Push-Pull Effect of Terpyridine Substituted by Triphenylamine Motive—Impact of Viscosity, Polarity and Protonation on Molecular Optical Properties

The introduction of an electron-donating triphenylamine motive into a 2,2′,6′,2′′-terpyridine (terpy) moiety, a cornerstone molecular unit in coordination chemistry, opens new ways for a rational design of photophysical properties of organic and inorganic compounds. A push-pull compound, 4′-(4-(di(4-tert-butylphenyl)amine)phenyl)-2,2′,6′,2′′-terpyridine (tBuTPAterpy), was thoroughly investigated with the use of steady-state and time-resolved spectroscopies and Density Functional Theory (DFT) calculations. Our results demonstrate that solvent parameters have an enormous influence on the optical properties of this molecule, acting as knobs for external control of its photophysics. The Intramolecular Charge Transfer (ICT) process introduces a remarkable solvent polarity effect on the emission spectra without affecting the lowest absorption band, as confirmed by DFT simulations, including solvation effects. The calculations ascribe the lowest absorption transitions to two singlet ICT excited states, S1 and S2, with S1 having several orders of magnitude higher oscillator strength than the “dark” S2 state. Temperature and viscosity investigations suggest the existence of two emitting excited states with different structural conformations. The phosphorescence emission band observed at 77 K is assigned to a localized 3terpy state. Finally, protonation studies show that tBuTPAterpy undergoes a reversible process, making it a promising probe of the pH level in the context of acidity determination.

In the present work, ICT processes in tBuTPAterpy are thoroughly investigated with steady-state and time-resolved optical spectroscopy and Density Functional Theory (DFT) simulations. The in-depth knowledge of the system's photophysical properties is of high significance for the rational design of new functional materials with tailored optical behavior.

Optical Spectra of tBuTPAterpy vs Constituent Building Blocks
The absorption and emission properties of the push-pull tBuTPAterpy molecule were compared to those of its constituent building blocks considering the combination of two possible pairs: (i) 2,2 :6 ,2 -terpyridine (terpy) and bis(4-tert-butylphenyl)aniline (tBuTPA), and (ii) 4 -phenyl-2,2 :6 ,2 -terpyridine (4 -Ph-terpy) and bis(4-tert-butylphenyl)amine (tBuDPA) (see Scheme S1 and Table S1). To minimize the impact of the solvent polarity, the absorption and emission spectra of tBuTPAterpy and its building blocks were measured in the apolar solvent n-hexane, and the results are shown in Figure 1. The UV-Vis spectrum of tBuTPAterpy in Figure 1a shows two well-resolved bands, with maxima at 291 nm and 364 nm, and with the high energy absorption profile fairly reproduced by the sum of the absorption bands of terpy and tBuTPA. The low-energy absorption band of tBuTPAterpy at 364 nm is absent in all model chromophores, meaning that it arises from the conjugation of the two building blocks terpy and tBuTPA into the extended molecule (tBuTPAterpy), most likely involving a charge transfer process from the electron-rich tBuTPA donor to the electron-deficient terpy acceptor. The fluorescence spectrum of tBuTPAterpy, in black in Figure 1b, cannot be ascribed to the emission of any of the building blocks. In fact, this emission band is distinctly red-shifted with respect to the others, suggesting that it originates from an electronic state that is more delocalized than the separated molecular moieties.

Solvent Polarity Effect
Steady-state electronic absorption and emission spectra of tBuTPAterpy were recorded in a wide range of solvents of different polarities in order to characterize their photophysical properties as a push-pull system. The spectra are shown in Figures 2-4 and summarized in Table 1 (see also Tables S2 and S3 and Figures   The UV-Vis spectrum of tBuTPAterpy in Figure 1a shows two well-resolved bands, with maxima at 291 nm and 364 nm, and with the high energy absorption profile fairly reproduced by the sum of the absorption bands of terpy and tBuTPA. The low-energy absorption band of tBuTPAterpy at 364 nm is absent in all model chromophores, meaning that it arises from the conjugation of the two building blocks terpy and tBuTPA into the extended molecule (tBuTPAterpy), most likely involving a charge transfer process from the electron-rich tBuTPA donor to the electron-deficient terpy acceptor. The fluorescence spectrum of tBuTPAterpy, in black in Figure 1b, cannot be ascribed to the emission of any of the building blocks. In fact, this emission band is distinctly red-shifted with respect to the others, suggesting that it originates from an electronic state that is more delocalized than the separated molecular moieties.

