Substituent effects on the lifetimes and reactivities of arylnitrenium ions studied by laser flash photolysis and photothermal beam deflection

Article abstract of DOI:10.1021/ja961518z

Photolysis of 5-substituted-N-tert-butyl-3-methylanthranilium ion produces transient N-tert-butyl-(2-acetyl-4-substituted) phenylnitrenium ions. This is confirmed by identification of the stable products, transient absorption experiments, and photothermal

Full text of DOI:10.1021/ja961518z

J. Am. Chem. Soc. 1996, 118, 8127-8135
8127
Substituent Effects on the Lifetimes and Reactivities of
Arylnitrenium Ions Studied by Laser Flash Photolysis and
Photothermal Beam Deflection
Rebecca J. Robbins, David M. Laman, and Daniel E. Falvey*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Maryland,
College Park, Maryland 20742
ReceiVed May 6, 1996X
Abstract: Photolysis of 5-substituted-N-tert-butyl-3-methylanthranilium ion produces transient N-tert-butyl-(2-acetyl-
4-substituted) phenylnitrenium ions. This is confirmed by identification of the stable products, transient absorption
experiments, and photothermal beam deflection experiments. Analysis of the stable photoproducts shows that the
major decay pathway for these species is addition of nucleophiles to the aromatic ring. The kinetics of the these
reactions were examined with the goal of determining how various ring substituents affect arylnitrenium ion stability.
Rate constants measured for the 4-phenyl and 4-methoxy derivatives are compared with those from previous work.
It is shown that a 4-phenyl group stabilizes the nitrenium ion to approximately the same extent as a 4-methoxy
group. Both of these substituents stabilize the arylnitrenium ions considerably more than 4-halogens or a 4-methyl
group. Time resolved photothermal beam deflection experiments were applied to the 4-cyano and 4-unsubstituted
derivatives, which gave no transient absorption spectra. The latter two compounds are shown to have lifetimes of
less than 100 ns.
Introduction
The mutagenic potency of a given arylnitrenium ion depends
upon a balance of reactivity and stability. The nitrenium ion
in question must be reactive enough to add rapidly and
irreversibly to the DNA molecule. On the other hand, it must
be stable enough to survive attack by other nucleophiles present
in the physiological environment such as water and glutathione.
Novak et al.14-16 have demonstrated that certain arylnitrenium
ions which possess para-π-electron donating groups (such as
4-biphenylylnitrenium ion and 2-fluorenylnitrenium ion) meet
this criterion. Apparently, the presence of an additional aromatic
ring in the position para to the nitrenium center confers sufficient
stability to the nitrenium ion that it can survive attack by water
yet still add to guanine bases. In contrast, simple phenylnitre-
nium ions which lack such substituents in the para-position have
very short lifetimes in aqueous solution. For example Fishbein
and McClelland17 estimate the lifetime of the unsubstituted
phenylnitrenium ion at ca. 200 ps.
Arylnitrenium ions are cationic reactive intermediates that
contain a divalent nitrogen atom having a phenyl or other
aromatic ring directly attached to the nitrogen (Ar-N-R+).1,2
Investigation into the behavior of arylnitrenium ions has been
motivated in part by the proposal that these species are
intermediates in physiological DNA-damaging reactions which
are responsible for carcinogenesis.3-7 The in ViVo mechanisms
of carcinogenesis are complex, and a full picture of this process
is still emerging. It has been proposed that aromatic amines
are enzymatically oxidized to hydroxylamines, which in turn
are enzymatically converted to hydroxylamine O-sulfate (or
O-acetate) esters.8 Heterolytic scission of the N-O bond9,10 is
thought to generate a transient arylnitrenium ion which co-
valently adds to guanine bases in DNA.4,5,11-13
X Abstract published in AdVance ACS Abstracts, August 15, 1996.
(1) Abramovitch, R. A.; Jeyaraman, R. In Azides and Nitrenes: ReactiVity
and Utility; Scriven, E. F. V., Ed.; Academic: Orlando, FL, 1984; pp 297-
357.
(2) Simonova, T. P.; Nefedov, V. D.; Toropova, M. A.; Kirillov, N. F.
Russ. Chem. ReV. 1992, 61, 584-599.
(3) Frederick, C. B.; Beland, F. A. In Polycyclic Aromatic Hydrocarbon
Carcinogenesis: Structure-ActiVity Relationships; Yang, S. K., Silverman,
B. D., Eds.; CRC: Boca Raton, FL, 1988; Vol. II, pp 111-139.
(4) Kadlubar, F. F. In DNA Adducts Identification and Significance;
Hemminki, K., Dipple, A., Shuker, D. E. G., Kadlubar, F. F., Segerba¨ck,
D., Bartsch, H. Eds.; University Press: Oxford, UK, 1994; pp 199-216.
(5) Kadlubar, F. F.; Beland, F. A. In Polycyclic Hydrocarbons and
Carcinogesis; Harvey, R. G., Ed.; American Chemical Society: Washington,
DC, 1985.
(6) Turesky, R. J.; Lang, N. P.; Butler, M. A.; Teitel, C. H.; Kadlubar,
F. F. Carcinogenesis 1991, 12, 1839-1845.
(7) Turesky, R.; Markovic, J. Chem. Res. Toxicol. 1994, 7, 752-761.
(8) King, C. M.; Phillips, B. Science (Washington, DC) 1968, 159, 1351-
1353.
(9) Novak, M.; Roy, A. K. J. Org. Chem. 1985, 50, 571-580.
(10) Novak, M.; Pelecanou, M.; Pollack, L. J. Am. Chem. Soc. 1986,
108, 112-120.
(11) Bosold, F.; Boche, G. Angew. Chem., Int. Ed. Engl. 1990, 29, 63-
64.
Work in this laboratory has employed anthranilium salts 118
as precursors for studies of substituted alkylaryl nitrenium ions
of general structure 2.19-22 On the basis of product studies and
LFP experiments, we have proposed the general photochemical
mechanism shown in Scheme 1. Direct irradiation of 1 gives
(13) Moore, P. D.; Rabkin, S. D.; Osborn, A.; King, C. M.; Strauss, B.
S. Proc. Natl. Acad. Sci. U.S.A. 1982, 79, 7166-7170.
(14) Novak, M.; Kennedy, S. A. J. Am. Chem. Soc. 1995, 117, 574-
575.
(15) Novak, M.; Kahley, M. J.; Eiger, E.; Helmick, J. S.; Peters, H. E.
J. Am. Chem. Soc. 1993, 115, 9453-9460.
(16) Davidse, P. A.; Kahley, M. J.; McClelland, R. A.; Novak, M. J.
Am. Chem. Soc. 1994, 116, 4513-4514.
(17) Fishbein, J. C.; McClelland, R. A. J. Am. Chem. Soc. 1987, 109,
2824-2825.
(18) Haley, N. F. J. Org. Chem. 1977, 42, 3929-3933.
(19) Chiapperino, D.; Anderson, G. B.; Robbins, R. J.; Falvey, D. E. J.
Org. Chem. 1996, 61, 3195-3199.
(20) Anderson, G. B.; Yang, L. L.-N.; Falvey, D. E. J. Am. Chem. Soc.
1993, 115, 7254-7262.
(21) Robbins, R. J.; Falvey, D. E. Tetrahedron Lett. 1994, 35, 4943-
4946.
(12) Humphreys, W. G.; Kadlubar, F. F.; Guengerich, F. P. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 8278-8282.
(22) Robbins, R. J.; Yang, L. L.-N.; Anderson, G. B.; Falvey, D. E. J.
Am. Chem. Soc. 1995, 117, 6544-6552.
S0002-7863(96)01518-1 CCC: $12.00 © 1996 American Chemical Society

8128 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Robbins et al.
