Ground-State Tautomerism and Excited-State Proton Transfer in 7-Hydroxy-4-methyl-8-((phenylimino)methyl)-2H-chromen-2-one as a Potential Proton Crane

Abstract

The tautomerism in the title compound as a potential long-range proton transfer (PT) switch has been studied by using the DFT and TD-DFT approaches. The data show that in aprotic solvents, the enol tautomer dominates, while the increase in the content of the keto tautomer (short-range PT) rises as a function of polarity of the solvent. In ethanol, due to specific solute–solvent stabilization through intermolecular hydrogen bonding, a substantial amount of the keto forms exists in solution. The irradiation leads to two competitive processes in the excited state, namely ESIPT and trans/cis isomerization around the azomethine bond as in other structurally similar Schiff bases. The studied compound is not suitable for bistable tautomeric switching, where long-range PT occurs, due to the difficult enolization of the coumarin carbonyl group.

1. Introduction

Excited-state intramolecular proton transfer (ESIPT) is a feature of organic molecules, containing strong proton-donating/accepting groups, which, due to the increased acidity/basicity in the excited state, exchange a proton over the short distance of pre-existing hydrogen bonding [1,2,3,4,5,6,7,8,9,10]. ESIPT-exhibiting molecules have become a field of active research in recent decades, due to their applications as light-emitting materials and laser dyes [11,12,13,14,15], optical sensors [16,17,18,19,20,21,22,23,24,25,26,27], organic light-emitting diodes [20,21,28,29,30,31,32,33,34,35,36], optical storage devices [37] and photo stabilizers [38]. The same process underlies a special class of bistable photo switches [39], called proton cranes, where the ESIPT is accompanied by intramolecular rotation [40], leading to the transport of a movable proton from one side of the molecule to another.

The overall process of switching in proton cranes is outlined in Scheme 1. The proton, which is located at the proton donor site of a PT frame (PTF), undergoes ESIPT to the proton crane unit (PCU) upon irradiation. In this way, the proton is separated from the PTF, in which two free proton acceptor sites remain. The competition between them is the driving force of the rotation of the PCU, which can lead to the delivery and release of the proton to the other side of the PTF. The overall process can be classified as long-range intramolecular proton transfer.

Depending on the extent of conjugation between PTF and PCU, and more importantly, on the possibility for tautomeric proton exchange between PTF and PCU, the switching mechanism can be very different [41,42]. In 1 [43,44], the ESIPT (i.e., the transition from I* to II*) leads to tautomeric structural re-arrangement, which changes the nature of the axle from a single to a double bond. This causes twisting around the axle to a structure where the PFT and PCU are perpendicular each other, and relaxation to the ground state occurs through conical intersection. As a result, II and III are populated simultaneously, leading (at suitable relative energies) to a population of I and IV, respectively. In 2 [42,45,46,47], the initial ESIPT leads to II*, which has a zwitterionic structure (i.e., intramolecular deprotonation occurs). The nature of the axle remains unchanged and the long-range PT occurs entirely in the excited state, which leads to IV*, where the PTF is protonated again. Then, the relaxation to the ground state populates IV. In 3 [48,49,50], there is also possibility for tautomeric re-arrangement in the first stage of the process, but due to the highly unstable zwitterionic nature of IV, the end state has never been reached [49,50].

7-hydroxy-quinoline is one of the most frequently studied examples for long-range intramolecular proton transfer [51,52,53,54,55,56,57,58,59,60,61] and is a very suitable structure for PTF, where the proton donor OH group and the proton acceptor N atom are far apart to provide conditions for truly intramolecular proton exchange. The enol tautomer is substantially more stable, while the NH tautomer has been experimentally observed only in protic organic solvents or in the presence of water and results from the intermolecular, solvent-assisted PT mechanism [43,61]. Previous computational studies suggested that the reversible long-range PT process could be observed for a series of 7-hydroxy-quinoline-based switches and oxazine [41,62] and other six-membered ring heterocycles [62], pyridines [62,63,64], the carbaldehyde group [43], and CO-BF₂ [65,66] as proton cranes units. Recently, we have reported several classes of proton cranes, where the azomethine group plays the role of a PCU, providing intramolecular PT to III, and in limited cases to IV, upon irradiation [67,68].

