Two-electron transfer stabilized by excited-state aromatization

The scientific significance of excited-state aromaticity concerns with the elucidation of processes and properties in the excited states. Here, we focus on TMTQ, an oligomer composed of a central 1,6-methano[10]annulene and 5-dicyanomethyl-thiophene peripheries (acceptor-donor-acceptor system), and investigate a two-electron transfer process dominantly stabilized by an aromatization in the low-energy lying excited state. Our spectroscopic measurements quantitatively observe the shift of two π-electrons between donor and acceptors. It is revealed that this two-electron transfer process accompanies the excited-state aromatization, producing a Baird aromatic 8π core annulene in TMTQ. Biradical character on each terminal dicyanomethylene group of TMTQ allows a pseudo triplet-like configuration on the 8π core annulene with multiexcitonic nature, which stabilizes the energetically unfavorable two-charge separated state by the formation of Baird aromatic core annulene. This finding provides a comprehensive understanding of the role of excited-state aromaticity and insight to designing functional photoactive materials.

I have carefully read the revised manuscript, SI, and the authors' careful responses to all of the referees.
Whilst I think that the data themselves are very nice, I still have concerns about their interpretation.
In the authors' response to referee 1's question about Mulliken population analyses, it becomes clear that the computational data and experimental results do not present a fully consistent picture. Since the Mulliken analysis doesn't show charge-transfer in T1, it is of no surprise that the calculated IR spectrum of the T1 state is also in poor agreement with experiment -it just seems like there is excessive delocalization in the DFT picture. The TRIR is unequivocal, based on the CN str, that there is some CT character. But, once we no longer trust the calculated T1 state's charge distribution, I fear that we cannot trust much else: we can't trust the bond length alternation plots in Fig 4 (or the HOMA/dihedrals), nor the calculated IR. Since much of the ms focuses on demonstrating the aromaticity in the CT T1 state based on computational metrics, this seems to me an unfortunate flaw. I would suggest that the poor agreement of Mulliken populations may arise from the choice of B3LYP as a functional, whereas a range-separated functional may be more appropriate for this CT case.
All of the referees were concerned by the poor match of the calculated and experimental T1 IR spectra. I think that the authors' explanation that the band at ca. 1400 cm-1 is an artefact is helpful. However, even excluding that band, the match is not great and does not feel good enough to provide such a key pillar of evidence for the claims of the ms. Perhaps another functional more suited to CT states may help? Alternatively, it would perhaps be possible (though certainly more complex and challenging) to divide the molecule into fragments and treat those individually (e.g. the M10A core dication, or M10A-thiophene dication). In such a circumstance, comparisons of the single-electron transfer case (ie M10A[-thiophene] cation doublet) to the dication could be helpful. Realistically this fragmentation approach may be too challenging, since the choice of how/where to break up the molecule is quite uncertain.
In their response to Referee 3's second query, the authors have provided a Figure showing their spectroelectrochemical setup. The problems of diffusion/spectral purity can be avoided by aiming the FTIR laser through the working electrode grid. The authors seem (based on Figure 3-1) to have aimed their beam through a region of electrolyte, rather than through the electrode. There should be no particular deleterious effect on the FTIR spectra from the presence of the grid, but much purer spectra (ie pure oxidation states) will be obtained. It should certainly be possible to measure pure dianion by holding a potential beyond that at which the dianion is generatedmeasurement of the monoanion may require multicomponent curve resolution, or may be practically impossible if the monoanion spectrum is a linear combination of those for the neutral and dianion states. I remain to be entirely convinced that two electrons transferred in the photophysical experiments, and not just one. Based on the authors' explanation, the ground-state spectrum of 'dianion' in Fig 2g (left,lower) is actually probably monoanion, or at best a mixture of neutral, dianion, and monoanion. It is certainly *not* pure dianion! Any dianion formed at the working electrode would disproportionate by the time it got to the region measured by the FTIR (shown in Fig 3-1). It would be good to repeat the spectroelectrochemical measurements to obtain pure spectra of monoanion and dianion, and use these to assign the TRIR spectrum in Fig 2g ( top, left). The most compelling assignment would come from the use of both peak positions and relative intensities/oscillator strengths.