Solvent Polarity Effect
Steady-state electronic absorption and emission spectra of tBuTPAterpy were recorded in a wide range of solvents of different polarities in order to characterize their photophysical properties as a push-pull system. The spectra are shown in Figures 2-4 and summarized in Table 1 (see also Tables S2 and S3 and Figures  The position of the tBuTPAterpy absorption maxima in Figure 2 is marginally affected by the solvent polarity, which, however, modulates the molar absorption coefficients of the entire spectrum and the width of the low-energy band (see also Figure  S6 in the SI). Specifically, by going from non-polar to polar solvents, a decrease of the extinction coefficients and an increase of the Full Width at Half Maximum (FWHM) is   To estimate the difference between the excited and ground state dipole moments = µe − µg), the correlation between the solvent polarity and the Stokes shift was analy using Lippert equations [73]:  Compared to the room temperature measurements, the fluorescence maximum tBuTPAterpy in n-hexane and cyclohexane red-shifts upon cooling to 77 K. An opposi trend occurs for solvents with an ET(30) larger than 33 kcal·mol −1 . For these media, th low-temperature emission maximum is systematically narrowed and blue-shifte compared to the room-temperature spectra. This finding is compatible with a strong ICT character of the emitting state upon solvent polarity increase because the rigidity the 77 K glassy matrix prevents the full stabilization of the solute through solve reorganization.
It is worth highlighting that the temperature decrease leads to higher emissio  The position of the tBuTPAterpy absorption maxima in Figure 2 is marginally affected by the solvent polarity, which, however, modulates the molar absorption coefficients of the entire spectrum and the width of the low-energy band (see also Figure S6 in the ESI). Specifically, by going from non-polar to polar solvents, a decrease of the extinction coefficients and an increase of the Full Width at Half Maximum (FWHM) is observed, indicating a low polar character of the Ground State (GS) [68][69][70][71].
In contrast, the fluorescence spectra of tBuTPAterpy reported in Figure 3 show a strong dependence of the position of the emission band on the solvent polarity [72]. The emission maximum progressively shifts to longer wavelengths with increase the solvent polarity, namely from 407 nm in n-hexane to 527 nm in acetonitrile. This significant redshift is accompanied by changes in the emission profile-from narrow and vibronically structured in apolar n-hexane and cyclohexane to structureless and very broad (FWHM of 4050 cm −1 ) in the more polar aprotic acetonitrile. Such solvatochromic behavior is typical of push-pull systems undergoing a photoinduced ICT process, which leads to the formation of a highly polar emitting state that is stabilized by polar solvents with respect to the neutral GS [70,71].
To estimate the difference between the excited and ground state dipole moments (∆µ = µ e − µ g ), the correlation between the solvent polarity and the Stokes shift was analysed using Lippert equations [73]: where: hν Em represents the emission energy of the compound in a particular solvent; hν Em o corresponds to the absorption and emission energies in vacuum, µ g and µ e are the dipole moments of the molecule in its ground and excited states, a o is the Onsager cavity radius, and f is defined as: Molecules 2022, 27, 7071 6 of 20 The plot of the Stokes shift of tBuTPAterpy against the orientation polarizability ( Figure 3, top inset) shows a very good linearity, spanning from 2887 cm −1 in n-hexane to 8609 cm −1 in acetonitrile, as expected for a push-pull compound. The large Stokes shift values in polar environments can be ascribed to a remarkable change in the dipole moment of tBuTPAterpy between the GS and the emitting excited state [74], with an estimated ∆µ value of 23.49 D, assuming the Onsager cavity radius of 6.79 Å determined by quantum chemical calculations.
The solvent also affects the excited state lifetime of tBuTPAterpy at room temperature (Table 1 and Table S2 and Figure S6), which gradually increases with the solvent polarity. Concerning the quantum yield of the system, it is very high in chloroform, ethyl acetate, tetrahydrofuran, dichloromethane, dimethylformamide, dimethylsulfoxide (0.7-0.84), and it is slightly attenuated in acetonitrile (0.63) and in the apolar solvents n-hexane (0.48), cyclohexane (0.54), toluene (0.64). Instead, a dramatic reduction occurs in methanol (0.02), which is probably related to conformational changes induced by H-bonding interactions between the terpyridine nitrogen atoms and the solvent molecules.