Scheme 1
Chart 1. Nucleophilic Addition Products
of the triplet state, and electron donating groups favor the singlet
state. With sufficiently electron-withdrawing groups it is
possible to reverse the state energies and make the triplet the
ground state. Recent experimental work on N-tert-butyl-(2-
acetyl-4-nitrophenyl) nitrenium ion provides indirect evidence
that this arylnitrenium ion might be a ground state triplet.32
The study described below was undertaken with the goal of
learning how various ring substituents affect the lifetimes and
reaction rates of singlet state arylnitrenium ions. In this regard,
it is interesting to contrast the behavior of arylnitrenium ions
with that of the arylcarbenium ions (eq 1), as there is
considerable information about the latter class of
intermediates.33-36 On one hand, it might seem that substituent
effects in the arylnitrenium ions would simply mirror those seen
in the arylcarbenium ions. On the other hand, there are some
significant differences in the mechanism by which the two
intermediates are trapped. Nucleophiles such as alcohols and
water add to the exocyclic atom of the arylcarbenium ions (eq
1a),36 but they add to the para and ortho ring atoms of
arylnitrenium ions (eq 1b).20,37-40 For this reason, steric effects
of para-substituents cannot be neglected in the nitrenium ions
as they often are in the carbenium ions. Resonance and
inductive electronic effects may be felt differently in the two
different species. Novak41 has suggested that the loss of
aromatic character in the transition state for addition to
arylnitrenium ions might cause them to respond differently to
the electronic properties of various substituents than do the
carbenium ions. Furthermore, a nitrenium ion center is much
more electronegative than the corresponding carbenium center
and therefore polarizes the π-electrons of the substituents more
strongly.
the singlet excited state (11*) which can either ring-open to
singlet nitrenium ion (path kso to form 12) or intersystem cross
to triplet excited state anthranilium (path kisc to form 31*). Once
generated, 12 can cyclize to regenerate ground-state anthranilium
ion (kc) or react intermolecularly (knuc) to give nucleophilic
addition products (e.g., 12-16, Chart 1), or intramolecularly
(km) to give rearrangement products (4). If triplet excited state
anthranilium (31*) is generated (by either triplet sensitization
or kisc), then its ring opening generates the triplet nitrenium ion
(32). The latter reacts via sequential H-atom abstractions (kH1
and kH2) to give the parent amine (5).
The nature of the electronic spin state of nitrenium ions has
been examined both theoretically and experimentally. The
parent nitrenium ion, NH2+, is a ground state triplet. Quantum
calculations23 and high-resolution photoelectron spectroscopy24
both give an energy gap (ES - ET) of ca. +30 kcal/mol.
Arylnitrenium ions are generally ground state singlets. Density
functional theory25 as well as Hartree-Fock26 calculations both
predict that phenylnitrenium ion PhNH+ is a ground state singlet
with an energy gap of ca. -20 kcal/mol. Semiempirical
calculations27-31 predict that most substituted arylnitrenium ions
also have singlet ground states. However, the magnitude of
the energy gap is dependent upon the ring substitution pattern.
Electron-withdrawing substituents shift the energy gap in favor
(23) Peyerimhoff, S. D.; Buenker, R. J. Chem. Phys. 1979, 42, 167-
176.
(24) Gibson, S. T.; Greene, J. P.; Berkowitz, J. J. Chem. Phys. 1985,
Previous LFP experiments on arylnitrenium ions 2 showed
that the 4-methyl derivative 2d was significantly more reactive
83, 4319-4328.
(25) Cramer, C. J.; Dulles, F. J.; Falvey, D. E. J. Am. Chem. Soc. 1994,
116, 9787-9788.
(26) Li, Y.; Abramovitch, R. A.; Houk, K. N. J. Org. Chem. 1989, 54,
(32) Srivastava, S.; Falvey, D. E. J. Am. Chem. Soc. 1995, 117, 10186-
10193.
(33) Ritchie, C. D. Can. J. Chem. 1986, 64, 2239-2259.
(34) Mayr, H.; Patz, M. Angew. Chem., Int. Ed. Engl. 1994, 33, 938-
957.
(35) Olah, G. A. Carbocations and Electrophilic Reactions; Wiley: New
York, 1973; p 148.
(36) Freedman, H. H. In Carbonium Ions; Olah, G. A., Schleyer, P. v.
R., Eds.; Wiley-Interscience: New York, 1973; Vol. IV, pp 1501-1578.
(37) Gassman, P. G.; Campbell, G. A. J. Am. Chem. Soc. 1971, 93,
2567-2569.
2911-2914.
(27) Falvey, D. E.; Cramer, C. J. Tetrahedron Lett. 1992, 33, 1705-
1708.
(28) Ford, G. P.; Herman, P. S. J. Mol. Struct. (THEOCHEM) 1991,
236, 269-282.
(29) Ford, G. P.; Herman, P. S. J. Am. Chem. Soc. 1989, 111, 3987-
3996.
(30) Glover, S. A.; Scott, A. P. Tetrahedron 1989, 45, 1763-1776.
(31) Campbell, J. J.; Glover, S. A.; Rowbottom, C. A. Tetrahedron Lett.
1990, 37, 5377-5380.

Substituent Effects on Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 8129
Table 1. Product Distributions (% Yield) from the Photolysis of
Anthranilium Ions 1
Scheme 2
ion
trap (concn)
5
4
12 13 14 15 16 17
1e
1e
1e
1f
H2O (9.2 M)
MeOH (3.8 M)
(none)
H2O (4.1 M)
MeOH (2.1 M) 23
23
91
65a
21
7
77
30
13 17
45 8
8
1f
1f
1g
1g
(none)
MeOH (3.5 M)
(none)
29 25
23
43
60
Scheme 3
a Average of three runs.
Scheme 4
toward nucleophiles than the halogenated derivatives 2b and
2c. This contrasts with the effects of such groups on the stability
of carbenium ions. With the latter, para-halogens act as net
destabilizers and para-alkyl groups act as stabilizers.42,43 In
order to further elucidate electronic and steric effects on the
reactions of arylnitrenium ions with nucleophiles, we have
synthesized photoprecursors (1) to 4-methoxy, 4-phenyl, and
4-cyano substituted N-tert-butyl-(2-acetylphenyl) nitrenium ions
(2). Product studies, laser flash photolysis, and photothermal
beam deflection measurements were undertaken on these
compounds, and the results are compared with the previously
studied nitrenium ions. It is shown that the CN group
destabilizes the arylnitrenium ions relative to the previously
studied examples. Phenyl and methoxy substituents stabilize
arylnitrenium ions to approximately the same extent as each
other and considerably more than the halogens.
Scheme 2. The 5-methoxy derivative (1e) was also synthesized
following Scheme 2.
2. Product Studies. UV irradiation of solutions of the
anthranilium ions 1 gives products derived from the arylnitre-
nium ions 2. There are basically three types of products: (1)
the parent amine 5; (2) an iminium ion, 4; and (3) a number of
products all derived from addition of nucleophiles to the
aromatic ring (12-16).
Irradiation of the methoxy derivative 1e gives the product
distribution listed in Table 1. There are two major products:
the ortho adduct 12e and the dihydroxy adduct 16 (eq 2). The
latter is observed only at relatively high water concentrations.
At low water concentrations or with CH3OH as the trap, only
12e is detected. We suggest that 16 is formed via the
mechanism depicted in Scheme 4. Initial attack at the para
position reversibly forms a hemiketal of the imine quinone (19).
The productive pathway involves a proton-catalyzed elimination
of methanol to give (following deprotonation) the imine quinone
20. The latter is an especially electrophilic imine quinone due
to the electron-withdrawing acetyl group. Michael addition of
water to the 3-position yields 21 which is a tautomer of the
stable product 16. Because 16 is not seen at low [H2O] or with
CH3OH as the trap, it is surmised that the first irreversible step
leading to 16 must involve an additional H2O molecule. One
possibility is that the net elimination of MeOH from 19 is more
sensitive to general acid catalysis than is the elimination of H2O.
Thus water, being more acidic than CH3OH, would be a more
effective catalyst for this reaction.
