Specific deuterium isotope effects on the intramolecular charge-transfer dynamics of 4-(dimethylamino)benzonitrile

The photoinduced charge-transfer (CT) dynamics of 4-(dimethylamino)benzonitrile (DMABN) are studied for methyl-deuterated isotopomers DMABN-d6 (with two CD3 groups) and DMABN-h3d3 (with one CH3 and one CD3 group) by steady-state and time-resolved fluorescence spectroscopy.

The asymmetrically deuterated species DMABN-h3d3 is found to exhibit less CT fluorescence in sufficiently polar solvents than the symmetric isotopomers of DMABN-h6 and DMABN-d6.

The results of fluorescence decay measurements are also reported, which suggest that the rate of backward electron transfer is increased upon asymmetric substitution of the dimethylamino group.


The photoinduced charge-transfer (CT) reaction of DMABN in polar solvents has been the subject of extensive experimental and theoretical investigations.1,2

This process is usually considered as a twisting motion of the dimethylamino (DMA) group with respect to the benzonitrile group to produce an emissive transient species called twisted intramolecular charge-transfer (TICT) state.3

After finding similar behavior in many other molecules containing electron donor and acceptor groups, the rotamerism accompanying charge separation is most widely accepted.

A wide variety of model compounds have been studied to explain the motion leading to the CT state.

The so-called pretwisted compounds, such as sterically hindered ortho-methyl-derivatives of DMABN, are known to emit preferentially or only CT fluorescence in a medium polarity solvent.3,4

The CT fluorescence is also observed in the static vapor5 and under supersonic jet conditions.6

Since the DMA group in these compounds is already twisted in the ground-state structure, direct excitation of the CT state becomes possible.

Other derivatives of DMABN, in which the amino group is rigid with respect the twist around the Namino–Cring bond, reveal only locally excited (LE) fluorescence.3,4

These results are considered as evidence of the TICT scenario.

On the other hand, neither the meta nor the ortho isomer of DMABN shows dual fluorescence even in strongly polar solvents.7

The specific ring-substitution behavior has been ascribed to the large energy gap between the S1 and S2 states as compared to that of para-DMABN, leading to suggest a planar CT state.7

Nonetheless, the nature of the structural change is not well characterized yet and under controversial discussion.

Recently, we have measured S1–S0 spectra of methyl-deuterated isotopomers of DMABN (DMABN-d6 and DMABN-h3d3) in supersonic jets using resonant two-photon ionization (R2PI) spectroscopy.8

The spectra of the deuterated isotopomers reveal low-frequency progression due to the torsional motion of the DMA group in the S1 state, essentially analogous to the protonated species DMABN-h6.

Analysis shows that in all isotopomers the DMA group is twisted by ≈26° with respect to the ring plane under the isolated molecular condition.

However, the DMABN-h3d3 spectrum exhibits additional bands with a less regular pattern, suggesting that the asymmetric deuteration of the DMA group induces extensive mixing of its torsional motion and other low-frequency motions in S1.

Possible candidates for these motions are predicted to be torsion of the methyl groups and pyramidarization at the amino nitrogen.

It is therefore interesting to compare the CT dynamics of the symmetric and asymmetric rotors in polar solvents.

Several examples of deuterium isotope effects on electron-transfer reactions have been reported, especially for those in the Marcus inverted region.9–11

It has been shown that in intermolecular complexes between methyl-substituted benzene donors and cyanoanthracene acceptors, the backward electron transfer rate decreases substantially when the methyl hydrogens of the donor are deuterated.9

However, no or insignificant isotope effect is observed upon ring deuteration.

An asymmetric substitution effect has been suggested for intramolecular electron transfer reactions in bianthryls.12,13

In this case, the so-called symmetry breaking is believed to be an important step for CT.

Likewise, such symmetry reduction can be achieved upon complexation of bianthryl with other solvent molecules.

It has been shown that the 1:1 complex with acetone formed in a supersonic jet reveals a red-shifted emission assignable to the CT fluorescence.14

The complex is likely to have the solvent molecule sticking to one of the anthracene moieties, undergoing symmetry breaking and charge separation.

On the other hand, no CT formation occurs for the 1:2 complex, which is explained by invoking a symmetric structure.

In this paper we report the results of the systematic study of the effect of deuterium substitution on the dual fluorescence behavior of DMABN in solvents of different polarity.

