3.1. Co-Sensitization of the Dyes in the Cells
We investigated the effect of co-adsorption of MK2 and D205 dyes on TiO2 by stationary absorption measurements of the photoanodes, current–voltage and IPCE measurements of the complete cells and femtosecond transient absorption measurements of whole cells treated as samples. Different ratios of MK2:D205 concentration, different types of TiO2 layers and different electrolytes (iodide- and cobalt-based) were tested. No common anti-aggregation additives (like chenodeoxycholic acid, CDCA) were used for sensitization in order to clearly observe the effects of the interaction of the two dyes in the mixtures.
Figure 1A,B presents examples of sensitization effects on the absorption properties of a relatively thin (2.5 μm) layer of TiO
2 at a D205:MK2 concentration ratio of 1:3.2 in the solution. As can be seen in
Figure 1A, the minimum D205 absorption in this solution is around 450 nm, while in this spectral region the absorption of the MK2 dye is close to the maximum. The complementary absorption spectra of the two dyes were one of the reasons for choosing them for co-sensitization studies. The absorption spectra recorded on TiO
2 (
Figure 1B) show some important differences with respect to those in solution. First of all, the absorption onset of both dyes in solution is finished at around 650 nm (
Figure 1A), while those on TiO
2 extend above 700 nm for D205 and even above 750 nm for MK2 (
Figure 1B). The difference is even more pronounced in the IPCE spectra, shown in
Figure S2 (in the Supplementary Materials).
Secondly, unlike in the sensitizing solutions, the maximum absorbance of MK2 alone on TiO
2 is almost twice that of D205. Taking into account the about 1.5 times-smaller extinction coefficient of MK2 compared to D205 (in Tol: 38400 M
−1cm
−1 vs. 59300 M
−1cm
−1, according to [
31] and our current measurements, respectively), this implies that about 3 times more MK2 than D205 molecules are adsorbed on the same TiO
2 layer (absorbance ratio × extinction coefficient ratio: 2 × 1.5). Higher dye loading of MK2 can probably be explained by the dye structures (shown in
Figure S1 and in insets in
Figure 1C,D). The structure of MK2 is more “linear”, with the −COOH anchoring unit at one end, while in D205 the anchoring unit is in the middle of the dye structure; therefore, D205 cannot be as densely packed on TiO
2 nanoparticles as MK2.
Interestingly, D205 dye is adsorbed much faster on TiO
2 than MK2 molecules. This is probably due to the difference in the values of the acid dissociation constant between the MK2 and D205 anchoring groups. Therefore, thirdly, the dye ratio on TiO
2 is much different from that in solution, in favor of D205. As can be seen in
Figure 1B, the absorption spectra of the dyes’ mixture on TiO
2 (1:3.2 sensitizing D205:MK2 ratio) are much more similar to the spectra of D205. The mixture absorbance on TiO
2 can be fitted by adding 0.84 times the contribution of D205 absorbance and 0.24 times the MK2 absorbance (
Figure 1B), which, taking into account different extinction coefficients, corresponds to a ratio of 69% to 31% of D205 and MK2 molecules, respectively, on TiO
2.
Table 1 summarizes the fitted contributions of absorbance of both dyes and different ratios of molecules on TiO
2 for different initial concentration ratios in solution.
Next, the solar cells were prepared in different configurations, and their photovoltaic parameters were compared. Special attention was paid to the character of changes in the parameters of the cells with the dye mixtures, relative to those in the cells with one type of dye.
Table 2 shows the exemplary parameters measured for the as-produced cells (measured 60 min after preparation): a 12 μm thick TiO
2 layer composed of mixtures of nanoparticles with thicknesses of 18 nm and 30 nm and larger scattering nanoparticles and filled with Co-Bpy electrolyte. The highest short-circuit current density (J
SC) was found for the cells with MK2. However, the cells with a mixture of D205 and MK2 exhibited a much higher fill factor (FF); therefore, their overall PCE was slightly higher than that of MK2 cells.