Results and Discussion
1. Synthesis of Substituted Anthranilium Ion Photopre-
cursors. Compounds 1e-g were prepared from the appropriate
3-substituted acetophenone derivative 6 according to the general
method depicted in Scheme 2. Nitration of 6 followed by
reduction produced the arylamine 8. Conversion of the latter
to the arylazide 9, followed by thermolysis, gave the anthranil
free base 10. The latter was alkylated by the generation of tert-
butyl cation in the presence of a strong acid to give 1.18 The
syntheses of anthranilium salts 1a,20 1b, 1c,22 and 1d21 have
been reported elsewhere.
The phenyl derivative 1f was prepared by a somewhat
different route (Scheme 3). In this case 2-nitro-5-bromoac-
etophenone 7b was first prepared. Stille coupling44 was used
to convert 7b into the 5-phenyl derivative 7f in 70% yield. The
latter was taken through to the anthranil 10f as indicated in
Scheme 2. Likewise the cyano derivative 1g was prepared by
first converting 7b into 2-nitro-5-cyanoacetophenone 7g in 97%
yield using a Pd-coupling procedure described by Kosugi.45 The
product from this reaction was converted into 1g following
(38) Gassman, P. G.; Campbell, G. A. J. Chem. Soc., Chem. Commun.
1970, 1970, 427.
(39) Fishbein, J. C.; McClelland, R. A. J. Chem. Soc., Perkin Trans. 2
1995, 653-662.
(40) Some nucleophiles do add to the nitrogen of nitrenium ions: most
notably guanosine bases of DNA. For example, see ref 14.
(41) Novak, M.; Kahley, M. J.; Lin, J.; Kennedy, S. A.; Swanegan, L.
A. J. Am. Chem. Soc. 1994, 116, 11626-11627.
(42) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195.
(43) McClelland, R. A.; Kanagasabapathy, V. M.; Banait, N. S.; Steenken,
S. J. Am. Chem. Soc. 1991, 113, 1009-1014.
(44) Stille, J. Angew. Chem., Int. Ed. Engl. 1986, 25, 508.
(45) Kosugi, M.; Kato, Y.; Kiuchi, K.; Migita, T. Chem. Lett. 1981, 69-
72.
One interesting feature of 2e is its apparent failure to
isomerize to the iminium ion 4 (km in Scheme 1). Photolysis
of 1e was carried out in nominally anhydrous CH3CN (i.e.,
distilled from CaH2). Nonetheless in all cases a 40-80% yield
of the ortho hydroxy adduct was detected. Thus, if km occurs
at all, it is much slower than the knuc reaction with adventitious
water. Similar behavior has been observed for the methyl
derivative, 2d.21 In contrast, the other nitrenium ions in this
series, 2a, 2b, 2c, 2f, and 2g all give significant yields of 4
under the same conditions.

8130 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Robbins et al.
Photolysis of the 4-phenyl derivative 1f affords several
products, the majority of which originate from para ring addition.
When 1f was irradiated in the presence of MeOH, a small
amount (7%) of ortho adduct 12f (nuc ) CH3O) is observed,
but the major product comes from para addition to give quinol
imine 13f (nuc ) CH3O) in 45% yield (eq 3). Proton-catalyzed
isomerization of the latter occurs slowly under the reaction
conditions, and 14f is observed in 8% yield (eq 4). The other
significant product is the parent amine 5f which occurs from
the parallel triplet pathway. Apparently with the phenyl
derivative the excited state intersystem crossing pathway (kisc
in Scheme 1) becomes competitive with the singlet state ring
opening (kso). In the absence of nucleophiles, the iminium ion
4f is formed.
Water gives products analogous to those from CH3OH, although
there are some differences in the distribution. With H2O,
somewhat higher yields of the ortho adduct 12f (nuc ) OH)
are seen. The major products still result from initial para
addition: namely, the quinol imine 13f (nuc ) OH) in 13%
yield and the addition/rearrangement product 14f (nuc ) OH)
in 17% yield. Small amounts of a product that results from
addition to the phenyl substituent (15, nuc ) OH) are also
observed.
Figure 1. Transient absorption spectra obtained from pulsed laser
photolysis (308 nm, 10 ns, 20-50 mJ/pulse) of anthranilium ions 1f
(panel A) and 1e (panel B) in N2-purged CH3CN. These spectra are
assigned to the corresponding arylnitrenium ions 2f and 2e (see text).
Irradiation of 1g in the absence of nucleophiles gives mainly
iminium ion 4g. In the presence of CH3OH, 1g behaves in a
similar fashion to the halogenated derivatives. The ortho adduct
12g (nuc ) CH3O) is observed as well as the methoxy
substitution product 17. This latter product is interesting; it
clearly involves addition of CH3OH and elimination of the
elements of HCN. By analogy to earlier studies of the bromo-
derivative, 1b,22 we attribute this product to an initial para-
addition followed by homolysis of the cyano group to give an
aminyl radical. The latter presumably abstracts a H-atom from
the solvent to give 17 (eq 5a). We considered the possibility
the the initial para adduct would decay by heterolysis (i.e.,
elimination of CN-, eq 5b). This process would form the para
methoxy substituted nitrenium ion 2e. This would lead to the
same products obtained from photolysis of the methoxy an-
thranilium ion 1e in CH3OH, namely 12e. Because 12e is not
observed, the heterolysis mechanism is excluded.
Figure 2. Decay waveforms from laser flash photolysis of 1f in CH3CN
containing 0.00-4.17 M H2O monitored at 470 nm, the absorption
maximum of the arylnitrenium ion 2f (see Figure 1 and text). Inset
shows the dependence of the pseudo-first-order decay rate constant,
kobs on [H2O].
ions 2 on the basis of the following considerations. (1) The
products described above are consistent with the formation of
1
singlet arylnitrenium ion 2. The ring addition products are
particularly characteristic of singlet arylnitrenium ions.37,38,46,47
(2) The lifetimes of the transients are unaffected by O2, a triplet
quencher.48 This argues against an excited triplet state of 1
being the signal carrier. (3) The lifetimes of the transient species
are unaffected by the addition of H-atom donors such as nBu3-
3. Detection of Nitrenium Ions by Laser Flash Photolysis
and Photothermal Beam Deflection. Pulsed laser photolysis
of anthranilium ions 1e and 1f gives the transient spectra shown
in Figure 1. The λmax for 1e is 500 nm (Figure 1b), whereas
for 1f it is 470 nm (Figure 1a). The transient spectra are
assigned to the singlet state of the corresponding arylnitrenium
(46) Gassman, P. G.; Campbell, G. A.; Frederick, R. C. J. Am. Chem.
Soc. 1972, 94, 3884-3891.
(47) Gassman, P. G.; Campbell, G. A. J. Am. Chem. Soc. 1972, 94,
3891-3896.
(48) Rosenthal, I. In Singlet Oxygen; Frimel, A. H., Ed.; CRC Press:
Boca Raton, FL, 1985; Vol. I, pp 13-35.

Substituent Effects on Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 8131
SnH (in concentrations as high as 0.1 M). The triplet state of
the nitrenium ions or aminyl radicals would be expected to react
rapidly with such traps.32 (4) McClelland16,49 has reported LFP
spectra of several similar phenyl-substituted arylnitrenium ions
which were generated using different photoprecursors. These
nitrenium ions also absorb in the region 450-500 nm. (5) The
observed transients react rapidly with nucleophiles with rate
constants of 105-1010 M-1 s-1 (Figure 2). This is the range
typical for substituted arylnitrenium ions.22,50 Furthermore the
rate constants vary in the expected way as the nature of the
nucleophile is varied. (These rate constants are discussed in
more detail in section 4, below.)
As expected, the lifetimes of 2e and 2f are considerably longer
than those of the previously studied examples. Nitrenium ions
2b-d have lifetimes of 119-440 ns.22 In contrast the phenyl
and methoxy stabilized have lifetimes of 30 and 600 µs,51
respectively. Earlier studies showed that in the absence of
nucleophiles (e.g., water, alcohols) the lifetimes of the 2-acetyl-
phenyl nitrenium ions are limited by a combination of the methyl
migration (km, Scheme 1) and thermal reversion to the anthra-
nilium ion (kc). Inasmuch as 2e yields no detectable iminium
ions 4, it can be assumed that this lifetime is limited primarily
by ring closuresi.e., 1/kc > 600 µs for 2e.
