Theoretical Study of the Photolysis Mechanisms of Methylpentaphenyldimetallanes (Ph3MM′Ph2Me; M, M′ = Si and Ge)

The mechanisms of the photolysis reactions are studied theoretically at the M06-2X/6-311G(d) level of theory, using the four types of group 14 molecules that have the general structure, Ph3M–M′Ph2Me (M and M′ = Si and Ge), as model systems. This study provides the first theoretical evidence for the mechanisms of these photorearrangements of compounds that contain a M–M′ single bond. The model investigations indicate that the preferred reaction route for the photolysis reactions is, as follows: reactant → Franck-Condon (FC) region → minimum (triplet) → transition state (triplet) → triplet/singlet intersystem crossing → photoproducts (both di-radicals and singlets). The theoretical findings demonstrate that the formation of radicals results from reactions of the triplet states of these reactants. This could be because both the atomic radius and the chemical properties of silicon and germanium are quite similar to each other and compared to other group 14 elements, their photolytic mechanisms are nearly the same. The results for the photolytic mechanisms that are studied in this work are consistent with the available experimental observations and allow for a number of predictions for other group 14 dimetallane analogues to be made.

The photochemistry of disilanes that have various substituents have attracted intense interest because they are used in the study of fundamental chemistry problems, as well as for their potential applications . It is established that the direct irradiation of aryldisilanes in solution results in the formation of transient silenes, which are formally derived from the disproportionation and recombination of the silyl free radicals that are developed in the homolysis of the Si-Si single bond (Scheme 1) [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41]. Although there has been much experimental study of the photoreactions of aryldisilanes and their derivatives, to the authors' best knowledge, no theoretical studies deal with the photochemical mechanisms of these silicon-based compounds.
The TD-DFT computations [44,45] also demonstrate that the excited states (kcal/mol) increase in the order (Figure 1):    Namely, the theoretical evidences indicate that the lowest-lying excited state for these four molecules (1-Si-Si, 1-Si-Ge, 1-Ge-Si, and 1-Ge-Ge) is the first excited triplet state (T 1 ). It is well known that spin-allowed absorption cross-sections are typically larger than those for spin-forbidden excitations. As a result, photo-excitation promotes these four molecules to a singlet excited electronic state (S 1 ), and the molecule subsequently relaxes, branching between the T 1 and S 0 states, so considerations start on the triplet state energy surface. This computational data strongly suggest that the photolysis reactions of 1 must firstly progress on the triplet states and must only include the σ → σ* transition. As a result, the triplet state surfaces are the main focus of this study.
The essential characteristic of the photochemical mechanisms of 1 is the location of the intersystem crossing in the excited-triplet and the ground-singlet electronic states. In this study, since triphenylsilyl and methyldiphenylsilyl radicals are observed as photoproducts of 1-Si-Si (Scheme 1) [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41], the Si-Si single bond in 1-Si-Si must be broken. Accordingly, a Si-Si σ bond cleavage mechanism is suggested for the conversion of 1-Si-Si to triphenylsilyl and methyldiphenylsilyl radicals, and this is shown in Figure 2(i) [48]. It is worthy of note that although the silicon-silicon σ bond cleavage does not require full optimization of 1-Si-Si, this at least shows that there is degeneracy between a HOMO (σ) and a LUMO (σ*) when the silicon-silicon single bond is lengthened. Increasing the M-M single bonds in 1-Si-Ge, 1-Ge-Si, and 1-Ge-Ge, can also lead to degeneracy of the S 0 and T 1 energy surfaces at about 4.0 Å, which are estimates, as shown in Figure 2(ii-iv), respectively. These results are used to explain the mechanisms for the photolysis reactions of 1-Si-Si, 1-Si-Ge, 1-Ge-Si and 1-Ge-Ge in the following study. In order to understand the similarities and differences between the four types of photolysis reactions (Figure 2), the reaction profiles are shown in Figure 1, which also contains the relative energies for all of the critical points with respect to the energy of the corresponding reactants 1. The optimized geometrical structures are all collected in Figure 3  The potential energy surfaces for the photolytic mechanisms (Scheme 2) that are studied in this work are analogous to each other. That is to say, these theoretical computations suggest that the reaction mechanisms for all of the photolysis reactions for 1 proceed, as follows: 1(S 0 ) + hν → (1-T 1 )* → 1-T 1 − Min → 1-T 1 -TS → 1-T 1 /S 0 → 2,3-S 0 → 4,5-S 0 and 6-S 0 . In other words, the reactant (1) is initially excited to its lowest-lying triplet state (T 1 ), which is denoted as (1-T 1 )*, by absorbing light. From this (1-T 1 )* point, the