Response to Reviewers' Comments
First of all, we are sincerely appreciative of the reviewers who have obviously taken considerable time out of their own busy lives to help us make this a better paper through their thoughtful consideration and advice. Our responses and specific revisions to the reviewers' comments are shown below.
[  TMTQ, Angew. Chem. Int. Ed. 2015, 54, 5888 (Ref 24 in the new version) and Chem. Eur. J. 2016, 22, 2793. Clearly, as a compound which could become important in pi-conjugated spin-bearing materials, TMTQ is starting to attract considerable interest, and the current manuscript makes a major contribution to the understanding of its excited state behaviour.
Our response: We sincerely appreciate the reviewer's comprehensive understanding of our scientific findings with his/her great interest. As mentioned in the reviewer's comments, we have also concentrated on revealing an effect of excited-state aromaticity and strongly believe that our study provides a new and crucial insight into the understanding of excited-state behaviors and their applications. Therefore, we are grateful to the reviewer's positive comments and full attention on our study. My view on the originality of this work remains unchanged. The revised manuscript is surely closer to meet the acceptance criteria of Nat. Comm. in terms of its potential significance and broad interest.
However, the spectroscopic analysis of the triplet state remains inconclusive.
Our response: We are glad that the reviewer recognized our efforts to deliver clearer description of our scientific findings. And we also sincerely thank the reviewer for letting us know what we have missed and need to explain more. First, for the reviewer's concern on our IR analysis based on TD-DFT calculations, we did not just suggest the exclusion of the most intense IR band around 1400 cm -1 simply for a spectral similarity. The IR band around 1400 cm -1 was also overestimated in the TD-DFT frequency results for the S1 state even though TMTQ showed less structural changes in the S1 state and the other IR bands in the lower energy region (1250~1400 cm -1 ) exhibited weak IR intensities. To explain this feature, we put Figure S11 in the previous version of supplementary information.
For the reviewer's another concern on the triplet state, we appreciate the reviewer's calling this issue to our attention and are sorry to have confused you with our rather insufficient explanation. In this study, we experimentally observed the two electron transfer process, as all reviewers conceded. Thus, what we experimentally observed and focused on is not the triplet state but the excited CT state. In our experimental results, all the excited state dynamics and processes occurred rapidly with the time constants of less than tens of picoseconds, which is too fast to consider the intersystem crossing between singlet and triplet states as long as there is no heavy-metal effect. We also described this point in our previous version of manuscript, "As mentioned, Baird's rule describes the reversed aromaticity in the excited triplet state, which reminds that the core 8π annulene of TMTQ should be triplet in the CT state. Furthermore, the excited CT state, that we proposed in which Baird aromaticity takes place, is a singlet excited state which resides at around +1.8 eV over the singlet ground electronic state. However, the true T1 state detected by heating from the ground state in reference 24 is at around +0.15 eV over the ground electronic state. However, the joint discussion on these different excited singlet and triplet states that might have confused the reviewer is argued in the following way: 1) after electronic structure calculations at the RAS-2SF-srPBE level of theory we conclude that the singlet CT state can be described by the coupling of two correlated triplets in which one of them is locally placed on the central annulene core (but this state is overall a singlet). It is on this local pseudo triplet that Baird aromaticity applies. This local pseudo triplet state on the core annulene might resemble the structure of the heating accessible true T1 state BUT the involved states are largely different!. 2) Based on the previous similitude on (1) and also based on the argument on line 9, page 11 in the revised version of the manuscript (see below) we attempt to use calculations on the true T1 state to QUALITATIVELY understand the vibrational properties of the excited CT state only.
In fact, the electronic structure of TMTQ in its singlet excited state manifold (not triplet manifold) was computationally analyzed with the multi-configurational calculation with RAS-2SF-srPBE functional as well as the TD-DFT calculation (see section 2.4 in Figure 4d in the revised manuscript).
