A Theoretical Study: Solvent Effects on the Electronic Excitations of (N,N)-bridged 4-Aminobenzonitriles

Rudolf Schamschule^1,+, Andreas B. J. Parusel¹, Gottfried Köhler^1,2

¹ Institut für Theoretische Chemie und Strahlenchemie, University of Vienna, Althanstr. 14, 1090 Vienna, Austria.
² Austrian Society for Aerospace Medicine - Institute for Space Biophysics, Lustkandlgasse 52, 1090 Vienna, Austria. ⁺ Corresponding author.

First presented at the first internet conference on photochemistry and photobiology

Keywords: Aminobenzonitriles, Charge transfer, Dual fluorescence, Solvent Effects

Contents

Abstract

1. Introduction
2. Computational Methods
3. Results and Discussion
        3.1. Results for N-azetidinyl-benzonitrile (4BN)
        3.2. Results for N-pyrolidinyl-benzonitrile (PYRBN)
        3.3. Results for N-piperidinyl-benzonitrile (PIPBN)
        3.4. Comparison to Experimental Data
4. Conclusion
5. Acknowledgements
6. References

N-azetidinyl-benzonitrile (4BN), N-pyrolidinyl-benzonitrile (PYRBN) and N-piperidinyl-benzonitrile (PIPBN)

Abstract

The influence of polar environments on the excited state properties is investigated for a series of four to six membered N,N-bridged 4-aminobenzonitriles. The self consistent reaction field model (SCRF), with an extension for the description of solvent effects on the electronically excited states, is used in a semiempirical AM1 based method. The rotational isomerization of the alkyl-amino group, a pseudo Jahn-Teller mechanism, and the rehybridization at the cyano carbon are investigated as possible explanations for the observed occurrence of dual fluorescence. Based on these solvent calculations, fluorescence wavelengths are compared to experimental data from literature. The RICT (Rehybridization by Intramolecular Charge Transfer) and WICT (Wagging Intramolecular Charge Transfer) models are rejected as explanations of the dual fluorescence phenomenon for this class of donor-acceptor compounds. The TICT (Twisted Intramolecular Charge Transfer) is found to explain the dual fluorescence of each compound for itself. A sufficiently small energy gap is found between the S₁ and S₂ excited states, making vibronic coupling, as proposed by the pseudo Jahn-Teller hypothesis, an accurate hypothesis for the description of the dual fluorescence behaviour of these compounds.

back to Contents.

1. Introduction:

The phenomenon of dual fluorescence has been widely studied for 4-N,N-Dimethylamino-benzonitrile (DMABN) and related electron donor-acceptor compounds. Besides the so-called normal fluorescence out of the locally excited S₁ state, a second, strongly red-shifted emission can be observed with increasing solvent polarity. Since discovery of this effect DMABN and derivatives by Lippert et al. [1], the strong polarity dependence of the spectral position of the second, red-shifted fluorescence has been well documented [2-5]. The increased quantum yield, when hydrocarbons are replaced by polar solvents [6,7], combined with the dependence on environment polarity, indicate a highly dipolar excited state achievable only by a charge transfer in combination with an internal geometric relaxation out of the locally excited (LE) S₁ state, and not reachable by direct excitation.

Figure 1a: Definition of the wagging angle , the twist angle , and the rehybridization angle in combination with the explanation of the WICT, TICT, and RICT hypotheses for 4BN. The definitions for the other compounds are analogous.

The wagging angle , defining the pyramidalization of the amino-group, and the twist angle , defining its out of the benzonitrile plane rotation, as well as the rehybridization angle representing the structural change of the molecules during the rehybridization of the cyano carbon are presented for 4BN in Figure 1a. The definitions of these angles are analogous for the other molecules.

