The 2-carbazolylnitrenium ion and its conjugate acid and base forms

The 2-carbazolylnitrenium ion and its 9-methyl analog have been investigated by laser flash photolysis. These electrophiles were generated by irradiation of 2-azido-9H -carbazole and 9- methyl-2-azido-9H -carbazole in solutions of 20% acetonitrile in water. Kinetic and spectral evidence is presented for three forms of the 2-carbazolylnitrenium ion, the monocation, its dication conjugate acid and a neutral quinonoid-like conjugate base obtained by loss of the 9-NH proton. The 9-methyl system, which cannot form the neutral species, exists as dication and monocation. Acidity constants relating the various forms and rate constants for their reaction with water, bromide ion and azide ion are obtained by kinetic analyses.


Introduction
2-Aminofluorene 1 and 2-amino-α-carboline 2 are representative members of a diverse group of arylamine carcinogens found, for example, in tobacco smoke, automobile exhaust, broiled meats and fish, and as side products of industrial processing. 1 The fluorene derivative 1 is a carbocyclic example that has seen extensive investigation since its carcinogenic properties were discovered over 50 years ago. 2 The carboline 2 is one of the more common carcinogenic and mutagenic heterocyclic amines derived from pyrolysis of proteins and amino acids found in cooked food. 3 feature common to carbocyclic and heterocyclic amines is a requirement for metabolic activation -oxidation of the arylamine 3 to an arylhydroxylamine 4, followed by conversion to an acetate or sulfate ester 5. 4 While there has been debate over the years, there is now clear evidence with both the carbocyclic and heterocyclic amines that such esters undergo spontaneous N-O heterolysis producing an arylnitrenium ion 6. 5,6 This reactive cation arylaminates and alkylates DNA to produce covalent adducts 7 that can lead to mutagenesis and carcinogenesis.

Scheme 1
Nitrenium ions are a class of reactive electrophiles that, in contrast to carbocations, have not been successfully characterized under superacid conditions.Since 1993 however, there have been a number of reports of their direct observation using laser flash photolysis (LFP). 8This approach allows for the their generation in solvents such as acetonitrile and water, and has provided direct kinetic information about the reactivity of nitrenium ions with the solvent and with added nucleophiles.The recent use of IR 9 and Raman 10 detection systems is also providing detailed structural information.An approach that our group has used is outlined in Scheme 1. 6 An aryl azide 7 serves as the photochemical precursor.Irradiation produces a reactive singlet nitrene 8 that in aqueous solutions is protonated on the picosecond time scale to provide the conjugate acid, the nitrenium ion 6.
This paper provides a kinetic analysis of the behavior of the nitrenium ions 11 and 12 obtained, respectively, upon irradiation of 2-azido-9H-carbazole 9 and its 9-methyl analog 10 (Scheme 2).These cations lie intermediate between the carbocyclic 2-fluorenyl nitrenium ion, and the nitrenium ion derived from 2-amino-α-carboline.The former cation has been studied by LFP from the precursor 2-azidofluorene. 6The Novak group have recently published a kinetic investigation of the latter using the indirect method of competition kinetics.5f An attractive feature of the carbazole system is that it provides an evaluation of the role of the central nitrogen.As shown in the resonance contributors 11b and 12b, this nitrogen in principle can provide conjugative stabilization to the positive charge.We also anticipated that when R = H, the proton might be removed to generate a neutral conjugate base.We have recently described neutral conjugate base forms of arylnitrenium ions in the case of benzidine derivatives.