The asymmetric isotopomer DMABN-h3d3 is found to reveal less CT fluorescence in moderately polar solvents than the symmetric ones (DMABN-h6 and DMABN-d6).

In less polar solvents (e.g., 1,4-dioxane), on the other hand, the fully protonated compound gives rise to red-shifted fluorescence with higher efficiency than the deuterated isotopomers.

We have also employed a time-resolved spectroscopic technique to investigate the origin of the specific deuteration effects.

The results suggest that the backward electron transfer is facilitated upon asymmetric deuteration of the DMA group.

These observations are interpreted within the framework of the TICT model.


Commercial DMABN (DMABN-h6) was obtained from Aldrich and purified by vacuum sublimation before use.

The methods of preparing the deuterium-substituted DMABN samples, (CD3)2N-C6H4-CN (DMABN-d6), and (CH3)(CD3)N-C6H4-CN (DMABN-h3d3) have been described elsewhere.15

The isotope content (>99%) was confirmed by NMR and mass spectroscopy analysis.

The DMABN-h3d3 sample showed traces (≈0.17%) of monomethylated impurities [e.g., (CD3)HN-C6H4-CN].

A series of solvents of different polarities [cyclohexane, tetrahydrofuran (THF), 1,4-dioxane, n-butyronitrile, acetonitrile, 1-butanol, 1-propanol, 2-propanol, ethanol, methanol, and water] were purchased from Kanto Chemical Co. These solvents were of spectroscopic grade and used as received.

The sample concentration was 1 × 10−4 M.

A Hitachi F-2000 spectrofluorometer was used for the steady-state fluorescence measurements.

The time-resolved fluorescence measurements were carried out by exciting the samples with the frequency-tripled output of a femtosecond laser (Spectra Physics Tsunami/Spitfire) at 266 nm and by recording the fluorescence with a streak camera (Hamamatsu Photonics).16

The system response function was found to be ≈30 ps.

Since the samples exhibit dual fluorescence, each fluorescence emission was collected separately through a monochromator.

The fluorescence decays were fitted with exponential functions by taking account of the response function of the streak camera.


The steady-state fluorescence spectra of three isotopomers of DMABN in 1,4-dioxane, ethanol, and acetonitrile at 25 °C are shown in Fig. 1.

In each solvent, the absorption and fluorescence excitation spectra are essentially identical for the three isotopomers.

The fluorescence spectra of the fully protonated species (Fig. 1) are in good agreement with previous measurements.4,17

In the case of weakly polar dioxane (dielectric constant ε = 2.2 at 25 °C), the relative LE fluorescence intensity is strongest for DMABN-d6.

With increasing the solvent polarity (ethanol and acetonitrile), the ratio for the asymmetric isotopomer DMABN-h3d3 becomes larger than those for -h6 and -d6.

The symmetric species give nearly the same intensity ratios.

This behavior cannot be ascribed to the monomethylated impurity described in the Experimental section, which is known to exhibit only LE fluorescence.17

The fluorescence spectra of the three isotopomers in cyclohexane (not shown) appear to be identical.

In all polar solvents studied, the relative intensity of the LE and CT fluorescence is found to be different for the three isotopomers.

Ratios of the CT and LE fluorescence yields (ΦCT/ΦLE) can be obtained by separating the respective emission bands.

The ΦCT/ΦLE ratios plotted against the dielectric constant of the solvent are shown in Fig. 2.

It can be seen that in 1-butanol (ε = 17.1) or more polar solvents the ratio for the -h3d3 isotopomer is considerably lower than for the symmetric isotopomers.

Furthermore, the ratios in protic solvents (propanols, 1-butanol, methanol, and ethanol) appear to increase monotonically with the polarity, but differ noticeably from those obtained with aprotic solvents of similar polarity [e.g., 1-propanol (ε = 20.1) and n-butyronitrile (ε = 20.4)].

Fluorescence yield ratios have also been obtained for the three isotopomers with an excitation wavelength of 306 nm (not shown).

The symmetry dependent behavior is essentially the same for the two different excitation wavelengths.

Time profiles of the LE and CT fluorescence emissions have been measured separately for three DMABN isotopomers in each solvent.

The selected spectral range chosen to follow the LE fluorescence kinetics is centered at 350 nm in dioxane and 355 nm in more polar solvents, whereas the CT fluorescence was measured with the observation window centered at 450 nm for dioxane and 490 nm for the other solvents.

Each time-resolved fluorescence intensity appears to consist of two components, with a picosecond decay followed by a nanosecond decay for the LE emission and a picosecond rise followed by a nanosecond decay for the CT emission.