The impact of the solvent polarity on the emission spectra of tBuTPAterpy was also investigated in rigid matrices formed at the liquid nitrogen temperature. Under these conditions, the solvent reorganisation effect is strongly reduced, limiting the spectral tuning of the ICT states [75]. As shown in Figure 4, the emission energy of tBuTPAterpy at 77 K is indeed slightly affected by the solvent polarity. In this condition, the estimated value of ∆µ is reduced to 14.02 D.
Compared to the room temperature measurements, the fluorescence maximum of tBuTPAterpy in n-hexane and cyclohexane red-shifts upon cooling to 77 K. An opposite trend occurs for solvents with an E T (30) larger than 33 kcal·mol −1 . For these media, the lowtemperature emission maximum is systematically narrowed and blue-shifted compared to the room-temperature spectra. This finding is compatible with a stronger ICT character of the emitting state upon solvent polarity increase because the rigidity of the 77 K glassy matrix prevents the full stabilization of the solute through solvent reorganization.
It is worth highlighting that the temperature decrease leads to higher emission intensities while the corresponding excited state lifetimes become shorter. For instance, the PL decay time of tBuTPAterpy in butyronitrile is twice longer at room temperature than that at low temperature ( Figure S7). This observation agrees with the hypothesis that the initially populated Franck-Condon (FC) state and the Lowest Energy Excited State (LEES) have different electronic characters.

Temperature and Solvent Viscosity Effects
To investigate the impact of the solute conformational changes on its photophysics, steady-state PL spectra were acquired in a methanol:ethanol mixture (1:4) at selected temperatures in the 80-290 K range ( Figure 5a). Additionally, the tBuTPAterpy room temperature emission was measured in several mixtures of two solvents having similar polarity, but different viscosity, glycerol (η = 954 cP; ε = 46.5)/methanol (η = 0.54 cP; ε = 32.7), and Time-Resolved Emission Spectra (TRES) were recorded in pure glycerol, the solvent with the highest viscosity, with the results respectively reported in Figure 5b,c. Figure 5a shows that the emission band of tBuTPAterpy remains centred at~430 nm and slightly changes in intensity in the temperature range from 80 to 110 K, i.e., when the intramolecular rotations of the solute are hindered and the solute-solvent interactions are weak. Upon temperature increase up to 210 K, a bathochromic shift of the band is observed along with a gradual drop of its intensity, suggesting that conformational changes in the electronically excited state of tBuTPAterpy become allowed, stabilizing the ICT emitting state. For temperatures higher than 230 K, the emission energy and its intensity remain unchanged.  Figure 5a shows that the emission band of tBuTPAterpy remains centred at ∼430 nm and slightly changes in intensity in the temperature range from 80 to 110 K, i.e., when the intramolecular rotations of the solute are hindered and the solute-solvent interactions are weak. Upon temperature increase up to 210 K, a bathochromic shift of the band is observed along with a gradual drop of its intensity, suggesting that conformational changes in the electronically excited state of tBuTPAterpy become allowed, stabilizing the ICT emitting state. For temperatures higher than 230 K, the emission energy and its intensity remain unchanged.
Stationary tBuTPAterpy PL spectra collected in methanol: glycerol mixtures of variable composition are reported in Figure 5b and show that the addition of glycerol up to 30% induces a small decrease of the emission intensity at 560 nm. Instead, in the 20:80 Stationary tBuTPAterpy PL spectra collected in methanol: glycerol mixtures of variable composition are reported in Figure 5b and show that the addition of glycerol up to 30% induces a small decrease of the emission intensity at 560 nm. Instead, in the 20:80 methanol: glycerol mixture, the emission band centered at 560 nm shifts towards higher energies, and a strong band appears at 450 nm. A further increase of the glycerol proportion up to 90% leads both emission bands to slightly red-shift and a significantly increase in intensity [72]. Finally, in pure glycerol (η = 954 cP at 25 • C), tBuTPAterpy is characterized by two emission bands centered at 458 nm and 594 nm, with intensities respectively higher and lower than in the 10:90 methanol: glycerol mixture, and shows an isoemissive point at 517 nm. The appearance of a second emission band upon solvent viscosity increase indicates the presence of two emitting states having different structural conformations.
