Bogdan Dereka*a,
Nikhil Maroli
b,
Yevgen M. Poronik
c,
Daniel T. Gryko
c and
Alexei A. Kananenka
b
aDepartment of Chemistry, University of Zurich, CH-8057 Zurich, Switzerland. E-mail: bogdan.dereka@chem.uzh.ch
bDepartment of Physics and Astronomy, University of Delaware, Newark, Delaware 19716, USA
cInstitute of Organic Chemistry, Polish Academy of Sciences, 01-224 Warsaw, Poland
First published on 28th August 2024
Microscopic electric fields are increasingly found to play a pivotal role in catalysis of enzymatic and chemical reactions. Currently, the vibrational Stark effect is the main experimental method used to measure them. Here, we demonstrate how excited-state symmetry breaking can serve as a much more sensitive tool to assess these fields. Using transient infrared spectroscopy on a quadrupolar probe equipped with nitrile groups we demonstrate both its superior sensitivity and that it does not suffer from the notorious hydrogen-bond induced upshift of the CN stretch frequency. In combination with conventional ground-state infrared absorption, excited-state symmetry breaking can be used to disentangle even weak specific hydrogen bond interactions from general field effects. We showcase this capability with the example of weak C–H hydrogen bonds in polar aprotic solvents. Additionally, we reveal for the first time symmetry breaking driven not by solvent but by the entropy of the pendant side chains of the chromophore. Our findings not only enhance our understanding of symmetry-breaking charge-transfer phenomena but pave the way toward using them in electric field sensing modality.
Elaborate theoretical models of ESSB have been developed in recent years.26 They treat the quadrupolar28–36 or octupolar37,38 molecules as a geometrically well-defined arrangement of excitonically coupled charge-transfer states in individual branches and a locally excited state of the core30,36 and take into account solute–solvent interactions,28,30,33,35 electron-vibration coupling,29 dynamic solvation31,32 and hydrogen bonding34 effects. These theoretical models successfully reproduce experimentally observed dependencies and allow for quantification of asymmetry.
Here, for the first time, we demonstrate that symmetry breaking can be driven by the pendant side chains of the quadrupolar molecule itself. We show that the outstanding sensitivity of the ESSB process to minute details in the immediate vicinity of the chromophore originates from the dependence of the asymmetry extent on the fluctuating microscopic electric fields. We further demonstrate that ESSB can serve as a useful tool to quantify these fields in various complex environments and outperforms the vibrational Stark effect (VSE) by its responsiveness. Further, our molecular probe is not susceptible to the interfering upshift of the CN stretch frequency induced by hydrogen bonding that commonly plagues the interpretation of nitrile VSE experiments. Our results showcase an important new application of symmetry-breaking charge transfer phenomena in strongly coupled excitonic systems where excitation does not fully collapse on a single branch. These systems can act in electric field sensing capacity with the extent of asymmetry serving as a primary metric which is directly connected to experimental observables. Given the increasing awareness of the catalytic importance of the microscopic electric fields at reaction sites,39 and the prospects of utilizing these fields as “smart reagents” in chemistry,40 our new tool to probe them will prove extremely useful in the future.
In the ground-state IR spectrum, there is a single CN stretch band that was ascribed to the antisymmetric vibration of the two nitriles (Fig. S1†). When symmetry is preserved in non-polar media, one excited-state absorption band (ESA1) is observed in the TRIR spectrum corresponding to the same antisymmetric stretching (Fig. 2, top). Upon ESSB, ESA1 band redshifts in frequency and the second band emerges at higher frequencies (ESA2) that corresponds to the symmetric stretching vibration of the two nitriles. The splitting between these two bands (ΔνESA) reports on the amount of asymmetry in the excited state and thus on the progress of symmetry breaking: the larger ΔνESA, the more symmetry-broken is the state (Fig. 2, bottom).