Thus, the time-resolve fluorescence intensity can be described by the following equationswhere I(t)LE is the LE fluorescence intensity with the fast and slow decay lifetimes τfast and τslow, and the amplitudes Afast and Aslow.

I(t)CT is the CT fluorescence intensity with the risetime τrise and decay lifetime τdecay, and the amplitudes Arise and Adecay.

The decay profiles of the LE fluorescence obtained for dioxane, ethanol and acetonitrile are compared in Fig. 3.

In each solvent, the profiles of the CT fluorescence are nearly identical for the three isotopomers.

The fluorescence decay lifetimes and preexponential factors calculated by fitting the data to eqns. (1) and (2) are summarized in Table 1.

As shown in Fig. 3, a small isotope effect is observed for the decay curves in dioxane.

The fast decay lifetime of the LE fluorescence τfast increases slightly upon deuterium substitution.

The value of 25 ps for the -h6 species is in good agreement with that reported by Schuddeboom et al.17

Pronounced isotope effects occur in the case of ethanol and acetonitrile.

The asymmetric deuteration leads to an increase in the magnitude of the slow decay component of the LE fluorescence.

The ratio of the amplitudes of the fast and slow decay components (Afast/Aslow) for -h3d3 is calculated to be 6.3 for ethanol and 22.2 for acetonitrile.

The corresponding values for the symmetric isotopomers are much larger (28.6 and 34.5 for -h6 and 76.9 and 83.3 for -d6).

As listed in Table 1, the lifetimes of the fast and slow decay components of the LE fluorescence (τfast and τslow) are not significantly influenced by the isotope substitution.

The result obtained for -h6 is in qualitative agreement with that reported by Changenet et al.18

In all cases, the amplitude ratio of the rise and decay components of the CT fluorescence Arise/Adecay is nearly equal to –1, which implies that the CT state is formed through the LE state decay.

Isotope dependent behavior obtained with methanol is very similar to the case of ethanol, except that the fast components τfast and τrise are much shorter than in ethanol.

For all isotopomers in the protic solvents (i.e., ethanol and methanol), the CT fluorescence risetime τrise is found to vary with the emission wavelength: it increases as the observation window is shifted to a longer wavelength region.

This implies that the CT fluorescence intensity distribution redshifts with time, which has been recognized previously18,19 and explained as a result of the longer solvent reorganization time in these solvents.

In this case, the risetime of the CT fluorescence τrise does not match the fast decay lifetime of the LE fluorescence τfast.

On the other hand, rapid equilibrium between the LE and CT states can be reached upon excitation in aprotic solvents, and thus τrise and τdecay of the CT fluorescence are nearly equal to the decay lifetimes τfast and τslow of the LE fluorescence, respectively.


We have observed that in solvents of moderate and high polarity, the intensity ratio of the CT to LE fluorescence (ΦCT/ΦLE) of DMABN decreases upon deuteration of one of the methyl groups.

This symmetry effect is contrasted with the case of the weakly polar solvent dioxane where the fully methyl-deuterated isotopomer gives the lowest ΦCT/ΦLE fluorescence ratio.

Moreover, the time-resolved fluorescence measurements reveal that the asymmetric substitution leads to a substantial decrease in the ratio of the fast to slow decay component of the LE fluorescence (Afast/Aslow) in polar solvents.

Here, we employ the kinetic scheme (Scheme 1) proposed by Grabowski et al3. to rationalize these observations.

In addition, the TICT model is invoked in which the CT stabilization occurs as a consequence of the torsional motion of the DMA group.

For DMABN in polar aprotic solvents, the excited-state equilibrium can be fully established within the lifetime, i.e., k1, k2kLE, kCT, where k1 and k2 are the forward and backward reaction rate constants, respectively, and kLE and kCT are the decay rate constants of the respective excited states.

In this high-temperature regime, the ratio of the LE fluorescence components Afast/Aslow can be associated with the ratio of the reaction rate constants k1/k2.

The reciprocal of the short decay lifetime τfast corresponds to the sum of the forward and backward reaction rate constants k1 + k2.

For DMABN in protic solvents, the risetime of the CT fluorescence is determined by the solvent reorganization time as described above.

In this case, the reaction kinetics can be deduced from the LE state decays.18

The rate constants k1 and k2 for the two aprotic solvents (dioxane and acetonitrile) derived based on the high-temperature approximation are listed in Table 2.

