Acid-base properties of arylnitrenium ions

Article abstract of DOI:10.1021/ja954248d

This study uses a combination of laser flash photolysis (LFP) and product analysis to show that singlet nitrenes from the irradiation of phenyl, 4-biphenylyl, and 2-fluorenyl azide can be trapped by protonation in aqueous solutions forming nitrenium ions. With phenyl azide, the phenylnitrenium ion is indicated by the formation of ring-substituted anilines in yields of up to 50% in 1 M acids. The acidity dependence furnishes the ratio k(H):k(exp) = 1.1, where k(H) refers to H⁺-trapping of singlet phenylnitrene and k(exp) to ring expansion of this species. With k(H) expected to be 2-4 x 10¹⁰ M^-1 s^-1, k(exp) is therefore estimated as 2-4 x 10¹⁰ s^-1. Protonation by solvent water also occurs, but even though the rate constant is of the order of 10⁹ s^-1, it constitutes a minor pathway in competition with the ring expansion. LFP studies in acids reveal a transient that is assigned the structure of N-protonated 4-hydroxy-2,5-cyclohexadienone imine, the intermediate formed by water addition to the para position of the phenylnitrenium ion. With 4-biphenylyl- and 2-fluorenylnitrene, ring expansion (and intersystem crossing) occurs more slowly and protonation by water is faster, with the consequence that there are substantial yields of nitrenium ion without added acids. These nitrenium ions are detected with ns LFP, and their formation from singlet nitrene is observed with ps LFP. Combining the LFP experiments with product analysis furnishes a pK(a) value of 16 for the 4-biphenylylnitrenium ion deprotonating to singlet nitrene in 20% acetonitrile. Thus singlet 4-biphenylylnitrene falls close to the category of a strong base in this solution. LFP experiments in acids show behavior consistent with N-protonation of the nitrenium ion forming an aniline dication. Kinetic analyses furnish pK(a) values of 0.1 (4-aminobiphenyl dication) and 0.6 (2-aminofluorene dication) in 20% acetonitrile with 1 M ionic strength. This and other pieces of evidence are consistent with these arylnitrenium ions being better regarded as 6-iminocyclohexadienyl carbocations. Overall, arylnitrenium ions (ArNH⁺) are very weak acids in water in their deprotonation to singlet nitrenes. They are also weak bases, accepting a proton to form the aniline dication - ¹ArN ? ¹ArNH⁺ ? (ArNH₂)²⁺.

Full text of DOI:10.1021/ja954248d

4
794
J. Am. Chem. Soc. 1996, 118, 4794-4803
Acid-Base Properties of Arylnitrenium Ions
Robert A. McClelland,* Mary Jo Kahley, P. Adriaan Davidse, and
Gordana Hadzialic
Contribution from the Department of Chemistry, UniVersity of Toronto,
Toronto, Ontario M5S 1A1, Canada
X
ReceiVed December 19, 1995
Abstract: This study uses a combination of laser flash photolysis (LFP) and product analysis to show that singlet
nitrenes from the irradiation of phenyl, 4-biphenylyl, and 2-fluorenyl azide can be trapped by protonation in aqueous
solutions forming nitrenium ions. With phenyl azide, the phenylnitrenium ion is indicated by the formation of ring-
substituted anilines in yields of up to 50% in 1 M acids. The acidity dependence furnishes the ratio kH:kexp ) 1.1,
+
where kH refers to H -trapping of singlet phenylnitrene and kexp to ring expansion of this species. With kH expected
1
0
-1 -1
10 -1
to be 2-4 × 10
M
s , kexp is therefore estimated as 2-4 × 10 s . Protonation by solvent water also occurs,
9 -1
but even though the rate constant is of the order of 10 s , it constitutes a minor pathway in competition with the
ring expansion. LFP studies in acids reveal a transient that is assigned the structure of N-protonated 4-hydroxy-
2
,5-cyclohexadienone imine, the intermediate formed by water addition to the para position of the phenylnitrenium
ion. With 4-biphenylyl- and 2-fluorenylnitrene, ring expansion (and intersystem crossing) occurs more slowly and
protonation by water is faster, with the consequence that there are substantial yields of nitrenium ion without added
acids. These nitrenium ions are detected with ns LFP, and their formation from singlet nitrene is observed with ps
LFP. Combining the LFP experiments with product analysis furnishes a pKa value of 16 for the 4-biphenylylnitrenium
ion deprotonating to singlet nitrene in 20% acetonitrile. Thus singlet 4-biphenylylnitrene falls close to the category
of a strong base in this solution. LFP experiments in acids show behavior consistent with N-protonation of the
nitrenium ion forming an aniline dication. Kinetic analyses furnish pKa values of 0.1 (4-aminobiphenyl dication)
and 0.6 (2-aminofluorene dication) in 20% acetonitrile with 1 M ionic strength. This and other pieces of evidence
are consistent with these arylnitrenium ions being better regarded as 6-iminocyclohexadienyl carbocations. Overall,
+
arylnitrenium ions (ArNH ) are very weak acids in water in their deprotonation to singlet nitrenes. They are also
1
1
+
2+
weak bases, accepting a proton to form the aniline dications ArN a ArNH a (ArNH2) .
Carcinogens such as 4-aminobiphenyl (1) and 2-aminofluo-
electron oxidation of the corresponding amine. These include
rene (2) are generally accepted to undergo a 2-fold metabolic
activation to O-acetate or O-sulfate esters 3, followed by N-O
heterolysis to an arylnitrenium ion (4, 5).1
diarylnitrenium ions bearing stabilizing o- and p-methoxy
6
substituents such as the bis(4-methoxyphenyl)nitrenium ion and
7
the 4-biphenylyl derivative 7. In the cases of the diarylnitre-
nium ions, only UV-visible absorption spectra were reported.
1
The ion 7 was sufficiently persistent for its H NMR spectrum
to be recorded.
The critical cellular target appears to be DNA, and indeed a
2
covalent adduct of guanine residues has been observed in ViVo
as well as in model systems.3 In part because of this relevance
to arylamine carcinogenicity a number of studies have been
carried out probing the chemistry of arylnitrenium ions when
No nitrenium ion has been characterized as a distinct species
under super acid conditions, except for the special case of
4
,5
formed as intermediates of solvolysis and related reactions.
+
cyanodiarylmethyl cations (Ar2C -CN) which were considered
to have significant nitrenium character (Ar2CdCdN ). As
implied by studies with nitrosobenzenes (e.g. 8), one problem
In contrast to carbenium ions, however, examples of the
nitrenium class that have been characterized by direct spectro-
scopic means are limited. There are three reports of relatively
stabilized nitrenium ions produced by electrochemical two-
+
8
(4) (a) Gassman, P. G. Acc. Chem. Res. 1970, 3, 26. (b) Gassman, P.;
Hartman, G. D. J. Am. Chem. Soc. 1973, 95, 449. (c) Novak, M.; Pelecanou,
M.; Roy, A. K.; Andronico, A. F.; Plourde, F. M.; Olefirowicz, T. M.;
Curtin, T. J. J. Am. Chem. Soc. 1984, 106, 5623. (d) Fishbein, J. C.;
McClelland, R. A. J. Am. Chem. Soc. 1987, 109, 2824. (e) Novak, M.;
Kahley, M. J.; Eigen, E.; Helmick, J. S.; Peters, H. E. J. Am. Chem. Soc.
1993, 115, 9453. (f) Novak, M.; Kahley, M. J.; Lin, J.; Kennedy, S. A.;
Swanegan, L. A. J. Am. Chem. Soc. 1994, 116, 11626. (g) Fishbein, J. C.;
McClelland, R. A. J. Chem. Soc., Perkin Trans. 2 1995, 663.
(5) See also: Abramovitch, R. A.; Jeyaraman, R. in Azides and Nitrenes;
ReactiVity and Utility; Scriven, E. F. V., Ed.; Academic Press: New York,
1984; pp 297-354.
X
Abstract published in AdVance ACS Abstracts, April 15, 1996.
(1) (a) Miller, J. A. Cancer Res. 1970, 30, 559; 1978, 38, 1479. (b) Miller,
E. C.; Miller, J. A. Cancer 1981, 47, 2327. Miller, E. C.; Miller, J. A.
EnViron. Health Perspect. 1983, 49, 3. (c) Garner, R. C.; Martin, C. N.;
Clayson, D. B. In Chemical Carcinogens, 2nd ed.; Searle, C. E., Ed; ACS
Monograph 182; American Chemical Society; Washington, DC, 1984; Vol.
1
, pp 175-276. (d) Kadlubar, F. F.; Miller, J. A.; Miller, E. C. Cancer
Res. 1977, 37, 805.
(
(
2) Schut, H. A. J.; Castongauy, A. Drug Metab. ReV. 1984, 15, 753.
3) (a) Underwood, G. R.; Price, M. F.; Shapiro, R. Carcinogenesis 1988,
9
, 1817. (b) Famulok, M.; Bosold, F.; Boche, G. Angew. Chem., Int. Ed.