The parameters for the six other configurations are summarized in
Tables S1–S6 (in Supplementary Information). The maximum relative errors (standard deviation/average value) of the photovoltaic parameters for the screen-printed cells were 1% for V
OC and 3% for FF, J
SC and PCE, while for doctor blade cells they were: 1% for V
OC, 6% for FF and J
SC and 10% for PCE. Particular photovoltaic parameters in each configuration depend on many factors that are well known from previous DSSC studies. For example, the open circuit voltage (V
OC) increases in different electrolytes in the order: I
−/I
3− (
Tables S1 and S2), Co-Bpy (
Table 2,
Tables S5 and S6) and Co-Phen (
Tables S3 and S4), in accordance with the more positive redox potential of the redox pair. Except for the iodide configuration without TBP, the J
SC of the cells with only D205 is significantly lower than that of those with only MK2, and the relative photocurrent (total_APCE [
34]) is also lower, indicating poorer efficiency of the absorbed photons’ conversion. The FF of D205 cells was always better than that of MK2 cells (probably due to the differences in molecular structures and better blocking of the TiO
2 surface from contact with redox pairs in the cells with adsorbed D205). Increasing the thickness of TiO
2 layer results in higher J
SC and higher PCE values, but the total_APCE becomes lower (
Tables S2–S6). However, what is most important is that the trends indicated in
Table 1 were confirmed for all other configurations in
Tables S1–S6: the use of a mixture of the dyes always results in the improvement of the J
SC with respect to the cells with D205, and in the improvement of FF with respect to the cells with MK2. In other words, the addition of MK2 to the cells with D205 does not affect their high FF, but it increases significantly their J
SC. On average, the photocurrent of the co-adsorbed mixtures is higher by 25 ± 8% (relative to that of D205 cells, standard error of the mean indicated), while the improvement in the fill factor is 10 ± 2% (relative to that of MK2 cells). The latter result is especially important because it proves that co-adsorbed active molecules can produce the same effect as commonly used passive anti-aggregation additives (e.g., CDCA). Such additives occupy an area on TiO
2 surface but, unlike the co-adsorbed sensitizing dyes, do not contribute to light absorption and electron injection into TiO
2.
The overall PCE of the cells with the dye mixtures is sometimes higher and sometimes lower than that of MK2 cells, depending on how much D205 cells show poorer performance than MK2 cells in a particular configuration. For example, for pure anatase TiO
2 (paste from Dynamo) and TBP in solution, the resulting shift of the TiO
2 conduction band edge (towards more negative potentials) reduces the driving force for electron injection, and the J
SC of the cells with D205 is then drastically lower than that of MK2 cells (
Tables S3 and S4). It should also be noted that D205 cells quickly lose their PCE due to the drop of the photocurrent on the time scale of hours, while MK2 cells keep their PCE constant or sometimes even improve their J
SC, probably due to the irradiation effects studied by us recently [
44]. For the cells with the D205:MK2 mixtures, a decrease in performance was also observed (but not as much as for the cells with D205 alone); therefore, after 4 h or more, the PCE of the MK2 cells was always the highest. However, their FF was always lower than that of the D205:MK2 cells.
In this section we will also present the results of femtosecond transient absorption studies of the cells with the dye mixture. The transient absorption of the cells was measured upon excitation at 470 nm and globally analyzed in a spectral range from 560 to 830 nm and in a temporal window of up to 3 ns.
Figure 2A,C shows the wavelength-dependent amplitudes of the fitted components (3-exponential fit with constant offset) and the corresponding time constants of the exemplary cells with only D205, only MK2 and their mixture, respectively. The relative error of the given time constants is less than 10%. The interpretation of the results for the cells with single dyes is performed along the same lines as in our earlier studies of indoline [
37,
38,
45] and carbazole dyes [
34,
35,
39,
40]; therefore, it will be only briefly summarized here. In the cells with MK2 only (
Figure 2A), the electron injection takes place on a fast time scale, within the first component (1 ps) and partially within the second component (14 ps). The third component of the fit for MK2 cells (520 ps) has the same spectral shape as that of the constant offset component; therefore, it is assigned to the partial recombination of the injected electrons from TiO
2 to the oxidized dye. The fastest component for D205 (
Figure 2B) of a time constant close to that of the IRF (0.24 ps) is due to the decay of the initially excited (Franck–Condon) state (also called the locally excited, or LE, state). The second (6 ps) and the third (230 ps) time constants were assigned to the decay of the subsequent charge transfer (CT) state of the indoline dye (hot and cold), respectively. Its decay is due to electron injection and undesired self-quenching of CT states. The residual spectrum (constant offset component) is proportional to the successfully separated charges and contains the contributions of the ground state depopulation (bleach, negative signal below 650 nm) and the absorption of oxidized dye and electrons into TiO
2 (positive signal above 650 nm).