The widely accepted TICT (Twisted Intramolecular Charge Transfer) hypothesis [2, 6-8] predicts the intramolecular relaxation of the planar vertically excited state (LE for locally excited) to a highly dipolar charge transfer state with perpendicular orientation of the amino (electron donor) and benzonitrile (electron acceptor) moieties as the cause for to the occurrence of the strongly redshifted fluorescence band. For the investigation of this TICT model, the change of dipole moment and excitation energy of the relevant electronically excited states are studied with increasing torsion angle of the amino group. In this TICT hypothesis, the twisting motion of the amino group (see Figure 1a) causes a maximized charge separation, by means of a mutually perpendicular conformation of the amino-donor and the benzonitrile-acceptor subunits. In this charge transfer state, an electron is transferred from the nitrogen lone pair into the unoccupied aromatic benzonitrile orbital.

The WICT model (Wagging Intramolecular Charge Transfer), proposed by Gorse and Pesquer [9,10], predicts a charge transfer state, where the amino nitrogen lone pair orbital is decoupled from the benzonitrile -orbitals by means of an amino wagging mode. Therefore the WICT model demands a strong deviation from planarity corresponding to large values for the wagging angle . This change of the pyramidalization of the amino nitrogen is the considered to be the sole relaxation mode responsible for the dual fluorescence. With the wagging angle as its characterizing coordinate, a change of the amino nitrogen from planar sp² to pyramidal sp³ hybridization is described as another possible hypothesis for the description of the observed dual fluorescence phenomenon.

Figure 1b: Energy diagram of the two lowest excited states. According to the pseudo Jahn Teller hypothesis, a sufficient small energy gap E is necessary for the occurrence of dual fluorescence. With a large energy gap between the S₁ and the S₂ state (left), no vibronic coupling occurs. With a sufficiently small energy gap (right), the characteristic double minima of the S₁ state is observed, resulting in vibronic coupling between these two states.

Vibronic coupling can be achieved by a sufficiently small energy gap between two interacting states S₁ and S₂ (see Figure 1b). I n combination with the presence of a promoting mode, such as nitrogen inversion, this coupling can result in a charge transfer state leading to the second fluorescence band. This so-called solvent induced pseudo Jahn-Teller effect model [4, 11-14] was first introduced by Zachariasse et al. Contrary to the WICT hypothesis, the amino wagging angle only varies between the ground state minima at =-35^o through the planar conformation (at =0^o to =35^o

In 1996, an in-plane bending of the cyano group was calculated [15-17] as another possible main relaxation coordinate (see Figure 1a). With this so-called RICT hypothesis (Rehybridization by Intramolecular Charge Transfer), a sp to sp² rehybridization of the carbon atom of the cyano group, causing a relaxation by bending of the cyano angle, is presented as the main stabilizing factor for the formation of the charge transfer state.

Scheme 1: Molecular structures of 4BN, PYRBN and PIPBN (left, middle and right).

Rate constants for PYRBN and PIPBN measured with laser techniques were presented by W. Rettig [18]. Further studies on the solvent dependent dynamics of the formation of the TICT state at different temperatures are known from literature for PYRBN and PIPBN [19]. To determine the excitation energy dependence of the fluorescence, low temperature studies in n-propanol were discussed by D. Braun and W. Rettig for PYRBN [20]. In 1994, the dual fluorescence was investigated by K. A. Zachariasse et al. [11], leading to the proposal of the pseudo Jahn Teller effect as a cause of the dual fluorescence phenomenon. The solvent polarity dependence of the dual fluorescence was measured for a series of three- to eight-membered (N,N)-bridged 4-Aminobenzonitriles [21].

In a recent paper, we reported on semiempirical electronic structure calculations in gas phase, describing the excited state potential energy surfaces of N-azetidinyl-benzonitrile (4BN), N-pyrolidinyl-benzonitrile (PYRBN) and N-piperidinyl-benzonitrile (PIPBN) [22] (see Scheme 1). Different possible structural relaxation modes were compared for these N,N-bridged aminobenzonitriles. Distinctive charge separation and large dipole moments were correlated to the occurrence of the red-shifted fluorescence. Based on these gas phase calculations, we rejected the WICT model as an explanation of the dual fluorescence phenomenon. The RICT model yielded high energy barriers against the cyano bending mode, but due to the resulting stabilized charge transfer states, the gas phase calculations were insufficient for a definite rejection of this RICT hypothesis. The TICT and the pseudo Jahn-Teller hypotheses were found as possible concurring relaxation modes for this class of donor-acceptor molecules.