Neither 1a nor 1g give a detectable transient absorption under
the conditions employed here. Earlier,20,22 we speculated that
2a might decay too rapidly to be detected with the nanosecond
apparatus. This was based on the reasoning that the lack of a
stabilizing substituent in the 4-position ought to leave this
particular nitrenium ion more open to attack by nucleophiles.
The lower π-delocalization of the positive charge in 2a
compared with 2b-d also would be expected to make the
cyclization process (kc) more rapid. The same reasoning would
also hold that the cyano-derivative 2g is too short lived to detect
with the nanosecond apparatus. However, it was of interest to
verify these assumptions by obtaining experimental lifetimes,
or at least upper limits for 2a and 2g.
To estimate the lifetimes of 2a and 2g, photothermal beam
deflection calorimetry (PDC) was employed. This technique
can measure the rates of fast photochemical reactions by
monitoring the time-dependence of heat release that follows the
excitation of a sample with a short pulse of light. Extensive
treatments of the fundamentals52-54 and practice55-58 of this
method are available elsewhere. Briefly, pulsed photolysis of
a solution of 1a or 1g produces heat that is released as a
consequence of both nonradiative relaxation of excited states
as well as reactions of short-lived intermediates (i.e., 2). The
heating results in a thermal expansion of the solvent, which in
turn lowers its index of refraction. The excitation beam
illuminates only a small region of the sample cuvette creating
Figure 3. Time-resolved photothermal beam deflection signal gener-
ated from pulsed laser photolysis (308 nm, 10 ns, 0.2-1.0 mJ/pulse)
of 1f in Ar-purged CH3CN. Laser pulse arrived at ca. 52 µs on the
time axis shown.
a local zone of low refractive index surrounded by the higher
index of refraction in the unilluminated regions of the sample.
This is a virtual lens which can be used to deflect the
propogation of a second laser beam. The deflection angle of
the latter can be monitored as a function of time using a fast
bi-cell detector. The rate at which the deflection angle varies
is exactly the same as the rate of the exothermic processes
producing the heat.
In order to determine if the PDC technique would give reliable
kinetic data for nitrenium ions, it was first applied to a species
that could be independently examined by conventional LFP.
Nitrenium ion 2f is sufficiently long-lived that its lifetimes and
reaction rates could be measured by both methods. Figure 3
shows an example of a PDC kinetic waveform obtained from
photolysis of 2f in CH3CN. The rise of the deflection signal
shows two parts, a fast jump followed by a slower exponential
growth. The fast jump is assigned to processes that occur
rapidly (i.e., <100 ns) following excitation. These include kso,
kisc, and nonradiative relaxation to the lowest excited singlet
state of 1. We attribute the slower, exponential rise to the decay
reactions of 2f. In the absence of nucleophiles these include
the ring-closure process (kc) and the methyl shift (km). This
rise fits well to first-order growth with a time constant of 32
µs. This is in reasonable agreement with the 30 µs lifetime for
2f measured by LFP.
To further verify that the observed PDC growth was due to
1
reactions of 2, the effect of nucleophiles on the rate constant
was tested. Addition of CH3OH causes an increase in the
risetime that is proportional to the concentration of added
CH3OH. A pseudo-first-order analysis gives knuc ) 2.6 ((0.2)
× 105 M-1 s-1, a value in good agreement with that derived by
conventional LFP (2.70 × 105 M-1 s-1).
(49) McClelland, R. A.; Davidse, P. A.; Hadzialic, G. J. Am. Chem. Soc.
1995, 117, 4173-4174.
(50) Anderson, G. B.; Falvey, D. E. J. Am. Chem. Soc. 1993, 115, 9870-
9871.
(51) In contrast to the previously studied examples 2b, 2c, and 2d, the
lifetimes of these longer-lived nitrenium ions are much more sensitive to
traces of water in the solvent. The values cited would vary by ca. 10%
from experiment to experiment.
(52) Braslavsky, S. E.; Heibel, G. E. Chem. ReV. 1992, 92, 1381-1410.
(53) Braslavsky, S. E.; Heinoff, K. In Handbook of Organic Photochem-
istry; Scaiano, J. C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. 1, pp
327-355.
(54) Tam, A. C.; Sontag, H.; Hess, P. Chem. Phys. Lett. 1985, 120, 280-
284.
Having determined that PDC could give reliable kinetic data
for the nitrenium ion systems, we next turned our attention to
the anthranilium ions that did not give detectable transient
absorptions. Anthranilium ions 1a and 1g were examined by
PDC, the results of these experiments appear in Figure 4. For
both derivatives, the thermal rise was indistinguishable from
the instrument response function (ca. 100 ns). This demonstrates
that the corresponding nitrenium ions 2a and 2g have lifetimes
that are <100 ns and verifies our earlier conjecture.
(55) Yeh, S.-R.; Falvey, D. E. J. Photochem. Photobiol. A: Chem. 1995,
87, 13-21.
(56) Poston, P. E.; Harris, J. M. J. Photochem. Photobiol. A: Chem.
1991, 60, 51-61.
4. Kinetic Studies on the Reactions of Arylnitrenium Ions
with Nucleophiles. Nucleophilic trapping experiments provide
further support for the assignment of the observed transient
intermediates to the nitrenium ions 2. The lifetimes of the
(57) Poston, P. E.; Harris, J. M. J. Am. Chem. Soc. 1990, 112, 644-
650.
(58) Laman, D. M.; Falvey, D. E. ReV. Sci. Instrum. 1996, (In Press).

8132 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Robbins et al.
In several cases, the behavior of some nonhydroxylic nu-
cleophiles was examined. The halides Br- and Cl- have been
shown to be much more reactive toward carbenium ions than
are the alcohols.61,62 This is true as well for the arylnitrenium
ions. The phenyl derivative 2f reacts at slightly less than the
diffusion limit with these halides, and the methyl 2d derivative
reacts at the diffusion limit with Br-. The electron-rich alkenes
such as ethylvinyl ether63 react rapidly with 2d, 2e, and 2f
having rate constants similar to or larger than to those of MeOH.
Dihydropyran63 showed similar quenching behavior with 2f, in
this case the knuc was 1.18 × 106 M-1 s-1. Interestingly with
these particular types of nucleophiles the phenyl derivative was
seen to be much more reactive than the methoxy derivative.
The unactivated alkene, cyclopentene, did not quench 2f even
at concentrations as high as 3 M. Likewise the electron-poor
alkene, cyclopentenone did not quench 2f.
Ring substituents affect the stabilities of the arylnitrenium
ions in the following order: CN ≈ H < CH3 < Br ≈ Cl <<
Ph ≈ MeO. This trend is qualitatively different from that seen
for arylcarbenium ions.36,42 In carbenium ions the MeO group
(σ+ ) -0.78) is considerably more effective at stabilizing the
positive charge than is the Ph group (σ+ ) -0.18). Further-
more, in the carbenium ions, the CH3 (σ+ ) -0.31) group is
more effective than Ph and much more effective than the
halogens (+0.15 for Br and +0.11 for Cl). Similar substituent
effects have been observed with different arylnitrenium ions
studied by different methods. Novak’s group41 measured the
relative trapping rates of the thermally generated N-acetyl-4-
substituted phenylnitrenium ions 21. In this case, the stabiliza-
tion toward addition of water to 21 followed the order H <
CH3 < OEt ≈ Ph. In these particular arylnitrenium ions, the
phenyl group is slightly more effective at stabilizing the
arylnitrenium ion than alkoxy substituents.