TRES of tBuTPAterpy recorded in pure glycerol as a function of the time delay (Figure 5c) also report the presence of an isoemissive point around 20 ns, suggesting that the two emissive states are populated in a sequential way upon a conformational change. In less viscous solvents, this structural modification probably occurs on shorter time scales, and it cannot be observed due to the limited time resolution of the TRES measurements (see Figure S8), calling for dedicated ultrafast investigations in order to characterize this population transfer process.

Protonation Effect
The effect of protonation on the absorption and emission properties of tBuTPAterpy was studied in titration experiments conducted with the use of trifluoroacetic acid (TFA) in CHCl 3 using the procedure previously described in [34]. The portions of TFA were selected in order to cover a broad range of titration steps from 1:1 to 1:1000. Then, the deprotonation experiment was conducted with the use of the strong basis triethylamine (TEA), adding 1000 equivalents to the 1:1000 tBuTPAterpy:TFA sample.
Upon the addition of TFA to the chloroform solution, naked eye color changes from pale yellow to orange-red are observed. The changes in the absorption and emission profiles of tBuTPAterpy are shown in Figure 6a-c and Figures S9 and S10 of the ESI. In the absorption spectra (Figure 6a), the gradual addition of TFA (1-1000 equivalents) leads to an intensity decrease of the band at 372 nm and to the formation of two new bands with maxima at 330 nm and 493 nm. In the literature, the band at 330 nm was ascribed to 1 π → 1 π* transitions of the protonated terpy unit [76,77]. Instead, the spectral red shift in the visible region was attributed to the enhancement of the tBuTPAterpy ICT character due to the electron-withdrawing increase of the terpy acceptor upon protonation [34]. Isosbestic points at 352 nm and 398-434 nm indicate the presence of multiple protonated-neutral forms in equilibrium between each other. Finally, the reversibility of the protonation/deprotonation processes was demonstrated by recovering the original absorption and emission spectra of neutral tBuTPAterpy after the addition of 1000-equivalents of TEA to the final mixture of tBuTPAterpy and TFA (1:1000) (Figures 6a,d and S9) [78].

Phosphorescence of tBuTPAterpy
The phosphorescence of tBuTPAterpy was measured and compared to the triplet emission of the model chromophores terpy and tBuTPA, as reported in Figure 7a,b. The 77 K steady-state emission spectrum of tBuTPAterpy was recorded with the addition of a The PL spectra in tBuTPAterpy titrated with TFA acid (Figure 6b and Figure S9) show significant quenching of the fluorescence band at 476 nm upon the increase of the acid concentration. Starting from the addition of 30 equivalents of TFA (inset in Figure 6b), the decrease in the emission band at 476 nm occurs together with the appearance of a red-shifted emission band. The spectra for mixtures tBuTPAterpy:TFA with acid fractions higher than 200 equivalents exhibit only one emission maximum (at 595 nm for a 1:200 mixture and at 624 nm for a 1:1000 one). The PL lifetime for the protonated form is 5.92 ns, and it is 1.4 times higher than in the neutral compound (see Figure S10 in the ESI). The presence of an equilibrium between neutral and protonated forms of tBuTPAterpy is supported by the observation of an isoemissive point around 520-530 nm, which is visible in the normalized steady-state emission spectra of the system upon the addition of TFA (1-1000 equivalents) ( Figure S9 in the ESI), as well as in the TRES collected for the 1:50 tBuTPAterpy:TFA mixture (Figure 6c). The excitation spectrum recorded for the protonated form displays a maximum at 435 nm (see Figure S9 in the ESI), which well overlaps with the isosbestic point of the absorption spectra.
Finally, the reversibility of the protonation/deprotonation processes was demonstrated by recovering the original absorption and emission spectra of neutral tBuTPAterpy after the addition of 1000-equivalents of TEA to the final mixture of tBuTPAterpy and TFA (1:1000) (Figure 6a,d and Figure S9) [78].

