| [page 27↓] |
Abstract
Absorption and emission properties of DMABN-F4, the tetrafluoro analogue of DMABN, have been investigated and compared with the parent compound. Unlike in DMABN, this new compound exhibits only a highly solvatochromic and strongly redshifted fluorescence CT and is characterized by the absence of an LE band even in nonpolar solvents. This evidences the faster formation of CT in the excited state as compared to DMABN. The low quantum yield values of DMABN-F4 suggest that the high rate of non-radiative decay takes place via internal conversion (IC) rather than intersystem crossing (ISC) as no phosphorescence is observed in rigid glass solvents at 77 K in contrast to DMABN. The emission transition moment and radiative rate constant values of DMABN-F4 in medium and highly polar solvents point to a forbidden emission in the excited state similar to DMABN. Electronic structure and twist potentials were also studied by quantum chemical calculations using ab initio and semiempirical methods. In contrast to DMABN, DMABN-F4 is found to be twisted by around 30-50°, but the photophysics are concluded to be analogous to DMABN with the addition of a very fast IC channel.
Donor-Acceptor substituted benzenes have been the focus for many years regarding the nature of dual fluorescence or of emissions with very large Stokes shifts. Lippert et al. [8]showed that DMABN emits a dual fluorescence consisting of two bands assigned to two different excited states: The A band for the "anomalous" emission from the 1La-type state, B band for the normal short wavelength arises from the 1Lb-type/CT state. The emitting species, also called A* and B* states, can be in thermal equilibrium.
The photophysics of electron donor-acceptor aromatic systems has been well explained with the help of the TICT model (“Twisted Intramolecular Charge Transfer”). According to this model, [1, 9, 16, 19] the untwisted dimethylamino group (electron donor) rotates after photoexcitation towards an orthogonal orientation of the donor group relative to the aromatic ring system. In nonpolar solvents and under jet-cooled conditions, DMABN [page 28↓]emits only from the near planar LE excited state, [22]whereas a second minimum at a twisted conformation is populated on the excited state energy surface in more polar solvents.
Recently, the effect of F-substitution in DMABN-derivatives has been investigated for 4-(azetidinyl)benzonitrile by Druzhinin et al [45]who stated that there is no indication of dual fluorescence, and that internal conversion is enhanced by the fluoro substituent.
DMABN-F4 which differs more from DMABN by the further increased strength of the acceptor unit leads to the expectation of an increased CT nature of the excited state. In the present chapter, DMABN-F4 is characterised spectroscopically and compared with DMABN. The investigation of the spectroscopic behaviour includes both polarity and temperature effects. The red shift of both absorption and more strongly fluorescence spectra can be ascribed to this increase of the acceptor nature quantifiable by an enhancement of the electron affinity (EA) by 0.78eV.
In order to compare and interpret the spectroscopic properties, quantum chemical calculations were performed using ab initio and semiempirical methods.
Figure 4.1: Structure of the molecules investigated | ||
The compounds with their structures and abbreviations are shown in Fig. 4.1 DMABN-F4 was synthesized by the reaction of pentafluorobenzonitrile with dimethylformamide using the procedure described in ref. [46] DMABN was a sample previously used. The absence of possible traces of impurities was confirmed by thin layer chromatography (TLC).
| [page 29↓] |
The experimental details about the absorption, fluorescence, lifetime and quantum yield measurements are described in chapter 3.
Semiempirical calculations for the compounds DMABN and DMABN-F4 were carried out with full geometry optimization and with a vibrational analysis of the optimized structures by the Newton algorithm using the AM1 program contained in the AMPAC 6.0 package [42, 44].