Based on the gas phase calculations presented in the previous paper, the effects of a polar solvent on the excited state properties are explicitly calculated for these molecules. The SCRF (self consistent reaction field) model [23] is used for the description of these solvent effects. Solvated excited state properties of N-azetidinyl-benzonitrile (4BN), N-pyrolidinyl-benzonitrile (PYRBN) and N-piperidinyl-benzonitrile (PIPBN) are presented as a means of describing the spectroscopic behavior of these compounds. To model the solvent polarity dependence of the fluorescence shift, calculated fluorescence wavelengths for acetonitrile and diethylether are compared to experimental data.

back to Contents.

In a previous paper, gas phase potential energy surface scans along internal relaxation coordinates were presented for these N-bridged aminobenzonitriles [22]. Based on these gas phase calculations performed with the semiempirical AM1 method [24], the effects of increasing solvent polarity are calculated with the SCRF (self consistent reaction field) [21] approach. In the SCRF model for ground and excited states, the solvent is treated as a polarizable dielectric continuum, characterized by its refractive index n and its dielectric constant . The so called reaction field, an electrostatic field, is generated by the polarization of the continuum. The polarization of the solute within the solvent cavity, itself influencing the field, is brought to self consistence within a self consistent field (SCF) iteration sequence. The solvated excited states are described by configuration interaction calculations with the CISD method with 10 active orbitals including all single and double excitations. Excitation energies and orbital analysis as well as dipole moments are used as a means of characterizing the properties of the relevant charge transfer states along the relaxation coordinates.

The torsion angle and the wagging angle of the amino moiety are presented as possible relaxation modes resulting in a red-shifted fluorescence band. The torsion and wagging angles, and , respectively, were varied in steps of 10^o with the values of the twist angle ranging from 0^o (planar conformation) to 90^o (twisted conformation). The wagging angle was scanned from 0^o with a planar sp² conformation up to 70^o with a sp³ hybridized amino nitrogen. Due to the absence of a sufficiently stabilized charge transfer state for large wagging angles , the potential energy scans for varying wagging angles are only briefly discussed to reject of the WICT model for these compounds. Additionally, the cyano bending angle is scanned from 180^o to 110^o. For a correct description of the changing hybridization of the cyano carbon, the C_cyano- N_cyano bond distance was varied linear with the cyano bending angle, as done by Domcke and Sobolewski [15]. The bond length varied from 1.14 Å for =180^o (sp hybridization, corresponding to a CN triple bond), to 1.30 Å for =120^o (sp² hybridization, CN double bond), and up to 1.33 Å for =110^o.

For all computations, the semiempirical AM1 Hamiltonian included in the VAMP program package, version 6.1 [25], is used. These calculations were performed on Indy workstations (MIPS R4400, Silicon Graphics Inc.).

back to Contents.

Results and Discussion:

In a previous publication, relaxed scans of the gas phase potential energy surface of the electronic ground state were used to illustrate the change in the orbital character as well as changes in the energies of the excited states along certain relaxation coordinates [22]. While gas phase calculations are a fairly reliable tool for the qualitative description of these electronic excitation processes, only explicit inclusion of solvent effects can lead to a quantitative description of energy barriers and solvent stabilization effects. The excited state energies are investigated in acetonitrile for selected excited states of interest for the description of the dual fluorescence phenomenon.

back to Contents.

The excitation energies and orbital compositions of the charge transfer states considered by the TICT (alkylamino rotation) and RICT (bending of the cyano group) hypotheses are presented for 4BN in Figure 2.