Phosphorescence of tBuTPAterpy
The phosphorescence of tBuTPAterpy was measured and compared to the triplet emission of the model chromophores terpy and tBuTPA, as reported in Figure 7a,b. The 77 K steady-state emission spectrum of tBuTPAterpy was recorded with the addition of a 10% dopant of ethyl iodide-a fluorescence quencher-in order to promote phosphorescence. Having a heavy iodine atom, ethyl iodide facilitates the intersystem crossing via a stronger spin-orbit coupling [79]. The results in Figure 7a show that the low-temperature phosphorescence band of tBuTPAterpy overlaps with its room-temperature ICT emission. In Figure 7b, the phosphorescence spectra of tBuTPAterpy, terpy and tBuTPA are respectively observed in the ranges: 460-650 nm, 425-650 nm and 400-550 nm. By comparing the phosphorescence of tBuTPAterpy with its model building blocks, we conclude that the triplet state of tBuTPAterpy is predominately related to the electron-acceptor terpy fragment. Also, low-temperature TRES of tBuTPAterpy in BuCN highlights the late appearance (>27 ns) of an emission band in the same wavelength range of the isolated terpy unit, suggesting the formation of a localised terpy triplet state (Figure 7c). Figure 8 shows the simulated absorption spectra of tBuTPAterpy in the wavelength range 240-500 nm that were obtained from the calculated vertical excitation energies and the relative oscillator strengths between singlet states. The position of the simulated band maxima agrees well with the experiment (differences within 10-25 nm) and has little dependence on the solvent polarity. An even better agreement between experiment and theory is found for the position of the band at~390 nm (differences of only a few nm).

Quantum Mechanical Calculations
In all solvents, the absorption band for wavelengths longer than 350 nm involves two electronic transitions with significantly different oscillator strengths. Specifically, the excitation to the lowest singlet state, S 1, has an oscillator strength several orders of magnitude higher than the excitation to the S 2 state. As such, the absorption band is dominated by an electronic transition to the lowest singlet state, S 1 , while the state S 2 acts as a "dark" state, not being involved in the light absorption process. The second band at shorter wavelengths (about 300 nm for the simulated spectra) arises from the transition to four further singlet states, three of which are characterized by a significant oscillator strength, while the transition to the S 3 state is one order of magnitude less intense than the others (Table S4). Except for S 6 , all these states result from an electronic excitation from the Highest Occupied Molecular Orbital (HOMO), which is localized on the central phenyl ring and on the aromatic rings of the amine substituents, towards the unoccupied π* antibonding orbitals located on different parts of the molecule ( Figure S11). highlights the late appearance (>27 ns) of an emission band in the same wavelength range of the isolated terpy unit, suggesting the formation of a localised terpy triplet state ( Figure  7c).  Figure 8 shows the simulated absorption spectra of tBuTPAterpy in the wavelength range 240-500 nm that were obtained from the calculated vertical excitation energies and the relative oscillator strengths between singlet states. The position of the simulated band maxima agrees well with the experiment (differences within 10-25 nm) and has little dependence on the solvent polarity. An even better agreement between experiment and theory is found for the position of the band at ~390 nm (differences of only a few nm). In all solvents, the absorption band for wavelengths longer than 350 nm involves two electronic transitions with significantly different oscillator strengths. Specifically, the excitation to the lowest singlet state, S1, has an oscillator strength several orders of magnitude higher than the excitation to the S2 state. As such, the absorption band is dominated by an electronic transition to the lowest singlet state, S1, while the state S2 acts as a "dark" state, not being involved in the light absorption process. The second band at shorter wavelengths (about 300 nm for the simulated spectra) arises from the transition to four further singlet states, three of which are characterized by a significant oscillator strength, while the transition to the S3 state is one order of magnitude less intense than the others (Table S4). xcept for S6, all these states result from an electronic excitation from the The Lowest Unoccupied Molecular Orbital (LUMO) and LUMO+1 are populated upon transition to the S 1 and S 2 electronic states, respectively. Even though the character of both excitations is essentially the same and can be classified as CT, the LUMO orbital extends over both the central phenyl ring and the terpyridine fragment, while the LUMO+1 orbital is located solely on the terpyridine motif ( Figure 9). The higher oscillator strength of the S 0 → S 1 excitation with respect to the S 0 → S 2 one should thus be related to the larger overlap between the two conjugated π-molecular orbitals involved in the former transition. In contrast, the absence of such an overlap in the case of the LUMO+1 orbital results in an almost zero value of the transition moment, leading to an inactive S 0 → S 2 excitation. Figure S12 shows that the calculated oscillator