Engl. 1989, 28, 337. (c) Novak, M.; Kennedy, S. A. J. Am. Chem. Soc.
995, 117, 574.
(6) (a) Svanholm, U.; Parker, V. D. J. Am. Chem. Soc. 1974, 96, 1234.
(b) Serve, D. J. Am. Chem. Soc. 1975, 97, 432.
(7) Riecker, A.; Reiser, B. Tetrahedron Lett. 1990, 31, 5013.
1
S0002-7863(95)04248-X CCC: $12.00 © 1996 American Chemical Society

Acid-Base Properties of Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 20, 1996 4795
with the super acid experiments is the formation of dications
in the presence of acids such as acetic, trifluoroacetic, and
trifluoromethanesulfonic acid and characterized the products as
being consistent with a phenylnitrenium ion intermediate.1
In experiments employing a biological end point, the Wild group
photolyzed a series of azides derived from mutagenic arylamines
and heterocyclic amines in the presence of bacterial strains.20
They showed that there was a short-lived mutagenic intermediate
and they suggested this to be the nitrenium ion. Finally there
is a 25-year old study employing conventional lamp photolysis
that showed that the irradiation of p-(dialkylamino)phenyl azides
9
(e.g. 9) which might be viewed as N-protonated nitrenium ions.
8,19
Within the past two years have appeared pathways for the
1
9 in aqueous solution resulted in quantitative yields of the
direct study of arylnitrenium ions using the method of laser flash
photolysis (LFP). The first report came from the Falvey group
and involved the photochemical ring opening in acetonitrile of
corresponding nitrenium ion 20, referred to in that work as a
21
quinonediimine 21.
1
0
1
0; this has been followed by two detailed studies of this
system probing aspects such as the electrophilicity of the
nitrenium ions 11 and their spin multiplicity.1
1,12
Arylnitrenes also undergo other reactions very rapidly,17 and
these present a potential problem when attempting to trap by
proton transfer. Somewhat to our surprise however we found
that 4-biphenylyl azide and 2-fluorenyl azide photolyzed in
aqueous solutions to produce high yields of the corresponding
nitrenium ions (4, 5) and these intermediates were detectable
with LFP. This result has been reported in a preliminary
Shortly after the first Falvey paper, we reported in collabora-
tion with the Novak group the N-acetyl-4-biphenylyl- and
N-acetyl-2-fluorenylnitrenium ions (14, 15), observed in aqueous
acetonitrile on irradiation of precursors 12 and 13.13
2
2
communication.
Subsequent to this the Platz group have
reported that 2,3,5,6-tetrafluorophenyl azides with a variety of
substituents at the 4-position also produce nitrenium ions, in
these cases when photolyzed in the presences of acids such as
2
3
sulfuric acid.
In this paper we provide full details of our studies of the
-biphenylyl and 2-fluorenyl azides, and also discuss the
4
photochemical behavior of the parent phenyl azide in aqueous
acids. We in particular address the question of the acidity and
basicity of the intermediate nitrenium ions, and show that these
can be quantitatively assessed in some cases using the LFP
kinetics.
Both the Falvey approach and our approach involved
photoheterolysissthe cleavage in the excited state of a N-leaving
group bond. Photoheterolytic systems have also been employed
successfully for the LFP generation and study of carbenium
1
4
Results
ions. A different approach that has also been successful with
carbenium ions has involved the protonation of photogenerated
Phenyl Azide: Products in Aqueous Acid. Solutions of
phenyl azide in aqueous solutions were purged with argon and
irradiated at 300 nm in a Rayonet reactor. The solutions
contained a mixture of HCl and NaCl so that the total ionic
strength and total chloride concentration was 1 M. Four
products were quantified by HPLC, 1H-azepin-2(3H)-one (22),
1
5,16
carbenes,
as exemplified in eq 6 by the protonation by
15
alcohols and water of diarylcarbenes 17.
4
-aminophenol (23), 4-chloroaniline (24), and 2-chloroaniline
This raises the possibility of obtaining arylnitrenium ions
through the protonation of arylnitrenes, species readily generated
photochemically from aryl azides.17 An examination of the
literature shows that there is precedence. The Takeuchi group
has investigated the thermolysis and photolysis of phenyl azide
(18) (a) Takeuchi, H.; Takano, K. J. Chem. Soc., Chem. Commun. 1982,
1
4
254. (b) Takeuchi, H.; Takano, K. J. Chem. Soc., Chem. Commun. 1983,
47. (c) Takeuchi, H.; Koyama, K. J. Chem. Soc., Perkin Trans. 1 1982,
1269. (d) Takeuchi, H.; Koyama, K. J. Chem. Soc., Perkin Trans. 1 1988,
1269. (e) Takeuchi, H.; Ihara, R. J. Chem. Soc., Chem. Commun. 1983,
175. (f) Takeuchi, H.; Koyama, K.; Mitani, M.; Ihara, R.; Uno, T.; Okazaki,
(
8) (a) Olah, G. A.; Prakash, G. K. S.; Arvanaghi, M. J. Am. Chem. Soc.
980, 102, 6640. (b) Olah, G. A.; Arvanaghi, M.; Prakash, G. K. S. J. Am.
Chem. Soc. 1982, 104, 1628.
Y.; Kai, Y.; Kasai, N. J. Chem. Soc., Perkin Trans. 1 1985, 677. (g)
Takeuchi, H.; Takano, K. J. Chem. Soc., Perkin Trans. 1 1986, 611. (h)
Takeuchi, H.; Hirayama, S.; Mitani, M.; Koyama, K. J. Chem. Soc., Perkin
Trans. I, 1988, 521-527. See also: (i) Abramovitch, R. A.; Cooper, M.;
Iyer, S.; Jeyaraman, R.; Rodrigues, J. A. R. J. Org. Chem. 1982, 47, 4819.
(j) Abramovitch, R. A.; Jeyaraman, R.; Yannakopoulou, K. J. Chem. Soc.,
Chem. Commun. 1985, 1017. (k) Abramovitch, R. A.; Hawi, A.; Rodrigues,
J. A. R.; Trombretta, T. R. J. Chem. Soc., Chem. Commun. 1986, 283.
(19) For another explanation see: Olah, G. A.; Ramaiah, P.; Wang, Q.;
Prakash, G. K. S. J. Org. Chem. 1993, 58, 6900.
(20) (a) Wild, D.; Dirr, A. Carcinogenesis 1988, 9, 869. (b) Wild, D.;
Dirr, A.; Fasshauer, I.; Henschler, D. Carcinogenesis 1989, 10, 335. (c)
Wild, D.; Dirr, A. Mutagenesis 1989, 4, 446. (d) Wild, D. EnViron. Health
Perspect. 1990, 88, 27. (e) Sabbioni, F.; Wild, D. Carcinogenesis, 1992,
13, 709.
1
(9) Olah, G. A.; Donovan, D. J. J. Org. Chem. 1978, 43, 1743.
(10) Anderson, G. B.; Falvey, D. E. J. Am. Chem. Soc. 1993, 115, 9870.
(11) Robbins, R. J.; Yang, L. L.-N.; Anderson, G. B.; Falvey, D. E. J.
Am. Chem. Soc. 1995, 117, 6544.
(
(
12) Srivastava, S.; Falvey, D. E. J. Am. Chem. Soc. 1995, 117, 10186.
13) Davidse, P. A.; Kahley, M. J.; McClelland, R. A.; Novak, M. J.
Am. Chem. Soc. 1994, 116, 4513.
(
14) McClelland, R. A. Tetrahedron In press.
(15) (a) Kirmse, W.; Kilian, J.; Steenken, S. J. Am. Chem. Soc. 1990,
1
(
12, 6399. (b) Chateauneuf, J. J. Chem. Soc., Chem. Commun. 1991, 1437.
c) Belt, S. T.; Bohne, C.; Charette, G.; Sugamori, S. E.; Scaiano, J. C. J.
Am. Chem. Soc. 1993, 115, 2200.
16) (a) Schepp, N. P.; Wirz, J. J. Am. Chem. Soc. 1994, 116, 11749.
b) Kirmse, W.; Strehlke, I. K.; Steenken, S. J. Am. Chem. Soc. 1995, 117,
007-7008.
17) Schuster, G. B.; Platz, M. S. AdV. Photochem. 1992, 17, 143.
(
(21) Baetzold, R. C.; Tong, L. K. J. J. Am. Chem. Soc. 1970, 93, 1347.
(22) McClelland, R. A.; Davidse, P. A.; Hadzialic, G. J. Am. Chem. Soc.
1995, 117, 4173.
(
7
(
(23) Michalak, J.; Zhai, H. B.; Platz, M. S. J. Phys. Chem. In press.

4
796 J. Am. Chem. Soc., Vol. 118, No. 20, 1996
McClelland et al.