Figure 4. Time-resolved photothermal beam deflection signal gener-
ated from pulsed laser photolysis of 1a (top panel) and 1g (bottom
panel) in Ar-purged CH3CN. Excitation pulse arrived at ca. 88 µs on
the time axis shown.
transient species are diminished when nucleophiles such as H2O,
alcohols, or halide ions are added. An example of this is
illustrated in Figure 2 which shows the decay of nitrenium ion
2f in the presence of varying concentrations of H2O. In most
cases the dependence of the observed first-order rate constant
for decay (kobs) on nucleophile concentration, [nuc], conformed
to pseudo-first-order behavior (eq 6). We define ko as the decay
rate constant in the absence of nucleophile and knuc as the
second-order rate constant for the reaction of 2 with a given
nucleophile.
We considered that the differences between the arylcarbenium
ions and the arylnitrenium ions might be due to steric effects.
In this case, one would expect that those substituents which
exerted the largest steric hindrance to be most sensitive to
changes in the size of the nucleophile. Specifically, the change
in knuc in going from MeOH to t-BuOH should be more dramatic
for a nitrenium ion with a large substituent than one with a
smaller substituent. For example the knuc values for cyclohexa-
dienyl cations show little or no dependence on the size of the
alcohol, varying by no more than a factor of 2 between MeOH
and t-BuOH.64 In contrast, these values vary by more than a
factor of 10 for the more sterically congested triarylvinyl
cations65 and the diarylmethyl carbenium ions.61,66
Comparison of the knuc values for various alcohols shows that
the steric hindrance from the various substituents studied here
is roughly constant. Figure 5 shows a free energy correlation
between the knuc values measured for the various arylnitrenium
ions 2a-f and the corresponding values for Ph2CH+.61 In each
case there is a good correlation. For the substituted systems
2b-f the slopes all fall in the range of 0.80 ( 0.09. In contrast,
kobs ) ko + knuc[nuc]
(6)
The knuc (CH3CN) derived from the slopes of the pseudo-
first-order plots for a variety of nucleophiles with both nitrenium
ions 2e and 2f are compiled in Table 2. Listed therein are also
knuc values for the previously studied nitrenium ion 2d.22 For
each nitrenium ion the trend of reactivity is conserved with
respect to the nucleophile. For example in the series of
hydroxylic nucleophiles, the trend MeOH > EtOH > i-PrOH
∼ H2O . t-BuOH holds true for all systems studied. For the
alcohols, knuc is determined by steric hindrance due to branching
at the R-carbon atom. Such a trend is typical for reactive
cationic species such as vinyl cations,59 alkene cation radicals,60
and diarylcarbenium ions.61 Water, lacking an inductively
electron-donating alkyl group, falls at approximately the same
value as i-PrOH.
(62) Dorfman, L.; Sujdak, R. J.; Bockrath, B. Acc. Chem. Res. 1976, 9,
352-357.
(63) Product studies showed a complex mixture of products when
nitrenium ion 2f was generated in the presence of this nucleophile.
(64) Steenken, S.; McClelland, R. A. J. Am. Chem. Soc. 1990, 112,
9648-9649.
(65) Kobayashi, S.; Kitamura, T.; Taniguchi, H.; Schnabel, W. Chem.
Lett. 1983, 1117-1120.
(66) Bartl, J.; Steenken, S.; Mayr, H.; McClelland, R. A. J. Am. Chem.
Soc. 1990, 112, 6918-6928.
(59) Ginkel, F. I. M. v.; Visser, R. J.; Varma, C. A. G. O.; Lodder, G.
J. Photochem. 1985, 30, 453-473.
(60) Johnston, L. J.; Schepp, N. P. J. Am. Chem. Soc. 1993, 115, 6564-
6571.
(61) Bartl, J.; Steenken, S.; Mayr, H. J. Am. Chem. Soc. 1991, 113, 7710-
7716.

Substituent Effects on Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 8133
Table 2. Rate Constants knuc (M-1 s-1) for Reaction of Arynitrenium Ions with Nucleophiles in CH3CN Solution Measured by LFP
MeOH
EtOH
i-PrOH
t-BuOH
H2O
EVEa
Br-
Cl-
2d
2e
2f
2.23 × 108 b
1.95 × 105
2.70 × 105
1.18 × 108 b
2.60 × 105
2.84 × 105
4.63 × 107 b
1.02 × 105
6.03 × 104
1.48 × 107 b
3.39 × 104
2.09 × 104
3.08 × 107 b
2.06 × 105
6.16 × 104
1.65 × 108
7.95 × 105 c
1.18 × 107
1.26 × 1010
8.47 × 109
d
1.19 × 109
a Ethylvinyl ether. b From ref 22. c Above 0.01 M the pseudo-first-order plot curved upward. The cited value was determined at concentration
below 0.01 M. d Nitrenium ion appears to be quenched, but the kinetics are complex.
are shown to be less stable than any of the other systems,
including the 4-halo. It is clear that simple quantitative models
derived for carbenium ions cannot be readily extrapolated to
the nitrenium ions.
Experimental Section
General Photolysis Conditions. Photolysis reactions were per-
formed in quartz or Pyrex vessels. In the cases where no nucleophilic
trap was present, the vessels and necessary syringes were treated with
a 5% v/v solution of Me2SiCl2 in CH2Cl2, rinsed, and oven-dried
overnight before use. The solutions were purged of O2 by passing N2
or argon through the solutions for 10-15 min. The light source was
a medium pressure 450 W Hg lamp.
Analysis of Steady-State Photoproducts. The product distributions/
yields were determined by 1H NMR spectroscopy of the photolysates.
The corresponding anthranilium salt 1 was irradiated in a Pyrex vessel
Figure 5. Correlation of the log(knuc) for reactions of various nitrenium
ions 2a-f with ROH in CH3CN plotted as a function of the
corresponding values for diphenylcarbenium ion (Ph2CH+). The values
for 2a-d were taken from ref 22, the values for Ph2CH+ were taken
from ref 61.
using a 450 W medium pressure Hg lamp. The sample solutions were
7-20 mM in starting material and purged with N2 prior to photolysis.
Following the irradiation, the solvent was evaporated under vacuum,
and the residue was dissolved in a known volume of CD3CN containing
either triphenylmethane, hexamethyldisiloxane, or trimethylsilane as
internal standards. Yields were determined by 1H NMR integration of
the product peaks and comparison of these values to the peak integration
of the internal standard. The observed peaks correspond to the
conjugate acids of the product amines. For comparison, 1H NMR
spectra of the conjugate acids of the authentic products were taken. A
drop of concentrated HClO4 was added to the NMR sample. Yields
are corrected for conversion of the starting material.
parent system 2a gives a significantly smaller slope of 0.44.
The latter example shows that ring additions to nitrenium ions
is indeed sensitive to steric hindrance. However, the nearly
equal slopes for the substituents CH3, Cl, Br, Ph, and MeO
suggest that these particular substituents all exert roughly the
same steric hindrance.
LFP Experiments. LFP experiments were done using a Questek
2120 excimer laser using Xe/HCl as the reagent gas. The latter provides
excitation 20-50 mJ light pulses of 10 ns duration at 308 nm. Kinetic
traces were obtained by monitoring a CW probe beam from a 350 W
Xe arc lamp which passed through the sample perpendicular to the
path of the excitation beam. Sample solutions were placed in a 50
mL, continuous stirring, flow-cell (20-30 mg in 40-50 mL of specified
solvent, 2-3 mM in anthranilium salt) or in quartz cuvettes (1 cm ×
1 cm × 5 cm). Both vessels were sealed with rubber septa prior to
bubbling N2 through the solutions for 10-15 min. For nitrenium ion
quenching experiments, the decay of transient generated upon irradiation
of 1d was monitored at 460 nm, whereas 1e and 1f derivatives were
observed at 470 and 500 nm, respectively. Pseudo-first-order rate
constants (kobs) were obtained for the decay of the transient using 5-15
different nucleophile concentrations [nuc]. The nucleophilic quencher
was added through the septum by means of a syringe. As in the
previous study,22 all of the nucleophile solutions contained 10% of
H2SO4 to prevent a ground-state addition of the nucleophile to 1.