Previously, we conducted a comprehensive investigation of ESSB in 1 in 55 solvents of different nature.3 In 19 dipolar solvents that could not participate in specific interactions with 1, the amount of asymmetry was found to increase with the solvent reaction field that can be quantified with a simple Onsager dipolar polarizability function, Δf = f(ε) − f(n2), where f(x) = 2(x − 1)/(2x + 1) (Fig. 3a, red markers). Each point in Fig. 3 is a result of the TRIR experiment and corresponds to the fully relaxed S1 state in respective solvent environment. The band splitting was accurately extracted from a lineshape analysis of the solvent-equilibrated excited state (details in Section S2.2, ESI, Fig. S7†). We note that this dependence for molecule 1 is the same as reported previously,3 except that we corrected the value of the Onsager function for several low polar solvents (e.g., vinyl acetate and di-n-pentyl ether) that were erroneous due to errors in dielectric constant values reported in literature (Fig. S5†). The theoretical prediction of the Ivanov ESSB model28 is shown as a green line. As a new result, we show the dependence of band splitting in 2 obtained in 22 dipolar solvents. What stands out in Fig. 3a is that while the dependencies and the degree of symmetry breaking are identical in highly polar solvents (Δf > ∼0.4), there are differences in medium polar (0.32 <Δf < 0.4) and strong deviations in low polar (Δf < ∼0.32) environments despite 1 and 2 featuring the same chromophore. In these media, dye 2 systematically displays more asymmetric electronic distribution in the excited state than 1. Fig. 3b shows the relaxed spectra of 1 and 2 obtained in solvents of identical polarity (Δf = 0.195), where the amount of asymmetry is significantly larger in 2 (Fig. S6† showcases a more dipolar tert-butyl methyl ether with Δf = 0.332). While experimental data for both molecules deviate from the Ivanov model in low dipolar solvents, the deviation is stronger for dye 2. We attribute this behavior to the influence of the pendant glycol side chains in 2. These chains are long polar tails located on the side phenyls (Fig. 1) and they can adopt many conformations, the vast majority of which will not be symmetric. This creates an asymmetric field in the vicinity of the chromophore that influences symmetry breaking together with the solvent reaction field. Originating from hexaethylene glycol, these chains are expected to provide similar field as other glymes (glycol ethers), such as diglyme (G2) or monoglyme (dimethoxyethane) solvents. Indeed, we observe that the G2 (Fig. 3a, vertical dashed line) data point is close to the position where the symmetry-breaking behavior of 1 and 2 begins to diverge. The difference increases notably upon decrease of the polarity of the surrounding environment: in less polar solvents the influence of the polar side chains is greater. One can view the effect of these pendant side chains in low polar environment as a type of preferential polar solvation forced upon a chromophore by covalent attachment of the glyme solvent molecules. As a result, it is a conflation of the solvent and side chain fluctuating fields that drives asymmetry in this case.
It is worth noting that the details of the chain conformational distribution that vary among different low polar environments determine the exact field and, in turn, the amount of asymmetry in the relaxed S1 state. For example, fine structure in the 0.12 < Δf < 0.30 region with a pronounced peak at Δf = ∼0.22 likely originates from a shift toward higher content of gauche conformations that have a higher dipolar moment and thus confer stronger field.47,48
While the CN band splitting ΔνESA parameter is a direct and quantitative structural measure of symmetry breaking in the chromophore, it can be challenging to determine it accurately for weak ESSB cases due to the difficulty of fitting the low-intensity ESA2 band on top of the wing of the intense ESA1 band. Thus, in Fig. 3c we plot ΔνESA versus the position of the ESA1 band maximum, νESA1, and observe an almost perfectly linear relationship between these two parameters (R2 = 0.97). This is not an intuitive result because ΔνESA is determined by the behavior of both ESA1 and ESA2 bands. Nonetheless, this relationship holds across all the environments free of strong specific interactions with the chromophore: not only dipolar but also quadrupolar solvents fall on the same line (orange markers, Fig. 3c). Moreover, even the solvents that belong to neither of the two classes: CCl4 (cyan) that is highly symmetric and formally hexadecapolar or cyclohexene (green) that does not contain any polar bonds, both obey the underlying dependence. The excited-state band position in truly non-polar media, like aliphatic hydrocarbons, deserves special attention.