Although multi-configurational calculations show higher accuracy in the electronic structure, they are not optimized to vibrational frequency analysis. Hence, TD-DFT frequency results of TMTQ, that show an excellent consistency with the experimental data in the ground state at the B3LYP-D3/6-311G(d,p) level, were obtained for the S1 and T1 states and compared with the experimental excited-state IR spectra.
It is here that we use the TD-DFT vibrational infrared calculations on the T1 state to QUALITATIVELY interpret the experimental excited state IR spectrum. This approach has been already used for the excited singlet state (Sn) in our recent study (Chem 2017, 3, 870-880) empirically suggested that, for the excited singlet state, the larger conformational changes, being more close to the T1-state optimized structures rather than S1-state ones, are expected.
Taking all these points into consideration, our IR analysis with the TD-DFT frequency results for the T1 state is considered to be reliable for interpreting our experimental IR data and revealing the conformational changes of TMTQ in the excited states. To deliver this information more perspicuously, we changed the related parts in the revised manuscript. Once again, we sincerely thank the reviewer for helping us communicate more clearly in our manuscript.
Regarding the controversy described in reference 51, we are aware that the triplet discussed in this article is the true T1 state (thermally accessible from the ground electronic state) which is characterized by a 15% Baird aromaticity contribution according to the authors (they defined it as a Hückel-Baird hybrid). In the same article, the authors described that for higher excited states the contribution from Baird aromaticity will progressively increase and this is actually our case in which the maximal Baird aromaticity contribution is in fact prepared by previous two electron transfer and generation of the CT state.

>>>(Page 11, line 7 in the revised manuscript)
In the comparative analysis between the experimental and calculated IR spectra (Fig.3), the IR spectrum calculated for the S0 state of TMTQ showed excellent consistency with that obtained experimentally. This reveals that the S0 optimized geometry derived from calculations and the true ground electronic state molecular structure of TMTQ are certainly close. To obtain valuable theoretical data for IR spectra in the excited states, TD-DFT calculations were carried out in the S1 and T1 states. These two lowest lying related excited states are well known to be structurally similar in closed-shell molecules. On the other hand, the accurate prediction of spectroscopic properties in the excited states of π-conjugated molecules remains a challenge for quantum chemical methods, a situation which is particularly difficult for the elucidation of excited singlet states and significantly ameliorated in the corresponding triplet state due to the distinctive electron-electron correlation. Hence, it is sometimes preferred to consider TD-DFT calculations of the excited triplet state to understand the homologue excited singlet state rather than conduct them on the singlet state itself. Going to the current results in our study, the experimental transient IR spectrum of TMTQ is compared with those obtained for the S0 and T1 states  from which we observe that, in line with the discussion above, the resemblance to the T1 state is better than that for the S1 state, from which we consider the former to qualitatively understand the changes in the transient IR experimental spectrum and use these results to guide and address the qualitative changes in molecular geometries and conformations in the excited two-electron CT state. Our response: We appreciate the reviewer's attention on our study and recognition of our experimental results. We have already discussed this point with the reviewer 2 and are sorry for our short explanation to confuse the reviewers. In our study, what we have discussed and experimentally detected is not the triplet state but the excited CT state. Our experimental measurements showed the rapid excited state dynamics with the time constant of less than tens of picoseconds, which is too fast to consider intersystem crossing between singlet and triplet states as long as there is no heavy-metal effect. This point is also mentioned in our previous version of manuscript, "As mentioned, Baird's rule describes the reversed aromaticity in the excited triplet state, which reminds that the core 8π annulene of TMTQ Here, we insist on the fact that we relied on the calculations on the T1 state ONLY for an interpretation of the vibrational spectra. Although multi-configurational calculations show a higher accuracy in the electronic structure analysis, they are poor in analyzing conformational changes and vibrational frequencies. On the other hand, the TD-DFT calculations provide highly accurate results for the ground state conformations and vibrational frequencies. In our study, the TD-DFT frequency results of TMTQ in the ground state from B3LYP-D3/6-311G(d,p) show an excellent consistency with the experimental data. In this regard, we compared the experimental excited-state IR spectra with the TD-DFT frequency results for the S1 and T1 states for an in-depth investigation of the experimentally observed IR spectral changes. In particular, the reproduced similar IR spectral features, the contrasting IR intensities between lower and higher energy region, in the TD-DFT frequency results for the T1 state assisted our qualitative interpretation of experimental IR data. In the same line, for more detailed information for the IR spectral changes, we also comparatively analyzed the TD-DFT optimized structures in the S0, S1 and T1 states with HOMA and dihedral angle analysis methods. All remaining experimental and computational data were described in the manuscript for the excited CT state. It must be highlighted that the excited CT and T1 states are different.
