Theoretical Investigations of the Ground State Properties of Stereoselectively Substituted Aminobenzonitriles

Institut für Theoretische Chemie und Strahlenchemie, University of Vienna, Vienna, Althanstr. 14, A-1090 Vienna, Austria

Keywords:

Ab-initio, Aminobenzonitriles, Density functional theory, DMABN, Donor-acceptor systems, Dual fluorescence, Twisted intramolecular charge transfer.

<! ############################################################################## >

Contents Abstract Introduction Computational Methods Discussion and Conclusion Acknowledgments References		DMABN

Abstract

4-( N,N-Dimethylamino)-benzonitrile (DMABN) is the classical model compound showing solvent induced dual fluorescence, a phenomenon attributed to formation of an intramolecular charge transfer state and concomitant rotational relaxation of the dimetylamino group about the amine - aryl single bond. This mechanism is supported by the fluorescence properties of various stereo-selectively substituted derivatives of DMABN and, therefore, ground state properties of such a set of model compounds are studied as a basis for a detailed examinationof the influence of substitution on excited state relaxation pathways. The coordinates examined in this study are torsion about the single bond connecting the donor amino group to the aryl moiety, the amine wagging vibration and bending of the cyano group(rehybridization). Ground state geometries and electrostatic properties as well as the energetics of these modes are obtained from semiempirical AM1 calculations and are compared to HF ab initio results using 3-21G* and 6-21G** split valence basis sets and to those obtained by the Becke3 density functional method with the Lee Young Parr correlation function and the D95** basis set.

back to Contents.

Introduction

Solvent induced dual fluorescence, i.e. the appearance of a second fluorescence emission band when hydrocarbons are replaces by polar solvents, is found for many organic donor- acceptor compounds [1,2]. Since the discovery of this phenomenon for 4-(N,N-Dimethylamino)-benzonitrile (DMABN) and derivatives by Lippert et al. [3], it has been well documented [4], that a variety of compounds composed of a donor and an acceptor group connected by a single bond show, besides the so-called normal fluorescence out of the vertically excited S1 state, a second, strongly red shifted emission in a polar environment. This emission originates from a highly dipolar excited state as its spectral position depends strongly on the polarity of the solvent. Furthermore, it was found that stereoselective substitution affecting the torsional motion of the amino group alters also the appearance of the second fluorescence band in a very specific manner [1,3,5-8]. These observations support the so-called TICT (twisted intramolecular charge transfer) model [1,2,6] as explanation for these effects, which predicts, that the planar vertically excited state relaxes to a charge transfer state with perpendicular orientation of donor and acceptor moieties (see fig. 1a).

Fig. 1: Definition of the rotational mode (1a.), the wagging mode (1b.), and the bending mode (by rehybridization) (1c.).

In this mutually perpendicular conformation of donor and acceptor subunits, one electron could be transferred from the nitrogen lone pair into the unoccupied aromatic benzene orbital and charge separation could thus be maximized. Although there is good experimental evidence supporting this model, it is still debated and one alternative, which was proposed by Zachariasse et al. [2] suggests two requirements for the occurrence of dual fluorescence: A sufficiently small energy gap between two interacting states and the presence of a promoting mode, such as the amino nitrogen inversion mode, coupling both states. The final highly dipolar state could then result from a solvent induced pseudo Jahn-Teller effect. The amino wagging mode is the most important motion, we refer to this model as WICT (W stands for wagging) hypothesis (see fig. 1b). As an additional alternative, Domcke and Sobolewski proposed in-plane bending of the cyano group as main relaxation coordinate (see fig. 1c) andsp to sp2 rehybridization (RICT) on the carbon atom of the cyano group should stabilize the charge transfer state [7,8].