Figure 2: Excited state energies and orbital analyses for 4BN in acetonitrile for varying cyano bending angle (bottom left) and for varying amino twist angle (bottom right). Orbital compositions are presented (top, from left to right) for the RICT state at =120^o and for the S₃ (the charge transfer state becoming the RICT state upon cyano bending) at =180^o and for the S₂ charge transfer state at =0^o and for the same state (TICT state) at =90^o

For the twisting relaxation mode, the lowest two excited states are of special interest. The S₁ state is of B₁ symmetry, consisting mainly of a benzonitrile ^* transition. The S₂ state, where an electron is transferred from the amino nitrogen lone pair to the benzonitrile moiety, is a HOMO to LUMO charge transfer state of A symmetry. A change in the character of the relevant orbitals towards increasing charge separation can be observed for growing twist angles. For =90^o, a localization of the HOMO on the nitrogen of the amino donor group can be observed, with a delocalization of the LUMO over the benzonitrile acceptor group. At about 20^o amino twisting, the energies of these two states cross, the HOMO to LUMO state (TICT state) is lowered below the ^* state, which in turn becomes the S₂ state for twisted conformations. The twisting of the amino group yields a strong diminished dipole moment for the ^* state, from 14.0 D for the untwisted conformation to 6.6D for the twisted conformation. The dipole moment of the n_amino^* charge transfer state is increased from 12.4D (untwisted conformation) up to a dipole moment of 21.5D for the twisted conformation (TICT state). In addition to these two states, a higher excited state is stabilized by twisting of the amino group. Due to the comparable properties (upon twisting of the alkylamino group) of this state and the TICT state, this second, energetically higher charge transfer state is referred to as TICT2 state. Due to the stabilizing effects of the polar solvent, the S₂ excited state is lowered close to the S₁ even for the untwisted (LE for locally excited) conformation of 4BN, with an energy difference of only 0.04 eV, which is almost isoenergetic. For increasing twist angles of the amino group, the high dipole moment of the solvated TICT charge transfer state results in a strong stabilization of this excited state. The energy barrier against the relaxation along this twisting mode is found to be only 0.02 eV (see Table 1), which is surmountable by vibronic excitation at room temperature, resulting in an unhindered twisting motion upon electronic excitation. This small barrier facilitates the relaxation along the twisting mode, strongly supporting the TICT mechanism as a possible description of the dual fluorescence phenomenon for 4BN. Another dipolar charge transfer excited state, the TICT2 state, is strongly stabilized for large twisting angles, with its energy decreased below the lowest ^* excited state for 90^o twisting.

The inversion angle was considered as another possible relaxation mode leading to an internal charge transfer state yielding the second, red shifted fluorescence. In previous gas phase calculations, it was found that, due to the absence of an energy stabilization, the wagging mode of the alkylamino group is unlikely to cause the dual fluorescence. Even with the explicit inclusion of the effects of a polar solvent and for large wagging angles ( > 60), intramolecular relaxation of these molecules along the inversion mode leads to increasing excited state energy with increasing wagging angle, with a drastic increase in the energy for angles greater than 50^o. The absence (for all three systems) of a stabilization of any charge transfer state below the S₁ state makes the WICT hypothesis unsuitable for all of the discussed N,N-bridged aminobenzonitriles. Due to this rejection of the WICT hypothesis for this class of molecules, this model is not further discussed in this paper.

The RICT hypothesis proposes the cyano bending mode, where the hybridization of the cyano carbon atom from sp to sp² is the sole relaxation mode responsible for the occurrence of dual fluorescence. Contrary to the TICT mode, where the energy of the S₂ state is decreased below energy of the former S₁ state, a higher state with B₁ symmetry (the so-called RICT state) is lowered in energy. Due to the high dipole moment of this charge transfer state, this excited state is lowered from the S₆ state in gas phase [22] to the S₃ in acetonitrile. Even though the dipole moment varies only from 24.7 D for the unbent conformation, up to 27.2D for the energy minimum at =130^o this state is drastically lowered in energy (below the energy of the S₁ state of the untwisted conformation). For larger angles, the dipole moment rises again, to 20.3D for =110^o accompanied by a corresponding increase of the excited state energy. Even though the energy lowering of this charge transfer state is in favor of this hypothesis, the prohibitively high energy barrier of 0.71 eV makes the intramolecular relaxation along the cyano bending mode impossible. The energy barriers against the relaxation along the twisting and cyano bending modes are presented in Table 1.

There is no energy stabilization caused by the relaxation along the amino wagging mode. Therefore no energy barriers are presented for the WICT model. For the cyano bending mode in PIPBN, there is only a small barrier against the cyano bending mode. Nevertheless, the RICT model does not offer a correct hypothesis for PIPBN, due to fact that the energy of the amino-cyano charge transfer state is not lowered below the energy of the locally excited (LE) state.