The results for the cells with the dye mixture (
Figure 2C) contain the contributions from both the D205 and MK2 cells. The most important finding can be deduced from the constant offset component, which is dominated by the signal of MK2 (compare
Figure 2A,B). However, in contrast to the cell with only MK2, the intercept of the residual spectrum with the zero-absorption axis is shifted to shorter wavelength (around 660 nm for the cell with the mixture instead of 680 nm for that with MK2) and the contribution of the constant offset component is greater. Such changes are characteristic of the MK2 cells upon addition of CDCA, as observed by us earlier. In the cells with MK2 and CDCA, the ratio of the residual to the initial signal amplitude at 750 nm (close to the absorption maximum of the oxidized dye) was significantly higher than that for the cell without CDCA due to the suppression of electron recombination [
35]. A similar tendency can be clearly observed when monitoring the transient absorption kinetics at 750 nm for the samples studied, which are shown in
Figure 2D. As can be seen, the residual signal at 3 ns for the cell with the dye mixture (black line) is almost twice as high as that calculated from the individual contributions from the cells with D205 or MK2 (green line), which should be explained by a higher relative population of the charge-separated species at 3 ns and a smaller contribution of fast electron recombination. Along with the above-mentioned increased fill factor of the cells with the dye mixture (with respect to that of MK2 cells), this is more evidence that the co-adsorbed active molecule (D205) can play a similar role in blocking electron recombination as the CDCA additive. We have also tried to obtain additional information by analyzing the time evolution of the transient absorption signals.
Figure S4 (in Supplementary Materials) shows the transient absorption spectra of D205, MK2 and co-sensitized cells by D205 and MK2 at selected time delays, as well as the dye mixture spectrum obtained by adding the individual contributions of D205 and MK2 best fitting the above spectra. The contributions from individual dyes change from 0.8 × D205 + 0.55 × MK2 at 0.6 ps to 0.7 × D205 + 0.7 × MK2 at 3 ns. This means that the contribution of the MK2 signal increases slightly over time with respect to that of D205, which could be an indication of partial electron transfer from the excited state of D205 to the excited state of MK2. Such a process is indeed possible, as the LUMO energy of D205 is slightly higher than that of MK2 (potentials −0.91 V vs. SHE [
34] and −0.89 V vs. SHE [
31], respectively). However, such a hypothesis requires verification with further evidence.
3.2. Solvatochromic Studies and Anionic Forms of the Dyes
The stationary absorption spectra of D205 in ACN at different dye concentrations and after DBU addition (to create the anionic form: D205
−) and of MK2 in Tol, are included in
Figure 1C,D. The positions of the longest-wavelength absorption maximum for both dyes, under different conditions, are given in
Table 3. In general, the position of the absorption maximum of D205 is influenced by the type of solvent used and often changes after DBU addition. Taking into account the two solvent parameters [
46]: α (the ability of the solvent to donate a hydrogen bond) and β (the ability of the solvent to accept hydrogen bond), it seems that the β parameter has a greater impact on the position of the absorption maximum of D205 in a given solvent. The longest-wavelength absorption maximum was found for DCM (546 nm at β = 0.10), while the shortest-wavelength one was for EtOH (527 nm at β = 0.75). Similar conclusions were formulated on the basis of D205 solvatochromic studies with more solvents used; in the solvents with lower H-bond ability, a shift of the absorption maximum to the longer wavelengths was observed [
47]. Given the structure of the D205 molecule, it seems most likely that a hydrogen bond is formed between a hydrogen atom from the carboxyl group of D205 and a molecule of a given solvent. A change in the concentration of the dye may result in a significant or only slight shift of the absorption maximum, which is largely dependent on the solvent used. Taking into account the change in the concentration of D205 in ACN from ~10
−4 M (ACN_c) through 10
−5 M (ACN) to 10
−6 M (ACN_d), it can be concluded that the blue shift of the absorption maximum (from 528 nm to 519 nm) is observed as a result of dilution of the solution (
Figure 1C and
Table 3). By diluting the D205 solution in ACN, the formation of D205