Photothermal Beam Deflection Experiments. The PBD apparatus
has been described elsewhere.58 Experiments without nucleophiles
present consisted of solutions of 1g in CH3CN (4.7 mg salt 1g in 3 mL
of solution). The solution was placed in quartz cuvette, sealed with a
septum, and purged with argon for 10 min. Solutions with CH3OH
present were 8.2 × 10-2 and 4.1 × 10-2 M in MeOH (0.010 and 0.005
mL, respectively) and were prepared in the same fashion as the solution
with no nucleophilic trap.
The above analysis indicates that the differences in substituent
effects betwen arylcarbenium ions and arylnitrenium ions must
have an electronic basis as well as a steric one. Novak, et al.41
have argued that differences in transition state aromaticity
account for at least part of this discrepency. In the transition
state leading to the carbenium ion adduct, the delocalized
π-electrons are relocalized into the phenyl ring, resulting in
increased aromatic stabilization. In the transition state leading
to the nitrenium ion adduct, the π-electrons are further delo-
calized onto the exocyclic bond, resulting in a loss of aromatic
stabilization. It was demonstrated that the knuc (measured by
competitive trapping) correlated well with the ab initio calculated
enthalpies of hydration.41
It is also possible that the increased electronegativity of the
nitrogen in the nitrenium ions compared with carbon in the
carbenium ions leads to increasing demand for π-electron
density. This is more readily satisfied by a polarizable
substituent such as Ph than the far less polarizable CH3O or
CH3 groups. Likewise, the halogens nonbonding electrons are
more easily donated into the lower energy vacant nitrenium p
orbital than they are in the higher energy carbenium ion p orbital.
Conclusions
Synthesis of Anthranilium Salts. Detailed procedures for the
preparation of synthetic intermediates 7f (mp 78-80 °C), 8f (mp 108-
110 °C), 7e (mp 120-122 °C), 8e (a yellow oil), 7g (mp 276 °C dec),
and 8g (mp 122-124 °C), along with spectroscopic data for each are
provided in supporting information.
5-Phenyl-3-methylanthranil (10f). Amine (8f) (1.01 g, 4.8 × 10-3
mol) were dissolved in 25 mL of glacial acetic acid to which 1 mL of
H2SO4 was added. Reaction mixture was cooled to 0 °C, and an
LFP and photothermal beam deflection investigations of the
4-substituted tert-butyl phenylnitrenium ions 2 reveal some
interesting effects of substituents on the stability of these species
toward attack by water and alcohols. It is shown that a 4-phenyl
substituent stabilizes the nitrenium center far more than a
4-methyl or 4-halo group and to almost the same extent as a
4-methoxy group. The 4-unsubstituted and 4-cyano systems

8134 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Robbins et al.
aqueous solution of NaNO2 (0.67 g, 9.7 × 10-3 mol in 2 mL of H2O)
was added in several aliquots. Reaction was allowed to stir for 1 h
before aqueous NaN3 (0.63 g, 1.0 × 10-2 mol in 2 mL of H2O) was
added. Gas evolution was observed. After addition was complete
reaction was stirred for 20 min at 0 °C. Reaction mixture was extracted
with Et2O (3 × 50 mL). Organic layer was washed with H2O until
aqueous layer was neutral and then concentrated under vacuum. The
residue was taken up in 30 mL of H2O and heated at reflux for 4 h.
Once cooled, the reaction mixture was extracted with Et2O (3 × 25
mL), dried with MgSO4, and concentrated under vacuum. 10f was
5-Cyano-3-methylanthranil (10g). Amine 8g (0.4 g, 2.3 mmol)
was treated with NaNO2 (0.4 g, 5.5 mmol) and NaN3 (0.4 g, 5.6 mmol)
at 0 °C. Workup was performed as previously indicated. Extraction
of cooled solution yielded a yellow residue which after recrystallization
with CH3OH gave 0.17 g (47%) of 8g as yellow crystals; mp 138-
1
139 °C; H NMR (CD3CN, 200 MHz) δ ) 8.21 (s, 1H), 7.61-7.56
(dd, J ) 9.3, 1.0, 1H), 7.41-7.36 (dd, J ) 9.3, 1.0, 1H), 2.82 (s, 3H);
13C NMR (CD3CN) 172.0, 156.7, 131.5, 119.5, 117.1, 115.8, 107.3,
12.6; MS, m/e (rel intensity, %) 158 (M+, 100), 130 (14), 129 (78),
116 (18), 104 (18), 103 (41); HRMS, m/e 158.04815 (M+ calcd for
C9H6N2O, m/e 158.04802); IR (CH2Cl2) 3960 (m), 3060 (s), 2990, (br,
s), 2600, 2315, 1652 (s), 1550, 1420 (vs), 1255, 895.
5-Cyano-3-methyl-N-tert-butylanthranilium perchlorate (1g). From
cyano substituted anthranil 10g (0.13 g, 7.9 × 10-4 mol) was obtained
1g (0.1 g, 37%), mp 168-170 °C (dec), as white crystals: 1H NMR
(CD3CN, 200 MHz) 8.64 (dd, J ) 0.9, 1.2, 1 H), 8.16-8.10 (dd, J )
1.2, 9.5, 1 H), 8.05-7.99 (dd, J ) 9.5, 0.9, 1 H), 3.08 (s, 3 H), 1.87
(s, 9 H); 13C NMR (CD3CN, 55.1 MHz) 178.2, 146.9, 141.1, 132.9,
119.7, 117.4, 114.2, 111.5, 72.3, 28.7, 14.0; MS, m/z (rel intensity, %)
214 (M+, 0.2), 159 (13), 158 (100), 129 (35), 103 (22), 63 (11), 59
(39), 56 (15); HRMS, m/z 214.10968 (M + 1 - ClO4 calcd for
C13H14N2O, m/z 214.11061); IR (CH2Cl2) 3950 (m), 3045 (br, s), 2995
(s), 2318, 1432 (s), 1299, 910, 800.
Isolation of Photoproducts. The anthanilium ions 1 were dissolved
to ca. 1 M in CH3CN, along with the indicated trapping molecule. The
solutions were purged with N2 for 15 min and then irradiated for the
indicated time with a 450 W medium pressure Hg lamp through a Pyrex
vessel. The reaction mixture was partitioned between Et2O and aqueous
NaHCO3, and the organic fraction was further extracted with H2O and
then dried over MgSO4. The solvent was evaporated, and the residue
was subjected to radial chromatography over silica gel using 20%
EtOAc/hexane.
1
obtained in 57% yield: mp 61-65 °C; H NMR (CDCl3) 7.60-7.38
(m, 8H), 2.82 (s, 3 H);13C NMR (CD3CN) 168.2, 157.4, 140.9, 136.2,
132.9, 129.9, 128.6, 127.7, 118.5, 116.9, 115.8, 12.2; MS, m/z (rel
intensity, %) 209 (100), 196 (19), 181 (42), 180 (80), 179 (11), 167
(16), 166 (18), 154 (20), 153 (13), 140 (15), 139 (20), 70 (13); HRMS,
m/z 209.08268 (M+ calcd for C14H11NO, m/z 209.08406); IR (CH2Cl2)
3700 (br, m), 3060 (br, s), 2999 (w), 2700, 2540, 2420, 1430, 1290,
1230 (s), 800.
5-Methoxy-3-methylanthranil (10e). Amine (8e) (2.04 g, 0.012
mol) was dissolved in 30 mL of glacial acetic acid to which 5 mL of
H2SO4 was added. Reaction mixture was cooled to 0 °C. NaNO2 (3.24
g, 0.049 mol) dissolved in 10 mL of H2O was added in several aliquots,
and the reaction was allowed to stir for 1 h. NaN3 (5.09, 0.078 mol)
was dissolved in 10 mL of H2O and was added dropwise to cold acidic
solution. Gas evolution was observed upon addition. After the addition
was complete, the reaction mixture was stirred for an additional 20
min at 0 °C. Reaction mixture was extracted with Et2O (3 × 50 mL).