-
4
Figure 1. Absorption spectra following 248 nm irradiation of 1 × 10 M solutions of phenyl azide in aqueous HClO
4
solutions. Solutions were
argon saturated and were irradiated under identical conditions of laser intensity and cell configuration. Ionic strength was 1.0 M maintained with
NaClO
4
, and the temperature was 20 ( 1 °C. Initial spectra (9) were obtained 100-200 ns after the laser pulse, intermediate spectra (0) at 5-5.1
µs, and final spectra (2) at 35-40 µs.
Table 1. Product Chemical Yield and Relative Quantum Yieldb
a
quantum yields were calculated as the ratio of the fractions of
phenyl azide consumed.
As also shown in Table 1, experiments were carried out in
selected HClO4 solutions. Only the two products 22 and 23
were measured here since the chloroanilines did not form in
these solutions.
for Disappearance of Substrate for Photolysis of Phenyl Azide in
Aqueous HCl + NaCl (or HClO
4
+ NaClO
4
) of Total Concentration
M (300 nm Irradiation, 5 × 10 M Substrate, Argon-Purged,
0 ( 1 °C)
-
5
1
2
yield, %
23d
[HCl]
22c
24e
25f
Φ
rel
Experiments were also carried out in the HCl:NaCl mixtures
with N-phenylhydroxylamine. This compound reacted thermally
under acidic conditions to form 23-25 as the only detectable
products, accounting for >96% of the reacted starting material.
In these solutions where the chloride ion concentration was kept
1
0
0
0
0
0
0
0
0
0
1
0
0
.00
.75
.50
.25
40.1
53.5
61.8
59.1
71.8
74.1
79.9
78.6
76.0
71.8
41.1
68.7
71.8
39.0
29.2
28.9
18.9
14.2
7.4
5.8
5.3
5.7
4.8
9.3
7.6
7.4
5.0
3.1
1.9
1.5
1.2
1.5
1.3
4.2
3.7
3.4
2.3
1.4
0.9
0.7
0.6
0.7
0.6
0.97
0.92
0.95
0.94
.10
2
4
.025
.010
.004
.001
.0003
.00
.10
constant, the yields of the three products were within
experimental error independent of acid concentration. The ratio
0.88
1.00
(
[24] + [25]):[23] was 0.458 ( 0.021, and the ratio [24]:[25]
was 2.35 ( 0.08.
g
Phenyl Azide: Laser Flash Photolysis in Aqueous Acid.
These experiments were carried out with HClO4 solutions of
concentration 0.001-1.00 M, and representative transient spectra
are shown in Figure 1. In solutions with [HClO4] > 0.1, one
major transient species absorbing in the region 270-320 nm
was observed. The maximum absorbance was ∼290 nm,
although this may not be the true λmax since the precursor is
55.2
18.6
6.8
g
g
.01
a
Yields based on the percent phenyl azide consumed, for conversions
b
of ∼50%. Relative to 0.001 M HCl:0.99 M HCl, and based upon the
c
d
fraction of phenyl azide consumed. 1H-Azepin-2(3H)-one. 4-Ami-
e
f
g
nophenol. 4-Chloroaniline. 2-Chloroaniline. HClO
4
.
2
5
beginning to absorb in this region. This transient decayed
(
25). Results are given in Table 1. Control experiments
demonstrated that the yields of the products were within
experimental error unchanged for irradiations corresponding to
5-60% conversion. The phenyl azide was also shown to be
thermally stable in all solutions for the times required for
irradiation and subsequent analysis.
Examination of Table 1 reveals that the products 22-25
accounted for 78-100% of the total phenyl azide consumed,
with the mass balance poorer in the more dilute HCl solutions.
Although yields changed with changing HCl concentration, the
ratio of chloroanilines to aminophenols([24] + [25]):[23]swas
within experimental error constant at 0.369 ( 0.026, as was
also true for the ratio of 4-chloro- to 2-chloroanilines[24]:[25]
with single exponential kinetics, with a rate constant of (1.3 (
5
-1
0
.1) × 10 s independent of acid concentration. As illustrated
by comparison of the first two panels of Figure 1, the absolute
magnitude of the signal did show an acid dependence, increasing
significantly with increasing acid concentration. This transient
continued to be observed in the more dilute acids, but its
intensity became weak, and as shown in the third panel of Figure
2
1
4
, it was overlaid with a broad signal in the region from 300 to
00 nm. This absorbance decayed with reasonable first-order
+
kinetics. The first-order rate constants were proportional to H
+
8
-1 -1
concentration, with kH ∼ 10 M
the range 0.01-0.05 M HClO4).
-Biphenylyl Azide: Products. Photolysis of this azide in
s
(based on three acids in
4
)
2.14 ( 0.07.
an aqueous solution containing 20% acetonitrile resulted in the
formation of 4-hydroxy-4-phenyl-2,5-cyclohexadienone (26) as
the major product. This material was isolated in good yield
from a scaled-up photolysis; a sample that had been prepared
Also given in Table 1 are quantum yields in selected acids
for the disappearance of phenyl azide relative to the quantum
yield in the solution containing 0.001 M HCl. These experi-
ments were carried out by simultaneously irradiating solutions
containing the same concentration of phenyl azide (and same
optical density at 300 nm) in a merry-go-round cuvette holder,
such that the solutions received identical amounts of the 300-
nm light. The irradiations were carried out for times corre-
sponding on average to 35% conversion, and the relative
(24) (a) For details of experiments where the reaction was carried out
with varying chloride concentration see ref 24b. (b) Fishbein, J. C.;
McClelland, R. A. Can. J. Chem. In press.
(25) For all three spectra of Figure 1 there was a large negative signal
associated with bleaching of the phenyl azide at wavelengths below 270
nm.

Acid-Base Properties of Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 20, 1996 4797
Table 2. 4-Biphenylyl Azide Photolysis. Acetonitrile:Water
Dependences of Yield of 4-Hydroxy-4-phenyl-2,5-cyclohexadienone
(
∆
26), Relative Quantum Yield for Consumption of Substrate,
OD, and First-Order Rate Constant for Decay of LFP Transient
at 460 nm
acetonitrile,a yield of
k
obs(460 nm),
× 10 s
b
c
∆OD(460 nm)d
6
-1
%
26, %
Φ
rel
1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
88 ( 2
0.224 ( 0.006
0.212 ( 0.011
2.81 ( 0.11
2.84 ( 0.04
3.03 ( 0.07
3.16 ( 0.03
3.37 ( 0.09
3.35 ( 0.09
3.43 ( 0.09
3.38 ( 0.12
2
3
4
5
6
7
8
9
84 ( 3 1.00
73 ( 2 0.88 ( 0.05 0.188 ( 0.008
66 ( 1 1.01 ( 0.12 0.163 ( 0.012
57 ( 1
0.139 ( 0.007
43 ( 2 1.24 ( 0.09 0.105 ( 0.005
28 ( 2 0.062 ( 0.003
13 ( 1 0.97 ( 0.07 0.028 ( 0.001
∼1
1.14 ( 0.15 trace
0
.0
e
2.69 ( 0.11
a
Percent by volume. Solutions for product studies also contained
b
0
3 3
.002 M CH COOH:0.002 M CH COONa. Measured 24 h following
Figure 2. Absorption spectra following 248 nm irradiation of 2 ×
irradiation, and based upon the amount of 4-biphenylyl azide consumed
-5
1
0
M solutions of 4-biphenylyl azide (9) and 2-fluorenyl azide (0)
c
for experiments carried out to ∼40% conversion. Relative to 20%
in 1:4 (v/v) acetonitrile-water. Optical density units in the main figure
have been normalized so that the maximum ∆OD is 1. The inserts
show the kinetic traces at 460 nm for 4-biphenylyl azide in 1:4
acetonitrile-water (A) and 4:1 acetonitrile-water (B).
acetonitrile, and based upon the fraction of azide consumed. d Optical
density at 460 nm measured immediately after completion of the laser
pulse minus the residual optical density after complete decay. The
latter was small (0.002-0.004) but finite. All solutions contained 20
µM of substrate and were irradiated under identical conditions.
e
Table 4. Rate Constants for the Formation and Decay of
Concentration of 4-biphenylyl azide was 1 µM.