7-hydroxy-coumarin and its derivatives (including 4, Scheme 2), also possess a proton donor group (O-H) at a large distance from the proton acceptor group (C=O), making direct intramolecular PT impossible. The substantially more stable enol form is observed in organic solvents, while the keto form appears in the presence of alcohols, water and organic bases as for 7-hydroxy-coumarin [69,70,71,72,73,74] as well as for 4 [74,75,76,77,78,79]. As reported for 4, concentration-assisted PT takes place in the gas phase [80] as well. Quite interestingly, in 7-hydroxy-4-(trifluoromethyl)coumarin in toluene, the addition of 1-methylimidazole leads to excited-state intermolecular long-range PT in the picosecond time scale, where the base acts as a proton-transfer shuttle from the hydroxyl group to the carbonyl [81].

In spite of its structural suitability as a PTF, 7-hydroxycoumarine and its derivative 4 have never been considered in the design of proton cranes. Very recently, the synthesis and spectral properties of 5 have been reported, along with some photochemical information [82,83]. This creates a solid background for detailed theoretical studies that can reveal the ground-state tautomeric properties and excited-state PT in 6 as simplified models of 5. Bearing in mind that the Ph rings in the Schiff bases are partially out of planarity, the effect of weak substituents like OMe on the tautomeric state is negligible [84]. In the current investigation, the tautomerism of 6 in the ground and excited state is studied in a variety of solvents, and the applicability of 7-hydroxy-coumarine Schiff bases as proton cranes is considered.

2. Computational Details

Quantum-chemical calculations in the ground state were performed using the Gaussian 16 C.01 program suite [85]. All structures (in both the ground and excited state) were optimized without restrictions, using tight optimization criteria and an ultrafine grid in the computation of two-electron integrals and their derivatives. The true minima were verified by performing frequency calculations in the corresponding environment. The implicit solvation was described using the Polarizable Continuum Model [86] (the integral equation formalism variant, IEFPCM, as implemented in Gaussian 16). The transition states were estimated using the Berny algorithm [87] and again verified by performing frequency calculations in the corresponding environment.

The M06-2X [88,89] density functional with the TZVP [90] basis set was used for the structure optimizations in the ground state. This fitted hybrid meta-GGA functional with 54% HF exchange was especially developed to describe main-group thermochemistry and non-covalent interactions. The use of M06-2X provides very good predictability of the ground-state tautomerism [68,91,92] in tautomeric compounds and proton cranes in solution, as well as the E/Z isomerization ratio in rotary switches.

The TD-DFT method [93,94] was used for singlet excited-state optimizations. CAM-B3LYP [95] with the TZVP basis set was used for the optimizations (NStates = 6). The selection of CAM-B3LYP was based on its better performance (in comparison with a variety of density functionals, including M06-2X) in describing electronic excitation energies, excited-state geometries, dipole moments and oscillator strengths in a variety of systems [96,97,98,99], including ESIPT ones [100], as well as from our own previous experience [68,101].

Bearing in mind that M06-2X systematically underestimates vertical transition energies [102], the UV-Vis spectral data were predicted by the B3LYP [103] functional (6-311 + G(2d,p) basis set) using the M06-2X optimized ground-state geometries. The corresponding spectral curves (presented as molar absorptivities) were constructed from the theoretically obtained band positions and oscillator strengths by using Gauss band shape [104] according to Equation (1) as follows:

$$\epsilon \left(\lambda \right)={\sum }_{i=1}^n{\epsilon }_{max,i}·e^{-\ln \left(2\right)·{{10}^{14}·\left(\frac{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{${\lambda }_{max,i}$}\right.-\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$\lambda $}\right.}{\frac{{\Delta \nu }_{1/2}}{2}}\right)}^2}$$

(1)

where n is the number of the transitions taken into account, ${\lambda }_{max,i}$ is the predicted position of the band of the ith transition in nm and ${\epsilon }_{max,i}$ is the calculated maximal intensity of the ith band according to Equation (2) from the oscillator strength $f_i$ [105].

$${\epsilon }_{max,i}=\frac{f_i}{7.04\times {10}^{-9}·{\Delta \nu }_{1/2}}$$

(2)

A half-band width (${\Delta \nu }_{1/2}$) value of 3500 cm⁻¹ was used for all bands.