− is preferred, similar to TPC1 dye [
48]. This is confirmed by the impact of using DBU. The addition of this organic base to a solution of D205 in ACN with a concentration of ~10
−5 M results in the formation of D205
−, accompanied by a shift of the absorption maximum from 528 nm to 523 nm. A pronounced blue shift after DBU addition was found in DCM, Tol and THF solutions. Moreover, the results obtained for D205 in DCM at concentrations in the range of ~10
−4 M to ~10
−6 M indicate that upon dilution of the solution, the absorption maximum is at the same wavelength (546 nm, spectra not shown in the manuscript). In the case of DCM, Tol and THF, it can be postulated that regardless of the dye concentration, the neutral species of the dye predominates. A completely different picture emerges when EtOH is used as a solvent of D205. The addition of DBU does not cause a significant shift of the absorption maximum (527 nm in EtOH and 528 nm after DBU addition). This may indicate the dominance of the anionic form of D205 in the pure EtOH. This is consistent with the hypothesis presented in the literature, which indicates that the anionic form of similar indoline dyes is favored in alcohols and ACN, while in alkyl halides (DCM) the neutral form of the dye predominates [
47,
49]. For the MK2 dye in Tol, after the DBU addition, a clear blue shift in the absorption maximum from 495 nm to 439 nm (
Table 3) was observed. A similarly large change in the absorption maximum position has been found for TPC1 (453 nm in DCM and 406 nm in DCM + DBU) [
48]. In addition, a large positive Mulliken charge (0.43) for the hydrogen atom in cyanoacrylic acid has been confirmed [
48]. Hence, it seems that for polar solvents with a high β parameter, including alcohols and ACN, the −COOH group is more easily deprotonated, and the anionic form of the dye is favored, unlike in solvents with low values of β (DCM, Tol). Much smaller changes in absorption upon the transition from the neutral to anionic form of D205 (~10 nm), compared to those noted for MK2 and TPC1 (~50 nm), indicate increased acidity of the hydrogen atom in the carboxyl group of D205 dye; therefore, it deprotonates more easily, and the dye attaches to the TiO
2 surface faster than MK2.
Next, we wanted to find out the impacts of the type of solvent used and DBU addition on the chemisorption of the D205 and MK2 dyes on the TiO
2 surface. The absorption spectra of D205 and MK2 from the solutions in DCM, ACN and EtOH, sensitized on the TiO
2 surface along with those recorded after the addition of DBU, are presented in
Figure S5. The same time of sensitization and the same D205 concentration (~10
−5 M) were applied. The greatest absorbance of the photoanodes was found after sensitization of D205 in ACN; absorbance was slightly lower in DCM, and it was the least for D205 dye from EtOH attached to TiO
2. This may indicate a beneficial effect of the equilibrium between the anionic and neutral forms of the dye on sensitization. The use of DBU in any case prevents effective chemisorption of D205 on the TiO
2 surface. A similar effect was observed for MK2 in Tol after the DBU addition. This is probably due to the interaction of the cation formed from the DBU with the TiO
2 surface.
Further confirmation of different anion–neutral form equilibria in different solvents comes from steady-state fluorescence and time-resolved studies. The fluorescence spectra of D205 (
Figure 3A and
Figure S6) in different organic solvents are significantly affected by DBU addition and the type of organic solvent used. In general, a blue shift of the fluorescence maximum was found in all solvents due to DBU addition (
Table 3). Only for EtOH does DBU addition not cause any changes in the position of the fluorescence maximum (627 nm). A distinct blue shift is observed for D205 in ACN in concentrations changed from 10
−4 M through 10
−5 M to 10
−6 M (637 nm, 628 nm and 621 nm, respectively). The value observed for the most diluted dye solution is the closest to that observed upon DBU addition to the D205 solution in ACN (623 nm). A clearly different result was found for D205 in Tol, for which the fluorescence maximum was located at a much shorter wavelength than in the other solvents (584 nm) and after DBU addition (571 nm). This may be related to the long-lived LE state in this solvent (instead of the CT state in the other solvents), which was confirmed by the transient absorption measurements (see below). An even more pronounced blue shift of the fluorescence maximum after DBU addition was found for MK2 in Tol (from 642 nm to 582 nm,
Figure 3B and
Table 3).