In the electronic ground state, both IR absorption and Raman scattering show a single peak at the same frequency5 signifying that symmetric and antisymmetric CN stretches, that follow mutually exclusive selection rules in these two spectroscopies, occur at the same frequency. This implies that the language of symmetric/antisymmetric stretch is not correct, because each C
N group can be treated as an independent local vibrator, and the nitrile vibrations do not couple to produce symmetric and antisymmetric combinations. In the electronic excited state in non-polar environment, only ESA1 is seen, and no symmetry breaking was detected.3 Based on these two observations it was concluded that in the absence of symmetry breaking in the S1 state, the two transitions occur at the same frequency too, and thus the band splitting is zero. However, given the linear trend in Fig. 3c, it is highly unlikely that non-polar solvents that exhibit no specific interactions with 1 and feature very low reaction fields will exhibit a drastically different ΔνESA vs. νESA1 dependence, producing data more than 20 cm−1 off the trendline. Instead, according to the observed linear relationship, the predicted splitting of symmetric and antisymmetric C
N stretches in the excited state would be ∼20.5 cm−1. This non-zero band splitting would be very difficult to observe experimentally, and it has a purely intramolecular origin. Upon photoexcitation, the two local C
N stretches couple and yield the antisymmetric and symmetric C
N modes whose frequencies are no longer identical. The emergent coupling in the S1 state originates from electrical anharmonicity. It is negligible in the S0 state, where each C
N stretch is rather weak, because each molecular branch has a small dipole moment (Fig. 1, μ1,2). However, the situation is quite different in the S1 state, where strong charge transfer creates a large dipole moment in each arm, even though the molecule as a whole has no dipole moment in the absence of symmetry breaking and retains its quadrupolar character. As a result, the variation of the molecular dipole moment upon (antisymmetric) vibration of two C
N's, ∂μ/∂Q, becomes substantial, which is why the antisymmetric C
N stretch intensity in the excited state is much stronger (compare intensities of the bleach vs. ESA1 in Fig. 3b). This strong modulation of the dipole moment upon vibrational displacement is more likely to be nonlinear, that is the electrical anharmonicity ∂2μ/∂Q2 is much larger, for an almost charge-separated excited state than in a weakly dipolar ground state. This is why we expect that molecules that undergo a strong increase in their dipole moment upon photoexcitation should generally exhibit this emergent coupling between the initially local vibrations. We note in passing that the original Ivanov model28 (Fig. 3a, green line) needs to be augmented to account for the inferred non-zero band splitting in non-polar environments.
In principle, one could measure an excited-state Raman CN stretch spectrum in non-polar environment and compare it to the TRIR experiment to obtain this intramolecular splitting in the absence of ESSB. In practice, however, it will be very challenging to observe due to the strong broadening of the Raman band when electronic excited-state absorption band is used for resonance enhancement as reported recently,10 and solubility limited if opting for the stimulated emission band resonance enhancement.
Fig. 3d combines the ΔνESA vs. ESA1 peak dependence for both 1 and 2. The large amount of experimental data (33 solvents for 1 and 22 solvents for 2) unambiguously show the identical behavior of both molecules following the same linear trend (R2 = 0.97 − 0.98, average slope ∼1.04), independent of the presence of the decorating side chains. The ESA1 band maximum can thus serve as a more accurate and robust metric of ESSB than ΔνESA due to the ease of its determination. Another advantage is that it allows for comparison of the data from different solvent classes using the same scale: Fig. 3d includes data for 1 in non-dipolar solvents that cannot be placed in Fig. 3a. The extreme sensitivity of this parameter is striking. For example, the difference between cyclohexane and cyclohexene is a single π-bond, but it is enough to cause a difference of 5 cm−1 for νESA1 in these two solvents. Experimental data points for 2 never reach the low symmetry breaking regime observed for 1 (minimum ΔνESA ∼32 cm−1 in 2 vs. ∼24 cm−1 in 1). It clearly indicates that deviating behavior of 2 in medium and especially low polar solvents (Fig. 3a) is not a unique characteristic of the modified chromophore but originates from the larger surrounding field due to the entropy of the pendant side chains. Thus, these observations provide first direct spectroscopic evidence of the symmetry breaking induced not by the solvent but by the intramolecular structural factor such as pendant side chains.