The study of equilibrium and transition structures for the compounds by ab initio calculations at Hartree-Fock theory HF and Density Functional Theory DFT levels were realized with different basis sets (6-31G(d), 6-311++(d,p), cc-pVDZ, and D95(d,p)) using Gaussian 98 [47]. Full optimization of the ground state including vibrational analysis was performed to detect stable minima and transition geometries. The twist angle between the compound fragments was determined as the torsional angle between the lone pair on the nitrogen atom and the benzene plane from the bisector between the optimized torsional angles of the carbon atoms of the dimethylamino group (see Scheme 2). The study of the fragment rotation in the S0-state was carried out by fixing the torsional angle of one carbon atom of the dimethylamino-group optimizing all other geometrical parameters.
The calculations of the transition energies and oscillator strengths for the ground state optimized geometries were carried out using configuration interaction (CI) for the optimized structures with inclusion of 10 unoccupied and 10 occupied orbitals (C.I. = 10) by ZINDO/s (CIS) included in Gaussian 98.
The absorption spectra of DMABN-F4 in various solvents of different polarity are depicted in Fig.1a. The corresponding spectra of DMABN are also presented for comparison (Fig. 4.2b). All spectra of DMABN-F4 as compared with the spectra of DMABN are shifted to the red (Table 4.1a). But in contrast to DMABN, in the spectrum of DMABN-F4 in hexane the weak shoulder is not visible, which is found at the red side of the main absorption maximum of DMABN and ascribed to absorption to the 1Lb-S1 state. It can be concluded that [page 30↓]the weak 1Lb absorption band of DMABN-F4 is hidden by the stronger 1La type band. It is even possible that the 1La state is S1 in this compound.
Figure 4.2: Absorption and normalised fluorescence spectra at room temperature of DMABN-F4 (a) and DMABN (b) in various solvents of different polarity. Hex = n-hexane; BCl = n-butyl chloride ; EOE =diethyl ether; ACN = acetonitrile. | ||
The molar extinction coefficient values for both compounds determined in n-hexane are approximately equal [ε (λ) = 28,911and ε (λ
) = 34,436 respectively for DMABN and DMABN-F4]. In view of the twisted ground state structure of DMABN-F4, this might indicate a different vibronic mixing of the two lowest singlet states, 1La and
1Lb.
Analogous to DMABN, the long wavelength absorption band of DMABN-F4 is shifted to the red by increasing the solvent polarity (Fig. 4.2a and Tables 4.1a and 4.1b).
Table 4.1a: Photophysical parameters of DMABN-F4 in Various Solvents at Room Temperature
Table 4.1b: Photophysical parameters of DMABN in Various Solvents at Room Temperature
In all solvents studied, DMABN-F4 possesses very weak fluorescence (quantum yield φf ≤ 0.002) with a broad emission band (Δν1/2 > 6000 cm-1) shifted unusually far from the long wavelength absorption band (Stokes shift Δνst ≥ 11000 cm-1) even in hexane (Table 4.1a and Fig. 4.2a) In contrast, in the latter solvent the emission of DMABN is relatively strong (φf= 0.17) and narrow (Fig. 4.2b and Table 4.1b). The absence of dual fluorescence and the indicated fluorescence properties (strong red shift) suggest a very fast formation of an emitting CT species in the excited singlet state of DMABN-F4 in contrast to DMABN. As one can see from Table 4.1a, quantum yield values decrease from low-polarity to high-polarity solvents with the exception of hexane, where an anomalously high k nr is found.
The fluorescence decay curves measured for DMABN-F4 in different solvents are monoexponential and similar for different wavelengths of the emission spectra with the fluorescence lifetime ranging between 0.21 and 1.19 ns (Table 4.1a). These measurements support the formation of only one emitting state in all solvents.
| [page 32↓] |
If we assume the direct formation of the emitting species without losses (i.e. validity of the Kasha rule), then the radiative kr and nonradiative knr rate constants can be calculated according to equations 4.1 and 4.2,
where kcorresponds to the sum of all nonradiative processes including triplet formation, internal conversion and the possible formation of nonemissive photochemical products. The kr values are extremely small and decrease from about 2 x 106 s-1 (the corresponding the radiative life time τr is 450 ns) in hexane to 0.5 x 106 s-1 (τr is 2100 ns) (Table 4.1c) in acetonitrile. Similarly low
kr values were observed for a sterically hindered DMABN analogue, TMABN [48, 49] in which two methyl groups are present in ortho positions of the DMABN, in the highly polar solvent propanol (5.1 x 106 s-1). This gives supporting evidence for the forbidden radiative transition from the excited state, which is typical for TICT states.