4
-Biphenylyl- and 2-Fluorenylnitrenium Ion in 1:4 (v:v)
a
Acetonitrile-Water (20 C)
ï
Table 3. 4-Biphenylyl Azide Photolysis. Yield of
4
-Hydroxy-4-phenyl-2,5-cyclohexadienone (26) in 1:4 (v:v)
Water-Acetonitrile, with Ionic Strength 0.5 Maintained
with NaClO
ionic
strengtha
constant
4-biphenylyl
2-fluorenyl
4
-
1
9
10
4
k
k
k
k
k
k
k
app, s
0
0
6.0 ( 0.1 × 10
1.1 ( 0.2 × 10
conditions
.1 M HClO
yield of 26, %
conditions
yield of 26, %
-1
6
s
s
s
s
, s
, s
, s
, s
2.84 ( 0.04 × 10 3.4 ( 0.2 × 10
-
-
-
1
1
1
6
4
0
4
86.2 ( 1.1
84.8 ( 1.2
82.5 ( 0.8
79.8 ( 0.9
77.6 ( 0.8
67.4 ( 1.4
61.3 ( 0.5
52.9 ( 0.6
0.050 M NaOH
0.100 M NaOH
0.150 M NaOH
0.200 M NaOH
0.250 M NaOH
0.300 M NaOH
0.400 M NaOH
0.500 M NaOH
43.1 ( 1.6
38.2 ( 2.1
28.5 ( 0.9
23.4 ( 0.4
21.7 ( 0.4
21.4 ( 0.8
20.8 ( 0.2
21.2 ( 0.7
0.5
1.0b
0.5
1.85 ( 0.07 × 10 2.2 ( 0.2 × 10
6
4
pH 4.5
pH 7.0
0
0
0
0
0
1.06 ( 0.02 × 10 1.3 ( 0.2 × 10
6
4
1.78 ( 0.04 × 10 2.0 ( 0.4 × 10
Az, M-1s-1
0.5
b
5.0 ( 0.2 × 10
9
4.0 ( 0.2 × 10
9
.002 M NaOH
.005 M NaOH
.010 M NaOH
.020 M NaOH
.030 M NaOH
OH(obs), M s-1
-
1
0.5
6.4 ( 0.3 × 10
7
1.05 ( 0.06 × 10
7
a
With NaClO
4
. b Solution contains 5% acetonitrile.
precursor 27 was also observed to a similarly small extent. In
fact an excellent correlation was found between the area of the
peak for the precursor 27 measured immediately after irradiation
and the area of the peak for 26 measured the next day. Other
peaks were observed in the HPLC, these being very small in
the solutions with high water content, but increasing in intensity
in the acetonitrile-rich solutions. Two of the HPLC peaks from
irradiation in 100% acetonitrile were matched to authentic
samples, one (∼10%) to 4-aminobiphenyl and the other (∼30%)
to 4,4′-diphenylazobenzene.
a
Measured 24 h following irradiation. For the experiments where
NaOH was present during irradiation, solutions were neutralized
immediately following, and stood for 24 h at pH ∼5.
in a different way was also obtained from the Novak group.26,27
Quantitative yields, as well as relative quantum yields for
disappearance of starting azide, were determined by HPLC and
are given in Tables 2 and 3 for photolyses carried out under a
variety of conditions. Yields of 26 were independent of
irradiation time for conversions up to 70%.
4
-Biphenylyl Azide and 2-Fluorenyl Azide: Laser Flash
The ketone 26 was not the initial product of photolysis.
Injections made immediately following the 0.5-1.5-min ir-
radiations resulted in HPLC traces with a peak with a different
retention time which disappeared over time with a 3-4-h half-
life (at pH 5). The peak for 26 was not present in the first
injection (or was very small), but grew in at the same rate as
the initial peak disappeared. Such behavior had been observed
previously by the Novak group (vide infra) with the precursor
for 26 being assigned as 4-hydroxy-4-phenyl-2,5-cyclohexadi-
enone imine (27).27 In the experiments carried out in NaOH
solutions (Table 3), the imine was present after irradiation, but
did not convert to the ketone. To analyze for 26, these solutions
were therefore neutralized to pH 5 immediately after irradiation.
Photolysis. Figure 2 shows the results of laser flash photolysis
experiments in aqueous solutions containing 20% acetonitrile.
With each azide, a relatively intense transient was observed,
with λmax around 460 nm for the biphenylyl system and 450
nm for fluorenyl. With each azide the absorbance decayed with
the same rate constant across the entire spectrum from 400 to
500 nm, while the absorbance below 350 nm showed little decay
on a 10-µs time scale. Rate constants were increased by the
addition of azide or hydroxide, and plots of kobs were linear in
the concentration of the added reagent. First-order rate constants
for decay in the solvent alone and second-order rate constants
for the quenching by azide and hydroxide are given in Table 4.
Experiments were carried out in which the biphenylyl azide
was irradiated under identical conditions in solutions of varying
acetonitrile content. These showed that the initial intensity of
the 460-nm transient decreased with increasing acetonitrile
In the experiments with high contents of acetonitrile the
amount of 26 observed was very small or negligible. The
(
26) Novak, M.; Helmick, J. S.; Oberlies, N.; Rangappa, K. S.; Clark,
W. M.; Swenton, J. S. J. Org. Chem. 1993, 58, 867-878.
27) Novak, M.; Kahley, M. J.; Eiger, E.; Helmick, J. S.; Peters, H. E.
J. Am. Chem. Soc. 1993, 115, 9453.
(compare for example the two inserts in Figure 2). Table 2
summarizes the acetonitrile dependence, for both the rate
constant for the 460-nm decay and the change in optical density,
(

4
798 J. Am. Chem. Soc., Vol. 118, No. 20, 1996
McClelland et al.
Scheme 1
Figure 3. Rate constants for decay of nitrenium ions in acidic solutions
1:4 acetonitrile-water; 20 °C). Solutions with pH <3 contained HClO
and solutions with pH 3.5-6.5 contained dilute acetate buffer; the ionic
strength in all solutions was 1.0 M, maintained with NaClO : (9)
-biphenylyl nitrenium ion; (2) 2-fluorenyl nitrenium ion; (O) N-acetyl-
-biphenylylnitrenium ion, generated from N-chloro precursor.
(
4
,
4
4
4
calculated as the optical density immediately after the laser pulse
minus the optical density at the conclusion of the decay. As
shown by the inserts to Figure 2, there is residual absorbance
at 460 nm, but the amount is small. The 460-nm transient was
not observed in solutions with acetonitrile content above 90%,
and there was in fact only a small change in optical density
above 400 nm in these solutions. The absorbance around 340
nm continued to be observed, and in fact seemed to grow in
intensity (although this was not studied in detail).
The LFP experiments discussed thus far involved excitation
with a laser with a ∼20 ns pulse width, and in all cases the
transients in the region above 400 nm were fully formed at the
completion of the laser pulse. The appearance of this absor-
bance was however observed using a laser system with
excitation at 266 nm and a ∼25 ps pulse. The transients seen
as the “initial” absorbances in the ns experiments grew in with
good exponential kinetics with rise times in 20% acetonitrile
of 90 and 160 ps for the 2-fluorenyl and 4-biphenylyl systems,
respectively (Table 4). A further feature of these experiments
was the absence of optical density immediately after the 25-ps
pulse in the region from 400 to 700 nm where observations
could be made with this equipment.
Similar experiments were carried out with both azides in basic
2
0% acetonitrile in order to see if base had any effect on the
initial intensity of the transient. These measurements could be
carried out to 0.15 M NaOH with the biphenylyl azide and 0.6
M NaOH with fluorenyl before the decay of the transient became
too rapid to reliably measure the initial absorbance. Within
experimental uncertainty the conclusion was that the addition
of base up to the above limits had no effect on the initial
intensity at 460 or 450 nm.
Shown in Figure 3 are the rate constants for the decay of the
two transients as a function of pH under acidic conditions. Each
system showed an increase in rate below pH 3-4, with signs
of leveling in the more concentrated acids. For comparison
purposes, rate constants were also obtained in the same solutions
for the N-acetyl-4-biphenylylnitrenium ion as generated from
Discussion
Scheme 1 summarizes the established chemistries of arylni-
trenes and arylnitrenium ions that are of relevance to this work.
On the nitrene side, the key reactions from the singlet are the
competing intersystem crossing to the triplet nitrene and ring
expansion to a 1,2-didehydroazepine (28, 29). Triplet nitrenes
react further to give azobenzenes and anilines by coupling and
hydrogen abstraction reactions, respectively, while dehy-
1
7
1
3
the N-chloro precursor.
These rate constants remained
constant from pH 1-7, with a slight decrease in the more acidic
solutions. For example, the rate constant in 1.0 M HClO4 was
2
9
droazepines react with nucleophiles to give azepines, in the
case of water an azepinone such as 22.
6
-1
2
3
.70 ( 0.05 × 10 s , while in neutral 1.0 M NaClO4, it was
1
8c,30
The dehy-
6
.35 ( 0.07 × 10 .
droazepines can also convert to triplet nitrene, and indeed in
the absence of nucleophiles this is often the major reaction.2
With singlet phenylnitrene, the predominant pathway at low
temperatures (e.g. 77 K) is intersystem crossing, while in
solutions at ambient temperature ring expansion to the dehy-
Since we will argue that the transient species from the azides
9,31
are in a different state of protonation in acids (vide supra), we
constructed a spectrum for the 2-fluorenyl system in 1.0 M
2
8
HClO4.
We found however that within the limits of our
apparatus this had a very similar form and λmax as the spectrum
constructed without acids. There did appear to be a slightly
lower initial intensity for the transient in the 1.0 M acid, the
initial optical density at 450 nm in 1.0 M HClO4 being about
17,32
droazepine dominates.
that this ring expansion occurs with a rate constant (kexp) of
Platz and co-workers have estimated
(29) See: Li, Y.-Z.; Kirby, J. P.; George, M. W.; Poliakoff, M.; Schuster,
G. B. J. Am. Chem. Soc. 1988, 110, 8092 and references therein.
70% of that in 1.0 M NaClO4 for experiments carried out under
(
30) (a) Vogel, E.; Erb, R.; Lenz, G.; Bothner, A. A. Liebigs Ann. Chem.