Our results call for revision a previous study of Kim and coworkers19 that used ultrafast broadband fluorescence upconversion spectroscopy to study the effect of the side chain length on charge transfer in a symmetry-breaking quadrupolar probe using two solvents. They compared methoxyethyl vs. n-octyl chains that have not only different lengths but also different polarity. Our results demonstrate that swapping non-polar aliphatic tails to methoxyethyl groups is equivalent to attaching a medium polar dimethoxyethane molecule to a chromophore that exposes it to a larger effective field, which has a profound effect on the extent of symmetry breaking. Based on the time-dependent fluorescence Stokes shift, Kim and coworkers19 concluded that the side chains have a greater effect on the excited-state dynamics in polar dichloromethane than in less polar toluene, and that the chains control only the extent and dynamics of torsional relaxation without affecting ESSB of the quadrupolar probe. This is the opposite of what we observe with time-resolved infrared spectroscopy, which directly monitors ESSB via structural dynamics of the quadrupolar probe by the ΔνESA metric. The occurrence and extent of symmetry breaking cannot be inferred from the fluorescence Stokes shift measurements alone, it is a too unspecific metric.2,9,23,33,49
To further substantiate our hypothesis that νESA1 acts as a proxy for the electric field, we conduct classical atomistic molecular dynamics (MD) simulations and quantify the field magnitude projected on the CN bond in the S0 state (Fig. 4a). Electric fields calculated separately on carbon and nitrogen atoms are shown in Fig. S9.† A quantitative (R2 = 0.96) linear dependence of the ESA1 band maximum on the electric field is very clear with the slope of ∼1.6 cm−1 (MV cm−1)−1.
Using νESA1 as a metric for the electric field we plot in Fig. 4b the dependence of the ground-state CN stretch band (νCN) of the core chromophore 1 on νESA1 in 29 solvents (19 of which are aprotic and 10 are protic). The data in all media separate into 2 groups. In non-hydrogen bonding ones, that engage only in non-specific interactions with the nitriles via electric field (Fig. 4b, red), νCN is linearly correlated with νESA1. Protic solvents additionally make hydrogen bonds with ADA that cause excess blueshift identified as a vertical offset from the underlying trendline (Fig. 4b, green). The magnitude of this excess shift is determined by the strength of the underlying hydrogen bond(s) as evidenced by a linear correlation with Kamlet–Taft α parameter60 that quantifies the H-bond donating strength of solvents (Fig. 4c). The piecewise linear form of this relationship stems from the distinctly different character of conventional protic vs. superprotic (polyfluorinated alcohols) solvents whose hydrogen-bond donating ability is stronger than that of water. They make strongly bound hydrogen-bonded complexes with peculiar characteristics and strongly perturb the electronic structure of ADA as described earlier.5,43 These results demonstrate that (i) ESSB monitored via the C
N stretch mode does not suffer from the non-monotonous blueshifting effect of hydrogen bonding, (ii) combination of ESSB and ground-state IR absorption spectroscopy is capable of separating hydrogen bonding from general field effects.
Fig. 4d shows the variation of νCN with the Onsager function in aprotic dipolar solvents (and protic MeOH for comparison) for 2. Evidently, the relationship is complex and non-monotonous. The frequency variation reaches ∼4 cm−1 decreasing non-linearly from 2226 cm−1 in the least dipolar solvents to 2222 cm−1 in DMSO. After DMSO, it increases again in several highly polar solvents. Given the hydrogen-bond donating ability of methanol, the high 2228 cm−1 frequency in MeOH is expected, but it is not clear why strongly polar aprotic acetonitrile (ACN), propylene carbonate (PC) and DMF follow similar behavior.