The most elegant feature of the TICT concept is, however, the characterization of the main reaction coordinate, in that case rotation of the amino group, by stereoselective substitution of the parent molecule DMABN. Long wavelength fluorescence is absent when the amino group is fixed coplanar to the aromatic plane (in some cases, however, efficient fluorescence quenching is observed in highly polar solvents [9]) but it is exclusively observed when the amino group is hold in a perpendicular orientation. For the characterization of the molecular dynamics when crossing to the surface of the charge transfer state, it is thus most important to examine substitution effects on all proposed reaction coordinates. Coupling of molecular motions and to solvent relaxational modes should facilitate charge transfer. Solvent effects can in a first approximation be described by the variation of the molecular dipole moment along the reaction pathway. For the description of the excited state, a clear picture of ground state properties is a necessary prerequisite. In this paper we report on a comparison of calculated ground state properties for a set of model compounds obtained from different computational methods. The various model compounds are displayed in scheme 1 and the full names are given in table 1.

Scheme1: Molecular structures of stereoselectively substituted aminobezonitriles.

Table 1: Abbreviations for the model compounds discussed in this paper

The optimized ground state structures are compared for all model compounds with special regard to the rotation and inversion motion of the amino group and the in-plane bending of the cyano group. Force constants were calculated and a vibrational analysis was performed. The molecular charge distribution was analyzed by Natural Population Analysis (NPA) [10,11] and the dipole moments were calculated.

back to Contents.

Computational Methods:

All calculations were performed within the GAUSSIAN94 [12] program package. Equilibrium geometries of the model compounds were determined and were confirmed by a subsequent calculation of force constants and a vibrational analysis. Calculations were either performed using the semiempirical AM1 hamiltonian [13], or the closed-shell Hartree-Fock (HF) method, as well as Becke\222s three parameter density functional [14] in combination with the Lee, Yang and Parr correlation functional [15] (Becke3LYP). The standard integration grid of 75 radial shells with 302 angular points per shell, resulting in about 7000 points per atom, was used for these calculations. Different basis sets are compared in view of calculation times and accuracy of the results in comparison to experimental data. The 3-21+G* and the 6-31G* split valence basis sets are used for calculations at the HF level, and compared to results obtained by Sobolewski and Domcke [7,8] using the 3-21G* basis set. The Becke3LYP density functional calculations were performed with the 6-31G* and the D95** basis sets.

HF calculations with the 3-21G* basis set yield a generally planar ground state equilibrium structure for unsubstituted DMABN [8], which is in contrast to the experiment. The 3-21+G* basis set, which is additionally augmented with diffuse functions, was examined as a relatively inexpensive possibility to explore the equilibrium structures of the ground states. Larger basis sets as the 6-31G* and the D95** basis sets were included in these calculations to find an optimal set which yields reliable equilibrium geometries, while not too expensive in view of calculation time, and to test the applicability of split valence basis set for thi problem.

Definitions for the wagging angle and the twisting angle are given in Fig.2.

Fig. 2: Definition of wagging angle and twist angle for DMABN. An analogous definition is for accepted the other compounds.
back to Contents.

Results and Discussion:

Rotation and inversion of the amino group are compared for the fully optimized geometries of the ground states of all model compounds. The wagging angle w, defining the pyramidalization of the amino group, and the twist angle a, defining its out of plane rotation, are presented for DMABN and TMABN in Table 2 for various methods and basis sets.

Table 2: Wagging angle and twist angle for DMABN and TMABN

¹⁾ Lit. [16-18]

The ground state geometry of DMABN (see Table2) calculated on the HF/3-21+G* level results planar, i.e. as well as are 0 , in accordance with the ground state equilibrium geometry by Sobolewski and Domcke using the 3-21G basis set (without polarization and diffuse functions) [7]. This ground state geometry is of C2v symmetry, irrespective whether the 3-21G basis set is augmented with polarization or diffuse functions. Compared to the experimental values for the pyramidalization angle of 11.9 [16] and 15.0 [17], the 3-21G is thus inadequate for the description of the delocalization of the electron lone pair in DMABN. Both 6-31G* (8.9 ) as well as D95** (12.8 ) basis sets give wagging angles w comparable to experimental results, and those obtained with the D95** set are slightly better than those with the 6-31G* set. The Becke3LYP/D95** approach gives a far too small angle w for DMABN. The AM1 result of 16.1 for the DMABN wagging angle is in moderately good agreement with the experimental data.