The excited-state lifetimes of D205/D205
− and MK2/MK2
− are presented in
Table 3 and were determined from the femtosecond transient absorption signal probed at 730 nm (for MK2 and D205 in Tol) or at 600 nm (for D205 in the other, more polar solvents). As reported before [
38,
39] and also confirmed below, 730 nm corresponds to the maximum of the transient absorption signal of the excited state of MK2 and the LE state of D205, while 600 nm corresponds to the transient absorption maximum of the CT sate of D205 in polar solvents. Examples of the transient absorption kinetics at 600 nm for D205 in ACN solutions at different dye concentrations and after DBU addition are included in
Figure S7. The excited state lifetimes change with D205 concentration in ACN (from 350 ps at ~10
−4 M to 480 ps at ~10
−5 M and finally to 520 ps at ~10
−6 M). After DBU addition to the solution, the excited-state lifetime is similar to that of ~10
−6 M without DBU. This confirms again the concentration dependence of the D205/D205
− equilibrium in ACN and the domination of the anionic form of D205 only in the diluted D205 solution in ACN. In contrast, almost no changes in lifetime were found in EtOH after DBU addition, confirming that the anionic form dominates at all dye concentrations. For both D205 and MK2 dyes, the fluorescence quantum yield (Φ
F) and radiative rate constant (k
r) values are slightly higher for the anionic form than for the neutral one (
Table 3). Moreover, Φ
F and k
r are also significantly higher for the LE state (D205 in Tol) than for the CT state (D205 in the other solvents studied), indicating the possible twisted character of the CT state. The CT-state lifetimes of the neutral and anionic forms of D205 show different dependence on the solvent parameters. The lifetime of D205
− increases with increasing β and decreases with increasing α and solvent polarity. In contrast, for the neutral form, the CT lifetime depends mainly on the polarity and is significantly reduced in the most polar ACN solvent.
Representative transient absorption spectra of D205 in DCM and after DBU addition and those of MK2 in Tol and after DBU addition, for selected time delays between the pump and pulse probe, are presented in
Figures S8 and S9, respectively. Transient absorption studies with the use of single solvents have been presented in our previously published articles for D149/D205 [
34,
38,
50] and MK2 [
35,
39]. The new contribution in our present study is the use of DBU, which induces the transition to the anionic form. At this point, we want to draw attention to the comparison between the transient absorption spectra of D205 in DCM (only neutral form) and the same solution after DBU addition (only anionic form). After excitation, the main positive bands, with maxima at ~720 nm and between ~1400 nm and ~1500 nm, were observed for D205_DCM and D205_DBU_DCM. Their positions can be assigned to the initially locally excited (LE) state of the neutral and anionic forms of D205. The negative feature of the spectrum, with a maximum at ~540 nm, is due to the ground state depopulation, while that at ~650 nm is due to the stimulated emission. On the sub-ps and single ps time scales, further evolution of transient absorption spectra was observed, with a red shift of the stimulated emission and changes in the shape of the positive signals (in VIS and NIR). Both for D205_DCM and D205_DBU_DCM, the transient absorption maxima were observed at ~600 nm and ~1400 nm, which can be attributed to the relaxed charge-transfer (CT) state. Generally, the transient absorption spectra of the neutral and anionic forms of D205 are quite similar. For D205, the transient absorption of CT state above 700 nm (in the range of stimulated emission) contains a more negative contribution from the neutral than from the anionic form. For MK2, the difference in the position of the negative stimulated emission band can be observed in the VIS range (more long-wavelength for the neutral form than for the anionic form), while the transient absorption amplitude in the band tail above 1000 nm in the NIR range was higher for the neutral from than for the anionic form.
To acquire more information about the solvation times and excited-state lifetimes, a global analysis of transient absorption spectra was performed.
Figure 4 shows the wavelength-dependent amplitudes of the fitted four exponential components and the corresponding time constants for D205 in ACN at different concentrations and after DBU addition, both in the VIS and NIR ranges. The results for the other solvents are given in
Figures S10–S15. The interpretation of the time constants follows that presented in our previous papers [
34,
38,
50]. In all solvents studied except for Tol, the two fastest components of the wavelength-dependent amplitudes are due to the ultrafast decay of the LE state and solvation dynamics. Solvation times increase in this order, from the shortest to the longest: ACN (below 1 ps), DCM, THF, EtOH and finally tert-Bu. In the latter solvent, the remarkably slow solvation with the longest component of >100 ps (which influences the third component in the global analysis) should be emphasized. The longest components represent the decay of the CT state (LE in the case of Tol). Their dependence on D205 concentration and DBU addition is fully consistent with the findings from the single-wavelength kinetics (in
Table 3). For ACN (
Figure 4) the lifetime increases from that for the most concentrated sample (dominated by the neutral form) to that of the most diluted one, which shows the same lifetime as for D205
− (the samples with DBU). For DCM (
Figure S11), whose neutral form is present for all D205 concentrations, the lifetimes are the same and significantly longer than upon addition of DBU. For EtOH with the domination of D205
−, the lifetimes with DBU and without it are the same (
Figure S12). Interestingly, the concentration-dependent measurements in tert-Bu reveal that the CT lifetime slightly increases upon dilution up to the values obtained for D205_tert-Bu_DBU (
Figure S15). This indicates that, despite having a higher β parameter (0.94 in tert-Bu vs. 0.75 in EtOH) not all dyes exist in the anionic form at moderate concentrations in tert-Bu. Most probably, the anionic form is additionally stabilized by the higher polarity of the solvent because its dielectric constant is significantly higher in EtOH (24.5) than in tert-Bu (10.9).