Comparison of the kinetic data in Table 2 suggests that the backward reaction rate constant k2 increases upon asymmetric deuteration, whereas the change in k1 is minor.

It is also important to note that the emission maxima of the LE and CT fluorescence are identical for all isotopomers (Fig. 1), supporting that the energies of the LE and CT states are not shifted by the symmetry reduction.

Therefore, we conclude that the activation barrier for CT changes upon deuteration of one of the methyl groups.

For DMABN-h6 in polar solvents, the activation energy of the backward reaction E2 is much larger than that of the forward reaction E1.

Thus, the ratios of ΦCT/ΦLE and Afast/Aslow increase by lowering the temperature in the so-called high-temperature region.20

The activation energies E1 and E2 have been obtained separately from Arrhenius-type plots of the decay rate constants.21

In both protic and aprotic solvents, the observed Arrhenius activation energy E1 appears to be similar to the activation energy for the solvent mobility Eη.

This implies the absence of intrinsic potential barrier for the forward reaction.21

For example, the activation energy E1 in butyronitrile is obtained to be 10.5 kJ mol−1, which is comparable to Eη (9.2 kJ mol−1).

On the other hand, the activation energy of the backward reaction E2 (15.9 kJ mol−1 in butyronitrile) is substantially larger than Eη, and a significant intrinsic barrier exists for this process.

On the basis of the barrierless reaction model, we propose that the pronounced asymmetric substitution effect on the CT dynamics in sufficiently polar solvents is due to vibrational coupling between the primary reactive mode, namely the twisting motion around the Namino–Cring bond, and other low-frequency motions (e.g., pyramidalization at the DMA group and torsion of the methyl groups).

Among these nonreactive motions, pyramidalization is believed to be important in determining the CT dynamics of DMABN.4,22

With very low pyramidalization, the activation energy for the CT reaction is lowered, leading to a barrierless reaction coordinate, as discussed above.

Therefore, asymmetric substitution of the DMA group exerts a minor influence on the forward reaction process.

In contrast, vibrational coupling with nonreactive modes will increase the density of states for the back CT reaction, thus lowering its barrier height.

The above interpretation is consistent with the absence of noticeable symmetry effect in dioxane.

As shown in Fig. 1, the protonated species DMABN-h6 reveals the largest ΦCT/ΦLE value among all isotopomers.

In the weakly polar solvent, the forward reaction may also have an intrinsic barrier.

In this case, deuterium isotope effects prevail.

The slower CT reaction rate for the deuterated species can be explained in terms of the torsional frequency of the DMA group, which decreases by ≈10% in going from -h6 to -d6.8

While DMABN is known to be pyramidal in the ground state,23 increased conjugation upon excitation would be expected to lead to a more planar (with respect to inversion) conformation in S1.

It has been suggested that its S1 inversion potential is very flat-bottomed.8

On the other hand, the torsional geometry is planar in S0,23 while twisted in S1 by about 26° with a small barrier to planarity.8

In this case, vibrational coupling between the torsion and inversion of the DMA group is expected to occur extensively upon excitation.

Evidence in support of such coupling can be obtained from vibrational analysis of the S1 ← S0 spectra of three isotopomers.8

Each spectrum, obtained in a supersonic jet, consists of a prominent low-frequency progression due to the DMA torsion.

In addition to this, bands assignable to the inversion and its combination bands with the DMA torsion are apparent in the spectrum of the asymmetric isotopomer.

The -h3d3 spectrum also reveals bands involving torsion of each methyl group (CH3 and CD3).

The increased coupling of the reactive DMA torsion with these nonreactive modes upon asymmetric substitution is expected to reduce the activation energy.

It is also possible to ascribe the observed symmetry effect to an increase of the preexponential (frequency) factor due to such mode coupling.

In conclusion, it is demonstrated that the fluorescence intensity ratio of the CT and LE emissions of DMABN decreases substantially upon deuteration of one of the methyl groups.

This asymmetric substitution effect can be observed only in polar solvents of medium and high polarity.

The time-resolved fluorescence measurement reveals that the reverse electron transfer is facilitated in the asymmetric isotopomer, whereas no significant change occurs for the forward reaction.

This behavior is explained by mode coupling of the torsional motion of the DMA group with other low-frequency motions, which occurs strongly upon asymmetry reduction.

However, more accurate decay data would be essential to determine unequivocally the origin of the asymmetric substitution effect.