Results and Discussion
The azides 9 and 10 were irradiated with 308 nm laser light in solutions of 20% acetonitrile in water.Transient spectra were constructed point-by-point at wavelengths greater than 350 nm where the azides did not absorb.Figure 1 shows spectra obtained with the methyl derivative 10 in 0.001 M HClO 4 .The behavior of this azide in pH 7 phosphate buffer was identical.This system is characterized by a relatively intense signal with two maxima.This absorbance decays on the hundreds of microsecond timescale resulting in the spectrum labeled Final (a) in Fig. 1.The absorbance changes follow single exponential kinetics, with the same rate constant over the entire spectral range.Irradiation with bromide present results in more rapid decay, and has the feature that there is less residual absorbance, as shown in the spectrum labeled Final(b).
The NH azide 9 (Figure 2) was found to have essentially the same behavior in 0.001 M HClO 4 , with very similar spectra and spectral changes, and also similar rate constants for the decay in solvent alone and with bromide present.With this azide however, the transient spectrum in pH 7 phosphate buffer was quite different, with little OD at 440 nm, and a peak with λ max at 360 nm that underwent a very slow decay.A detailed spectrum was also constructed in 0.1 M HClO 4 , and was also found to be different.A similar difference between 0.001 M HClO 4 and 0.1 M HClO 4 was also apparent with the NMe system.
Figures 1 and 2. Transient absorption spectra obtained upon 308 nm irradiation of 2-azido-9methyl-9H-carbazole (Fig. 1) and 2-azido-9H-carbazole (Fig. 2).Carbazole concentrations were 5 x 10 -5 M in 20% acetonitrile.Fig. 1 was obtained in 0.001 M HClO 4 .The spectrum labeled "Initial" was obtained 200 ns after the laser pulse, and the spectrum "Final(a)" at the completion of the exponential decay.The spectrum "Final(b)" was obtained at the completion of the exponential decay with 0.04 M NaBr present.The spectra in Fig. 2 were the "Initial" spectra obtained 200 ns after the laser pulse in solutions with the pH indicated.
The effect of pH was studied in detail for the NH system.A monitoring wavelength of 450 Figures 3 and 4. Behavior of transients as a function of pH. Figure 3 plots the optical density at 450 nm measured 200 ns after the laser pulse for solutions of the same concentration of 2-azido-9H-carbazole (5.0 x 10 -5 M) irradiated at 308 nm under identical conditions.Figure 4 plots the first-order rate constants for the decay of the transients obtained with 2-azido-9H-carbazole (squares) and 2-azido-9-methyl-9H-carbazole (crosses).Data were obtained in 20% acetonitrile at 20 o C and an ionic strength of 0.1 M maintained with sodium perchlorate.
Figure 4 shows the effect of pH on the rate constants for the decay of the transients in the 20% acetonitrile solutions.The data for the N-methyl system were obtained by monitoring at 450 nm throughout.The same wavelength was employed for the NH system below pH 5, with a monitoring wavelength of 370 nm above pH 5. The pH was maintained with HClO 4 (below pH 3.5) and with acetate, phosphate and carbonate buffers.Dilute buffer solutions were employed; buffer dilutions showed that the buffer had no effect, outside of the experimental error of the measurement.The rate constants for the N-methyl system were essentially unchanged from pH 1 to pH 8.The NH system showed a small dependence below pH 2, with a significant decrease in rate constant above pH 6. Rate constants above this pH were in fact measured with a lamp flash photolysis apparatus.Two other observations were that the decays followed single exponential kinetics throughout and the rate constants did not depend upon the concentration of the precursor.
Both bromide and azide (as their sodium salts) provided significant accelerations of the decay, with the observed rate constants linear in the concentration of the added salt.Secondorder rate constants were obtained as the slopes of such plots for 4-5 different concentrations.These data are plotted as a function of pH in Figures 5 and 6.
The quenching by the nucleophiles bromide and azide support the conclusion that the 2azidocarbazoles 9 and 10 are undergoing photochemistry similar to that observed with 2azidofluorene, 6b,10 with the nitrenes protonating in 20% acetonitrile.The spectral behavior, and the dependence on pH are also consistent with this idea.While nitrenium ions 11 and 12 will be the initial intermediate, the kinetic and spectral data require a more complex system.As shown in Scheme 3, there is evidence of an acid-base equilibrium between the cations 11 and 12 and their conjugate acid dications 13 and 14 respectively.With the NH system, there is a second acid-base equilibrium involving a neutral compound 15.All the species in principal can react with water and with the added nucleophiles, although it turns out that not all rate constants can be determined.

