2-Azidoimidazolium Ions Captured by N-Heterocyclic Carbenes : Azole-Substituted Triazatrimethine Cyanines

1,3-Disubstituted 2-azidoimidazolium salts (substituents = methyl, methoxy; anion = PF6) reacted with N-heterocyclic carbenes to yield yellow 2-(1-(azolinylidene)triazen-3-yl)-1,3-R2imidazolium salts (azole = 1,3-dimethylimidazole, 1,3-dimethoxyimidazole, 4-dimethylamino-1methyl-1,2,4-triazole; R = methyl, methoxy; anion = PF6). Crystal structures of three cationic triazenes were determined. Numerous interionic C–H ̈  ̈  ̈F contacts were observed. Solvatochromism of the triazenes in polar solvents was investigated by UV-Vis spectroscopy, involving the dipolarity π* and hydrogen-bond donor acidity α of the solvent. Cyclovoltammetry showed irreversible reduction of the cations to uncharged radicals. Thermoanalysis showed exothermal decomposition.


Results and Discussion
Since the 2-azido-1,3-dimethylimidazolium ion is not yet described, we attempted to prepare it by azidation of the respective carbene using tosyl azide (Figure 1).Surprisingly, it was found that the resulting tosyltriazene reacted with the intermediate carbene, yielding the disubstituted triazene 1.The primary product mixture consisted of lithium toluenesulfinate and triazene and was easily separated.The desired 2-azidoimidazolium salt could not be isolated but will be attempted in further work.However, this result inspired us to prepare other triazenes, or triazatrimethine cyanines, from stable azidoazolium salts and N-heterocyclic carbenes.Thus, several methods to generate the carbenes were tested, and sodium hydride in acetonitrile was discovered to give agreeable results and allowed the synthesis of another symmetrically substituted triazene 2 and the unsymmetrically substituted triazene 3.
The hexafluoridophosphate anions were chosen because they readily yield non-hygroscopic, crystalline solids.The role of fluorine, which exhibits the lowest polarizability and highest electronegativity of the halogens in crystal structures of organic compounds has been reviewed [18].
After some controversy, it has been concluded that short C-H ... F contacts between oppositely charged molecules are genuine interionic hydrogen bonds [19].
The crystal structures of the new cationic triazatrimethine cyanines 1-3 were determined and are discussed below.

Crystal Structures
The C10H16N7 cation in triazene 1 contains 41.9% N and may be termed "nitrogen-rich".A crystal of 1 was refined as a two-component twin.Positional disorder of the fluorine atoms of PF6 was observed, and a ratio of 3:2 was obtained by refinement as a free variable.In addition, the small crystal was non-merohedrally twinned by 180 degrees about the reciprocal axis 0 0 1.
The imidazole-substituted triazene system in 1 is heavily distorted.The angle between the planes of the two imidazole rings is 59.3°, the N5-N6 bond is rotated out of the plane of the first Surprisingly, it was found that the resulting tosyltriazene reacted with the intermediate carbene, yielding the disubstituted triazene 1.The primary product mixture consisted of lithium toluenesulfinate and triazene and was easily separated.The desired 2-azidoimidazolium salt could not be isolated but will be attempted in further work.However, this result inspired us to prepare other triazenes, or triazatrimethine cyanines, from stable azidoazolium salts and N-heterocyclic carbenes.Thus, several methods to generate the carbenes were tested, and sodium hydride in acetonitrile was discovered to give agreeable results and allowed the synthesis of another symmetrically substituted triazene 2 and the unsymmetrically substituted triazene 3.
The hexafluoridophosphate anions were chosen because they readily yield non-hygroscopic, crystalline solids.The role of fluorine, which exhibits the lowest polarizability and highest electronegativity of the halogens in crystal structures of organic compounds has been reviewed [18].After some controversy, it has been concluded that short C-H¨¨¨F contacts between oppositely charged molecules are genuine interionic hydrogen bonds [19].
The crystal structures of the new cationic triazatrimethine cyanines 1-3 were determined and are discussed below.