strength of the HOMO → LUMO transition is drastically reduced upon the increase of the dihedral angle between the plane of the phenyl ring and the plane of the substituent. This effect can be rationalized in terms of a decoupling of the π-electronic system between the central phenyl ring and the substituents, which changes the form and the local symmetry of the HOMO and LUMO orbitals. Specifically, when the plane of the terpyridine rings is perpendicular to the plane of the phenyl ring, the π-orbitals of the phenyl-terpyridine units become completely decoupled from each other, and the oscillator strength drops to almost zero. Furthermore, the LUMO orbital localizes on the terpyridine motif and increases in energy, becoming the LUMO+1 orbital, as shown in Figure 10. As a consequence, the S 1 and S 2 states change in their relative energy order, and the oscillator strength of the S 0 transition towards both states becomes very small. the plane of the phenyl ring, the π-orbitals of the phenyl-terpyridine units become completely decoupled from each other, and the oscillator strength drops to almost zero Furthermore, the LUMO orbital localizes on the terpyridine motif and increases in energy becoming the LUMO+1 orbital, as shown in Figure 10. As a consequence, the S1 and S states change in their relative energy order, and the oscillator strength of the S0 transition towards both states becomes very small.  Since the S1 excitation is characterized by a significant value of the oscillator strength, the position of the first absorption band maximum is mostly determined by this electronic transition. In the calculations, the negligible solvent dependence of the S1 transition energy agrees very well with the experimental observations, suggesting that both the optimized GS and the FC region of the S1 state have comparable solvation energies [72]. This is confirmed by the Polarizable Continuum Model (PCM) energy stabilization of the GS (FC Since the S 1 excitation is characterized by a significant value of the oscillator strength, the position of the first absorption band maximum is mostly determined by this electronic transition. In the calculations, the negligible solvent dependence of the S 1 transition energy agrees very well with the experimental observations, suggesting that both the optimized GS and the FC region of the S 1 state have comparable solvation energies [72]. This is confirmed by the Polarizable Continuum Model (PCM) energy stabilization of the GS (FC S 1 ) state compared to the energy of the isolated system, corresponding to 2.6 kcal/mol (7.1 kcal/mol), 5.3 kcal/mol (7.1 kcal/mol) and 7.6 kcal/mol (13.9 kcal/mol) for n-hexane, chloroform and acetonitrile, respectively. Since for all solvents both GS and FC S 1 states are stabilized by a similar degree, the energy gap between them does not significantly depend on the solvent polarity.
In order to characterize the properties of the emissive states, the electronic and geometric configurations of the system were computed, performing structural optimizations of its relaxed lowest singlet (S 1 ) and triplet (T 1 ) states by using TD-DFT and Unrestricted Density Functional Theory (UDFT), respectively. The results for the relaxed S 1 state are shown in Figure 11, while the theoretical values of emission wavelength, oscillator strength and orbital character are reported in Table S3 for both S 1 and T 1 states. The de-excitation from the S 1 state preserves the same CT character (Figure 11b) of the vertical excitation from the relaxed S 0 structure (HOMO π R1/Ph → LUMO π* R2/Ph ), and it is independent of the solvent employed in the PCM model. The structural changes of the system as a result of the S 1 optimized geometry compared to the initial GS structure are mostly related to the spatial arrangement of the amine and terpyridine fragments with respect to the plane of the central phenyl ring (Figure 11a). In the case of the terpyridine group, the dihedral angle between the plane of the substituent and the plane of the phenyl ring becomes close to zero, i.e., both fragments of the molecule lay in a common plane. At the same time, the amino group undergoes a larger torsion with respect to the plane formed by the two other fragments. Therefore, in the relaxed S 1 geometry, the LUMO orbital is localized on the flattened structure. Instead, the HOMO orbital is almost completely localized along the vertical plane of the amine group with respect to the central phenyl ring. With respect to the ground state, the S 1 energy can be stabilized, among other processes, by lowering the energy of the LUMO orbital, i.e., by fully coupling the π-electron system between the terpyridine motif and the central phenyl ring. This is achieved when both fragments of the molecule lie in a common plane. At the same time, upon excitation, the depopulation of the HOMO facilitates the rotation of the amine group. It is worth noting that the oscillator strength for the vertical transition from the relaxed S 1 state is halved compared to the vertical excitation of the absorption process occurring from the relaxed S 0 state. According to the present analysis, in the relaxed S 1 state, the decoupling of the HOMO π-electron system that results from the rotation of the amino group (Figure 11b) is the most likely cause of the electronic dipole moment reduction for the excited-to-ground state transition.