As the ground state of the DMABN molecule is planar, = 0.0 , additional test calculations for various basis sets are presented for TMABN (see Table2). The AM1 approach gives = 68.4 , and this is again in very good agreement with the experimental value of ~70 . The twist angle calculated with the HF /6-31G* (77.6 ) and the Becke3LYP/D95** (66.4 ) approaches are of comparable good quality. Again the 3-21+G* basis set fails utterly in the description of delocalization and yields a twist angle = 90 for TMABN. Although the overlap of the amino lone pair orbital with the orbitals of the phenyl ring is severely hindered in TMABN, conjugation seems nevertheless to be of great importance for this pretwisted system, causing this large deviation from the sterically favored twist angle of 90.0 . The resulting geometry for TMABN is a compromise between sterical hindrance, pushing the amino group towards = 90 twisted conformation, and optimizing electronic conjugation effects of smaller rotational angles.

As the 3-21G and the 3-21+G* basis sets are inadequate for the description of the ground states of these model compounds, no further calculations were carried out with these basis sets.

Table 3: Wagging angle and twist angle for CBQ, CTPHI, JULCN, and MIN

For the other model compounds CBQ, CTPHI, JULCN and MIN, the results (see Table3) from HF / 6-31G* and Becke3lyp / D95** approaches are compared to these from semiempirical AM1 calculations. The 6-31G* basis set was chosen for the HF calculations, because the results are nearly as accurate as those obtained with the larger D95** basis set, with a smaller demand on computation time. In the DFT calculations using the Becke3LYP method, considering also electron correlation, only the more accurate D95** basis set is applied. The AM1 results are also presented as a relatively cheap reference method, making potential energy surface scans feasible also in the excited states. No experimental data for the relevant geometrical parameters are available for CBQ, CTPHI, JULCN and MIN.

In CBQ, the amino lone pair is fixed within the plane of the phenyl ring, rotation is severely hindered. Due to these strong steric effects by the two bridges, pyramidalization of the amino group is strongly increased. The calculated twist angle results identical 90.0 for all methods (AM1, HF/6-31G* and Becke3LYP/D95**), the wagging angle is given between 31 and 33 .In this case the results of all these methods are in excellent agreement. For compounds in which the amino group is kept nearly coplanar to the phenyl ring, differences between the three methods are generally much smaller than for the parent compound DMABN.

Donor and acceptor orbitals are orthogonal in the minimum geometry of CBQ. In contrast to CBQ, electron delocalization becomes important, when the amino group and the aryl moiety are kept in a nearly coplanar arrangement. For these compounds, CTPHI, JULCN and MIN, each approach yields similar results, with only one exception. This torsion of the amino group in the indoline derivative MIN results much larger from semiempirical calculation than from that on ab initio HF level. This stems from a different ground state geometry as in this case the five membered ring lies within the aromatic plane, the methyl group off this plane. The symmetrically bridged compounds (CTPHI and JULCN) result symmetric versus a plane normal to the benzene ring. Pyramidalization of the amino group is similar for all three compounds and to TMABN, but slightly smaller for the six membered bridges in comparison to five membered rings. JULCN has a larger flexibility and rotation of the amino group is also easier.

The most important characteristic of an intramolecular charge transfer state is the change in the magnitude and orientation of the molecular dipole moment upon excitation as it reflects the change in the charge distribution and contributes essentially to the stabilization of this state in polar solvents. Ground state dipole moments for these model compounds are compiled in table 4.

Table 4: Comparison of the Dipole Moments in Debye for all model compounds

Due to the large torsion angle of the amino group in TMABN delocalization of the lone pair is considerably smaller than in DMABN, and this results in a reduced dipole moment. The trend is reproduced by the HF as well as the DFT method. The Becke3LYP/D95** method yields, however, generally larger values for the dipole moment than the HF/6-31G* approach. The length of the dipole decreases generally as the angle of rotation a increases and this trend is independent on the method. The direction of the dipole moment is generally from the aromatic ring towards the amino nitrogen but slightly inclined to the molecular axes, due to the pyramidalization of the amino group. This is depicted for DMABN in fig. 3. Planarization of the amino group decreases likewise the permanent dipole moment.

