeds in the microsecond timescale and, therefore, is sufficiently slow. The reverse cis-trans isomerization in the excited state proceeds much faster, namely, in the nanosecond timescale and can take place even under the conditions of steady-state photoexcitation. Transformation to DHP-1 from any of the S 1 minima is unfavorable (denoted by red cross in Figure 10b) because the excited (DHP-1)* is~5.5 kcal/mol higher than the excited (cis-1)* minimum, and the excess energy received by the molecule upon primary flash excitation is dissipated.
The absorption bands of cis-1 and trans-1 are very close and overlap when measured (calculated 420 and 412 for trans-1 and cis-1, respectively). The emission proceeds in the nanosecond timescale, and the fluorescence spectra of both cis and trans isomers also overlap (calculated at 455 and 472 nm for trans-1 and cis-1, respectively).
The rate of phototransformation to the triplet state was not evaluated. However, it is worth noticing that T 2 lies only~6 kcal/mol lower than S 1 both in the transand cis-S 1 minima, and the nature of the T 2 state is different from the nature of the S 1 state (see Figure  S3 of Supplementary Materials). According to the El-Sayed rule, ISC is efficient in this case, leading to further evolution of the triplet molecule.
For the 1,3-diazadistyrylbenzene (2) and its electrocyclization product, the corresponding structures calculated by the same method are shown in Figure 11. The trans isomer is planar both in the ground and in the excited states. The cis isomer is nonplanar; the C-C=C-C torsion angle is~8 • in the ground state and~14 • in the excited state. These torsions are even closer to the same angles in the DHP (~8 and 16 • in the ground and excited states, respectively) than in 1. The ground-state DHP is a local minimum, while the excited-state DHP is a conical intersection. Trans-cis photoisomerization proceeds via another conical intersection at~90 • C-C=C-C torsion angle. The minimum energy path from both trans and cis minima to the CI includes two saddle points at~128 and 57 • C-C=C-C torsion angle. Figure 12 shows the energy level diagram of diazadistyrylbenzene 2 (a), its possible phototransformations (b), and the absorption maxima of the phototransformation products (c). On the basis of the calculated data, we may infer the following picture of the photoprocesses in 2.
After the laser pulse, the mixture consists of trans-2, cis-2, and, possibly, 3 trans-2. When photoinduced absorption is measured, trans-2 absorbs at 314-372 nm (350 nm in the experiment) with further relaxation to the minimum of the S 1 state ((trans-2)*). From this minimum, the molecule can either emit light at 409 nm (455 nm in the experiment), transform to the triplet state through the S 1 -T 3 intersystem crossing with further phosphorescence or nonradiative relaxation, or undergo a trans-cis isomerization. Without excess excitation energy provided by the laser pulse, this process proceeds in the microsecond timescale and, therefore, is sufficiently slow. The reverse cis-trans isomerization proceeds slower by order of magnitude, which may result in the accumulation of the cis-2 isomer.
From the (cis-2)* minimum, the molecule can transform to the excited (DHP-2)*, which has almost the same energy as the excited (cis-2)* minimum. The (DHP-2)* is a conical intersection and serves as a funnel for the nonradiative relaxation of (cis-2)*. The ground-state DHP-2, which absorbs in the visible region, is not accumulated due to its fast oxidation to the corresponding phenanthrene-2 (Phen-2), whose absorption band (calculated 358-361 nm) overlaps with those of the trans-2 and cis-2.
torsions are even closer to the same angles in the DHP (~8 and 16° in the ground and excited states, respectively) than in 1. The ground-state DHP is a local minimum, while the excited-state DHP is a conical intersection. Trans-cis photoisomerization proceeds via another conical intersection at ~90° C-C=C-C torsion angle. The minimum energy path from both trans and cis minima to the CI includes two saddle points at ~128 and 57° C-C=C-C torsion angle.
ISC in the (trans-2)* can proceed from the S 1 to the closest triplet level, which is T 3 . However, both S 1 and T 3 states are formed by the same orbitals (see Figure S3 of Supplementary Materials), and this makes ISC inefficient according to the El-Sayed rule (denoted by red cross in Figure 12b). This explains the absence of photoinduced absorption at 500-600 nm in 2.
The absorption bands of cis-2 and trans-2 are even closer than in 1 and overlap (calculated 372 and 368 nm for trans-2 and cis-2, respectively). The emission proceeds in the nanosecond timescale and the fluorescence spectra of both cis and trans isomers also overlap (calculated at 409 and 422 nm for trans-2 and cis-2, respectively).  Figure 12 shows the energy level diagram of diazadistyrylbenzene 2 (a), its possible phototransformations (b), and the absorption maxima of the phototransformation products (c). On the basis of the calculated data, we may infer the following picture of the photoprocesses in 2.