The formation of full charge transfer (i.e. a TICT state) in DMABN-F4 has been recently confirmed by time resolved absorption spectroscopy [50]. The transient absorption spectrum in acetonitrile at 1 ps delay showed a band around 360 nm. It was attributed to the CT state by its similarity with that reported for DMABN at 100 ps. The much faster appearance time of the CT state of DMABN-F4 suggests that the CT formation is a quasi-barrierless process in both polar and non polar solvents in this molecule and that TICT state formation is strongly favored with respect to DMABN.
The radiative transition moment values Mf in Table 4.1c as calculated from eq. 4.3 decrease from hexane to acetonitrile in the case of DMABN-F4 in contrast to DMABN and TMABN, where the Mf values are independent of polarity [49, 51]. The small magnitude of Mf in DMABN-F4 is typical for the twisted structure of a TICT-state. In the case of TMABN, the smaller values of Mf can be interpreted by sterical hindrance, which leads to a narrowing of the angular distribution around 90° as compared to the unhindered compound DMABN and therefore to more strongly forbidden emission. It is remarkable that Mf of DMABN-F4 in acetonitrile is the smallest value ever reported for the TICT fluorescence of an aniline derivative, lower even than for TMABN or other twisted model compounds of DMABN.
Figure 4.3: Low temperature effects on the fluorescence spectra of DMABN-F4 in n-butyl chloride. Down head arrow indicates the decreasing of temperature. The data points in the range 600-630 were omitted (second order of excitation wavelength). | ||
Table 4.2: Temperature Dependence of the Photophysical Data of DMABN-F4 in BCl
| [page 34↓] |
Figure 4.4: Fluorescence spectra of a) DMABN-F4 in BCl and BCN at 77 K and b) DMABN in BCl at room temperature and at 77 K. | ||
The fluorescence study of DMABN-F4 at lower temperatures was done in n-butylchloride in order to study the relaxation processes in a glassy matrix. These measurements detected a weak red shift of the emission maximum from 501 to 535 nm and a decrease of the fluorescence quantum yield by more than a factor of 2 when the solvent is cooled from room temperature to 173 K (Figs. 4.3 and 4.4, Table 4.2). Further cooling until 77 K did not allow quantum yield measurements, but the fluorescence band is found to be weak and shifted somewhat to the blue. However, even with the rigid glass matrix at 77 K the LE emission and phosphorescence are absent. In contrast to this, DMABN at 77 K possesses only the LE emission at 342 nm and phosphorescence is observed at 411 nm with a highly structured band (Fig. 4.4b). The small fluorescence intensity and the redshifted spectrum of DMABN-F4 at 77 K gives evidence that the emission is forbidden and that there is some relaxation even in a highly polar glassy matrix.
The calculations indicate that in contrast to planar DMABN,the derivative with fluorine atoms DMABN-F4 possesses a somewhat twisted equilibrium geometry in the ground state (torsional angle between the fragments is 35 – 50 degrees, depending on the calculation method, see Fig. 4.5 and Table 4.3). The reason is a stronger sterical interaction between the two methyl groups of N(CH3)2 and the fluorine atoms in the benzene ring. The [page 35↓]interaction is caused by the longer C-F bond (close to 1.3 Ǻ for all calculation methods) in DMABN-F4as compared to the C-H bond length close to 1.08 Ǻ in DMABN. The pyramidalization of the dimethylamino group of DMABN-F4 is predicted very differently depending on basis set and method used. For the conditions of Fig. 4.5, it is practically nonpyramidal in the equilibrium structure (near sp2 hybridization). Similar strong variations of the pyramidalization depending on the method can be observed for DMABN (Table 4.3).