Electronic S1 ← S0 absorption transition behaves analogously (Fig. S10a†). Therefore, we use the correlation of the two methods to identify a clear linear trend (Fig. S10b†) with two notable outliers (CHCl3 and MeOH) where the vibrational transition occurs at frequency higher than predicted by the trendline. The deviation points to hydrogen-bonding interactions between these solvents and the nitrile groups of ADA. We note that chloroform cannot be identified as an outlying point in Fig. 4d alone.
Given the complex dependence of νCN on Δf this factor cannot solely explain its variation. Instead, we use a combination of Δf and f(n2) with the latter describing electronic polarizability to elucidate the observed variation in νCN. Multilinear regression on these two predictors accounts for most of the variance in the data (Fig. S10c†). Regression coefficients point to 10 times larger importance of the electronic polarizability f(n2) compared to that of the dipolar reorientation Δf. The mean absolute value of the residuals for electronic transition amounts to 48 cm−1 with most of the values being in 10–30 cm−1 range. It highlights an often forsaken property of the Onsager model that the dipolar field works well only in media with similar electronic polarizability, that is, where the refractive index does not vary appreciably. In the current case, it is the dispersion interactions not the dipole–dipole ones that play a dominant role in ground-state vibrational solvatochromism of νCN.
MeOH, ACN and PC in Fig. S10c† are still the most deviating solvents. Is it possible that aprotic solvents, such as acetonitrile and propylene carbonate, engage in hydrogen bonding interaction with ADA? To answer this question, we turn to another method that has been shown to be sensitive to electric fields and correct for possible hydrogen bonding effects – 13C NMR spectroscopy. 13C Chemical shift (δ13C) of the nitrile carbon atom shifts upfield to lower δ in more polar solvents and hydrogen bonding does not disrupt this trend,58 thus we could use the correlation between δ13C and ground-state IR band maximum to identify possible hydrogen bonding interactions (Fig. 4e, Section S2.4 in ESI†). Indeed, we observe a distinct linear correlation between these two parameters with most of the solvents falling onto the trendline and a few outliers above it, where νCN is higher as expected due to hydrogen bonding. We have already identified MeOH and CHCl3 as hydrogen-bonding partners (Fig. S10b†), and they are showing the strongest deviation here as well (4.5 and 4.0 cm−1). Additionally, we observe that PC and ACN display smaller deviations (2.4 and 1.5 cm−1 respectively) attributable to H-bond interaction. Therefore, we could identify very weak C–H hydrogen bonds between ADA and propylene carbonate, acetonitrile and chloroform that are present at room temperature in liquid solution and lead to only a minor shift of a few wavenumbers in νCN. A few recent reports have shown some evidence of C–H hydrogen bond formation in chloroform61 and propylene carbonate.62
Since the core chromophore 1 displays good sensitivity and clear separation of hydrogen-bonding vs. electric field effects, Fig. 4f plots νCN vs. νESA1 dependence for compound 2. In agreement with the results for 1, most aprotic solvents follow the identical trend as in Fig. 4b featuring the same slope (Fig. S8†). Moreover, in full agreement with the IR-NMR correlation (Fig. 4e), we can separate the hydrogen bonding effect in protic MeOH, as well as in aprotic but C–H hydrogen-bonding PC, ACN and CHCl3 in much the same way. This demonstrates that the NMR experiment can be substituted with TRIR to separate the impact of H-bonding.
Thus, we can rationalize complex behavior of νCN shown in Fig. 4d. Onsager dipolar reaction field and electronic polarizability both lead to the redshift of the peak frequency of the ground-state CN stretch transition with the latter being the dominant contributor. Even weak nonorthodox hydrogen bonding interactions lead to the noticeable blueshift of the C
N stretch whose magnitude is comparable to the nonspecific effects. Importantly, all these effects are rather weak causing shifts up to a few cm−1 at most. According to the Onsager model, a molecule like ADA with zero dipole moment does not polarize solvent around itself and thus does not generate reaction field. Individual dipolar branches or nitrile groups themselves can weakly polarize solvent locally, which, in turn, inflicts the reaction field back onto them, but no long-range effects are expected in this case. Molecules 1 and 2 do not manifest meaningful differences in their ground-state IR spectra (Fig. S3†).