Fig. 3: Direction of the ground state dipole moment of DMABN.

A vibrational analysis was performed for the equilibrium geometries and the vibrational frequencies of the three considered modes are assumed to be a measure for the hindrance of the respective deformation. The results are compiled in table 5.

Table 5: Energies of the vibrational modes corresponding to the twisting motion (rotational mode) and the inversion motion (wagging mode) in cm-1.

For DMABN rotation of the amino group has the lowest quantum energy within the vibrational spectrum but the value for the umbrella inversion motion is only slightly larger. Distortion from the symmetric, pyramidal optimum conformation is thus rather feasible and both modes, rotation and wagging, is excited at ambient temperatures in solvents. The energy of the rotational mode increases gradually with increasing steric hindrance and becomes maximal for CBQ, as it is more than 10 times larger than for DMABN. In the planar five membered ring system (CTPHI) the hindrance for rotation is considerably larger than for the larger ring system (JULCN), as the energy of the mode is nearly twice the energy found for JULCN. This shows clearly the higher flexibility of the six over the five membered ring system, a property also shown for the inversion motion. For all three compounds, however, wagging is energetically preferred in comparison to rotational motion.

The results show, that a description on the 6-31G* level is necessary to yield appropriate geometries for the ground states of these systems and at least equally good basis sets should be necessary for a description of the excited state. However, the Becke3LYP DFT approach using the D95** basis set gives nearly equally good results, with the only exception of the pyramidalization of DMABN which results far too small. The correct description of DMABN is thus the key for the understanding of these donor-acceptor molecules.

back to Contents.

Acknowledgments:

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

back to Contents.

References:

• [1] Z. R. Grabowski, K. Rotkiewicz, A. Siemiarczuk, D. J. Cowley, W. Baumann, Nouv. J.Chim, 3, 443 (1979).
• [2] W. Rettig, Angew. Chem. 98, 969 (1986).
• [3] E. Lippert, W. Lüder, F. Moll, W. Nägele, H. Boos, H. Prigge, I. Seibold-Blankenstein, Angew. Chem. 73, 695 (1961).
• [4] L. Serrano-Andres, M. Merchan, B. Roos, R. Lindh, J. Am. Chem. Soc. 117, 3189 (1995).
• [5] K. Rotkiewicz, K. H. Grellmann, Z. R. Grabowski, Chem. Phys. Lett. 19, 315 (1973).
• [6] K. A. Zachariasse, Th. Von der Haar, U. Leinhos, W. Kühnle, J. Inf. Rec. Mat. 21, 501
• [7] A. L. Sobolewski, W. Domcke, Chem. Phys. Lett. 259, 119 (1996).
• [8] A. Sobolewski, W. Domcke, Chem. Phys. Lett. 250, 428 (1996).
• [9] G. Köhler, G. Grabner, K. Rotkiewiez, Chem. Phys. 173, 275 (1993).
• [10] NBO, Version 3.1, E. D. Glendening, A. D. Reed, J. E. Carpenter, F. Weinhold, (included in the Gaussian94 package).
• [11] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 88, 899 (1988).
• [12] Gaussian 94, Revision B.3, M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B.G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, C. Y. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon, C. Gonzalez, and J. A. Pople, Gaussian Inc., Pittsburgh PA, 1995.
• [13] M. Dewar et al, J. Am. Chem. Soc. 107, 3902 (1985).
• [14] A. D. Becke, J. Chem. Phys. 98, 5648 (1993).
• [15] B. Miehlich, A. Savin, H. Stoll, H. Preuss, Chem. Phys. Lett. 157, 200 (1989).
• [16] A. Heine, R. Herbst-Irmer, D. Stalke, W. Kühnle, K. A. Zachariasse, Acta Cryst. B50,363 (1994).
• [17] O. Koyimoto, H. Yokoyama, Y. Ooshima, Y. Endo, Chem. Phys. Lett. 179, 455.
• [18] W. Rettig, K. Rotkiewicz, J. Phys. Chem. 97, 13500 (1993).

back to Contents.