Crystal Structures
The C 10 H 16 N 7 cation in triazene 1 contains 41.9% N and may be termed "nitrogen-rich".A crystal of 1 was refined as a two-component twin.Positional disorder of the fluorine atoms of PF 6 was observed, and a ratio of 3:2 was obtained by refinement as a free variable.In addition, the small crystal was non-merohedrally twinned by 180 degrees about the reciprocal axis 0 0 1.The imidazole-substituted triazene system in 1 is heavily distorted.The angle between the planes of the two imidazole rings is 59.3 ˝, the N5-N6 bond is rotated out of the plane of the first imidazole ring by 25.9 ˝, and the N7-N6 bond is rotated out of the plane of the second imidazole ring by 32.4 (Figure 2a).Although the mesomeric cyanine molecule is symmetrically disubstituted, the bond lengths in the triazene bridge do not reflect this symmetry, presumably due to packing effects.
The heteroaromatic C2-H, C7-H and C8-H donate hydrogen bonds to F4, F3 and F1, whereas H atoms of one methyl group form contacts to F1 and F6 (Figure 2b).imidazole ring by 25.9°, and the N7-N6 bond is rotated out of the plane of the second imidazole ring by 32.4° (Figure 2a).Although the mesomeric cyanine molecule is symmetrically disubstituted, the bond lengths in the triazene bridge do not reflect this symmetry, presumably due to packing effects.The heteroaromatic C2-H, C7-H and C8-H donate hydrogen bonds to F4, F3 and F1, whereas H atoms of one methyl group form contacts to F1 and F6 (Figure 2b).Crystal data and structure refinement details are summarized in Table 1.Short contacts and interactions are listed in Table 2. Selected bond lengths, angles and torsions of the triazene bridge are compiled in Table 3.
The imidazole-substituted triazene system in 2 is far less distorted than in 1.Thus, the angle between the planes of the two imidazole rings is 14.8°, the N5-N6 bond is twisted out of the plane of the adjacent imidazole ring by 7.6°, and the N7-N6 bond by 16.3°.The methoxy groups of both imidazole rings adopt syn conformations, although in opposite directions (Figure 3a).
The PF6-groups in 2 and 3 were orientationally disordered among two positions.SADI commands were employed to restrain the corresponding P-F and F-F distances.For the refinement of the anisotropic displacement parameters, SIMU commands were used for atoms closer than 1.0 Å. Crystal data and structure refinement details are summarized in Table 1.Short contacts and interactions are listed in Table 2. Selected bond lengths, angles and torsions of the triazene bridge are compiled in Table 3.
The imidazole-substituted triazene system in 2 is far less distorted than in 1.Thus, the angle between the planes of the two imidazole rings is 14.8 ˝, the N5-N6 bond is twisted out of the plane of the adjacent imidazole ring by 7.6 ˝, and the N7-N6 bond by 16.3 ˝.The methoxy groups of both imidazole rings adopt syn conformations, although in opposite directions (Figure 3a).
The PF 6 -groups in 2 and 3 were orientationally disordered among two positions.SADI commands were employed to restrain the corresponding P-F and F-F distances.For the refinement of the anisotropic displacement parameters, SIMU commands were used for atoms closer than 1.0 Å.The unsymmetrically disubstituted triazene system in 3 is the least distorted of the three compounds.The angle between the planes of the two heterocyclic rings is only 4.8 ˝, the N5-N6 bond is rotated out of the imidazole plane by 10.5 ˝, and the N7-N6 bond out of the triazole plane by 13.5 ˝(Figure 4a).Again, the methoxy groups adopt syn conformation.In contrast to 1, although the cyanine molecule is unsymmetrically disubstituted, the corresponding bond lengths in the triazene bridge are almost equal.A comparison of the bond lengths in the triazenes with accepted values [20] reveals the delocalized nature of the bonds in these systems.The N¨¨¨N distances here are longer than trans N=N bonds (1.22 A), but shorter than typical planar N-N bonds (1.40 A).In addition, the C¨¨¨N distances are longer than C=N bonds (1.28 A) but shorter than isolated C-N= bonds (   In the structure of 3, the heteroaromatic C7-H of the triazole donates a hydrogen bond to F5, whereas the C3-H of the imidazole forms bifurcated hydrogen bonds to F3 and F6.Aliphatic hydrogen atoms form contacts to F1, F3 and F5, as well as to O1 and N7 (Figure 4b).In the structure of 3, the heteroaromatic C7-H of the triazole donates a hydrogen bond to F5, whereas the C3-H of the imidazole forms bifurcated hydrogen bonds to F3 and F6.Aliphatic hydrogen atoms form contacts to F1, F3 and F5, as well as to O1 and N7 (Figure 4b).In the structure of 3, the heteroaromatic C7-H of the triazole donates a hydrogen bond to F5, whereas the C3-H of the imidazole forms bifurcated hydrogen bonds to F3 and F6.Aliphatic hydrogen atoms form contacts to F1, F3 and F5, as well as to O1 and N7 (Figure 4b).

UV-Vis Spectroscopy
Typically, donor-acceptor-substituted conjugated systems exhibit solvatochromism.The solvatochromism of such systems arises from the different stabilization of their electronic ground and excited state by differential solvation of these two states, according to their different molecular and electronic structure.Transition energies ET were calculated from the wavelengths of the absorption maxima (Table 4) according to the equation ET/kJ•mol −1 = hcN/λ = 119625/(λ/nm).A linear solvation energy relationship was established by least-squares fitting the data to the solvatochromic equation involving the parameters π* [21] and α [22] which represent the dipolarity/polarizability and hydrogen-bond donor capabilities of the solvent.The hydrogen-bond acceptor basicity β [23] did not significantly contribute to the relationship.The following equations were derived by multiple linear regression (Figure 5), which adequately describe the solvatochromic behavior of the triazenes.