molecule relaxes to a local minimum (1-T 1 -Min) near to the S 0 geometry. When the M-M bond distance is increased, the transition state (1-T 1 -TS) is obtained, which is still located on the triplet state surface. The experimental observations [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41] show that the photolysis reaction for 1-Si-Si is non-adiabatic [49]. The reaction begins from the triplet surface and advances ultimately along the ground singlet state pathway. Therefore, the intersystem crossing point of the T 1 and S 0 surfaces must play a key role in the mechanistic photolysis reactions of 1. As illustrated in Figure 1, the spin crossover from the triplet state (T 1 ) to the ground singlet state (S 0 ) states occurs in the region of the T 1 /S 0 intersection (1-T 1 /S 0 ). The results given in Figure 1 show that relaxing through 1-T 1 /S 0 may create the first photoproducts that have doublet radicals (i.e., 2,3-S 0 ; Ph 3 M· and ·M Ph 2 Me; path (I)). When the dispropotionation (path (II)) and the recombination (path (III)) processes are complete, the other singlet photoproducts are produced. These are 4,5-S 0 (Ph 3 M-H and H 2 C=M Ph 2 ) and 6-S 0 .    The computational results outlined in Figure 1 allow for the following conclusions to be drawn: (1) The theoretical studies using the M06-2X/6-311G(d) level of theory show that after the direct irradiation of aryldisilanes, arylgermasilanes, and aryldigermanes of the general structure, Ph 3 M-M Ph 2 Me (1), they occupy the lowest excited triplet state and then proceed to M-M bond homolysis via the intersystem crossing, to acquire two doublet radicals, 2 and 3 (overall singlet). After the subsequent dispropotionation and recombination procedures, they finally generate a mixture of singlet photoproducts (4)(5)(6). It should be emphasized that this is the first theoretical verification that radicals are formed as a result of reactions of the triplet states of Ph 3 M-M Ph 2 Me (1). (2) Figure 1 clearly shows that the reactants 1 have an excess of energy, because of the vertical excitation to the Frank-Condon point at the triplet state (1-T 1 )*, which is larger than the subsequent activation barriers. That is to say, when Ph 3 M-M Ph 2 Me (1) absorbs light, this molecule has sufficient internal energy to overcome the energy barriers to produce various types of photoproducts (2-6) without any difficulty. This theoretical finding for the 1-Si-Si molecule is in good agreement with the experimental evidence [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41]. Since there are no pertinent experimental or theoretical reports on the photolysis reactions of the compounds, 1-Si-Ge, 1-Ge-Si, and 1-Ge-Ge, this result is a prediction. (3) The theoretical observations that are shown in Figure 1 indicate that the relative energies of two monoradical species (Ph 3 M·and·M Ph 2 Me) are higher than those of the other singlet photoproducts (4)(5)(6). The quantum yields of the former are estimated to be lower than those of the latter. This prediction is again consistent with the available experimental observations for the 1-Si-Si molecule, in which it is reported that the formation of the corresponding silyl radicals (2 and 3), 1,1-diphenylsilene (5), and the silatriene (6) give respective yields of ca. 20%, 35%, and 42% [30]. Therefore, in the photolysis reactions for compounds 1-Si-Ge, 1-Ge-Si and 1-Ge-Ge, the quantum yields for their singlet photoproducts 4-6 should be larger than those of the other two doublet radicals, 2 and 3 [50]. That is to say, the direct irradiation of these compounds produces the triplet species, which then mainly results in M-M bond homolysis to generate the formation of a radical and other singlet photoproducts (similar to 4-6), via radical disproportionation and recombination. Indeed, some earlier experimental studies of 7d by Mochida and Hayashi [51] and a related those for series of molecules (8b-8d), as reported by Mochida and Gaspar [52,53] and their co-workers, show similar results that are in reasonable agreement with these theoretical predictions.
In conclusion, this M06-2X/6-311G(d) study presents the first theoretical explanation for the photolytic mechanisms of group 14 dimetallanes that contain a M-M single bond. This work demonstrates that the spin crossover plays a central role in the photolysis reactions of these compounds, from the excited triplet state to singlet ground state. In principle, the theoretical evidence that is given by this study demonstrates that both the potential energy surfaces and the order of the photochemical reactivity are similar for these group 14 dimetallanes, regardless of whether silicon or germanium elements are concerned. This may be because silicon and germanium are quite analogous to each other, in terms of both their atomic radius (117 pm and 122 pm for silicon and germanium, respectively) and their chemical properties [54,55]. Therefore, their photolytic behavior should be analogous, as outlined in Figure 1.
It is hoped that this work will stimulate further study of the photolysis reactions of substituted group 14 dimetallanes that have a M-M σ bond.