In order to remove all glacial acetic acid, the organic layer was washed
with H2O until the aqueous layer ran neutral. Organic solvent was
then removed under vacuum and a reddish oil remained. 2-Azido-5-
methoxy-acetophenone 9e was carried on to the next step without
further purification: 1H NMR (200 MHz, CD3CN) δ ) 7.16-6.98 (m,
3H), 3.89 (s, 3H), 2.60 (s, 3H). The azide 9e was taken up in 30 mL
of H2O and refluxed for 5 h. Aqueous media required that aliquots be
removed from reaction mixture and extracted with Et2O in order to
monitor progress by TLC. Once cooled, the reaction mixture was
extracted with Et2O (3 × 25 mL), dried over MgSO4, and concentrated
under vacuum. A colorless oil remained which was identified as 10e
(40% yield): 1H NMR (200 MHz, CD3CN) δ ) 7.14-7.06 (d, J )
7.13, 1H), 6.57-6.54 (d, J ) 7.13, 1H), 6.57-6.54 (d, J ) 7.13, 1H),
(d, J ) 9.07, 1H), 3.92 (s, 3H), 2.73 (s, 3H); 13C NMR (CDCl3) 165.5,
152.5, 148.0, 123.6, 121.1, 117.3, 111.3, 105.4, 55.6, 12.0; MS, m/z
(rel intensity, %) 163 (100), 162 (25), 161 (8), 148 (11), 120 (14);
HRMS m/z 163.06218 (calcd for C9H9NO2, m/z 163.06332); IR
(CH2Cl2) 3960 (m), 3040 (s), 2930 (s), 2200, 2420, 2320, 1650 (s),
1550 (s), 1430, 1200, 1189, 1060, 895, 780.
Preparative Photolysis of 1f in CH3CN. Salt (1f) (34.0 mg) was
dissolved in 20 mL of freshly distilled CH3CN and irradiated for 90
min. The solvent was evaporated, and the residue was taken up in
1
CD3CN and analyzed by H NMR. The mixture contained 29% 5f
and 25% 4f of another product assigned to rearrangement product.
Iminium salt 5f based on the similarity of its 1H NMR spectrum to
previous examples:20,22 (CD3CN) 7.73-7.79 (m, 3 H), 7.59-7.46 (m,
5 H), 3.75-3.74 (m, 3 H), 2.70 (s, 3 H), 2.44 (br s, 6 H);
Preparative Photolysis of 1f in the Presence of CH3OH. Irradia-
tion of 1f in CH3CN, in the presence of CH3OH (2.1 M), gave four
products 23% 17f, 7% 12f, 45% 13f (nuc ) CH3O), and 8% 14f (nuc
) CH3O). Quinone imine 13f could not be isolated, presumably it
rearranges to 14f upon workup. 2-Acetyl-4-phenyl-N-tert-butyl aniline
(5f): mp 93-95 °C; 1H NMR (400 MHz, CD3CN) δ ) 9.26 (s, broad,
1H), 8.06 (d, J ) 2.4, 1H), 7.64-7.60 (m, 3H), 7.59-7.58 (m, 2H),
7.43-7.39 (m, 1H), 7.30-7.26 (m, 1H), 2.61 (s, 3H), 1.44 (s, 9H);
13C NMR (CD3CN) 201.9, 150.2, 141.0, 133.4, 132.3, 129.6, 127.0,
126, 118.9, 115.5, 51.2, 29.6, 28.4; MS, m/z (rel intensity, %) 267 (72),
253 (20), 252 (100), 234 (49), 220 (27), 211 (57), 210 (16), 196 (61),
167 (30), 91 (44), 56 (47); HRMS, m/z 216.16243 (M+ calcd for C18H21-
NO, m/z 267.16232); IR (CH2Cl2) 3740 (b), 2995 (s), 2840, 2702, 2420,
2348, 1640 (s), 1420, 1280, 1315, 900. 2-Acetyl-6-methoxy-3-phenyl-
N-tert-butylaniline (12f, nuc ) CH3O): mp 72-74 °C; 1H NMR (400
MHz, CD3CN) δ )7.66.7.64 (m, 2H) 7.55 (d, J ) 2.1, 1H), 7.46-
7.42 (m, 2H), 7.34-7.32 (m, 1H), 7.29-7.28 (d, J ) 2.1, 1H), 3.88 (s,
3H), 2.62 (s, 3H), 1.23 (s, 9H); 13C NMR (CD3CN) 203.9, 154.1, 141.4,
140.1, 132.9, 130.2, 56.1, 55.3, 31.3, 30.0; MS, m/z (rel intensity, %)
297 (66), 283 (11), 282 (50), 264 (10), 241 (100), 226 (62), 198 (14);
HRMS, m/z 297.17313 (M+ calcd for C18H21NO, m/z 297.17288); IR
(CH2Cl2) 3740 (s), 3700 (w), 3080, 2990 (s), 2318, 1610 (s), 1420,
1280, 990, 890, 788. 2-Acetyl-4-methoxy-3-phenyl-N-tert-butyl-
aniline (14f nuc ) CH3O): mp 93-95 °C; 1H NMR (400 MHz, CD3-
CN) δ )7.40-7.34 (m, 3H) 7.25-7.22 (m, 3H), 7.07-7.05 (d, J )
9.07, 1H), 7.03-7.00 (d, J ) 9.07, 1H), 3.64 (s, 3H), 1.63 (s, 3H),
1.24 (s, 9H);13C NMR (CD3CN) 206.9, 149.74, 138.7, 138.0, 132.8,
131.4, 129.9, 128.7, 128.3, 120.3, 115.5, 56.8, 52.1, 32.1, 30.2; MS,
m/z (rel intensity, %) 297 (100), 282 (82), 264 (11), 242 (15), 241
(62), 240 (18), 210 (25), 183 (31), 182 (10), 180 (15), 153 (12), 57
(20); HRMS, m/z 297.17294 (M+ calcd for C18H21NO, m/z 297.17288);
IR (CH2Cl2) 3905 (s), 3760 (w), 3058, 2980 (s), 2700, 1720 (s), 1415.
5-Methoxy-3-methyl-N-tert-butylanthranilium Perchlorate (1e).
Methoxy anthranil 10e (0.58 g, 3.58 × 10-3 mol) was dissolved in
CH3NO2 and treated with perchloric acid and tert-butyl alcohol as
indicated above, yielding 0.69 g (60%): mp 136-138 °C dec; 1H NMR
(400 MHz, CD3CN) δ ) 7.59-7.56 (dd, J )7.6, 1.7, 1 H), 7.44-7.43
(dd, J )7.6, 1 H), 7.42 (d, J )1.7 1H), 4.08 (s, 3 H), 2.99 (s, 3 H),
1.87 (s, 9 H); 13C NMR (CD3CN) 173.3, 144.1, 140.8, 128.6, 122.4,
118.4, 114.8, 72.4, 57.3, 29.1, 13.0; HRMS m/z 220.1343 (M + 1 -
ClO4 calcd for C13H18NO2, m/z 220.1338); MS, m/z (rel intensity, %)
M+ 220 (0.3), 163 (100), 162 (50), 148 (31), 144 (19), 121 (18), 120
(62), 107 (10), 78 (20), 65 (13), 51 (14); IR (CH2Cl2) 3700 (m), 3050
(br), 2995, 2315, 2420, 1610 (w), 1420 (br), 1550 (s), 1250 (br), 1102,
900 (s).
5-Phenyl-3-methyl-N-tert-butylanthranilium perchlorate (1f) (1.11
g, 54%) was obtained from 10f (1.17 g, 5.63 × 10-3 mol) using the
method indicated above. Recrystallization with CH3OH yielded tan
crystals: mp 260-261 °C (dec); 1H NMR (CD3CN, 400 MHz) 8.36-
8.26 (dd, J ) 9.4, 1.63, 1 H), 8.26 (s, 1 H), 7.99-7.97 (d, J ) 9.4, 1
H), 3.08 (s, 3 H), 1.88 (s, 9 H); 13C NMR (CD3CN, 110 MHz) 175.8,
146.9, 142.1, 141.9, 138.5, 130.3, 129.9, 128.2, 121.1, 112.8, 70.9,
28.7, 13.6; CIMS, m/z (rel intensity, %) 267 (M + 1, 8), 252 (11), 209
(92), 196 (15), 180 (100), 179 (15), 167 (16), 166 (21), 153 (28), 139
(27), 115 (13), 56 (57); HRMS, m/z 267.16207 (M + 1 - ClO4 calcd
for C18H20NO, m/z 267.16237); IR (CH2Cl2) 3960 (m), 3705, 3060 (s),
2990 (s), 2700, 2430, 2310, 1650 (s), 1458, 1417 (br), 1090.