In addition, the excited singlet state (Sn) calculations are still a challenging issue in the quantum chemical calculations. Thus, the TD-DFT optimized structures for the Sn states are incomplete.
Moreover, the recent study (Chem 2017, 3, 870-880) empirically suggested that, for the excited singlet state, the larger conformational changes, being more close to the T1-state optimized structures rather than S1-state ones, are expected. Taking these points into consideration, our IR analysis with the TD-DFT frequency results for the T1 state is considered to be reliable for QUALITATIVELY interpreting our experimental IR data and revealing the conformational changes of TMTQ in the excited state. To deliver this information more precisely, we modified the related parts in the revised manuscript. Once again, we sincerely thank the reviewer for helping us communicate more clearly in our manuscript. >>>(Page 11, line 7 in the revised manuscript) In the comparative analysis between the experimental and calculated IR spectra (Fig.3), the IR spectrum calculated for the S0 state of TMTQ showed excellent consistency with that obtained experimentally.
This reveals that the S0 optimized geometry derived from calculations and the true ground electronic state molecular structure of TMTQ are certainly close. To obtain valuable theoretical data for IR spectra in the excited states, TD-DFT calculations were carried out in the S1 and T1 states. These two lowest lying related excited states are well known to be structurally similar in closed-shell molecules. On the other hand, the accurate prediction of spectroscopic properties in the excited states of π-conjugated molecules remains a challenge for quantum chemical methods, a situation which is particularly difficult for the elucidation of excited singlet states and significantly ameliorated in the corresponding triplet state due to the distinctive electron-electron correlation. Hence, it is sometimes preferred to consider TD-DFT calculations of the excited triplet state to understand the homologue excited singlet state rather than conduct them on the singlet state itself. Going to the current results in our study, the experimental transient IR spectrum of TMTQ is compared with those obtained for the S0 and T1 states (Fig. S10-12) from which we observe that, in line with the discussion above, the resemblance to the T1 state is better than that for the S1 state, from which we consider the former to qualitatively understand the changes in the transient IR experimental spectrum and use these results to guide and address the qualitative changes in molecular geometries and conformations in the excited two-electron CT state. Our response: We deeply appreciate the reviewer's full attention and supportive advice on our study.