To rationalize both sensitivity enhancement and the lack of notorious hydrogen bonding blueshift, we need to make a distinction between the mechanisms behind the field sensitivity of these two effects. The sensitivity of the VSE is determined by the change of the bond dipole moment upon vibrational excitation due to mechanical anharmonicity.53 This change of the dipole moment is the major factor that underlies vibrational response to the surrounding field. The electronic structure of the vibrating fragment, namely its charge distribution, remains constant. In contrast, ESSB operates in the excited electronic state and during this process the electronic structure of the excited molecule varies in response to the stronger field. The total amount of charge on both nitriles is conserved but it rebalances between the two of them during the symmetry breaking process. It is a result of this uneven distribution of the charge that ESA1 downshifts and ESA2 upshifts, because the CN frequency sensitively redshifts upon charge accumulation (reduction) and blueshifts (but much less) upon oxidation.66,67 Thus, stronger surrounding field creates more lopsided charge distribution on the nitriles. Hydrogen bonding acts similarly: its primary effect is not so much the blueshift of the C
N frequency, but enhancement of the charge redistribution as this specific interaction helps to drive it further.5 As such, the blueshifting influence of the hydrogen bonding is dwarfed by the impact it produces on the electronic structure of the probe. This is a great advantage of symmetry-breaking charge transfer phenomena in such a strongly coupled system, where interbranch electronic coupling is ∼1800 cm−1:42 it prevents a complete localization but puts the system into the regime where the electronic structure of the molecule sensitively responds to the tiniest perturbations of the field. This electronic structure variation amplifies the vibrational response. This is why we can easily differentiate a very small structural change of 2 π-electrons between cyclohexane and cyclohexene via a 5 cm−1 ESA1 shift, which is comparable to the entire solvatochromism range of dicyanobenzene VSE probes.56 It is also the reason that we are able to assess the equilibrium fields present around the non-excited chromophore: their magnitude determines the extent of ESSB that the molecule undergoes upon photoexcitation.
In contrast, the linear dipolar Stark effect alone could not explain the vibrational solvatochromism of our probes (Fig. 4d). We demonstrated an outsized role that electronic polarizability plays in determining the response of these probes due to their low Stark rate dictated by both non-dipolar nature of the entire molecule and by the small dipole moment in each arm in the ground state. Essentially, as the molecular arms are low polar in the ground state, the dipole–dipole interactions are outweighed by the dispersion interactions that are typically much smaller. These observations highlight the importance of electronic polarizability in determining the peak shift of the CN mode that is typically not accounted by the VSE models.
Additionally, even weak hydrogen-bonding interaction with solvents that are usually considered aprotic and non-hydrogen-bonding (such as acetonitrile and propylene carbonate) in combination with the other two factors contributes to the complex nonlinear and non-monotonous behavior depicted in Fig. 4d. In agreement with the previous studies, combination of IR solvatochromism with other experimental methods, such as UV-vis and especially NMR spectroscopy proves useful in disentangling hydrogen-bonding from non-specific electric field effects. We note that in NMR only the chemical shift of the nitrile carbon yields a meaningful metric. Even the ipso carbon shows a more diffuse dependence, and chemical shifts of other carbons are not correlated with the field (Fig. S15†). Therefore, it is important to correctly identify the nitrile peak in 13C NMR spectrum if this field dependence is to be followed. Combination of vibrational solvatochromism measurements with ESSB disentangles field and hydrogen bond contributions in the same way as IR-NMR combination (Fig. 4e and f).