Table 2: Energy gap <E [eV] between the lowest excited states for the untwisted conformation of 4BN, PYRBN and PIPBN.

The solvent induced pseudo Jahn Teller theory has been widely discussed as an alternative hypothesis for the description of the dual fluorescence phenomenon for 4-(N,N- Dimethylamino)-benzonitrile (DMABN). For 4BN, the explicit inclusion of solvent effects (calculated in acetonitrile) results in an energy gap between the S₁ and the S₂ states of less than 0.04 eV. This small energy difference permits vibronic coupling between these two states. Combined with the nitrogen inversion as a promoting mode, this pseudo Jahn-Teller hypothesis is found to be a second possible explanation for the emission of the red-shifted fluorescence.

The barrier for the amino twisting mode for 4BN implies the reduction of the normal fluorescence band to small intensity in the spectrum of 4BN. This is found to be in excellent agreement with experimental data [21], which shows an LE transfer band combined with a smaller charge transfer emission.

back to Contents.

Figure 3: Excited state energies and orbital analyses for PYRBN in acetonitrile for varying cyano bending angle (bottom left) and for varying amino twist angle (bottom right). Orbital compositions are presented (top, from left to right) for the RICT state at =120^o and for the S₃ (the charge transfer state becoming the RICT state upon cyano bending) at =180^o and for the S₂ charge transfer state at =0^o and for the same state (TICT state) at =90^o

For PYRBN, the excitation energies and orbital compositions are presented in Figure 3 for varying rotation angles and cyano bending angles . For this five-membered N,N-bridged aminobenzonitrile, the lowest two excited states are of special interest for the twisting relaxation mode. The S₁ state is of found to be of B₁ symmetry, consisting mainly of a benzonitrile ^* transition. For the S₂ state an electron is transferred from the amino nitrogen lone pair to the benzonitrile moiety resulting in this HOMO to LUMO charge transfer state of A symmetry. A change in the character of the relevant orbitals towards increasing charge separation can be observed for larger values of . For the twisted conformation ( =90^o), a strong localization of the HOMO orbital on the nitrogen of the amino donor group can be observed, with a delocalized of the LUMO over the benzonitrile acceptor group. The energy crossing of the charge transfer state below the ^* state is found between 30^o and 40^o twist angle. For the conformation with the twisted amino group, a strongly diminished dipole moment of 6.1 D is found for the ^* state, decreased from 14.0 D for the conformation with the untwisted amino subunit. For increasing twist angles of the amino group, the high dipole moment of the solvated TICT charge transfer state results in a strong stabilization of this excited state. The dipole moment of the n_amino^* charge transfer state increases from 12.0 D found for the untwisted conformation up to a dipole moment of 22.2 D for the perpendicular conformation (twisted amino group). For PYRBN, the higher charge transfer state relevant for the amino rotation (TICT2 state) is lowered only insubstantially, and does not cross below the lowest ^* state even in the fully twisted conformation. Due to the comparable properties (upon twisting of the alkylamino group) of this state and the TICT state, this second, energetically higher charge transfer state is referred to as TICT2 state. Induced by the stabilizing effects of the polar solvent, the S₂ excited state is lowered close to the S₁ even for the untwisted (LE for locally excited) conformation of 4BN, with an almost isoenergetic energy difference of only 0.03 eV. There is an energy barrier of 0.12 eV against the relaxation along this twisting mode. This relatively high barrier hints towards a large intensity for the normal fluorescence, with a much smaller charge transfer band than for PIPBN, due to the higher energy barrier. This predicted fluorescence behavior is in agreement with experiments [21].