965, 682, 1. (b) Doering, W. E.; Odum, R. A. Tetrahedron 1966, 22, 81.
31) Polymerization of the dehydroazepine also occurs, and accounts for
identical conditions (substrate concentration, laser dose, etc.).
1
(
(
28) This was not possible with 4-biphenylyl azide, since the decay in 1
the poor mass balances in product analyses frequently observed following
M HClO4 was too fast.
aryl azide photolysis.17

Acid-Base Properties of Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 20, 1996 4799
1
0
11
1
0 -10 s-1 at room temperature.32 This same group has also
the one that relates singlets. Based on the energy differences
3
0
established that in the case of phenyl azide the singlet phen-
ylnitrene is the branching point for the triplet and dehy-
droazepine.
Arylnitrenium ions are formed by ground state heterolysis
reactions, either from O-protonated hydroxylamines 31 under
acidic conditions or from hydroxylamine esters such as the
pivalate 33. These nitrenium ions react with water at the
position para to the nitrogen to give adducts such as 27 and 32.
In cases such as the parent system, rapid tautomerization gives
the p-aminophenol. In cases such as the 4-biphenylyl system
where this tautomerization is blocked, the imine can be observed
but it is not stable. Several reactions occur, including imine
cited above there would be an approximately 10 difference.
Trapping of Singlet Phenylnitrene by Protonation. Ir-
radiation of phenyl azide in 1 M HCl results in 40% 22, and
slightly greater then 50% in total of the anilines 23, 24, and 25.
The azepinone is explained by the nitrene chemistry in Scheme
1, through water addition to the ring expanded dehydroazepine
28. The ring-substituted anilines however are not the products
from a nitrene, but are diagnostic of the presence of the
phenylnitrenium ion. These same three products are formed in
the ground state reaction of N-phenylhydroxylamine in acid, in
this case in quantitative yield. Moreover, the chloroanilines:
aminophenol ratios and the p-chloroaniline:o-chloroaniline ratios
3
3
3
4
3
5
40
hydrolysis and migration of R or OH to an adjacent carbon.
are very similar whether phenylhydroxylamine or phenyl azide
Nitrenium ions can also be trapped by external nucleophiles
is the precursor.
such as chloride and azide to give products such as 24, 25, and
The question then arises as to the immediate precursor of
the nitrenium ion in the azide photolysis, that is, whether it is
indeed the singlet nitrene or the excited singlet state of the
phenyl azide. In the latter case, nitrenium ion would form with
3
4.
A singlet nitrene is the conjugate base of a singlet nitrenium
ion, and thus the two intermediates can in principle be linked
by protonation/deprotonation. This is shown in the highlighted
protonation concurrent with loss of N . In their study involving
2
+
box in Scheme 1, with the protonation occurring by H and by
fluorinated phenyl azides as precursors of nitrenium ions, Platz
water, and the microscopic reverse deprotonation by water and
by hydroxide ion, respectively. The acidity constant for the
singlet nitrenium ion deprotonating to the singlet nitrene is thus
given by the following:
and co-workers produced evidence for a singlet nitrene precursor
by demonstrating agreement between lifetimes of this species
as estimated by trapping by acid and independently by trapping
2
3
with pyridine. In the present case the excited singlet can be
ruled out as a precursor on the basis that the relative quantum
yields for disappearance of the phenyl azide are within
experimental error unchanged from dilute acid to 1.0 M acid.
This occurs in spite of an increase in the yield of nitrenium-
derived products from 6-8% to 52%. If these products were
derived from the excited state, the increasing importance of the
protonation pathway would have resulted in an increase in the
quantum yield, in fact an almost doubling in the 1 M acid.
Thus we conclude that the precursor to the phenylnitrenium
ion is singlet phenylnitrene and that this species can be trapped
by protonation. On this basis eq 9 can be written for the
d
w
d
OH
k
k
k
Kw
p
1
+
1
K ( ArNH a ArN) )
)
(8)
a
p
H
k
w
This equation of course refers to the interconversion of the
two species in their singlet electronic states. The ground state
of most arylnitrenes is the triplet, calculated for example in the
case of phenylnitrene to be ∼20 kcal/mol more stable than the
36
singlet. Most arylnitrenium ions on the other hand are ground
state singlets.3 This is seen experimentally in their nucleophilic
addition chemistry, a reaction pathway argued to be a charac-
7
+
dependence of the products on H concentration,
38
teristic of a singlet. Theoretical calculations also show singlet
3
9
ground states; a recent high-level calculation for example
places the singlet phenylnitrenium ion 21 kcal/mol below the
triplet.3 An interesting consequence of the different multiplici-
ties of the ground state nitrene and nitrenium ion is that the
acidity constant of a triplet arylnitrenium ion deprotonating to
a triplet arylnitrene is many orders of magnitude greater than
p
H
p
w
k
k
+
[
H ] +
9f
(
)
(
)
kexp
kexp
F
)
(9)
aniline
p
p
k
k
H
+
w
[
H ] +
+ 1
(
)
(
)
kexp
kexp
(
32) Leyva, E.; Platz, M. S.; Persy, G.; Wirz, J. J. Am. Chem. Soc. 1986,
1
08, 3783.
where Faniline is the fraction of the reacted phenyl azide that is
accounted for by the aniline products 23, 24, and 25. The
assumptions that are made are that there is no reversal of the
protonation, ring expansion, or intersystem crossing, that ring
expansion is more important than intersystem crossing (kexp >
(
33) Marcinek, A.; Leyva, E.; Whitt, D.; Platz, M. S. J. Am. Chem. Soc.
1
993, 115, 8609.
(
34) (a) As in the Bamberger rearrangement. For recent references see:
(
b) Sone, T.; Tokudo, Y.; Sakai, T.; Shinkai, S.; Manabe, O. J. Chem. Soc.,
Perkin Trans. 2 1981, 298. (b) Sone, T.; Hamamoto, K.; Seiji, Y.; Shinkai,
S.; Manabe, O. J. Chem. Soc., Perkin Trans. 2 1981, 1596. (c) Kohnstam,
G.; Petch, W. A.; Williams, D. L. H. J. Chem. Soc., Perkin Trans. 2, 1984,
1
7,32,41
kISC),
and that the three anilines account for all of the
4
23.
35) For examples see refs 27, 34c, and: Biggs, T. N.; Swenton, J. S. J.
Am. Chem. Soc. 1993, 115, 10416.
36) (a) Kim, S.-J.; Hamilton, T. P.; Schaefer, H. F. J. Am. Chem. Soc.
992, 114, 5349. (b) Hrovat, D. A.; Waali, E. E.; Borden, W. T. J. Am.
intermediate nitrenium ion that is formed. We do not quanti-
tatively account for all of the products from the phenyl azide,
especially in the more dilute acids, but assume that this is caused
(
(
3
1
by other reactions involving the dehydroazepine.
The HCl product data in Table 1 were fit to eq 9 to provide
1
Chem. Soc. 1992, 114, 8698. (c) For experimental evidence, see: Travers,
M. J.; Cowles, D. C.; Clifford, E. P.; Elison, G. B. J. Am. Chem. Soc. 1992,
p
H
-1
p
w
values of k :kexp ) 1.1 ( 0.1 M and k :kexp ) 0.07 ( 0.01
1
14, 8699. (d) Interestingly, the singlet state that has the lowest energy is
(
see Figure 4). As will be shown singlet arylnitrenes are
an open shell species.
(
37) The N-tert-butyl(2-acetyl-4-nitrophenyl)nitrenium ion recently stud-
relatively strong bases, stronger in fact than ammonia and
1
2
ied by Srivasta and Falvey is an exception.
(
38) Anderson, G. B.; Yang, L. L.-N.; Falvey, D. E. J. Am. Chem. Soc.
993, 115, 7254.
39) (a) Ford, G. P.; Scribner, J. D. J. Am. Chem. Soc. 1981, 103, 4281.
b) Glover, S. A.; Scott, A. P. Tetrahedron 1989, 45, 1763. (c) Li, Y.;
(40) The two ratios for [24]:[25] are within experimental error the same.
The ratios for ([24] + [25]):[23] (0.369 ( 0.026 for PhN3 and 0.458 (
0.021 for PhNHOH) may be statistically different. A possible explanation
is that the phenylnitrenium ion, which is very short-lived,2 reacts in a
different state of ion pairing, or perhaps in a different vibrational state,
when formed from the two precursors.
