The loss of carbon dioxide from activated perbenzoate anions in the gas phase: Unimolecular rearrangement via epoxidation of the benzene ring

Article abstract of DOI:10.1021/jo060730a

The unimolecular reactivities of a range of perbenzoate anions (X-C ⁶H₅CO₃^-), including the perbenzoate anion itself (X = H), nitroperbenzoates (X = para-, meta-, orrtho-NO ₂), and methoxyperbenzoates (X = para-, meta-OCH₃) were investigated in the gas phase by electrospray ionization tandem mass spectrometry. The collision-induced dissociation mass spectra of these compounds reveal product ions consistent with a major loss of carbon dioxide requiring unimolecular rearrangement of the perbenzoate anion prior to fragmentation. Isotopic labeling of the perbenzoate anion supports rearrangement via an initial nucleophilic aromatic substitution at the ortho carbon of the benzene ring, while data from substituted perbenzoates indicate that nucleophilic attack at the ipso carbon can be induced in the presence of electron-withdrawing moieties at the ortho and para positions. Electronic structure calculations carried out at the B3LYP/6-311++G(d,p) level of theory reveal two competing reaction pathways for decarboxylation of perbenzoate anions via initial nucleophilic substitution at the ortho and ipso positions, respectively. Somewhat surprisingly, however, the computational data indicate that the reaction proceeds in both instances via epoxidation of the benzene ring with decarboxylation resulting-at least initially-in the formation of oxepin or benzene oxide anions rather than the energetically favored phenoxide anion. As such, this novel rearrangement of perbenzoate anions provides an intriguing new pathway for epoxidation of the usually inert benzene ring.

Full text of DOI:10.1021/jo060730a

The Loss of Carbon Dioxide from Activated Perbenzoate Anions in
the Gas Phase: Unimolecular Rearrangement via Epoxidation of the
Benzene Ring
David G. Harman, Aravind Ramachandran, Michelle Gracanin,† and Stephen J. Blanksby*
Department of Chemistry, UniVersity of Wollongong, Wollongong NSW, 2522, Australia
blanksby@uow.edu.au
ReceiVed April 5, 2006
-
The unimolecular reactivities of a range of perbenzoate anions (X-C6H5CO3 ), including the perbenzoate
anion itself (X ) H), nitroperbenzoates (X ) para-, meta-, ortho-NO2), and methoxyperbenzoates (X )
para-, meta-OCH3) were investigated in the gas phase by electrospray ionization tandem mass spectrometry.
The collision-induced dissociation mass spectra of these compounds reveal product ions consistent with
a major loss of carbon dioxide requiring unimolecular rearrangement of the perbenzoate anion prior to
fragmentation. Isotopic labeling of the perbenzoate anion supports rearrangement via an initial nucleophilic
aromatic substitution at the ortho carbon of the benzene ring, while data from substituted perbenzoates
indicate that nucleophilic attack at the ipso carbon can be induced in the presence of electron-withdrawing
moieties at the ortho and para positions. Electronic structure calculations carried out at the B3LYP/6-
311++G(d,p) level of theory reveal two competing reaction pathways for decarboxylation of perbenzoate
anions via initial nucleophilic substitution at the ortho and ipso positions, respectively. Somewhat
surprisingly, however, the computational data indicate that the reaction proceeds in both instances via
epoxidation of the benzene ring with decarboxylation resultingsat least initiallysin the formation of
oxepin or benzene oxide anions rather than the energetically favored phenoxide anion. As such, this
novel rearrangement of perbenzoate anions provides an intriguing new pathway for epoxidation of the
usually inert benzene ring.
Introduction
olefins,4 and the oxidation of nitrogen, sulfur, and phosphorus
compounds,5,6 are routinely achieved using peroxy acids. Base-
promoted oxidations by peroxy acids have also been reported,
including the oxidation of sulfoxides to sulfones7,8 and the
epoxidation of electron-deficient olefins.9 Under basic condi-
tions, the latter reaction has been rationalized as a two-step
Peroxy acids, RC(O)O2H, are used extensively in organic
chemistry as reagents for synthesis.1,2 Transformations, such as
the Baeyer-Villiger oxidation of ketones,3 epoxidation of
* To whom correspondence should be addressed. Ph: ++61 2 4221 5484.
Fax: ++61 2 4221 4287.
† Current address: Heart Research Institute, 145-147 Missenden Road,
Camperdown, New South Wales, 2050, Australia.
(1) Organic Peroxides; Ando, W., Ed.; John Wiley: Chichester, UK,
1992.
(4) Rao, A. S. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: New York, 1991; Vol. 7, p 357.
(5) Gilchrist, T. L. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Ed.; Pergamon Press: New York, 1991; Vol. 7, p 735.
(2) Organic Peroxides; Swern, D., Ed.; Wiley-Interscience: New York,
1972.
(3) Krow, G. R. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: New York, 1991; Vol. 7, p 671.
(6) Uemura, S. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: New York, 1991; Vol. 7, p 757.
(7) Curci, R.; Modena, G. Tetrahedron 1966, 22, 1227.
(8) Curci, R.; Giovine, A.; Modena, G. Tetrahedron 1966, 22, 1235.
10.1021/jo060730a CCC: $33.50 © 2006 American Chemical Society
Published on Web 09/22/2006
7996
J. Org. Chem. 2006, 71, 7996-8005