Scheme 3
Based upon Scheme 3, the observed rate constants for each nucleophile take the form of equations 1 and 2, where Nu refers to w, Br or az.For the solvent reaction (Fig. 4), one feature is that there is not much difference between kw +2 and kw + .In fact with the NMe system there is hardly any change between acid and neutral solution, and the presence of the dication is only deduced because of a spectral change and in particular because of the kinetic behavior with bromide (see below).The line drawn in Fig. 4 for NMe is based on a fit to eq 1 setting Ka(1) at the value determined from the bromide data.

Eqn 1 Eqn. 2
The pKa values for the NH system can be compared with ones calculated from the absorbance data shown in Fig. 3. Eq 3 applies in this case, where OD(13), OD (11) and OD (15)  are the optical densities for the dication, cation and neutral forms under the conditions of this experiment.The line drawn in Fig. 3 is based on a fit to this equation.The acidity constants obtained in this fit were pKa(1) = 1.8±0.2 and pKa(2) = 5.74±0.08, in good agreement with values from the kinetic analysis.

Eqn. 3
Quenching by bromide could only be observed in solutions with a pH less than 5.5-6.At this point the kinetic traces began taking the form of a double exponential, and by pH 8, no quenching occurred.One interpretation is that bromide adds reversibly. 12The pH dependence can be explained by a mechanism involving conjugate acid-base forms of the initial bromide adduct, as shown in Scheme 4 where the position of bromide attachment is based on results in the carboline system.5f The formation of the protonated adduct 17 under acidic conditions would shift the equilibrium to the adduct side (and also provide for rapid rearomatization).