The calculated vertical de-excitation energies S 1 → S 0 , without taking into account the state-specific equilibrium solvation of the excited state, show a negligible dependence on the solvent polarity and correlate quite well with the experimental data for low-temperature PL measurements. When including the contribution of the state-specific equilibrium solvation in the PCM model, the calculated vertical de-excitation energies S 1 → S 0 reproduce the relative bathochromic shift observed in the room temperature PL measurements (Table  S4, results indicated by the superscript a), even though the predicted emission energies are systematically underestimated. This inaccuracy can be related to the limitations of the PBE0 functional. However, this functional is expected to correctly describe the electronic structure and the energies of low-lying excited states, as in the case of similar molecular systems [80], since it captures the relative energy shift of the emission experiment for the n-hexane, chloroform, and acetonitrile series. Indeed, even though in tBuTPAterpy the differences between the calculated emission wavelengths and the experimental results range from~77 nm to~109 nm, the relative shift of the simulations correlates well with the experiment. As such, the observed bathochromic shift can be interpreted in terms of the "classic" non-equilibrium state effect of the solvent occurring in the electronic transition between the relaxed excited state S 1 and the ground state S 0 . Overall, the difference in the equilibrium and non-equilibrium solvent effects originates from the electronic structure of the S 1 state with respect to the S 0 state. Compared to the ground state, the geometry relaxation of the S 1 state leads to a discernible change in the form of the orbitals involved in electronic excitation. The HOMO → LUMO excitation, in combination with the geometry relaxation, increases the electron density polarization, for which one electron remains on the HOMO orbital localized on the amine group, and the excited electron occupies the LUMO orbital located on the remaining part of the molecule. Since the excited electron density increases the difference of the solvent equilibrium interaction around the molecule in the relaxed S 1 and GS, it also causes a stronger stabilization of the S 1 state upon solvent polarity increase. group, the dihedral angle between the plane of the substituent and the plane of the phenyl ring becomes close to zero, i.e., both fragments of the molecule lay in a common plane. At the same time, the amino group undergoes a larger torsion with respect to the plane formed by the two other fragments. Therefore, in the relaxed S1 geometry, the LUMO orbital is localized on the flattened structure. Instead, the HOMO orbital is almost completely localized along the vertical plane of the amine group with respect to the central phenyl ring. With respect to the ground state, the S1 energy can be stabilized, among other processes, by lowering the energy of the LUMO orbital, i.e., by fully coupling the πelectron system between the terpyridine motif and the central phenyl ring. This is achieved when both fragments of the molecule lie in a common plane. At the same time, upon excitation, the depopulation of the HOMO facilitates the rotation of the amine group. It is worth noting that the oscillator strength for the vertical transition from the relaxed S1 state is halved compared to the vertical excitation of the absorption process occurring from the relaxed S0 state. According to the present analysis, in the relaxed S1 state, the decoupling of the HOMO π-electron system that results from the rotation of the amino group (Figure 11b) is the most likely cause of the electronic dipole moment reduction for the excited-to-ground state transition.
(a) (b) Figure 11. (a) tBuTPAterpy geometry changes as a result of the excitation and structural relaxation of the S1 state; (b) Character of the electronic transition for the S1 excited state in its optimised geometry. H and L are used as abbreviations for HOMO and LUMO, respectively. The calculated vertical de-excitation energies S1 → S0, without taking into account the state-specific equilibrium solvation of the excited state, show a negligible dependence on the solvent polarity and correlate quite well with the experimental data for lowtemperature PL measurements. When including the contribution of the state-specific equilibrium solvation in the PCM model, the calculated vertical de-excitation energies S1 The theoretical analysis presented above allows us to qualitatively explain the experimental results showing the existence of two emitting conformers, one having higher oscillator strength and a blue-shifted emission, the other having lower transition probability and a red-shifted emission. As shown in Figure 5a,b, the tBuTPAterpy emission band decreases in intensity and red-shifts upon reduction of the conformational hindrance of the solution by temperature increase and viscosity decrease, respectively. These observations can be rationalized by accounting for the conformational-dependent energy stabilization of the S 1 state, which involves the rotation of the amine and terpyridine units with respect to the central phenyl ring. As discussed above, this structural modification red-shifts the emission wavelength and decreases the oscillator strength compared to the emission process occurring from the FC S 1 state, which takes place from a molecular structure corresponding to the S 0 energy-stabilized structure.