Loss of CO2 from Perbenzoate Anions in the Gas Phase
SCHEME 1
mechanism initiated by nucleophilic attack of the percarboxylate
anion, RC(O)O2-, on the olefin.9,10 Recent calculations by
Washington and Houk,11 however, suggest that stepwise addition
of oxygen to the alkene occurs only in the presence of extremely
electron-deficient (e.g., NO2- and CN-substituted) alkenes, and
the reaction generally proceeds via a concerted mechanism.
Indeed, the concerted oxygen transfer by the percarboxylate
anion has been exploited in organic synthesis to increase
stereoselectivity12,13 and regioselectivity14 in the epoxidation of
some olefins as well as in the preparation of acid-sensitive
epoxides.15,16 Peroxy acids have pKa values17 of ca. 8.3, and
thus their conjugate basesspercarboxylate anionssare readily
formed under alkaline conditions. Despite their importance in
synthetic chemistry, there is limited understanding of the
structure and reactivity of percarboxylate anions. Computational
studies have shown that anti and syn conformations (Scheme
1a and b, respectively) exist as minima on the potential energy
surface with the syn conformation generally preferred for the
solvated anions by between 1 and 25 kJ mol-1 (depending on
R).11 Porter and co-workers have also proposed an isomeric
oxydioxirane structure (Scheme 1c) which calculation suggests
is only 5-50 kJ mol-1 (depending on the nature of R) more
energetic than the percarboxylate anion.18 Interconversion
between the two isomers is predicted to occur over a relatively
low barrier, and as such, the oxydioxirane anion may play a
role in the observed oxygen scrambling in percarboxylate anions.
Recent calculations, however, suggest that the oxydioxirane
anion does not participate directly in the base-promoted epoxi-
dation of alkenes by peroxy acids.11
FIGURE 1. (a) The ESI-MS spectrum of a 10 µM aqueous methanol
-
solution of C6H5CO3H. (b) The ESI-MS/MS spectrum of the C6H5CO3
anion, m/z 137. Both spectra were obtained using a QuattroMicro
(Waters, Manchester) triple quadrupole mass spectrometer.
unimolecular reactivity of percarboxylate anions is unique. In
particular, we focus on a novel rearrangement of the perbenzoate
anions that precedes fragmentation and results in loss of carbon
dioxide. The mechanism for this rearrangement is elucidated
via isotopic labeling studies, comparison of substituted homo-
logues, and quantum chemical calculations.
Results and Discussion
Electrospray Ionization of Peracids: The ESI-MS spectrum
of an aqueous methanol solution of perbenzoic acid is given in
Figure 1a and is representative of the spectra obtained from the
other peracids examined in this study. It is significant that the
peak at m/z 137 that corresponds to the perbenzoate anion,
C6H5CO3-, is relatively weak compared to the intense signal
for the benzoate anion, C6H5CO2-, at m/z 121. This difference
in ion abundance may be due to the preferential ionization of
benzoic acid present in the sample. Preferential deprotona-
tion of benzoic acid would be anticipated based on its greater
acid strength in both solution (i.e., pKa[C6H5CO2H] ) 4.2
compared to pKa[C6H5CO3H] ) 8.3)17 and the gas phase (cf.,
In the present study, we have generated and detected a range
of perbenzoate anions in the gas phase by electrospray ionization
mass spectrometry (ESI-MS) of solutions of peroxybenzoic
acids. Activation of these ions by collision with argon gas results
in fragmentation and provides insight into the fundamental
unimolecular reactivity of perbenzoate anions in the absence
of solvent effects or counterions. While the unimolecular
fragmentation of deprotonated alkyl peroxides has previously
been reported,19-21 the data presented here suggest that the
∆
acidH298[C6H5CO2H] ) 1423 ( 12 kJ mol-1 compared to the
only peracid measured, ∆acidH298[HCO3H] ) 1551 ( 9 kJ
mol-1).22,23 Adding authentic benzoic acid to the electrospray
solution increases the relative abundance of the m/z 121 ion;
however, it fails to quench the perbenzoate signal at m/z 137,
even in 1:1 mixtures of the two acids, as would be expected if
ionization of the former were actively suppressing the latter.
Alternatively, some perbenzoic acid molecules may undergo
one-electron reduction during the electrospray process to form
benzoate anions and hydroxyl radicals. This second possibility
seems more likely as the oxidative dissociation of the O-O
bond to form the benzoate anion and hydroxyl radical is
thermodynamically favorable (electronic structure calculations
give ∆Hrxn ∼ -190 kJ mol-1), and indeed a decrease in the
(9) Zimmerman, H. E.; Singer, L.; Thyagarajan, B. S. J. Am. Chem. Soc.
1959, 81, 108.
(10) Apeloig, Y.; Karni, M.; Rappoport, Z. J. Am. Chem. Soc. 1983,
105, 2784.
(11) Washington, I.; Houk, K. N. Org. Lett. 2002, 4, 2661.
(12) Ye, D. Y.; Fringuelli, F.; Piermatti, O.; Pizzo, F. J. Org. Chem.
1997, 62, 3748.
(13) Fioroni, G.; Fringuelli, F.; Pizzo, F.; Vaccaro, L. Green Chem. 2003,
5, 425.
(14) Fringuelli, F.; Germani, R.; Pizzo, F.; Santinelli, F.; Savelli, G. J.
Org. Chem. 1992, 57, 1198.
(15) Fringuelli, F.; Pizzo, F.; Germani, R.; Savelli, G. Org. Prep. Proc.
Int. 1989, 21, 757.
(16) Fringuelli, F.; Germani, R.; Pizzo, F.; Savelli, G. Tetrahedron Lett.
1989, 30, 1427.
(17) Ripin, D.; Evans, D. pKa Data, December 22, 2005, http://
daecr1.harvard.edu/pKa/pka.html.
(20) Schalley, C. A.; Schroeder, D.; Schwarz, H.; Moebus, K.; Boche,
G. Chem. Ber. Recueil 1997, 130, 1085.
(21) Domingues, P.; Simoes, M. M. Q.; Cardoso, A. M.; Cavaleiro, A.
M. V.; Cavaleiro, J. A. S.; Johnstone, R. A. W.; Ferrer-Correia, A. J. Rapid
Commun. Mass Spectrom. 1999, 13, 93.
(18) Porter, N. A.; Yin, H. Y.; Pratt, D. A. J. Am. Chem. Soc. 2000,
122, 11272.
(19) Blanksby, S. J.; Bierbaum, V. M.; Ellison, G. B.; Kato, S. Aust. J.
Chem. 2003, 56, 459.
(22) Caldwell, G.; Renneboog, R.; Kebarle, P. Can. J. Chem. 1989, 67,
661.
(23) Bowie, J. H.; DePuy, C. H.; Sullivan, S. A.; Bierbaum, V. M. Can.
J. Chem. 1986, 64, 1046.
J. Org. Chem, Vol. 71, No. 21, 2006 7997