A conclusion following from this section is that the spectroscopic studies of D205 and D205 + DBU (a strong organic base) in solution have revealed the existence of an equilibrium between the neutral and anionic forms of D205 in different solvents. D205− has slightly blue-shifted (~10 nm) absorption and emission maxima with respect to those of the neutral form. D205 exists mainly in the neutral form in toluene, tetrahydrofuran and dichloromethane solvents, while in ethanol (having strong polarity and proton-accepting properties), it exists mainly in the anionic form. In acetonitrile (having only weak proton-accepting properties) and tert-butanol (of low polarity), a concentration-dependent equilibrium between D205 and D205− is observed, with more anions in more-diluted solutions. The changes in the absorption position as a function of the solvent polarity and the differences between neutral and anionic forms have been confirmed by TD-DFT calculations. The domination of anions in the D205 sensitizing solution was found to have a negative influence on the dye’s adsorption ability on a mesoporous TiO2 electrode. The highest dye loading was observed to take place from the acetonitrile solution; that from the ethanol solution was almost twice as low, while no adsorption was observed from the solutions containing D205 + DBU.
Femtosecond transient absorption studies in solution yielded information about the characteristic time constants of the complex D205 deactivation processes upon excitation. In non-polar toluene, the LE state decays with an average time of 200 ps (similar for neutral and anionic forms). In polar solvents, the subsequent CT state is formed within solvation times that vary: 0.5 ps in acetonitrile, 1.5 ps in dichloromethane, 3 ps in tetrahydrofuran, 30 ps in ethanol and up to >100 ps in tert-butanol. The radiative rate constant of the LE state is about 4 times higher than that of the CT state. The lifetime of the CT state of D205− decreases in more-protic solvents (from 500–600 ps in dichloromethane and acetonitrile to 350 ps in ethanol), while the lifetime of the CT state of D205 neutral form decreases with increasing solvent polarity (from 700 ps in dichloromethane to 350 ps in acetonitrile).
3.3. Theoretical Studies and Anionic Forms of the Dyes
To elicit more information about the electronic structures of D205 and MK2 in their neutral (D205/MK2) and anionic (D205
−/MK2
−) forms in the ground and excited states, density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations were performed. Geometries of D205 in a vacuum, D205
/D205
− using a CPCM solvation model with EtOH, DCM, ACN and Tol as solvents were optimized using the b3lyp functional together with a 6–31G(d) basis set in the Gaussian 09 software [
51]. This theoretical method was previously used for the characterization of D149 and D205 dyes (D149 belongs to the same family of indoline dyes and has a similar structure to D205) [
36,
52,
53,
54,
55,
56]. The geometry of MK2 molecules was optimized in the vacuum using four different theoretical methods: b3lyp/6–31G(d), b3lyp/6–31G(d,p), pbe1pbe/6–31G(d) and cam-b3lyp/6–31G(d).
Figure 5,
Figures S17 and S18 show the absorption spectra of D205, D205
−, MK2 and MK2
− in different organic solvents (D205, D205
− in ACN, EtOH, DCM and Tol; MK2, MK2
− in Tol). The experimental absorption spectra were compared to the theoretical results of the predicted transition energies and oscillator strengths f. The theoretical results were determined at the b3lyp/6–31G(d) level of theory.