UV-Vis Spectroscopy
Typically, donor-acceptor-substituted conjugated systems exhibit solvatochromism.The solvatochromism of such systems arises from the different stabilization of their electronic ground and excited state by differential solvation of these two states, according to their different molecular and electronic structure.Transition energies E T were calculated from the wavelengths of the absorption maxima (Table 4) according to the equation E T /kJ¨mol ´1 = hcN/λ = 119625/(λ/nm).A linear solvation energy relationship was established by least-squares fitting the data to the solvatochromic equation involving the parameters π* [21] and α [22] which represent the dipolarity/polarizability and hydrogen-bond donor capabilities of the solvent.The hydrogen-bond acceptor basicity β [23] did not significantly contribute to the relationship.The following equations were derived by multiple linear regression (Figure 5), which adequately describe the solvatochromic behavior of the triazenes.
For 1 : E T {kJ¨mol -1 " 291.7 `3.98 π* `8.61 α pn " 7, R 2 " 0.992q For 2 : E T {kJ¨mol -1 " 313.2 ´9.21 π* ´5.08 α pn " 6, R 2 " 0.960q For 3 : E T {kJ¨mol -1 " 317.96 `3.32 π* `8.15 α pn " 6, R 2 " 0.960q The triazenes 1 and 3 exhibited negative solvatochromism (shift of the absorption maximum to shorter wavelengths on changing the solvent from dichloromethane to water) whereas, amazingly, triazene 2 displayed positive solvatochromism.No correlation was found with less polar solvents.The values in DMSO for 2 and 3 were not included in the calculation because a change of band shape rendered them not comparable.The π* and α values were taken from the compilation of Marcus [24] The values in DMSO for 2 and 3 were not included in the calculation because a change of band shape rendered them not comparable.The π* and α values were taken from the compilation of Marcus [24].

Cyclic Voltammetry (CV)
The three electrophoric systems have in common that no reverse peaks were observed.Presumably, the first electrochemical process involved an irreversible reduction to an uncharged radical species.The resulting radical evidently was unstable on the electrochemical time scale and underwent subsequent decomposition.The cathodic peak current (Ip)c was approximately proportional to the square root of the scan rate, and the cathodic peak potential (Ep)c shifted to negative values with an increasing rate (Figure 7).The peak potential and peak current values found in the voltammetric measurements are summarized in Table 5.

Cyclic Voltammetry (CV)
The three electrophoric systems have in common that no reverse peaks were observed.Presumably, the first electrochemical process involved an irreversible reduction to an uncharged radical species.The resulting radical evidently was unstable on the electrochemical time scale and underwent subsequent decomposition.The cathodic peak current (I p ) c was approximately proportional to the square root of the scan rate, and the cathodic peak potential (E p ) c shifted to negative values with an increasing rate (Figure 7).The peak potential and peak current values found in the voltammetric measurements are summarized in Table 5. Thermoanalysis showed exothermal decomposition (maxima at 282 for 1, 180 for 2 and 186 °C for 3) with considerable mass loss (Figure 6).