The RICT state, caused by the cyano hybridization mode, is decreased below energy of the S₁ excited state of the unrelaxed vertically excited LE state: Due to the high dipole moment of this state, This excited state is lowered from the S₇ state in gas phase [22] to the S₃ by explicit inclusion of solvent effects of acetonitrile with the SCRF model. The dipole moment varies from 26.0 D for the untwisted conformation, to 14.7D for the energy minimum at =130^o from there on rising up to 19.8 D for this state with =110^o. The energy barrier against this intramolecular relaxation mode is 0.14eV, more than five times as much as the energy supplied by thermal effects at room temperature. The lowering of the dipole moment for cyano bending angles around =120^o is a strong indication against the charge transfer nature of the observed state. Even though the energy lowering of this charge transfer state is in favor of the RICT hypothesis, the high energy barrier of 0.14 eV and the low dipole moment reject the RICT hypothesis for PYRBN.

For PYRBN, the widely discussed solvent induced pseudo Jahn Teller theory is supported by the minimal energy gap of 0.03 eV between the S₁ and the S₂ states, permitting vibronic coupling between these two states. Combined with the nitrogen inversion as a promoting mode, this pseudo Jahn-Teller hypothesis is found to be a possible explanation for the emission of the red-shifted fluorescence. Even though the charge transfer state is lowered below the ^* state in the untwisted conformation (see Figure 3) for PYRBN, the energy gap of less than 0.03 eV makes this inversion of the states inconsequential, making these two states nearly isoenergetic.

back to Contents.

Figure 4: Excited state energies and orbital analyses for PIPBN in acetonitrile for varying cyano bending angle (bottom left) and for varying amino twist angle (bottom right). Orbital compositions are presented (top, from left to right) for the RICT state at =120^o and for the S₁ (the charge transfer state becoming the RICT state upon cyano bending) at =180^o and for the S₃ charge transfer state at =0^o and for the same state (TICT state) at =90^o

For the energetic minimum of the electronic ground state of PIPBN, the six-membered alkylamino substituent is found to be in a chair conformation. The excitation energies and orbital compositions of the charge transfer states considered by the TICT (alkylamino rotation) and RICT (bending of the cyano group) hypotheses are presented in Figure 4 for PIPBN. The TICT state, with a twisting angle of 90^o, is strongly stabilized below the S₁ state of the vertically excited conformation. As with 4BN, a TICT2 state is lowered energy for large twist angles crosses under the locally excited state for =90^o. The S₁ state of B₁ symmetry consists of a benzonitrile ^* transition. The S₂ state consists of a HOMO/LUMO transition, with a partial electron transfer from the amino nitrogen lone pair to the benzonitrile moiety. The relevant orbitals change towards an increasing charge separation for a twist angle of =90^o, resulting in a localized HOMO on the amino nitrogen and a LUMO delocalized over the benzonitrile acceptor group. The charge transfer state crosses below the lowest ^* state around =50^o. The dipole moment of the ^* state is lowered from 11.9D to 6.2D for =0^o to =90^o respectively. This twisting of the amino group yields an increase of the dipole moment for the n_amino^* charge transfer state (TICT state) from 13.6D (=0^o up to 22.2D (=90^o According to these calculations, there is no energy barrier against twisting of the alkylamino subunit, strongly favoring the TICT hypothesis as an explanation for the dual fluorescence of PIPBN.

Due to the high dipole moment of this state, this excited state is lowered from the S₈ state found in gas phase calculations [22] below the ^* state by explicit inclusion of acetonitrile by means of the SCRF method, making this state. As the energy differences between this N_amino-C_cyano charge transfer state and the ^* state and the n^* state are only 0.05 eV and 0.03 eV, respectively, these states are considered to be nearly isoenergetic, making this reversed order of the excited state negligible. Even though the dipole moment varies only from 27.6 D for the untwisted conformation, to 19.4 D for =130^o this state is not lowered sufficiently in energy (below the energy of the S₁ state of the untwisted conformation). For larger angles, the dipole moment rises again, to 20.7 D for =110^o accompanied by a corresponding increase of the excited state energy. As for PYRBN, due to the lowering of the dipole moment for cyano bending angles of =120^o this excited state seems to be no charge transfer state. Contrary to 4BN and PYRBN, the so-called RICT state is not lowered in energy with increasing cyano beding angles . This effect is due to a more lowered energy of the vertically excited state, caused by the lowering of another excited state, causing vibronic coupling of this state with the S₁ and the S₂ states. In some part this different behaviour of the RICT state of PIPBN can also be caused by a insufficientactive space (CISD=10), which could not be increased sufficiently with the used method. The state relevant for the relaxation along the cyano bending mode (RICT) is not lowered under the energy of the S₁ excited state of the vertically excited conformation, making a relaxation along the RICT mode impossible for PIPBN.