Boxer and coworkers have recently reported that the magnitude of the CN vibrational transition dipole moment can also be used to disentangle hydrogen bonding from general field effects.68 Thus, we quantified this observable, and it is plotted vs. νESA1 – our preferred field metric – in Fig. S4a.† The correlation between these two parameters is weak (R2 = 0.51), and, surprisingly, we found no relationship between νCN and the magnitude of the transition moment (Fig. S4b†). Thus, more systematic investigations are needed to confirm the generality of such an approach.
Additionally, we convincingly demonstrate the non-innocent nature of the pendant side chains introduced to solubilize the molecule. Entropy of these side chains drives ESSB in low polar media to a significant extent. Given the range of the extra band splitting provided by the side chains between 2 and 14 cm−1 depending on the solvent (Fig. 3a), the average slope of ∼1.04 for the ΔνESA − νESA1 dependence (Fig. 3d), and ∼1.6 cm−1 (MV cm−1)−1 tuning rate of νESA1 with the field, we can estimate that they provide anywhere from 1 to 9 MV cm−1 field enhancement with the larger influence in low polar environments, where their effect can be dominant (e.g., in organic carbonates, Δf ∼ 0.2). This is the first study that identifies ESSB driven by the structural dynamics of the probe-carrying molecule itself. Our results can be compared with a recent investigation of ESSB in cyclohexane–acetone solvent mixtures, where the diffusion-limited arrival of polar component was observed to initiate ESSB process.8 That kind of diffusion-limited symmetry breaking was made possible by the lack of preferential solvation of the probe molecule by a polar solvent component. Instead, here, we have essentially tethered a polar chain to our quadrupolar probe and observed how it augments ESSB process.
Finally, we notice that using electric fields to describe symmetry breaking provides a compelling and physically meaningful language that unifies all sorts of interactions that were previously observed to affect ESSB process. We replace a chemical picture of particular chemical interactions, each of which requires its own description and metrics (e.g., quadrupole moment for quadrupolar solvents, Kamlet–Taft α for H-bonding ones etc.) by the unified electrostatic one, where all possible non-specific and specific interactions are accounted for on the same footing and contribute toward a total symmetry breaking extent.
Electric fields are being increasingly invoked to explain a variety of catalytic processes: from the catalytic proficiency of enzymatic active sites54,55 to the underlying principle behind directed evolution69 to catalysis of (non-redox) chemical reactions40 or directing known chemistry to the unknown routes.70 As a result of intense research efforts in this direction, there is an increasing need for expanding our toolbox of instruments capable of sensitively probing microscopic fields at various molecular sites. Our work paves a new way in this direction, and we hope to stir further interest from both experimentalists and theorists alike.
We identified a structurally unequivocal and quantitative metric of symmetry breaking – the peak position of the excited-state antisymmetric CN stretch band, νESA1, that is both sensitive and robust. In conjunction with atomistic MD simulations, we quantified electric fields projected onto the C
N bond vector and showed that they are linearly correlated with νESA1 metric featuring outstanding sensitivity of ∼1.6 cm−1 (MV cm−1)−1. Using a combination of IR, NMR, and electronic spectroscopies, we reveal that unlike frequency shifts in stationary IR spectra, νESA1 does not suffer from non-monotonous behavior of the C
N stretch peak upon hydrogen bonding. Instead, combining νESA1 measurements with ground-state vibrational response allows for separation and quantification of a general electric field from, even very weak, hydrogen-bonding effects. We identified such weak C–H hydrogen bonds that the probe forms with aprotic polar solvents: chloroform, acetonitrile, and propylene carbonate in liquid solution at room temperature.
Additionally, we demonstrated that ESSB can be driven by the entropy of the pendant side chains attached to the probe. This is one of the unique examples where symmetry breaking is driven by an intramolecular structural factor and not by the microscopic solvent environment. Thus, we demonstrate that remote functionalization of the chromophore with solubilizing or charged groups may affect its electronic structure and can even dominate the response in some environments. This understanding will help fine-tuning the engineered internal microscopic electric fields that are considered as means of regulating chemical and catalytic activity.63,71
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4sc04797d |
This journal is © The Royal Society of Chemistry 2024 |