Materials and Methods
The synthesis of perchlorates of 1,4-and 1,3-diazadistyrylbenzenes (1 and 2) is beyond the scope of this study and will be published elsewhere. 1 H-NMR spectra of 1 and 2 can be found in Supplementary Materials, Figures S1 and S2.
The absorption spectra were recorded on the Agilent 8453 spectrophotometer. The luminescence spectra were recorded on the Varian Eclipse spectrofluorimeter. Fluorescence lifetimes were measured on the Fluotime 300 spectrofluorimeter. The difference absorption spectra of photoproducts and kinetics decay were measured by means of the nanosecond laser flash photolysis apparatus [22]. Irradiation was performed using the third harmonic of the Nd-YAG laser Solar (λ = 355 nm). The dissolved oxygen was removed by the argon bubbling of the solution. Stationary irradiation of diazadistyrylbenzenes solutions was performed using a DKSSh-150 lamp with a UFS-6 glass filter. For the measurement of fluorescence quantum yields, 9,10-diphenylanthracene (quantum yield 0.92 in ethanol [23]) was used as a standard. The accuracy of quantum yield measurement was 10%. All measurements were carried out in MeCN (high purity grade, grade 0, Cryochrome) at room temperature.
Kinetic spectroscopy allows us to draw conclusions about the presence of certain photoreaction products from indirect data (photoinduced absorption spectra and kinetic curves). Since the concentration of the diazadistyrylbenzenes is very low, it is not possible to accumulate and isolate (in general, unstable) photoreaction products in their pure form.

Calculation
The structures of the cis and trans isomers of diazadistyrylbenzenes 1 and 2 and their electrocyclization products (dihydrophenanthrenes DHP-1 and DHP-2) in the ground S 0 and singlet excited S 1 states, as well as the transition states of the trans-cis photoisomerization on the S 1 potential energy surfaces and DHP oxidation products, phenanthrenes Phen-1 and Phen-2, were optimized by the DFT with PBE0 functional and 6-31+G(d,p) basis set using FireFly program package [24] partially based on GAMESS(US) code [25]. The solvent effect (MeCN) was included through the dielectric polarizable continuum model (D-PCM) [26].
The absorption and emission spectra were calculated by TDDFT with the same functional, basis set, and solvent model. The absorption spectra were calculated after geometry optimization of the ground state, while the emission spectra were calculated after geometry optimization of the π-π* excited state. The vibrational frequencies and normal modes were calculated for all the studied stationary points of the ground and S 1 excited states. The triplet energies were calculated by TDA [27].
The radiative lifetime τ r was calculated as follows: where k r is the fluorescence rate constant, f 0i is the oscillator strength, and ν i0 is the frequency (cm −1 ) of the transition. The trans-cis and cis-trans photoisomerization times τ tc were calculated using the Arrhenius equation: where E A is the activation energy calculated for the corresponding transition state on the S 1 potential energy surface, c is the speed of light, and ω 0 is the frequency of the reorganizational mode [28][29][30].
Since the energy profile of the isomerization and subsequent electrocyclization goes through conical intersection regions, the applicability of TDDFT needs special discussion. It is known that TDDFT is inapplicable to cases with quasidegeneracy. In the quasidegeneracy region, the excited state geometry optimization procedure fails because the gap between the ground and excited state (or the excitation frequency) becomes zero, and Casida equations do not have positive eigenvalues [31]. However, for our qualitative conclusions, we neither need the exact energy of the minimum energy CI nor its exact structure. For our purposes, it is sufficient to know that both (DHP-1)* and (DHP-2)* lie in the region of the conical intersection. If this intersection is peaked [32], it acts as a reaction funnel. By TDDFT, we may roughly demonstrate that the near-CI region for (DHP-1)* lies higher than the local cis and trans minima of the S 1 state, while for (DHP-2)*, this region has lower energy and, therefore, a barrierless path from the cis form to the CI is possible in 2.
The calculations were also repeated using BHandHLYP functional to demonstrate that the results are qualitatively the same and only slightly depend on the weight of exact exchange in the functional ( Figure S4 of Supplementary Materials) This methodology has been benchmarked against multireference (XMCQDPT2/CASSCF) calculation in [33].

Conclusions
It was found that the photoprocesses of 1,4-and 1,3-isomers of diazadistyrylbenzenes differ substantially. The phototransformation of the 1,4-diazadistyrylbenzene derivative involves trans-cis isomerization and intersystem crossing to the triplet state. In this case, the formation of DHP-1 is energetically unfavorable. For 1,3-diazadistyrylbenzene, trans-cis isomerization is followed by the electrocyclization from the cis-form to generate DHP-2 and further fast oxidation with atmospheric oxygen to give phenanthrene, while intersystem crossing is not efficient. The considered photoprocesses of diazadistyrylbenzenes can be implemented in the construction of photoactive supramolecular systems on their basis with the use of macrocyclic compounds.