Figure 4.5: The planar transition, twisted equilibrium and perpendicular transition structures of DMABN-F4 and the planar equilibrium geometry of DMABN with some geometrical characteristics calculated by DFT (B3LYP/6-311++G(d)). | ||
For comparable methods, both semiempirical and ab initio calculations demonstrate a smaller sp3 hybridization of the dimethylamino group for the relaxed [page 36↓]geometries of DMABN-F4 than for DMABN (from AM1, equilibrium pyramidalization angles are around 13° and 29° deg respectively, Table 4.3).
| [page 37↓] |
Figure 4.6: The ground state potentials of DMABN-F4 (1) and of DMABN (2) calculated by DFT (B3LYP/6-31G(d)) method. The torsion angle is determined according to figure 4.7. The inflection in the potential for DMABN at around 15° is due to different methyl group conformations being the most stable ones (see Scheme 4.8). | ||
The intramolecular fragment rotation away from the equilibrium geometry to the planar and perpendicular geometry of DMABN-F4 (Table 4.3, Fig. 4.6) increases the potential energy. Both planar and perpendicular geometries possess a Cs symmetry and correspond to a saddle point (Fig. 4.7) (one negative Eigenvalue in the Hessian matrix).
| [page 38↓] |
Figure 4.7: Planar, equilibrium and perpendicular structures of the dimethylamino group for DMABN-F4 in the ground state (top). Determination of the dimethylamino-group twist angle α and the dimethylamino group pyramidalization angle β (bottom): n is a vector perpendicular to the plane of the aromatic ring, nα is the bisector vector of the two N-methyl bonds of the dimethylamino-group, nβ is a vector perpendicular to the CNC plane of the dimethylamino group. | ||
| [page 39↓] |
Figure 4.8: conformation of methyl-group of dimethylamino-group for DMABN-F4 and DMABN in the ground state. | ||
A study of the possible conformations of the two methyl group of the dimethylamino N(CH3)2fragmentshows that several conformers are important (Fig. 4.8) that only conformer syn1 with Cs symmetry for the planar and twisted geometries has the lowest energy, and the Hessian matrix indicates a saddle point of first order. The other conformers syn2 and syn3 of the symmetries C2v and Cs possess saddle points of second order (Table 4.4). Depending on the twist angle, the relative energy of the conformers can change. This is the reason for the inflection in the potential for DMABN at around 15° (Fig. 4.6).
The excited-state dipole moments μ
e of DMABN-F4 and DMABN were determined from the solvatochromic slopes by applying the Mataga equation (4.4) [52, 53, 54]. The slopes from the corresponding solvatochromic plots of the emission maxima against the solvent polarity parameter, can be used to calculate the excited state dipole moment. Different values result from different assumptions regarding the Onsager radius a. These are compiled in Table 4.5a for the cases where ‘a’ was calculated from (1) the mass-density formula (eq. 4.5) [55](2) the molecular volume as calculated using the HF method and (3) using the Lippert approach [52].In order to gain a more reliable basis, the Onsager radii for different fluorosubstituted benzenes and dimethylaminobenzenes were determined from the experimental densities and the molecular weight (eq. 4.5). As one can see from Table 4.5b, addition of one or more fluorine atoms to benzene do not induce any drastic changes in the value of the Onsager radius. Thus, it is not surprising that both DMABN and DMABN-F4 are predicted to have similar Onsager radii. Hence, all the above mentioned three methods resulted in approximately of the same excited state dipole moment values for the CT state, around 13 D for DMABN-F4 and 18 D for DMABN. For the usual point dipole Onsager model, because of similar Onsager radii, the resulting significantly different CT dipole [page 40↓]moment values are entirely a consequence of the difference of the experimental solvatochromic slopes.
| [page 41↓] |
In the above equations, μ e and μ g are the excited and ground state dipole moments respectively, h is Planck’s constant, c is the velocity of light, M is the molecular mass, NA is Avagadro’s number and n and ε are the refractive index and dielectric constant, respectively.