According to the reviewer's careful guidance, we checked the TD-DFT calculation results from various range-separated functionals, CAM-B3LYP and a series of M06 functionals (M06, M06-L, M06-2X and M06-HF). Mulliken population and C≡N stretching IR frequency results showed a similar change in charge distribution and no significant difference was observed in the calculation results with the rangeseparated functionals ( Figure R1-R3). On the other hand, the IR spectral data in the C=C stretching region with the range-separated functionals showed an inconsistency with the experimental results ( Figure R4). Based on these results, it is considered that the calculation results with B3LYP-D3/6-311G(d,p) is most suitable for IR data analysis under the current circumstance. Figure R1. The experimental and simulated IR spectra for C≡N stretching mode of TMTQ. Figure R2. Mulliken population analysis of TMTQ. Figure R3. The calculated C≡N stretching IR spectra of TMTQ in the S0, S1, T1 and dianion states. Our response: We are grateful to the reviewer's concern with a sincere advice on our work, "Two- Figure R4. The experimental ground state (a) and calculated (b-c) IR spectra of TMTQ in the range of 1250~1700 cm -1 . electron transfer stabilized by excited-state aromatization". According to the reviewer's kind guidance, we deliberated the computational analysis of individual fragments. As the reviewer mentioned, "Realistically this fragmentation approach may be too challenging, since the choice of how/where to break up the molecule is quite uncertain", this approach required a careful choice of fragments because a wrong choice of fragments can give false information. In particular, TMTQ is an unusual molecule because the dicyanomethyl groups at both sides cause a unique quinoidal structure, the alternation pattern between C-C and C=C bonds, in the ground state ( Figure 4b in the manuscript). Thus, in our analyses, the ground state fragments without one or both of dicyanomethyl groups did not reflect the distinctive quinoidal character of mother system, TMTQ. Taking these points into consideration, we analyzed the M10A core dication fragments, where the triplet dicationic character well reflects the quinoidal nature of TMTQ system with its aromaticity change ( Figure R5). Here, upon structural planarization with enhanced aromatic nature, the C=C stretching IR intensity of M10A dication becomes significantly attenuated, which is well matched with our experimental observations and provides a reliable support for our IR data interpretation. Therefore, we put this information in our revised supplementary information as Figure S13. Once again, we appreciate the reviewer's considerate advice to improve our manuscript.

>>>(Page 13, line 4 in the revised manuscript)
For more deliberate interpretation for the IR spectral changes, we have also analysed the C=C stretching IR bands of triplet M10A dication fragment in various degrees of structural distortion (Fig. S13), where the quinoidal nature of TMTQ system with its aromaticity change was well reflected by the triplet dicationic character. The TD-DFT frequency results showed that, upon the structural planarization with enhanced aromatic nature, the C=C stretching IR intensity of M10A dication becomes significantly attenuated, which is well matched with our experimental observations. >>>( Figure S13  grid, but much purer spectra (ie pure oxidation states) will be obtained. It should certainly be possible to measure pure dianion by holding a potential beyond that at which the dianion is generatedmeasurement of the monoanion may require multicomponent curve resolution, or may be practically impossible if the monoanion spectrum is a linear combination of those for the neutral and dianion states. I remain to be entirely convinced that two electrons transferred in the photophysical experiments, and not just one. Based on the authors' explanation, the ground-state spectrum of 'dianion' in Fig 2g (left,lower) is actually probably monoanion, or at best a mixture of neutral, dianion, and monoanion. It is certainly *not* pure dianion! Any dianion formed at the working electrode would disproportionate by the time it got to the region measured by the FTIR (shown in Fig 3-1). It would be good to repeat the spectroelectrochemical measurements to obtain pure spectra of monoanion and dianion, and use these to assign the TRIR spectrum in Fig 2g (top, left). The most compelling assignment would come from the use of both peak positions and relative intensities/oscillator strengths..

Our response:
It is well known that most of the electrochemical reductions of tetracyano π-conjugated compounds show only one two-electron reduction wave in the cyclic voltammetry experiment (J. Am. Soc. Chem., 2002, 124, 12380-12388). In addition, in most of these compounds, in the infrared spectroelectrochemical experiment (like that in TMTQ), the spectral evolution goes from the neutral directly to the dianion without traces of radical anion. We firmly think this is our case here. No disproportionation reaction occurs in our reduction experiment. The particular setup and configuration of the used thin layer spectroelectrochemical cell, in which we carry out the experiment, definitively helps to avoid the disproportionation reaction. The optical path of the IR beam in the experiment crosses mainly two differentiated parts: 1) the solution which is outside the volume where the reduction takes place (i.e., defined by the Nernst diffusion layer and that consequently gives rise to the spectrum of the neutral molecule; and 2) the Nernst diffusion layer region of the cell in which the electrochemical reaction forms the dianion without intermediate oxidation states. In these conditions, we are confident that the spectra of the reduced species we obtained correspond exclusively to the dianion without traces of the radical anion.