1
(
4b
(
Abramovitch, R. A.; Houk, K. N. J. Org. Chem. 1989, 54, 2911. (d) Ford,
G. P.; Herman, P. S. J. Mol. Struct. (Theochem.) 1991, 236, 269. (e) Falvey,
D. E.; Cramer, C. J. Tetrahedron Lett. 1992, 33, 1705. (f) Cramer, C. J.;
Dulles, F. J.; Falvey, D. E. J. Am. Chem. Soc. 1994, 116, 9787.
(41) This is demonstrated in the present experiments by the relatively
high yield of the azepinone 22 for the non-nitrenium products.

4
800 J. Am. Chem. Soc., Vol. 118, No. 20, 1996
McClelland et al.
4-Biphenylyl Azide: Products. The relevant data concerning
the nitrenium ion in this system come from the study of the
2
7
pivalate 33 by the Novak group. This ester was found to
undergo relatively rapid solvolysis in 5% acetonitrile-water
resulting in two products. One, formed in 7% yield, retained
the pivaloyl group, and was attributed to internal trapping at
the stage of a nitrenium-pivalate ion pair. The other, formed
in 91% yield, was the cyclohexadienone 26. This was not an
initial product; a precursor with a half-life of several hours was
observed in the HPLC analysis and attributed to the imine 27.
The addition of sodium azide to the solvolysis caused no
increase in the rate of reaction of 33, but resulted in the
formation of the azido adduct 34 in place of 26. This is evidence
that the intermediate 4-biphenylylnitrenium ion forms in a rate-
limiting step and subsequently partitions between water and the
azide ion. Quantitative analysis of the ratio of 34 to 26 provided
Figure 4. Yield of aniline products {%(23 + 24 + 25):100} as a
function of HCl concentration. The points are experimental. The line
has been drawn according to eq 9 using the parameters given in the
text.
+
alkylamines. These amines are protonated by aqueous H with
10
-1 -1 42
p
H
rate constants of 2-4 × 10
lie within this range. This means that kexp is also 2-4 × 10
with s units), precisely in the range previously estimated by
M
s , and thus k must also
1
0
-
1
(
the Platz group, except with tighter limits.
One result that surprised us was the formation of a small
amount of nitrenium-derived products at very low acid con-
centration, indicating that the solvent was sufficiently acidic to
3
-1
the trapping ratio kaz:kw as (2.9 ( 0.2) × 10 M .
p
w
p
H
trap the singlet nitrene. Quantitatively k :k is 0.06 ( 0.02 M,
The formation of 26 in high yields, coupled with the HPLC
observation of its precursor, shows that there is substantial
formation of the same nitrenium ion in the photolysis of
p
9 -1
which means that k is 1-2 × 10 s . The magnitude of this
w
rate constant indicates that singlet phenylnitrene is relatively
basic, as shown by the following calculation of the upper limit
4
-biphenylyl azide. Assuming that free 4-biphenylylnitrenium
on the acidity constant of the phenylnitrenium ion. This
ion converts quantitatively to 26, as implied by the Novak result,
88% of the photolysis in 10% acetonitrile goes through the
nitrenium pathway. This is substantially higher than observed
with phenyl azide, and moreover it occurs without added acid,
i.e. by a pathway involving protonation by water. As would
be expected for such a pathway, the yield of 26 does decrease
with decreasing water content, and by 90% acetonitrile there is
p
calculation uses the lower limit for k and an upper limit for
w
d
OH
10
-1 -1
k
of 4 × 10
M
s , based on data for ammonium ions
and hydroxide. From eq 8, pKa( PhNH a PhN) is then
42
1
+
1
calculated to be greater than 12.4.
Transients Observed upon Phenyl Azide Photolysis in
Acids. The transient that can be recognized from previous
17
46
studies of phenyl azide is the dehydroazepine 28. This species
results in a broad absorbance in the 300-400-nm region and is
clearly present in the more dilute acids of our experiments. This
intermediate is of course the precursor of the azepinone 22. The
dehydroazepine transient however was only seen in solutions
more dilute that ∼0.1 M acid, in spite of the fact that 22
continued to form in reasonable yield in stronger acids.
Although detailed kinetic experiments were not performed, the
explanation appears to lie in the reaction of the dehydroazepine
little or no protonation. Under these conditions, the photolysis
appears to follow a similar course to that seen previously in a
2
9
study in cyclohexane solvent, with formation of the azo and
amino derivatives.
4-Biphenylyl- and 2-Fluorenylnitrenium Ions. There is
substantial evidence that the transient with λmax at 460 nm in
the LFP experiments with 4-biphenylyl azide is the nitrenium
ion.
(i) Such an intermediate is consistent with the formation of
+
being H -catalyzed, with the decay becoming too rapid in the
2
6. It can be noted that experiments with varying laser
more concentrated acids for detection on the >20 ns time scale
of our apparatus. This catalysis is not surprising since the
dehydroazepine is a ketenimine, and such compounds do hydrate
in a reaction that is efficiently catalyzed by acids.43
intensities demonstrated that the transient absorbance was linear
in the laser dose. Moreover, 26 was demonstrated to form in
the same yield by HPLC analysis following laser irradiation.
Thus, the laser photochemistry is the same as that observed on
steady-state photolysis.
The new feature of our experiments is the transient at lower
wavelengths. This species increases with increasing acidity in
a manner similar to the aniline products, implying that it is an
intermediate on the nitrenium side of Scheme 1. The phenylni-
trenium ion itself can be ruled out since its lifetime in water is
only 100-200 ps.2 Our assignment therefore is the cyclo-
(
ii) More importantly, there is an excellent linear correlation
47
(
not shown) between the initial absorbance at 460 nm and
the percent yield of 26 (Table 2), unequivocally demonstrating
that the transient is a precursor of the product.
4b
hexadiene 32, or more particularly in the acid solutions of our
(45) For previous LFP studies involving protonated aromatics see
Steenken, S.; McClelland, R. A. J. Am. Chem. Soc. 1990, 112, 9648.
Mathivanan, N.; Cozens, F.; McClelland, R. A.; Steenken, S. J. Am. Chem.
Soc. 1992, 114, 2198. Lew, C. S. Q.; McClelland, R. A. J. Am. Chem. Soc.
1993, 114, 2198. Zhang, G.; Shi, Y.; Mosi, R.; Ho, T.; Wan. P. Can. J.
Chem. 1994, 72, 2388. McClelland, R. A.; Cozens, F.; Steenken, S.
Tetrahedron Lett. 1990, 3821.
studies, its conjugate acid 35.44 This species is the cyclohexa-
dienyl cation (e.g. 36) that is formed by protonating p-
aminophenol at the carbon para to the amino group. What is
being observed therefore in the decay of this transient is the
deprotonation by water to form the aminophenol.45
(46) In addition to a decreased rate of protonation associated with the
lower water concentration, an added factor may be that there is insufficient
water to efficiently solvate hydroxide ion, so that the ion pair ArNH ‚ OH
cannot separate.
+
-
(
42) Eigen, M. Angew. Chem., Int. Ed. Engl. 1964, 3, 1.
(43) McCarthy, D. G.; Hegarty, A. F. J. Chem. Soc., Perkin Trans. 2
1
980, 579.
44) (a) The pKa value for protonated benzoquinone imine is 3.7.44b (b)
Novak, M.; Martin, C. A. J. Org. Chem. 1991, 56, 1585.
(47) The experiments measuring this quantity reported in Table 2 were
carried out in such a way that the ∆OD values represent relative quantum
yields for formation of the transient.
(

Acid-Base Properties of Arylnitrenium Ions
iii) The transient is quenched by azide ion, with kaz:kw
J. Am. Chem. Soc., Vol. 118, No. 20, 1996 4801
(
Scheme 2
calculated from the directly measured rate constants (2.8 ( 0.1)
3
-1
×
10 M identical to the ratio found by the Novak group.
This identity is important in an additional manner since it
establishes that the flash photolysis intermediate is a ground
state species.
(iv) A transient with a very similar λmax is observed for the
N-acetyl-4-biphenylylnitrenium ion generated by a completely
different photoreaction (eq 5).13
(
v) The 4-biphenylylnitrenium ion 7 (eq 2), which was
intersystem crossing that results in “nitrene products”. This is
followed by processes occurring on the ns time scale whereby
that fraction of nitrenium ion that has formed in the first step
reacts with water and hydroxide leading to 26. In competition
there is deprotonation by hydroxide to the singlet nitrene which
on this time scale rapidly partitions back to nitrenium ion and
on to the nitrene products.
7
characterized by NMR, exhibits a λmax of 502 nm. The 460-
nm λmax observed for the parent is thus consistent, especially
considering that 7 has a dimethylamino substituent in the second
ring and this might be expected to increase λmax.
A similar transient, albeit two orders of magnitude longer-
lived, is observed on irradiation of 2-fluorenyl azide, and can
be assigned to the corresponding nitrenium ion. Although
product evidence is lacking with this system, the similarity
with the spectrum for the 4-biphenylylnitrenium ion, the efficient
quenching by azide ion, and the close similarity to the spectrum
of the N-acetyl-2-fluorenylnitrenium ion13 place little doubt on
this assignment.