The calculations in acetonitrile yield the amino-cyano charge transfer state as nearly isoenergetic with the lowest ^* state and the lowest n^* state. This other state is lowered due to its large dipole moment for the vertical excitation conformation. Negligible energy gaps of only 0.05 eV between the (S₁ and the S₂) and 0.03 eV (between the S₂ and the S₃) are small enough to favor vibronic coupling between these states. As for 4BN and PYRBN, the solvent induced pseudo Jahn-Teller theory is an acceptable explanation for PIPBN.

back to Contents.

The absence of a barrier for the amino twisting mode for PIPBN implies the reduction of the normal fluorescence band to small intensity in the spectrum of PIPBN. This is found to be in excellent agreement with experimental data [21], which shows an intense charge transfer band with only a small shoulder found for the LE emission.

For a comparison with experimental data, further excited state calculations with the explicit inclusion of solvent effects were performed for diethylether as a solvent. For these calculations in diethylether, only the resulting wavelengths were calculated for the TICT (=90^o RICT (=120^o and LE (=0^o =180^o state, with no regard to the energy barrier associated with these relaxation modes. The calculated normal fluorescence bands of 354 nm, 351 nm and 357 nm for the TICT states of 4BN, PYRBN and PIPBN, respectively, in diethylether and acetonitrile are in excellent agreement with the experimental value of 350 - 360 nm in diethylether. As experimental data gave no second fluorescence for 4BN in diethylether, only the normal fluorescence out of the vertically excited state is presented for this molecule. For the second, red shifted fluorescence, the fluorescence bands are predicted in the correct wavelength region for PYRBN and PIPBN.

back to Contents.

Gas phase calculations are a valuable tool for the description of relaxation mechanisms leading to the emission of a red-shifted fluorescence band, as distinctive charge separation and a large dipole moment calculated are correlated to the solvent induced energy lowering of these excited states,. This makes gas phase calculations a valuable tool for the description of the favored relaxation mechanisms. Nevertheless, the explicit inclusion of solvent effects is necessary for the correct description of energy barriers and energetic stabilization caused by solvent effects.

The gas phase calculations presented in a previous paper [22] gave no sufficient energy lowering of a charge transfer excited state even for large wagging angles . Due to the absence of an energetically lowered WICT state for these N,N-bridged alkylaminobenzonitriles, the SCRF model was used to describe the solvent induced energy lowering of the excited states in acetonitrile. As in gas phase, no lowering of a charge transfer (CT) state below the S₁ excited state was observed even for large wagging angles. Therefore the WICT model, proposing large amino wagging angles and the amino inversion as the promoting mode of the stabilization of a charge transfer state as, is rejected for this class of compounds.

The RICT hypothesis is rejected for these molecules due to the high barrier against the cyano bending mode for 4BN and PYRBN, and due to the absence of an energy lowering below the vertically excited S₁ state for PIPBN.

Twisting of the amino group (TICT) is found to be possible for 4BN and unfavorable for PYRBN. For all molecules, another higher charge transfer (TICT2 state) is significantly lowered in energy for larger twist angles . The calculations predict different amino twist mode energy barriers, resulting in varying intensities for the CT and LE fluorescence bands, which is in accordance with the TICT hypothesis. Even though the TICT hypothesis has strong merits for this class of compounds, the large intensity of the charge transfer band for PYRBN can not be explained in a satisfying way by the TICT model alone.

Due to the small energy differences of 0.04 eV, 0.03 eV and 0.05 eV between the S1 and the S2 excited states for 4BN, PYRBN and PIPBN, respectively, vibronic coupling, as proposed by the pseudo Jahn-Teller effect is found as a good description for these N,N-bridged aminobenzonitriles.

The calculated occurrence dual fluorescence in acetonitrile for 4BN, PYRBN, and PIPBN is in accordance with experimental data. Calculated fluorescence wavelengths for PYRBN and PIPBN agree with the experiment, the calculated fluorescence wavelengths for the normal fluorescence in 4BN, PYRBN and PIPBN are in excellent agreement with experiments.

back to Contents.