The evidence that we get only the dianion species is that the infrared spectra of radical anion and dianions of similar molecules are well differentiated such as in the case of TCNQ (tetracyano quinodimethane) or TCNE (tetracyanoethylene, Angew. Chem. Int. Ed., 2006, 45, 2508-2525. So even trace amount of radical anion of TMTQ that could be formed would be clearly observed and distinguished. In addition, we agree with the referee in the fact that if the volume with the neutral compound (outside the Nernst diffusion layer region) and that with the dianion (inside the Nernst diffusion layer region) would mix then a disproportionation reaction would occur. However, this "mixture" does not happen neither in the voltammetry cyclic experiment nor in our spectroelectrochemical experiment.
For the dianion and CT state, it is obvious that the dianion species of TMTQ obtained by electrochemical reduction differ from the CT state obtained after photoexcitation in the fact that the CT state bears a dication in the central core and the dianion does not. The electron withdrawal effect of this central dication over the electron density on the external cyano groups (to which the CN stretching frequency is very sensitive) in the CT state would alter the amount of electron transfer or delocalization in this CT state compared with the case of the dianion. However, both frequencies in the CT and electrochemical dianion are the same. The possible difference is cancelled by the radical delocalization in the CN groups (not existing in the dianion) that is favored in the CT state due to the multiconfigurational character.
In response to my queries, the authors provided a new set of comments and revisions . This time, I feel that the discussion reached the necessary level of completeness. Even if some assertions may be speculative at this point, the work provides a unified spectroscopic picture of a really interesting molecule, and may be recommended for publication.
Reviewer #4 (Remarks to the Author): The authors have dealt with most of my concerns, except the final one (relating to spectroelectrochemistry).
In the authors' previous response to referees, they presented a figure showing their spectroelectrochemical cell setup, indicating that the light beam passed outside the working electrode.
This setup is not, in my view, optimal. It is preferred to direct the light source through the Pt grid electrode. The design of a typical OTTLE cell is such that the depth of analyte solution on either side of the working electrode (bounded by the cell windows) is much smaller than the thickness of the diffusion layer. In this way, 'pure' spectra can be obtained when the beam passes through the electrode.
In their recent response, the authors state that: (a) The beam passes through the 'neutral' region of the cell, giving spectrum of neutral molecules, and... (b) the beam passes through the 'reduction' region of the cell (within the electrochemical diffusion layer), giving spectrum of the dianion, but... (c) it is impossible for these solutions to mix. If they did mix, then the authors concede that disproportionation would occur The authors' assertion that this 'mixture' is impossible does not appear to be well founded, but I would appreciate clarification on this point. If they have indeed measured the spectrum using the experimental setup shown in the earlier response to referees (Fig 3.1 in that response), then their beam clearly passed through a region in which reduced species from the working electrode are diffusing into the neutral solution. This diffusion doesn't just 'stop' so some admixture is guaranteed: the concentration profile of reduced species will reduce asymptotically to zero with distance from the electrode; there's no interface as such between reduced and bulk material.
The authors' reference relates to TCNE, a much smaller compound which I would expect to have different electrochemical properties, since the CN groups are much more closely linked. Nonetheless, the paper doesn't really support their argument: the CN stretch frequency which they report for TMTQ2-(2170 cm-1) is in the region reported for the monoanion of TCNE (2150 cm-1 -2200 cm-1).
It should be possible to get a pure spectrum of dianion using spectroelectrochemistry, with no contamination from 'neutral' species. Alternatively, other methods can be used, such as chemical reduction.
I suggest the following means to conclusively resolve this issue: 1. Re-measure the spectroelectrochemistry of TMTQ, passing the probe beam through the Pt working electrode 2. Generate the dianion of TMTQ chemically, by addition of an excess of suitable reducing agent I suggest that it is important to have a pure spectrum of the dianion since much of the interpretation relies on having made the dianion by double CT in the TRIR experiment.
It is, of course, possible that the spectrum of the monoanion is the same as a linear combination of the dianion and neutral species, if the M10A linker does not permit strong electronic communication between the CN groups (i.e. Robin-Day Class 1 molecule). Then the monoanion would not have a distinct spectrum.