3. Results and Discussion

The possible tautomeric forms of 5 and 6 are given in Scheme 2. According to the data, obtained by Liubimov and co-authors [82], the E tautomer of 5, absorbing in the near UV range (340 nm), is substantially more stable in aprotic solvents like toluene. The KE form has an absorption maximum around 450 nm, the intensity of which arises in protic solvents. The same behavior is typical for the tautomeric Schiff bases derived from 2-naphthol (7, Scheme 3) [84] and 7-hydroxy-quinoline (8) [68], which feature intramolecular hydrogen bonding and increased aromaticity in the OH-containing part.

The absorption spectra of the possible tautomers of 6, as a model of 5, are shown in Figure 1 and Figure S1 in toluene and acetonitrile, respectively. In line with the experiment, the E tautomer absorbs around 330–340 nm, which corresponds very well to the experimental data in these two solvents [82,83]. The existence of the keto tautomers (KE, KK and K) should be manifested by the appearance of red-shifted absorbance, each of them with their own signature. According to the relative stabilities of the tautomers, shown in Figure 2 and, in detail, in Table S1, the energy difference between E and KE suggests a very small amount of the latter is presented in toluene and acetonitrile, with more in the latter. The energies of KK, and especially of K, are too high to be observed in solution. According to the spectrum of 5 in toluene [82], a very low intensive tail is observed in the range 400–500 nm, while in acetonitrile [83], the red-shifted band is well defined at 450 nm, which corresponds to the suggested spectra of KE and KK. As discussed above, due to the high relative energy (>6 kcal/mol in respect of E, Table S1), the KK tautomer should be excluded. In the case of 6, the appearance of KE has an individual feature, which differs from the E to KE transitions in 7 and 8. While in 7 and 8, the increase in the content of the KE form leads to a decrease in the intensity of the band assigned to the enol [68,84], in the case of 6, as very clearly seen from Figure 1 and Figure S1, the increase in the keto form should also lead to a rise in the absorption at 340 nm, due to the overlapping between the S₀-S₁ transition of E and the S₀–S_n transitions in KE, with the latter having larger oscillator strength. Actually, this is observed for 5, comparing the molar absorptivities at 340 nm in toluene and ethanol [82]. The intensity of the band in toluene is ~23 000 M⁻¹·cm⁻¹, which nicely corresponds to that in Figure 1. The same value in ethanol, where the content of KE is substantially higher, is much larger. Based on the predicted theoretical data for the KE maximum in ethanol with molar absorptivity around 9 000 M⁻¹·cm⁻¹ (Figure S2) and the measured value in ethanol (4 500 M⁻¹·cm⁻¹), it can be estimated roughly that the amount of 5KE is between 40 and 50%. This is surprisingly high compared to the predicted relative energies in ethanol as a medium (

Table 1. Relative energies of the 1:1 and 1:2 complexes of 6 with ethanol molecules in the ethanol environment. The corresponding values for the single tautomers in the same environment are given in brackets.

Structure	Relative Energetics [kcal/mol] (M06-2X/TZVP)		Structure	Relative Energetics [kcal/mol] (M06-2X/TZVP)
	∆E	∆∆G		∆E	∆∆G
E + EtOH	0.0 (0.0)	0.0 (0.0)	E + 2EtOH	1.34	0.24
KE + EtOH	0.99 (1.80)	1.36 (2.50)	KE + 2EtOH	0.0	0.0
KK + EtOH	3.4 (6.0)	3.9 (6.8)	KK + 2EtOH	0.93	2.63
K + EtOH	17 (30)	15 (29)	K + 2EtOH	spontaneously goes to KK

), suggesting that this tautomeric state can be only attributed to a specific solute–solvent stabilization.

As seen from Table 1, the relative stability of the tautomers in ethanol (as implicit solvation) suggests around 5% of KE, which is much lower than the observed value. Even a very simple explicit solvation model with one molecule solvent attached changes the situation substantially. The relative stability of KE increases to almost 15%, which shows the reasons for substantial stabilization in alcohols. The increase in the number of attached solvent molecules would probably shift the equilibrium further to the keto tautomer, as shown in the case of two ethanol molecules being attached, where the solvent stabilization of KE is substantial. Although the structure of the KK + EtOH and K + EtOH complexes, shown in Table 1, suggest that a channel for the exchange of the proton between KK and K potentially exists, the energy of the latter is too high to be really populated in solution. It is especially interesting to mention that this solvent channel actually works in the direction of the more stable KK, as seen from the 1:2 complex of K. Of course, the data, collected in Table 1, represent a very simplified model of the reality in solution, but they show the trends in the change in the tautomeric state as a result of the specific solvation.