For the discussion of the CT structure and dipole moment, it became important to know more about the properties and relative energies of the acceptor orbitals involved. Although fluorine substitution is expected to lead to an overall lowering of the energy of the acceptor orbitals, a closer look into the reported literature shows that this lowering very strongly depends on the substitution pattern, and that orbitals of different symmetry can exchange their energetic position. We therefore undertook to calculate the orbital energies of both the highest two occupied orbitals and the lowest two unoccupied orbitals by different methods and to compare them to the available experiments (Table 4.6). Especially useful is a look at their energetic difference (Table 4.7), which changes from negative to positive values if the orbitals exchange their energetic position. Both the theoretical and the experimental results (Table 4.6) show that insertion of the fluorine atoms into the benzene ring increases the acceptor property of this fragment. The HF ab initio calculations (Table 4.6) for a series of compounds containing fluorine are in rather good agreement with the experimental ionization energies, but the electron affinities are poorly represented. On the other hand, DFT (B3LYP/6-31G(d)) shows a much better correspondence to the LUMO energies than HF. Fig 4.9 shows that depending on the substitution pattern and the number of fluorine atoms, the [page 42↓]orbitals of different symmetry (labeled with respect to the C2 symmetry point group) exchange their energetic position. Fig. 4.10 depicts the HOMO and LUMO energies calculated with HF and DFT: HF fits much better to the experimental ionization energy, whereas the correspondence of DFT LUMO energies and the experimental electron affinities is much better. But both methods agree in the prediction of the relative changes of orbital energies with the fluorine substitution pattern.
|
Table 4.7: The energy difference Δε (eV) = ε(B) - ε(A), of the molecular orbitals of benzene with different substituents and with different symmetry as calculated by the following methods: HF/6-31G(d), AM1, DFT (B3LYP/6-31G(d)) and compared to experimental values as far as available. Upper rows: difference of the first two LUMOs, lower rows: difference of the two highest occupied orbitals.
Figure 4.9 and Tables 4.6 and 4.7 show that in some compounds (1 fluoro and 1,4 difluorobenzene), the orbitals of B symmetry are higher lying than the orbitals of A symmetry. In the twisted geometry, the electron transfer from dimethylamino group (the [page 45↓]donor orbital transforming as B) can be to either of the two lowest LUMO orbitals of the acceptor. This could have the consequence that also the energetic position of the two possible TICT states interchanges, and the lowest TICT state could become of B symmetry instead of A symmetry as in DMABN.
Figure 4.9: Correlation diagram of the energies of the occupied and unoccupied orbitals of fluorinated benzenes and benzonitrile as calculated by HF (values see table 6) | ||
Figure 4.10: The comparison of the orbital energy for HOMO (εhomo) and LUMO [(εlumo) calculated by different methods (HF/6-31G(d) and DFT (B3LYP/6-31G(d))] with the negative experimental values of electron affinity EA and ionisation potential IP for different compounds. The compounds are defined in table 6. | ||
The results for 1,2,4,5 tetrafluorobenzene, however, clearly show that the B orbital is lower than the A orbital for both the HOMO and the LUMO manifold, and this effect is enhanced for the LUMO by introducing a further cyano group (but weakened for the occupied orbitals of the acceptor). From the DFT-LUMO energies for 1,2,4,5-tetrafluorbenzene and its cyano derivative, we can conclude that the four fluorine atoms lower the LUMO energy by 0.78 eV.