Substituent Effects on Arylnitrenium Ions
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 8135
Preparative Photolysis of 1f in the Presence of H2O. Salt 1f (26.1
mg, 7.2 × 10-2 mmol) was dissolved in 15 mL of CH3CN and 1 mL
of a 10% HClO4 in H2O solution (3.1 M H2O). Irradiation for 15 min
gave several products. Ortho addition product (12f nuc ) OH) was
obtained in 21% yield. Reduction product 17f was obtained in 30%
yield. 14f (nuc ) OH) was obtained in 17% yield and 15 (nuc )
OH) in 8%. 2-Acetyl-4-phenyl-6-hydroxy-N-tert-butylaniline (12f,
nuc ) OH) was characterized as follows: 1H NMR (400 MHz,
CD3CN) δ ) 7.76-7.64 (m, 2H), 7.54-7.53 (d, J ) 2.1, 1H), 7.47-
7.34 (m, 3H), 7.30-7.29 (d, J ) 2.1, 1H), 2.62 (s, 3H), 1.10 (s, 9H);
13C NMR (CD3CN) 214.5, 132.0, 129.5, 128.0, 126.9, 121.0, 119.9,
119.4, 116.0, 105.0, 103.6, 68.9, 35.5, 29.0; MS, m/z (rel intensity, %)
283 (10, M+), 228 (16), 264 (11), 227 (100), 212 (20), 206 (28), 154
(11), 130 (13), 115 (10), 105 (64), 77 (16), 57 (22); HRMS, m/z
283.15275 (M+ calcd for C18H21NO2, m/z 283.15723). 2-Acetyl-3-
phenyl-4-hydroxy-N-tert-butylaniline (14f, nuc ) OH) was charac-
terized as follows: 1H NMR (400 MHz, CD3CN) δ ) 7.41-7.25 (m,
5H), 6.99-6.97 (d, J ) 8.9, 1H), 6.88-6.85 (d, J ) 8.9, 1H), 5.0
(broad s, 1H), 1.64 (s, 3H), 1.21 (s, 9H); 13C NMR (CD3CN) 219.0,
147.0, 138.9, 131.6, 130.1, 129.3, 128.7, 127.7, 127.0, 126.9, 122.0,
119.4, 30.6, 27.2; MS, m/z (rel intensity, %) 283 (95, M+), 269 (36),
268 (100), 250 (36), 227 (93), 226 (33), 213 (24), 194 (20), 183 (12),
159 (16), 130 (17), 57 (26); HRMS, m/z 283.15668 (M+ calcd for
C18H21NO2, m/z 283.15723). 2-Acetyl-4-(4-hydroxy-phenyl)-N-tert-
butylaniline (15f, nuc ) OH): 1H NMR (400 MHz, CD3CN) δ )
7.97-7.96 (d, J ) 2.3, 1H), 7.56-7.52 (m, 1H), 7.44-7.42 (d, J )
8.5, 2H), 7.07-7.05 (m, 2H), 6.85-6.83 (d, J ) 8.5, 2H), 2.59 (s,
3H), 1.43 (s, 9H); 13C NMR (CD3CN); MS, m/z (rel intensity, %) 283
(95, M+), 269 (20), 268 (100), 250 (30), 228 (12), 227 (75), 226 (18),
184 (17); HRMS, m/z 283.15621 (M+ calcd for C18H21NO2, m/z
283.15723);
m/z (rel intensity, %) 251 (27, M+), 195 (32), 180 (63), 152 (7), 69
(100); HRMS, m/z 251.1516 (M+ calcd for C12H21NO3, m/z 251.1522);
IR (CH2Cl2) 3922 (w), 3690 (s, br), 2440 (s, br), 2300 (w), 1735 (s),
1510 (s), 1230, 985.
Preparative Photolysis of 1e in the Presence of H2O. Compound
1e (39.8 mg, 1.3 × 10-1 mmol) was dissolved in 20 mL of CH3CN
and 2.2 mL of a 10% HClO4 in H2O (4.9 M in H2O). The solution
was irradiated for 35 min. Two products were obtained: (1) the ortho
adduct, 2-acetyl-6-hydroxy-4-methoxy-N-tert-butylaniline (12e, nuc )
OH), and (2) a dihydroxy adduct, 2-acetyl-3,4-dihydroxy-N-tert-
butylaniline (16, nuc ) OH). 2-Acetyl-6-hydroxy-4-methoxy-N-tert-
butylaniline (12e, nuc ) OH) was obtained in 23% yield: 1H NMR
(400 MHz, CD3CN) δ ) 6.57-6.55 (d, J )2.7, 1H), 6.51-6.49 (d, J
) 2.7, 1H), 3.75 (s, 3H), 2.54 (s, 3H), 1.02 (s, 9H); 13C NMR (CD3CN)
201.0, 147.2, 138.5, 132.2, 128.9, 125.3, 125.0, 30.3, 21.5, 21.3, 14.2;
MS, m/z (rel intensity, %) 237 (38, M+), 222 (40), 182 (12), 181 (88),
166 (100), 152 (16), 138 (13), 57 (14); HRMS, m/z 237.13631 (M+
calcd for C13H19NO3, m/z 237.13649); IR (CHCl3) 3685 (w), 3619 (br),
2938 (s, br), 2895 (br), 2400 (w), 1600 (s), 1410 (s), 1182, 1033.
2-Acetyl-3,4-dihydroxy-N-tert-butylaniline (16, nuc ) OH) was
obtained in 87% yield as an orange oil: 1H NMR (400 MHz, CD3CN)
δ ) 6.67-6.64 (d, J )10.2, 1H), 6.63-6.61 (d, J ) 10.2, 1H), 2.49
(s, 3H), 1.46 (s, 9H); 13C NMR (CD3CN) 182.5, 134.3, 124.8, 1237.,
121.6, 120.2, 106.8, 57.6, 28.0, 27.1; MS, m/z (rel intensity, %) 223
(4, M+), 181 (100), 166 (70), 149 (30), 138 (20), 111 (14), 97 (19), 57
(76), 56 (76); HRMS, m/z 223.11918 (M+ calcd for C12H17NO3, m/z
223.12085).
Acknowledgment. We are grateful to Profs. Mike Novak,
Chris Cramer, and Susan Graul for helpful advice. Mr. Dominic
Chiapperino is thanked for editorial assistance. We are grateful
for finanical support from the National Science Foundation.
Preparative Photolysis of 1e in the Presence of CH3OH. Solution
of 1e (86.1 mg, 2.7 × 10-4 mol) in 20 mL of CH3CN, in the presence
of CH3OH (2.3 M) was irradiated for 30 min. Only one product (12e)
was isolated in 66% yield. 2-Acetyl-4,6-dimethoxy-N-tert-butyl-
aniline (12e nuc ) CH3O) was a yellow oil: 1H NMR (400 MHz,
CD3CN) δ ) 6.66-6.65 (d, J ) 2.8, 1H), 6.63-6.62 (d, J ) 2.8, 1H),
3.76 (s, 3H), 2.57 (s, 3H), 1.05 (s, 9H); 13C NMR (CD3CN) 204.8,
156.8, 156.2, 136.6, 130.4, 104.9, 103.3, 56.2, 55.7, 30.4, 30.2; MS,
Supporting Information Available: Synthetic procedures
and spectroscopic data for synthetic intermediates 7e, 8e, 7f,
8f, 7g, and 8g (4 pages). See any current masthead page for
ordering and Internet instructions.
JA961518Z