The discussion up to now has been based on the theoretical simulations, which logically raises the concern of how reliable these conclusions are, especially because there are no quantitative data about the position of the tautomeric E–KE equilibrium in 5 and 6 to confirm the reliability of using the M06-2X/TZVP approach for the relative stability prediction. On one hand, it was previously proved that the M06-2X/TZVP level of theory predicts, in a reliable way, the tautomeric state and important structural characteristics of azodyes and Schiff bases in aprotic solvents [91]. On the other hand, experimental data are available for 7 and 8, where the same theoretical approach was used [68]. The corresponding information to compare 6–8 is collected in

Table 2. Theoretically predicted ∆E and ∆G values compared with the experimentally determined ∆G values at room temperature in toluene and acetonitrile (in brackets).

Structure	Relative Energetics [kcal/mol]
	6		7			8
	M06-2X/TZVP		M06-2X/TZVP		[84]	M06-2X/TZVP		[68]
	∆E	∆G	∆E	∆G	∆Gexp	∆E	∆G	∆Gexp
E	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)	0.0 (0.37)	0.0 (0.10)	(0.52)
KE	2.8 (1.76)	3.5 (2.4)	1.70 (0.55)	1.80 (0.60)	1.42 * (0.35)	0.52 (0.0)	0.62 (0.0)	(0.0)
KK	7.3 (6.0)	8.0 (6.7)	9.0 (7.1)	9.8 (8.3)	-	1.9 (0.70)	2.05 (0.63)	- (1.0)
K	32 (30)	32 (29)	-	-	-	7.6 (4.8)	7.5 (4.7)	-

* in cyclohexane.

.

As seen from the data in Table 2, there is reasonable agreement between the theoretically obtained ∆E values and the experimental Gibbs free energies for 7 and 8. Of course, as it was underlined above, this agreement is limited to aprotic solvents, since the implicit solvation model does not account the specific solute–solvent interactions with the protic solvents. This means, for instance, that the theoretical results for alcohols, obtained by using the PCM solvation only, have to be interpreted with care. The information given in Table 1 is a very convincing example.

The data collected in Table 2 allow to make conclusions about the effect of the PTF in compounds 6-8. In all three compounds in toluene, the enol is the most stable form, but the relative stability of the KE form increases from 6 to 8, from an almost negligible amount in 6 to almost 1/3 of the total population in 8. The KK form is strongly destabilized in 7, due to steric hindrance between the tautomeric NH and the C₈H proton, and in 6, due to the lack of stabilizing hydrogen bonding as in 8. The K tautomer does not exist in 7 and is very strongly destabilized in 6. The latter is a result of the difficult enolization of the carbonyl group in the coumarin backbone and/or by the strong stabilization of the enol form within the 7-hydroxy-coumarin core. In structurally similar compounds, where 7-hydroxy-coumarin plays a role of potential PTF, and where benzothiazole [106,107] and other conjugated nitrogen-containing heterocycles [22,108] are attached as PCU, there is no evidence of the existence of KK and K switched forms.

According to the data reported by Liubimov and co-authors [82], in toluene 5 exhibits a single emission band around 545 nm. The isolated PTU, 4, also show a single band, but it is substantially blue shifted (at around 380 nm), which is attributed to the local emission of the only existing in this solvent enol tautomer. Therefore, the substantial Stokes shift (11 000 cm⁻¹) in 5 suggests an ESIPT process from E* to KK* as the authors stated. Additional irradiation does not lead to spectral changes in 5 in in this solvent.