We thank the Fonds zur Förderung der wissenschaftlichen Forschung, (P-11880-CHE) in Austria for generous financial support.

back to Contents.

1. E. Lippert, W. Lüder, F. Moll, W. Nägele, H. Boos, H. Prigge, I. Seibold-Blankenstein, Angew. Chem. 73, 695 (1961).
2. Z. R. Grabowski, K. Rotkiewicz, A. Siemiarczuk, D. J. Cowley, W. Baumann, Nouv. J. Chim, 3, 443 (1979).
3. E. Lippert, W. Rettig, V. Bonacic-Koutecky, F. Heisel and J. A. Miehe, Adv. Chem. Phys. 68, 1 (1987).
4. W. Schuddeboom, S. Jonker, J. Warman, U. Leinhos, W. Kühnle, K. Zachariasse, J. Phys. Chem. 96, 10809 (1992).
5. L. Serrano-Andrés, M. Merchán, B. Roos, R. Lindh, J. Am. Chem. Soc. 117, 3189 (1995).
6. K. Rotkiewicz, K. H. Grellmann, Z. R. Grabowski, Chem. Phys. Lett. 98, 315 (1973).
7. W. Rettig, V. Bonacic-Koutecky, Chem. Phys. Lett. 62, 115 (1979).
8. W. Rettig, Angew. Chem. 98, 969 (1986); Angew. Chem. Int. Ed. Engl. 25, 971 (1986); W. Rettig, Volume 169 of Topics Curr. Chem., Springer Verlag, Berlin, 1994.
9. A. Gorse, M. Pesquer, J. Mol. Struct. (THEOCHEM), 281, 21 (1993).
10. A. Gorse, M. Pesquer, J. Phys. Chem, 99, 4039 (1995).
11. K. A. Zachariasse, Th. Von der Haar, U. Leinhos, W. Kühnle, J. Inf. Rec. Mat. 21, 501 (1994).
12. U. Leinhos, W. Kühnle, K. A. Zachariasse, J. Phys. Chem., 95, 2013 (1991).
13. K. A. Zachariasse, T. van der Haar, A. Hebecker, U. Leinhos, W. Kühnle, Pure Appl. Chem. 65, 1745 (1993).
14. K. A. Zachariasse, M. Grobys, T. van der Haar, A. Hebecker, Yu. V. Il\222ichev, O. Morawski, I. Rückert, W. Kühnle, J. Photochem. Photobiol. A: Chem. 105, 373 (1997).
15. A. L. Sobolewski, W. Domcke, Chem. Phys. Lett. 250, 428 (1996).
16. A. L. Sobolewski, W. Domcke, Chem. Phys. Lett. 259, 119 (1996).
17. A. L. Sobolewski, W. Sudholt, W. Domcke, J. Phys. Chem. in press.
18. W. Rettig, J. Phys. Chem. 86, 1970 (1982).
19. W. Rettig, Ber. Bunsenges. Phys. Chem. 95, 259 (1991).
20. D. Braun, W. Rettig, Chem. Phys. Lett 268, 110 (1997).
21. K. A. Zachariasse, M. Grobys, Th. Von der Haar, A. Hebecker, Y. V. Il'ichev, O. Morawski, I. Rückert, W. Kühnle, J. Photochem. And Photobiol. A: Chemistry 102, 59 (1996).
22. R. Schamschule, A. B. J. Parusel, G. Köhler, Internet J. Chem., 1998, 1, 5.
23. G. Rauhut, T. Clark, T. Steinke, J. Am. Chem. Soc. 115, 9174 (1993).
24. M. Dewar, E. Zobisch, E. Healy, J. Stewart, J. Am. Chem. Soc. 107, 3902 (1985).
25. G. Rauhut, A. Alex, J. Chandrasekhar, T. Steinke, W. Sauer, B. Beck, M. Hutter, P. Gedeck, T. Clark, VAMP6.1, Oxford Molecular Ltd., Oxford, 1997.

back to Contents.