Cyclic Voltammetry (CV)
The three electrophoric systems have in common that no reverse peaks were observed.Presumably, the first electrochemical process involved an irreversible reduction to an uncharged radical species.The resulting radical evidently was unstable on the electrochemical time scale and underwent subsequent decomposition.The cathodic peak current (Ip)c was approximately proportional to the square root of the scan rate, and the cathodic peak potential (Ep)c shifted to negative values with an increasing rate (Figure 7).The peak potential and peak current values found in the voltammetric measurements are summarized in Table 5.
Cyclic voltammetric (CV) experiments were performed in CH3CN at room temperature under Ar atmosphere using an EG&G 264 A polarographic analyser/stripping voltammeter (Princeton Applied Research, Oak Ridge, TN, USA).All measurements were carried out employing a three-electrode configuration consisting of a Pt electrode (disk, 2 mm diameter) as a working electrode and an (Ag/0.01M AgNO3 in CH3CN) electrode as the reference (EG&G Micro-Cell).A Pt wire in an electrode bridge tube filled with supporting electrolyte and separated from the sample by a porous frit served as the counter electrode.A solution of 0.1 M Bu4PF6 in CH3CN served as the supporting electrolyte.Sample solutions were prepared by dissolving triazene (2.5 mg) in electrolyte (25 mL).A solution of ferrocene in electrolyte was used for calibration.All formal potentials are referenced to the ferrocene/ferrocenium redox couple.
Single crystal diffraction intensity data were recorded by φ and ω scans with a Bruker D8 Quest Photon 100 (Billerica, MA, USA) (compound 1), or by ω scans with an Oxford Diffraction Gemini-R Ultra (Oxford Diffraction Ltd., Abingdon, Oxfordshire, UK) (compounds 2 and 3) diffractometer using MoKα radiation.The structures were solved by direct methods and refined by full-matrix
Cyclic voltammetric (CV) experiments were performed in CH 3 CN at room temperature under Ar atmosphere using an EG&G 264 A polarographic analyser/stripping voltammeter (Princeton Applied Research, Oak Ridge, TN, USA).All measurements were carried out employing a three-electrode configuration consisting of a Pt electrode (disk, 2 mm diameter) as a working electrode and an (Ag/0.01M AgNO 3 in CH 3 CN) electrode as the reference (EG&G Micro-Cell).A Pt wire in an electrode bridge tube filled with supporting electrolyte and separated from the sample by a porous frit served as the counter electrode.A solution of 0.1 M Bu 4 PF 6 in CH 3 CN served as the supporting electrolyte.Sample solutions were prepared by dissolving triazene (2.5 mg) in electrolyte (25 mL).A solution of ferrocene in electrolyte was used for calibration.All formal potentials are referenced to the ferrocene/ferrocenium redox couple.
Single crystal diffraction intensity data were recorded by ϕ and ω scans with a Bruker D8 Quest Photon 100 (Billerica, MA, USA) (compound 1), or by ω scans with an Oxford Diffraction Gemini-R Ultra (Oxford Diffraction Ltd., Abingdon, Oxfordshire, UK) (compounds 2 and 3) diffractometer using MoKα radiation.The structures were solved by direct methods and refined by full-matrix least-squares techniques on F 2 .Programs CELL_Now and TWINABS (Bruker) were used for cell search of twin components and absorption correction.CCDC 1428586-1428588 contain the supplementary crystallographic data for this paper.These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html.A solution of lithium bis(trimethylsilyl)amide in THF (5.4 mL 1 M, 1.3 equiv) was added to a suspension of 1,3-dimethylimidazolium hexafluoridophosphate (1.0 g, 4.1 mmol) in THF (10 mL) at 0 ˝C.The resulting solution was stirred for 30 min at 0 ˝C, then tosyl azide (0.41 g, 0.5 equiv) in THF (1 mL) was added.The mixture was stirred for 4 h at room temperature, then cooled at 0 ˝C again.The orange precipitate was collected by filtration and partitioned between CH 2 Cl 2 (15 mL) and H 2 O (15 mL).The aqueous phase was extracted once more with CH 2 Cl 2 (10 mL), and the combined organic phases were dried over MgSO 4 and taken to dryness under reduced pressure to yield the triazene 1 as an orange powder (0.25 g, 35%).Single crystals were obtained from H 2 O/acetone.M.p. 201 ˝C (decomposition). 1 H NMR (DMSO-d 6 , 300 MHz, δ): 3.68 (s, 12H), 7.39 (s, 4H) ppm. 13

Conclusions
In the growing field of N-heterocyclic carbene (NHC) catalysis, one of the most recent and exciting concepts is the thermally triggered release of the catalytically active species [27].The new azole-substituted triazenes could become of interest as metal-free, labile NHC progenitors capable of switching on latent reactions.More research is needed with regard to solvent and temperature applied in order to liberate an active NHC.

Figure 2 .
Figure 2. (a) molecular structure of the cation; (b) interactions in the crystal structure of 1.

Figure 2 .
Figure 2. (a) Molecular structure of the cation; (b) interactions in the crystal structure of 1.

Figure 3 .
Figure 3. (a) molecular structure of the cation; (b) interactions in the crystal structure of 2.

Figure 3 .
Figure 3. (a) Molecular structure of the cation; (b) interactions in the crystal structure of 2.

Figure 3 .
Figure 3. (a) molecular structure of the cation; (b) interactions in the crystal structure of 2.

Figure 4 .
Figure 4. (a) molecular structure of the cation; (b) interactions in the crystal structure of 3.

Figure 4 .
Figure 4. (a) Molecular structure of the cation; (b) interactions in the crystal structure of 3.

Table 1 .
Crystal data and structure refinement for 1-3.

Table 4 .
Solvent parameters and absorption maxima of the triazenes 1-3 in polar solvents. .

Table 5 .
Cathodic peak potential (E p ) c and cathodic peak current (I p ) c of the triazenes 1-3 in CH 3 CN/0.1 M Bu 4 PF 6 as function of scan rate (potential values vs. ferrocene couple).