Thus we can conclude that the properties of the lowest TICT state should be the same for DMABN and DMABN-F4 regarding symmetry, and hence also the charge distribution should be similar.
| [page 46↓] |
Figure 4.11: Schematic diagram showing the state energies of DMABN and DMABN-F4 in the gas phase, as calculated by ZINDO/s. | ||
This lowering of the TICT energy has also consequences for the reaction enthalpy of the B* → A* reaction. In the gas phase, it is calculated as being uphill (endothermic) for both compounds, but significantly less for so for DMABN-F4. Thus, for a given solvent polarity, where the reaction is exothermic in DMABN, the exothermicity is therefore much larger for DMABN-F4 (Figure 4.11)
The semiempirical ZINDO/s calculations show that the long wavelength absorption region of DMABN and DMABN-F4consists of two transitions (Table 4.8). The position and intensity of these transitions are characterized by an allowed S2 transition (4.22 eV, f = 0.49 and 4.59 eV, f = 0.64 for DMABN-F4 and DMABN, respectively) with charge transfer (CT) nature (dipole moment 12.4 D similar to DMABN) and a polarization along the long molecular axis (1La-type according to Platt nomenclature). Thus, there is a clear redshift of the 1La-type state as expected from the increased acceptor strength also shown by the experimental spectra (Fig. 4.2). The forbidden S1 transition (f = 0.01) with a somewhat lower energy (3.93 eV in DMABN-F4) is of 1Lb nature and does not lead to a structure in the absorption spectrum (Fig. 4.2a) in contrast to the similar S1 transition (4.27 eV, f= 0.02) of DMABN which leads to a weak shoulder on red side of the absorption spectrum. The absence of structure in the absorption spectrum of DMABN-F4 in hexane is in contrast to the calculated S1-S2 energy difference (Table 4.8), which is even larger for DMABN-F4 (0.52 eV) than for DMABN (0.32 eV). We consider that the energy lowering of the 1Lb-type state in DMABN-F4 is probably smaller than predicted by the calculations in Table 4.8.
|
Table 4.8:Comparison of the transition energy ΔE, oscillator strength f, dipole moments and configuration interaction analysis for the long wavelength absorption transitions for DMABN and DMABN-F4 as calculated by ZINDO/s for the optimized ground state equilibrium geometry and two further twist angles. (full optimization at the different fixed twist angles using DFT (B3LYP/6-311++G(d)).
The stronger acceptor in DMABN-F4 is also reflected in the energy of the calculated (and observed) TICT state: In DMABN, it is calculated at 5.36 eV, in DMABN at 4.60 eV, i.e. 0.76 eV lower, very close to the value predicted by the lowering of the acceptor orbital energy (0.78 eV, Table 6). As a consequence, in contrast to DMABN, the lowest TICT state of DMABN-F4 calculated for 90° corresponds to S2, whereas it is S4 in DMABN in the gas phase according to the ZINDO/s calculations.
| [page 48↓] |
On the basis of the solvatochromic measurements, the resulting CT dipole moments of DMABN (18 D) and DMABN-F4 (13 D) were found to be strongly different, with that of DMABN-F4 being anomalously low. On the other hand, the calculations indicate that the dipole moments should be similar (16.7 and 16.8 D, respectively, Table 8), and the transient absorption experiments [50] also indicate that the electronic structure is similar and full charge separation has occurred.
This discrepancy is solved if the assumption of equal Onsager radii a is dropped, because the solvatochromic slope is proportional to the ratio μ e /a 3. An increased a for DMABN-F4 will lead to a correspondingly increased μe. In fact, we can estimate that an increase from a = 3.7 to a = 3.8 is sufficient to result in similar dipole moments for the TICT states of both DMABN and DMABN-F4. On the other hand, using the molecular volumes estimated by quantum chemical calculations and from the densities and molecular weights (Table 4.5b), we concluded that the solvent cavity volume excluded by the solvent should be very similar for the two compounds. Using Onsager's point dipole approximation, which is at the basis of the solvatochromic treatment according to Mataga, we are therefore forced to conclude for different CT dipole moments from the different experimental solvatochromic slopes.