Precursor of the 4-Biphenylyl- and 2-Fluorenylnitrenium
Ions. As in the case of the phenyl azide, we will argue that
the precursor of these nitrenium ions is the singlet nitrene. Two
pieces of evidence can be cited to rule out the excited state of
the azide.
The assumption that is made in this treatment is that reactions
involving singlet nitrene are faster than reactions involving
48
p
w
-
nitrenium ion, i.e. that (k + kISC + k ) > (k + k [ OH] +
exp
w
OH
d
-
k
[ OH]). Evidence that this is the case will be seen in the
OH
numbers derived below. Also noteworthy is the observation in
the ns LFP experiments that the initial optical density of the
nitrenium ion is unchanged even in quite concentrated base.
This indicates that the quantity of nitrenium ion is still being
determined by the initial partitioning of the singlet nitrene, and
that on the time scale of this partitioning, reactions involving
hydroxide ion and nitrenium ion are unimportant.
(
i) In the biphenylyl system the relative quantum yield for
disappearance of the substrate is essentially unchanged from
0% to 90% acetonitrile, in spite of a decrease in the yield of
In terms of this model eqs 11 and 12 are derived for the rate
constant for decay of the nitrenium ion and the fractional yield
of the product 26
2
the nitrenium ion from 84% to almost zero. If the excited state
were the precursor the loss of this photochemical pathway would
surely have had an effect.
kdecay ) k + k [ OH] + (1 - P)k [-OH] (11)
-
d
OH
w
OH
kw + kOH[-OH]
(
ii) Picosecond spectroscopy with fluorene has recently
showed a transition attributed to the singlet excited state near
00 nm.49 Unless the azide substituent perturbs the system
F ) P
(12)
2
6
(
kw + k [ OH] + (1 - P)k [-OH]
-
d
OH
)
7
OH
sufficiently to move the transition out of the observable region,
the absence of such a transition with 2-fluorenyl azide implies
that even after 25 ps the singlet excited state has lost N2.
The rate constant for the appearance of the nitrenium ion
p
p
w
where P ) k
:(kexp + kISC + k
quenching with hydroxide ion, but the slope (kOH(obs)) is equal
to kOH + (1 - P)k and does not represent solely nucleophilic
quenching. Equations 11 and 12 contain four independent
). Equation 11 predicts a linear
w
d
OH
measured in the ps experiment represents the reactions of the
p
w
d
singlet nitrene, and is therefore equal to kexp + kISC + k . In
variables, P, kw, kOH, and k . Values for P (0.84) and kw
OH
6
-1
the case of biphenylyl this can be broken down further since
(1.85 × 10 s ) were taken from product and LFP data in
p
w
p
w
-
neutral solutions, and the term in [ OH] in the denominator of
k :(kexp + kISC + k ) is known from the yield of the nitrenium
d
p
eq 12, (kOH + (1 - P)k ) was taken as the value of kOH(obs)
product 26. Thus for 20% acetonitrile k is calculated as 5 ×
OH
w
7
-1 -1
9
-1
9
-1
(6.4 × 10 M s ). This left one unknown, kOH, and the
experimental data were fit to eq 12 to obtain its value. The fit
showing the good agreement of the model to the experimental
data is shown in Figure 5. The value of kOH so obtained was
1
0 s and (kexp + kISC) as 1 × 10 s . These numbers show
why the amount of nitrenium ion produced by solvent proto-
nation is so much greater with the biphenylyl system when
compared to phenyl. The rate of protonation is in fact only
7
-1 -1
1
.4 × 10 M s , and from this value and kOH(obs), the value
2.5-5 times greater for biphenylyl, but the lifetime of the singlet
d
OH
8 -1 -1
biphenylylnitrene with respect to ring expansion and intersystem
crossing is 20-40 times longer.
of k was obtained as 3.1 × 10 M s .
p
w
d
OH
The ratio k :k is 16 and is equal to the basicity constant
Acidity of the 4-Biphenylylnitrenium Ion. In this section
Kb for the singlet 4-biphenylylnitrene in 20% acetonitrile-water.
p
w
d
OH
we will combine this value of k with a value of k calculated
1
1
+
-
from data obtained in base. The analysis is based on Scheme
where the singlet nitrene reacts by processes occurring on
4-PhC H N + H O a 4-PhC H NH + OH (13)
6 4 2 6 4
2
the ps time scale that involve partitioning between protonation
to the nitrenium ion and a combination of ring expansion and
This means that a hypothetical solution containing 16 M
hydroxide ion is required for a 1:1 mixture of the singlet nitrene
and nitrenium ion at equilibrium. In 1 M hydroxide, there is
only about 5% nitrene. Using pKw ) 14.8 in 20% acetonitrile
(see Experimental Section), the acidity constant for the nitrenium
(48) (a) With this azide, HPLC analysis immediately after irradiation
shows a single major peak which we assign to the imine that is the fluorenyl
analog of 27. Unlike 27, however the fluorenylimine reacts to give several
new products which we have not identified. Similar observations have been
made in solvolysis reactions proceeding through this nitrenium ion.4 (b)
Novak, M. Personal communication.
-
16
ion (eq 8) is calculated as 10
M. In our preliminary report
8b
we suggested that the singlet nitrene was approaching the
2
2
category of a strong base in water. These absolute numbers
show the extent to which this is indeed true.
(
49) McGowan, W. M.; Hilinski, E. F. J. Am. Chem. Soc. 1995, 117,
9
019.

4
802 J. Am. Chem. Soc., Vol. 118, No. 20, 1996
McClelland et al.
Figure 5. Yield of 4-hydroxy-4-phenyl-2,5-cyclohexadienone (26) as
a function of NaOH concentration following irradiation of 4-biphenylyl
azide. The line has been drawn on the basis of eq 12 using the
6
-1
7
-1 -1
parameters P ) 0.84, k
w
) 1.85 × 10 s , kOH )1.4 × 10 M
s ,
d
8
-1 -1
and k ) 3.1 × 10 M
s .
OH
A reviewer has suggested that the deprotonation reaction
produces the ground state triplet nitrene. Such a possibility can
Figure 6. Corrected rate constants51 for the decay of the 4-biphenylyl-
and 2-fluorenylnitrenium ions in HClO solutions, with ionic strength
4
1
be added to Scheme 2, through an additional pathways Ar-
+
-
3
3 d
OH
NH
+
OH f ArN, with a rate constant defined as k ).
1.0 maintained with NaClO . Curves are drawn according to eq 15
4
Equation 11 is then modified by adding a further term in
hydroxide ion, k [ OH], and the same term is added to the
using the parameters given in the text.
3
d
OH
-
Table 5. Acidity Constants for Aniline Dications, and Rate
Constants for Reaction of Nitrenium Ions with Water (20%
Acetonitrile-Water, 20 °C, 1 M NaClO4)
denominator of eq 12. The analysis proceeds as above and
d
OH
results in the same value of kOH plus the quantity (1 - P)k
+
3
d
7
-1 -1
k
) 5.0 × 10 M s , where the contributions from the
OH
Ar )
two deprotonation pathways cannot be separated. With this
d
OH
constant
4-biphenylyl
0.1
2-fluorenyl
0.6
model, the value of k and acidity constant calculated above
+
pK
a
+
(ArNH
2
)2
are upper limits, corresponding to the situation where the
2
(ArNH)2+, s
-1
-1
7
6
3
d
OH
k
6.0 × 10
1.9 × 10
w
k
term makes a negligible contribution. If deprotonation
+
+
6
4
1.1 × 10
1.3 × 10
k (ArNH) , s
w
were to be occurring to the triplet nitrene, the singlet nitrene
must be an even stronger base than we have calculated.
Basicity of the 4-Biphenylyl- and 2-Fluorenylnitrenium
Ions. Our model to account for the kinetic behavior in acids
+
+
-1
6
4
a
3
.4 × 10
(4.4 × 10 )
k (ArNAc) , s
w
a
4
Not measured in 1 M NaClO . This number is calculated from
4
-1
the observed rate constant, 7.7 × 10 s in 0.5 M NaClO
4
, with the
assumption that the effect of ionic strength is the same as it is on the
(Figure 3) involves a mechanism where the nitrenium ion is
N-acetyl-4-biphenylylnitrenium ion.
further protonated to a more reactive dication, as illustrated in
5
0
eq 14. The behavior of the N-acetyl substituted cation is
are shown in Figure 6, and the parameters obtained in these
fits are given in Table 5.
This treatment produces pKa values of 0.1 for the 4-amino-
biphenyl dication and 0.6 for the 2-aminofluorene dication. In
terms of the conjugate bases, the nitrenium ions, this means
that these are weakly basic species. They are however suf-
ficiently basic to be significantly protonated in 1.0 M aqueous
acid, especially the fluorenyl derivative. In terms of superacids
where NMR spectroscopic detection of nitrenium ions might
be possible, this will mean that the species that will be observed
inevitably will be a dication. Equation 3 given in the introduc-
tion shows an example from the Olah group.
consistent with this interpretation. The amide-type nitrogen of
this ion will not be protonated, at least in the solutions involved
here, and indeed the rate constants do not increase. There is in
fact a 20% decrease from dilute to 1 M acid which we attribute
to a specific salt effect associated with replacing NaClO4 with
HClO4.