ISSN 1424-6376
Page 78 © ARKAT USA, Inc Scheme 4 The rate constants for bromide (Figure 5) provide a clear indication of the presence of the dication and monocation reacting forms, with a significant difference in reactivity between the two.The data for the NMe compound were fit to eq 1 to give k Br+2 = (5.9±0.5)x 10 8 M -1 s -1 , k Br+ = (5.0±0.2) x 10 6 M -1 s -1 and pK a (1) = 1.85±0.05.For the NH compound, the data were fit to eq 2 with pKa(2) set to 5.77, the average of the values determined above, and with kBr 0 ignored since the data do not extend to a region where this is important.The fit provided k Br+2 = (3.6±0.2) x 10 8 M -1 s -1 , k Br+ = (1.6±0.1)x 10 6 M -1 s -1 and pK a (1) = 1.92±0.05.The value of pK a (1), which is most precisely determined from the bromide data, is in good agreement with the less precise values calculated from the water rate constants and the spectral data.It can be seen both in Figure 5 and in the rate constants and pKa values that there is a close similarity between the NMe and NH systems.
Azide ion proves to be a very efficient quencher for both the NMe and NH systems.The data for the NMe compound show a plateau at higher pH with a downward break below pH 5.This can be explained by reaction of the azide ion with the monocation 12, with the downward break below pH 5 associated with protonation of the nucleophile.We have previously measured a pKa(HN3) of 4.46 in 20% acetonitrile, ionic strength 0.1 M. 13 The curve drawn for the NMe compound in Fig. 6 is based upon a correction of the observed rate constants based on this pKa; the rate constant for kaz + is calculated as (2.8±0.2) x 10 9 M -1 s -1 .The data for the NH compound parallels that of the NMe in the acidic region (pH 4.5-5.5),but there is then a significant decrease in rate constant, with a plateau in base at a much lower level.This again indicates the difference between the NMe and NH associated with the ionization of the latter to the neutral 15.With thecorrection for the azide protonation, the data were fit to give k az+ = (2.7±0.2) x 10 9 M -1 s 1 , k az0 = (1.0±0.2) x 10 6 M -1 s -1 and a pKa(2) of 5.83±0.08 in good agreement with values obtained in previous analyses.It can be noted that a contribution from reaction of azide ion and the dication was ignored in both the NMe and NH analyses.This can be done because even in the most acidic solution there is less than 1% of the dication present and the monocation forms are very reactive.In fact the two rate constants for k az + are close to the diffusion limit of 5 x 10 9 M -1 s -1 found for the combination of azide ion and arylnitrenium ions under conditions of ionic strength 0.1 M in 20% acetonitrile.6b That the values are slightly smaller is consistent with the kinetic stability of 11 and 12.With water rate constants of 4 x 10 3 and 5 x 10 3 s -1 , they lie intermediate between the 2fluorenylnitrenium ion, 6b k w = 2 x 10 4 s -1 and the 4'-methoxybiphenylylnitrenium ion, 6c k w = 1 x 10 3 s -1 .It is in this region of reactivity that the azide-nitrenium combination is changing from diffusion control to activation controlled.6d It is now clear that arylnitrenium intermediates have a rich acid-base chemistry.They can be ISSN 1424-6376 Page 79 © ARKAT USA, Inc protonated on the formal N + center under acidic conditions to give arylamine dications.In cases where there is a conjugating NH group, loss of the proton can occur to give a quinonoid-type neutral conjugate base.Table 1 provides a comparison of the results for the 2carbazolylnitrenium ion with those previously obtained for the 2fluorenylnitrenium ion, the 2-(αcarbolinyl)nitrenium ion and the very stable nitrenium ion derived from benzidine.In a comparison with the fluorene derivative, it can be seen that the introduction of the central NH group in the carbazole has had little effect on the lifetime of the nitrenium monocation form (kw + ).However, as seen in the values for pK a (1), the basicity is increased by 1.3 log units.Moreover the carbazole dication is over 200-fold longer lived (k w +2 ).This is likely a reflection of the ability to place the two positive charges on the two nitrogen atoms, as shown in structure 13.
Comparing the carbazole and carboline system, one sees that the introduction of the pyridine nitrogen in the latter has significantly destabilized the monocation.This is particularly apparent in the pK a (2) values, which differ by almost three log units.Not surprisingly the carboline system showed no evidence for the dication.The benzidine cations are much more stable.This is seen in both the pKa values and the lifetimes.One interesting feature of the comparison with the carbazole is that there is the same difference, a factor of 4, between pK a (1) and pK a (2).In other words, in both systems, there is only a region of about 4 pH units where the nitrenium monocation structure predominates.As noted before, 11b the benzidine dication is considerably less reactive towards water than the monocation.The opposite is true in the fluorene system.We would argue that the carbazole derivatives are just at the changeover, and thus k w +2 and k w + are very similar.

Experimental Section
General Procedures.The details of the flash photolysis experiments have been published previously. 6Amino-9H-carbazole, prepared following literature methods, 14,15 was dissolved in 100 mls of ethanol, and after treatment with concentrated sulfuric acid (5.0 ml) with cooling in ice-water bath, amyl nitrite (6.0 ml) was added in small portions.The reaction mixture was stirred for 20 min at 35-40 o C. The aryldiazonium salt was precipitated by diluting the reaction mixture with ether (200 ml), the resultant precipitate being washed repeatedly with ether until the ethereal washings gave no coloration.The precipitate was then treated at 0 o C with a saturated solution of sodium azide (5.0 g) and sodium acetate (10 g).After stirring the reaction mixture for 1 h at room temperature, the product was extracted into ether, which was washed with water , 2.5 % sulfuric acid, water, 10 % sodium hydroxide and finally with water.The ethereal extract was dried over magnesium sulfate and the solvent was evaporated under reduced pressure.The solid so obtained was recrystallized from ethanol: water to give 2-azido-9H-carbazole, mp 180-182 o C,