A possibility for justifying differently sized a-factors for the two compounds would be by assuming different solute-solvent interactions for the same value of the a-factor. If the solute-solvent response would be weaker in DMABN-F4 than in DMABN, this would correspond to an increased a-factor in the usual solvatochromic equation. This is equivalent to say that the Onsager point dipole approximation is not valid for this comparison.
We can justify such a point of view by considering the different charge distribution in the two compounds: The pi and pi* orbitals are very similar, hence the CT charge distribution due to the pi-electronic transitions should be similar (see the calculations, Tab. 4.8). But the fluorine atoms on the acceptor are negatively charged (in both ground and excited state) whereas the H-atoms in DMABN are positively charged. The negative charge on the fluorine atoms will prepolarize the solvent dipoles already in the ground state such that upon excitation to the charge transfer state, their relaxation possibility is reduced which is equivalent to say that the solute-solvent interaction is effectively weakened.
This explanation of the dipole moment discrepancy by a break-down of the Onsager point dipole theory is a first attempt to explain the unusual solvatochromic behaviour [page 49↓]of highly fluorinated charge transfer compounds. Further experiments are needed to verify this explanation.
The value of k nr is considerably larger for DMABN-F4 than for DMABN, especially in n-hexane (Table 4.1c). We can conclude that we have an additional nonradiative decay channel: Its nature could be either population of a triplet state (intersystem crossing ISC) or population of a transient nonemissive singlet species and/or nonradiative photochemistry through a conical intersection (internal conversion IC). As there is no permanent photochemical product these reaction paths have to lead back to the ground state of the starting material. As we do not observe phosphorescence, the triplet path is not probable. The number of fluorine substituents seems to be an important factor in the enhancement of this IC path as already stated by Druzhinin et al. [45]. In the tetrafluorinated DMABN, this IC path is further enhanced as compared with the monofluorinated derivatives.
A possible reaction path that can be discussed is the folding (relaxation to nonplanarity) of the benzene ring in the excited state. Recent calculations of DMABN indicate that even in this nonfluorinated compound, the TICT state possesses a nonplanar benzene ring [56]. The case of 1,2,4,5 tetrafluorobenzene further exemplifies the effect of the fluorine atoms which enhance the tendency for nonplanar folding: Whereas benzene shows no sign of a deviation from planarity in the emissive excited state, the emissive state of tetrafluorobenzene shows a folding (butterfly motion) resulting in a double minimum potential, as deduced from laser-induced fluorescence spectra [57]. If there is a similar folding in DMABN-F4 (and to some extend even in DMABN which also has a nonnegligible contribution of IC), it could lead to a conical intersection along the reaction coordinate leading either to the Dewar isomer (folding of two four-membered rings) or to the benzvalene isomer (folding of two three-membered rings). Such conical intersections are known to play an important role in the nonradiative photochemistry of benzene and its derivatives [58, 59, 60].
| [page 50↓] |
DMABN-F4 is spectroscopically closely related to DMABN but characterized by a higher acceptor strength. The strongly redshifted emission and absence of the short wavelength B-band in the fluorescence spectrum of DMABN-F4 at 77 K indicates the ultrafast formation of a CT structure in the excited state, probably linked to the pretwisted ground state geometry and the increased acceptor strength. The low fluorescence quantum yield values and absence of phosphorescence of DMABN-F4 suggest that the high rate of non-radiative decay takes place through internal conversion rather than intersystem crossing. A possible intersystem crossing reaction path could be the folding (butterfly motion) of the benzene ring either towards a Dewar or a prefulvene deformation. Results from time-resolved measurements indicate that the emission of DMABN-F4 is strongly forbidden consistent with the formation of a twisted intramolecular charge transfer state (TICT) with high dipole moment. The relatively small solvatochromic slope for this CT emission, as compared to expectations from quantum chemical calculations, indicates the possibility for a breakdown of the Onsager point dipole approximation for highly fluorinated CT compounds.
© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme. | ||
DiML DTD Version 4.0 | Zertifizierter Dokumentenserver der Humboldt-Universität zu Berlin | HTML generated: 07.07.2005 |