Tables S7–S9 include the calculated [b3lyp/6–31G(d)] singlet excitation energies with the corresponding oscillator strengths,
f, starting from the optimized geometry of the ground states of D205/D205
− and MK2/MK2
−, using the CPCM solvation model with ACN, EtOH, DCM (D205) and Tol (D205 and MK2) as solvents. In general, a satisfactory correspondence between the results of experimental stationary absorption and theoretical calculations is observed for both the neutral and anionic forms of both studied dyes. A common feature of both dyes in their neutral and anionic forms is the π,π* character of S
0→S
1 transition. As follows from a comparison of the theoretically determined S
0→S
1 transition energy with the longest-wavelength absorption band of D205, in the polar solvents (ACN, EtOH), the calculated energy is red-shifted with respect to the absorption maximum in a given solvent. This trend is more pronounced for the neutral form of D205 (~800 cm
−1) compared to D205
− (~300 cm
−1). However, in the non-polar or low-polar solvents (Tol, DCM), the opposite tendency is noted, with differences in the range from ~100 cm
−1 to ~400 cm
−1. A comparison of the calculated energy of S
0→S
1 electronic transitions between D205 and D205
−, in the same solvent, indicates a clear blue shift of the maximum for the anionic form. The results are consistent with those obtained for D149 dye [
49]. Higher energy transitions for D205 and D205
−, regardless of the solvent used in the CPCM model (S
0→S
2 and subsequent), mostly show the π,π* character. Some π,π* transitions are accompanied by n,π* transitions with low oscillator strength. A comparison of the determined transition energies S
0→S
1 and S
0→S
2 indicates significant differences between D205 and D205
−, regardless of the solvent used. For D205, ΔE ≈ 5000 cm
−1, while for D205
− it is in the range from ~1000 cm
−1 to 2600 cm
−1. This is primarily due to the marked shift of the energy of the S
0→S
2 transition to lower values for the anionic form compared to that of the neutral form.
For the neutral form of MK2 in the gas phase, DFT and TD-DFT calculations were carried out using different functionals (b3lyp, pbe1pbe and cam-b3lyp) in combination with two basis sets: 6–31G(d) and 6–31G(d,p). The results are shown in
Table S7. Regardless of the calculation method used, the S
0→S
1 transition had the π,π* character. Maintaining the same basis set, 6–31G(d), but changing the functional from b3lyp to pbe1pbe results in a clear red shift (to 544.8 nm); however, the use of the cam-b3lyp functional induces a clear blue shift of the S
0→S
1 maximum (to 388.8 nm). The use of a slightly larger basis set in 6–31G(d,p) instead of 6–31G(d) has a rather slight impact on the energy levels of the HOMO and LUMO, as well as on the S
0→S
1 electronic transition for MK2 in the gas phase. A comparison of the results obtained for MK2 within the CPCM solvation model with Tol (λ = 491.1 nm) and with the experimental absorption spectrum in this solvent (λ = 495 nm) implies that the DFT and TD-DFT calculations for MK2 reflect the experimental results in the best way when using the b3lyp functional.
Figure S18 shows the shape of the frontier molecular orbitals of D205 and D205
− obtained within the CPCM solvation model with ACN as a solvent at the b3lyp/6–31G(d) level of theory. Generally, due to the deprotonation of the carboxyl group in D205, the distributions and locations of the HOMO and LUMO do not change significantly. The HOMO of D205 and D205
− is located mainly on the indoline part of the molecules. This agrees with the results obtained for D149 [
52] and D205 in the gas phase [
34]. The LUMO of D205 and D205
− is located mainly on the rhodamine part of the molecules. The S
0→S
1 electronic transition between HOMO and LUMO can be considered as an intramolecular charge-transfer transition due to a significant separation between the HOMO and the LUMO distributions.
The shapes of the HOMO and the LUMO of MK2 in its neutral (MK2) and anionic (MK2
−) forms, determined at the b3lyp/6–31G(d) theory level using the CPCM solvation model with Tol as a solvent, are shown in
Figure S19. The S
0→S
1 electronic transition for both forms of MK2 takes place to a large extent between these orbitals. In the case of MK2, the distributions and locations of the HOMO and LUMO are distinctly different. The HOMO is evenly located on the π-conjugated system (oligothiophenes), in part also on carbazole, which is an electron donor group. On the other hand, its LUMO is clearly concentrated near the cyanoacrylic acid moiety, which is the dye’s anchoring group to the nanoparticle surface. For MK2
−, the HOMO is practically not found within the carbazole part. Most of this orbital is located on the oligothiophenes and on the cyanoacrylic acid anchoring group. In addition, its LUMO is largely located on oligothiophene units and to a small extent on the anchoring part.