In Figure 3, the ground- and first excited-state potential energy surfaces are sketched. According to these, and also to the general understanding of long-range PT in similar compounds [41], the excitation of the enol form through an energetically high Franck–Condon state could lead either spontaneously to conical intersection (CI) relaxation in the C=N bond isomerization region, yielding a mixture of ground-state E and Ecis, or ballistically, through ESIPT, to KE*. Then, the excited KE* can emit or, by twisting around the C-C bond (the axle in Scheme 2), can reach the twisting conical intersection region, relaxing as a mixture of KE and KK. As shown above, KE almost immediately restores to the predominant E, while KK can be accumulated due to the large relaxation barrier back (~25 kcal/mol in respect of KK). Only in such a consequence of events could an emission from KK be observed. Bearing in mind that no spectral changes are observed upon irradiation, the accumulation of KK is less probable to be assumed. Therefore, the observed red-shifted emission could be only attributed to KE*. The large Stokes shift of this emission probably comes from the substantial stabilization of KE* in respect of the enolic Franck–Condon state, as seen from the figure. Schiff base 8 also exhibits an ESIPT emission in toluene around 490 nm with a Stokes shift of 7400 cm⁻¹ [68], which comes from it being lower in energy, compared to here, in the Franck–Condon state. An emission in cyclohexane around 480 nm, attributed to the ESIPT process [109], is measured for 7, but there are no theoretical data to make a more detailed interpretation.

The situation dramatically changes in both 4 and 5 going to an ethanol solution [82]. As seen from Figure 4, the emission spectrum of 6 depends on the excitation wavelength; a single emission band with a maximum at 530 nm is observed when excited at 461 nm (spectrum 3), whereas two emission bands with maxima at 442 nm and 530 nm are observed when excited at 375 nm (spectra 2 and 5). According to the discussion above, in the ground state, a substantial amount of KE exists in ethanol, expressed by the red-shifted band at around 450 nm. Therefore, the emission observed upon excitation at 461 nm can be attributed to the KE* as it is observed in toluene.

Since, in the region around 380 nm, both E and KE are absorbing, the excitation in this region leads to an emission at 530 nm, attributed to the existing KE, and to an emission at 442 nm (missing in toluene), which probably comes from the enol form, where the tautomeric OH group is engaged in intermolecular hydrogen bonding with the solvent. From the existing data, it is impossible to make a conclusion as to whether the intramolecular O-H…N bonding is completely or partially broken by the solvent. As seen from Figure 4, upon irradiation with an external light source, substantial spectral changes are observed, associated with a rise in the emission at 440 nm. Basically, the irradiation can lead to two processes, as shown in Figure 3—trans/cis isomerization to Ecis or switching—leading to the accumulation of KK. The latter process should be more probable in toluene, where switching relaxation back from KK to KE through TS(KE-KK) is expected to be slower. The ground-state twisting transition state is very polar, and its energy decreases from 32 kcal/mol in toluene (Table S1) to 26 kcal/mol in ethanol, considered as an implicit solvent. The real stabilization is probably higher because of the increased possibilities for specific solute–solvent interactions in the transition state, in which the NH proton does not participate in intramolecular hydrogen bonding. It is much more probable to assume trans to cis isomerization with the accumulation of some amounts of Ecis. This form (as seen from Figure S2) is blue shifted compared to E, and the intramolecular hydrogen bonding is broken, which allows additional stabilization by the solvent. The relaxation back to E proceeds through the inversion transition state (as predicted by the theoretical calculations, Figure 3), in which energy does not depend on the solvent polarity. The process of cis/trans isomerization has been already described in structural analogues of 6 by Traven and co-authors [110,111]. A very recent study by Tang and co-workers describes the breakage of the CO-O bond in alcohols under strong external irradiation [112] in 4. The same process is theoretically predicted by Krauter at al [71], which creates an additional possibility in the interpretation of the data from Figure 4.

4. Conclusions

The tautomerism of 7-hydroxy-4-methyl-8-((phenylimino)methyl)-2H-chromen-2-one has been studied by using the density functional theory in ground and excited states, and the theoretical data were considered in comparison with the existing experimental results. Although the studied compound can potentially exist as a mixture of four tautomers, the enol form dominates in the ground state in aprotic solvents. Small amounts of the short-range proton-transferred keto tautomer are observed as a function of the polarity of the solvent. In ethanol, the content of the keto tautomer is substantial as a result of its specific stabilization by the solvent. According to the theoretical data, the long-range PT could not be observed due to the difficulty of the enolization of the carbonyl group in the coumarin frame. Upon excitation, two competitive processes could be expected, namely ESIPT and trans to cis isomerization, around the azomethine bond. However, the existing experimental data and their theoretical interpretation suggest that the latter is much more probable.