Materials and Methods
tBuTPAterpy was obtained according to the procedure previously described in [63]. The analytical data ( 1 H and 13 C NMR spectroscopy, FT-IR technique, HR-MS and elemental analysis) for tBuTPAterpy are reported in ESI and are in good agreement with those reported in ref. [63]. All solvents were of spectroscopic grade, commercially available and used without further purification.
Steady-state electronic absorption measurements were carried out with an Evolution 220 (ThermoScientific, Waltham, MA, USA) spectrophotometer. An FLS-980 fluorescence spectrophotometer (Edinburgh Instruments, Livingston, UK) was used to collect: steady-state emission spectra, TRES and PL lifetime measurements, the latter made using a time-correlated single photon counting (TCSPC). The experimental details of the PL measurements are briefly described in the ESI.
All calculations were obtained employing the DFT [81] and TD-DFT [82], using the hybrid PBE0 functional [83,84] and the def2-SVP basis set [85]. Solvation effects on the system's geometry and electronic structure were included by using a PCM [86]. Three solvents of increasing polarities were considered in the PCM model: n-hexane, chloroform and acetonitrile. At the DFT level, the geometry of the ground state (S 0 ) was fully optimized without any structural parameter constraints. The optimized geometry of the ground electronic state was employed in TD-DFT calculations to compute the energies of twenty singlet vertical excitations. The geometry of the lowest singlet excited state (S 1 ) was fully optimized using TD-DFT. Full geometry optimization of the lowest triplet state (T 1 ) was also performed using the UDFT formalism. The vertical de-excitation energy of the triplet state was determined as the energy difference with respect to the energy of the ground state in the optimised T 1 geometry.

Conclusions
By the conjugation of terpy and tBuTPA, new absorption and emission features appear in tBuTPAterpy. The intramolecular charge transfer character of these new bands is determined based on a solvent polarity investigation. At room temperature, the large red-shift and broadening of the emission band upon solvent polarity increase highlights the charge transfer nature of the lowest in energy excited state. At the same time, the negligible dependence of the absorption maximum suggests different characters between the initially populated Franck-Condon state and the lowest energy-emitting state. This conclusion agrees well with the emission measurements at 77 K, which show only minor changes as a function of the solvent polarity. The observation of two emission bands in the PL spectra collected for a series of mixtures of methanol: glycerol of variable composition suggests the presence of two different emitting states. This hypothesis is further supported by the detection of an isoemissive point in the TRES spectra in pure glycerol. Rigidochromic and viscosity effects indicate that the two emitting states are related to conformational changes of the solute. Finally, tBuTPAterpy shows reversible acidochromic properties in chloroform solution. DFT calculations reveal a dominant role of S 0 → S 1 vertical transition in the absorption band at the lowest energy, which has a charge transfer character. A second S 0 → S 2 transition is predicted at similar energies but is not observed in the absorption spectrum because of its much smaller oscillator strength compared to the S 0 → S 1 transition. The striking difference in their transition moment values is a consequence of the different overlap between the orbitals involved in the electronic excitation. Moreover, by calculating the potential energy curves for the three lowest singlet states as a function of θ and φ dihedral angles, an exchange of the order of the S 1 and S 2 states is observed. By comparing the results for vertical transitions starting from the energy minimum of the S 0 and S 1 states, the same HOMO π R1/Ph → LUMO π* R2/Ph character is found. Calculations including state-specific equilibrium solvation effects in the PCM provide a theoretical explanation for the substantial solvatochromic spectral shift of the emission band. The two emissive states observed in the experiments correspond to the energy-stabilized S 1 and Franck-Condon S 1 states, differing in the relative orientation of the amine-phenyl-terpyridine units. The results reported in this manuscript indicate that tBuTPAterpy possesses an intriguing and complex photophysics that is highly sensitive to external conditions. The strong solvatochromism of the presented molecule could be used in applications for bioimaging and biosensing, such as microenviromental polarity and viscosity sensors. Furthermore, the reversible acidochromism of this molecule makes it a potential candidate for volatile acids sensors. Further studies should be devoted to the characterization of the ultrafast dynamics of this molecule upon intramolecular charge transfer excitation.