Minor comment: annulene is misspelled in the caption to Fig 3.

Response to Reviewers' Comments
First of all, we are sincerely appreciative of the reviewers who have obviously taken considerable time out of their own busy lives to help us make this a better paper through their thoughtful consideration and advice. Our responses and specific revisions to the reviewers' comments are shown below. (c) it is impossible for these solutions to mix. If they did mix, then the authors concede that disproportionation would occur The authors' assertion that this 'mixture' is impossible does not appear to be well founded, but I would appreciate clarification on this point. If they have indeed measured the spectrum using the experimental setup shown in the earlier response to referees (Fig 3.1 in that  Our response: We are glad that the reviewer recognized our efforts to deliver clearer description of our scientific findings. And we also sincerely appreciate the reviewer's supportive comment and considerate advice for letting our manuscript more reliable and obvious. As the reviewer mentioned, we also agree with the importance of IR spectra of TMTQ dinaion because its C≡N stretching IR band plays a critical role in our scientific discovery, 'Two-Electron Transfer Stabilized by Excited-State Aromatization'. In this regard, according to the reviewer's kind advice, we have carefully remeasured the IR spectrum of electrochemically reduced TMTQ in the region of 2140~2240 cm -1 , for the C≡N stretching IR bands. To solve the residual problem, we have reduced the length of the path traversed by the IR beam. With this, there is a lack of absorbance but also we are able to reduce the volume of the solution volume outside the diffusion layer, and obtained the clear IR spectrum manifesting the pure TMTQ dianion obviously. Figure 1 below shows that recording the IR spectra under application of an overpotential of ≈ -0.8 V, we achieved almost full disappearance of the C≡N stretching IR band from neutral TMTQ which takes place in parallel with the progressive emergence of the band at 2170 cm -1 of TMTQ dianion. There is a clear isosbestic point between the spectra of the neutral and dianion species which reveals that the conversion goes through an equilibrium of these two species converting one into the other. Thus, the blue spectrum in Figure 1b is now clearly due to a >90-95% of the conversion of neutral TMTQ into the dianion thus corresponding to its almost pure IR spectrum. To clarify more this situation we also display in Figure 1c, the UV-Vis-NIR absorption spectra of TMTQ upon the electrochemical reduction carried out in the same conditions as the IR spectroelectrochemistry of Figure 1a. Again full conversion of the neutral into the dianion with a clear isosbestic point is observed with a main band at 560 nm for the dianion. We have also scanned the UV-Vis-NIR spectroelectrochemical response of TMTQ in the oxidation part and this is shown in Figure 2. During oxidation, the neutral species converts into an oxidized species which is characterized by a band around 470 nm ( Figure 2a). Interestingly, the position of this band concurs with the PIA (photoinduced absorption) band in the transient absorption (TA) spectra, which was assigned arising from the two-electron CT state (see red arrows in Figure 2a and 2b). This suggests that this transient species have a core with electron deficient character which justifies the resemblance of the bands between the TA spectra and UV-Vis-NIR spectra of the oxidized form of

TMTQ.
From these spectroscopic comparisons, the CT state of TMTQ can be clearly described with the shift of two π-electrons from the central M10A donor to both DT acceptors. To deliver this information, we change Figure 2g and put additional figures, Figure S7, in the supplementary information. of TMTQ in toluene (b), CH 2 Cl 2 (c), and CH 3 NO 2 (d). The inset plots are the decay profiles at 660 nm. e,f. The FT-IR (e) and transient IR (f) spectra of TMTQ in toluene (top), CH 2 Cl 2 (middle), and CH 3 NO 2 (bottom). The transient IR spectra were plotted within ~100 ps time window (from purple to dark red colored lines). g. The transient IR spectra of TMTQ in CH 2 Cl 2 . The spectra in the regions of 2140−2240 and 1275−1575 cm -1 are assigned to IR bands for C≡N and C=C stretching vibrations, respectively. The FT-IR spectra of neutral and electrochemically produced dianion TMTQ are inset.