Previous experimental evidence has suggested that arylnitre-
nium ions are perhaps better regarded in terms of carbenium
According to this model,
3
4c
+
w
2+
2+
+
ion resonance contributors (e.g. 37).
The above basicities
k K (ArNH ) + k [H ]
a
2
w
are fully consistent with such a structure, since these are what
might be expected for an imine with a nearby positive charge.
Further evidence is seen in the remarkably little difference in
the absorption spectra of N-H and N-acetyl nitrenium ions, i.e.
the chromophore is the cyclohexadienyl cation. Also consistent
are the relatively small increases in the water rate constants
kobs
)
(15)
2+
+
K (ArNH ) + [H ]
a
2
and rate constants are predicted to increase at some acidity as
2
w
+
the k pathway starts occurring but then plateau as the
equilibrium shifts to the dication. To fit the data to eq 16, values
(
Table 5) on substituting NH in a nitrenium ion with the
of kobs were first corrected for the specific salt effect observed
5
with the N-acetyl compound.51 The fits to the experimental data
electron-withdrawing N-acetyl group. Even full protonation
has little effect on the spectrum and results in rate increases of
only two orders of magnitude.
(
(
50) The constant k+ is the same as kw employed previously.
51) Fits of eq 15 to the experimental data were significantly poorer if
w
this was not done. The corrected rate constants were obtained by multiplying
kobs in each acid by the ratio kobs(N-acetyl, 1 M NaClO4):kobs(N-acetyl),
where kobs(N-acetyl) refers to the same acid.
(52) (a) This is seen with other nitrenium ions.52b (b) Novak, M.; Kahley,
M. J.; Lin, J.; Kennedy, S. A.; Swanegan, L. A. J. Am. Chem. Soc. 1994,
116, 11626.

Acid-Base Properties of Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 20, 1996 4803
The reactions of nitrenium ions with heteroatom nucleophiles
also show that they are reacting as carbenium ions. This is
seen not only in the nature of products, but in the reactivity
patterns. Thus nitrenium ions react with azide at the diffusion
azide, the detector wavelength was 250 nm, and the eluting solvents
were 0.05 M 1:1 HOAc-NaOAc and acetonitrile. The products were
eluted in the sequence 26, 27, and 4-biphenylyl azide by use of a
gradient program beginning with 40% acetonitrile for 12 min, then
changing to 80% acetonitrile over 5 min. Quantitative measurements
were based on the comparison of peak areas with areas obtained with
authentic samples (except for the unstable 27 which was not analyzed
quantitatively). The photolyses were carried out using a Rayonet
Photochemical Reactor equipped with lamps operating at 300 nm. In
the case of the biphenylyl system experiments were also performed
following one pulse of irradiation in the LFP apparatus. These showed
no difference from results obtained with the Rayonet. Solutions
contained 20-100 µM substrate, and were directly injected into the
HPLC for analysis. Irradiation times were of the order 0.5-1.5 min.
In the case of 26 which formed slowly over time, injections were made
the following day. For the experiments involving 4-biphenylyl azide
in NaOH, solutions were irradiated for 0.5 min, then an amount of
HCl equivalent to the NaOH was added to neutralize the solution, along
with a quantity of 1:1 HOAc-NaOAc buffer to set the pH at ∼5. This
solution was then allowed to stand overnight before analysis of 26.
The pH in the mixed acetonitrile-solvents was measured using a
combination electrode that had been calibrated with standard aqueous
9
-1 -1
limit with rate constants of 5 × 10 M s , the same as
5
3
carbenium ions.
The changes seen in kw on increasing
acetonitrile content in the mixed aqueous solvents (Table 2)
are also identical to what has been observed previously with
5
4
carbenium ions.
Summary. These studies have shown that singlet phenylni-
trene, 4-biphenylylnitrene, and 2-fluorenylnitrene are relatively
+
strong bases capable of being intercepted by protonation by H
and even water to give nitrenium ions in competition with other
reaction channels such as intersystem crossing and ring expan-
+
sion. In the case of phenylnitrene the H trapping competes
with a ring expansion that occurs with a rate constant of 2-4
1
0
-1 -1
×
10
M
s . Protonation by solvent water also occurs, but
9
-1
even though the rate constant is of the order of 10 s , it
constitutes a minor pathway in competition with the ring
expansion. In the cases of 4-biphenylyl- and 2-fluorenylnitrene,
ring expansion and intersystem crossing occur an order of
magnitude more slowly, water protonation is faster, and there
are substantial yields of nitrenium ion without added acids. In
these cases the nitrenium ions are sufficiently long-lived that
they are detected by ns LFP. A combination of ps LFP, ns
LFP, and product analysis provides pKa ) 16 for the 4-biphen-
ylylnitrenium ion deprotonating to singlet nitrene in 20%
acetonitrile. Thus singlet 4-biphenylylnitrene falls close to the
category of a strong base in this solution. The LFP experiments
under acidic conditions show kinetic behavior consistent with
protonation to a dication. Analysis of the kinetic curves
furnishes the pKa values for the dications as 0.1 (4-biphenylyl)
and 0.6 (2-fluorenyl). This and other pieces of evidence are
argued to be consistent with these arylnitrene ions being better
regarded as imine-substituted cyclohexadienyl cations. Overall,
buffers, and a small empirical correction factor “Q ” was added to the
s
measured pH. This factor was the value required to adjust the measured
pH of a solution of 0.01 M HCl to a value of 2.00. The pK
acetonitrile was obtained from the measured pH of a 0.01 M NaOH
solution, according to the equation pK
.00 + Q
Phenyl azide, 4-biphenylyl azide, and 2-fluorenyl azide were known
compounds and were prepared by standard routes. 4-Chloroaniline and
w
in 20%
w
) pHmeasured(0.01 M NaOH) +
2
s
.
2
4
-chloroaniline were commercial samples and were distilled before use.
-Aminophenol hydrochloride was used as received. 4-Hydroxy-4-
phenyl-2,5-cyclohexadienone (26) was obtained by irradiating at 300
nm a solution of 0.2 g of 4-biphenylyl azide in 1 L of 20% acetonitrile-
water until HPLC analysis indicated less than 5% azide. The solution
was then allowed to stand overnight to allow 27 to convert to 26, at
which time the solvent was completely removed, first with a rotary
evaporator and then with a freeze-dryer. The remaining sample was
+
recrystallized from ethanol-water and had NMR and melting point
arylnitrenium ions (ArNH ) are very weak acids in water when
identical with a sample provided by Mike Novak.2
6,27
1H-Azepine-
deprotonating to singlet nitrenes and are also weak bases,
accepting a proton to form the aniline dication.
2
0
8
5
(3H)-one (22) was prepared by 300-nm irradiation of a solution of
.2 g of phenyl azide in 2 L of 0.005 M HCl. After HPLC indicated
0% conversion of the phenyl azide, the solvent was concentrated to
0 mL, and extracted with dichloromethane. The dichloromethane layer
Experimental Section
Flash photolysis experiments were carried out in the standard fashion
with ca. 20 ns pulses at 248 nm (ca. 60 mJ per pulse) from a Lumonics
excimer laser (KrF emission) and with ca. 25-ps pulses at 266 nm (ca.
4
was washed with sodium bicarbonate and dried (MgSO ), and after
filtration the solvent was removed. The sample of 22 was then isolated
by semipreparative HPLC with a reversed-phase C18 column and an
eluting solvent of 30% methanol-70% water. The NMR of this
4
mJ per pulse) from a Continuum YG-601-C Nd/YAG laser.
Product analyses were performed with a Waters HPLC with UV-
samplesδ(CDCl ) 2.92 (d, 2H), 5.55-5.65 (m, 1H), 5.8-5.9 (m, 1H),
3
visible detector, using a µ-Bondapak C18 reversed phase column. In
the case of phenyl azide, the detector wavelength was 235 nm, and the
eluting solvents were 0.001 M HCl and methanol. The products were
eluted in the sequence 23, 24, 25, 22, and then phenyl azide by use of
a gradient program beginning with 10% methanol for 10 min, then
changing to 70% methanol over 15 min. In the case of 4-biphenylyl
6.2-6.3 (m, 2H), 8.3 (broad s, 1H)swas identical with those reported
in the literature.1
8a,30
Acknowledgment. Continued financial support of the Natu-
ral Sciences and Engineering Research Council of Canada is
gratefully acknowledged. Dr. Mike Novak is thanked for his
sample of 26 and for a number of helpful discussions. Dr. Fran
Cozens is gratefully acknowledged for the ps LFP experiments.
(
53) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N.; Steenken,
S. J. Am. Chem. Soc. 1991, 113, 1009.
54) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N.; Steenken,
S. J. Am. Chem. Soc. 1989, 111, 3966.
(
JA954248D