Harman et al.
TABLE 1. The Tandem Mass Spectra of Additional Precursor
Ions Examined in This Study (i.e., those not given as figures; all
spectra reported in this table were recorded on a triple quadrupole
mass spectrometer except where indicated)
SCHEME 2
product ions
precursor ion
m/z (% abundance)a
[C6H5CO2H - H]-
121(100), 77(59), 17(2)
143(47), 127(56), 126(2), 125(26),
99(5), 97(100),c 95(41), 71(3),
69(12), 55(2), 45(12), 17(5)
138(16), 122(100), 93(23),
77(7), 65(6)
[cyclo-C6H11CO3H - H]-
[C6H513CO3H - H]-
abundance of m/z 137 with respect to m/z 121 is observed with
increasing electrospray voltage. Electrochemical processes of
this kind have previously been observed at the electrospray
capillary.24 In either case, the result is a poor yield of the target
perbenzoate anion, suggesting that negative ion electrospray
ionization is a poor methodology for detection of small quantities
of perbenzoic acids in an analytical context. The ion abundances
achieved were, however, sufficient to examine the collision-
induced fragmentation of the anions in the triple quadrupole
mass spectrometer.
[13C-1-ring-C6H5CO3H - H]-
138(8.0), 122(100), 94(15),
78(7), 66(4)
[13C-1-ring-C6H5CO3H - H]-
13C-C6H5O-
f
94(37), 66(100)b
a Ion abundance is given relative to the base peak. b This MS3 spectrum
was recorded on a ThermoFinnigan LTQ linear quadrupole ion trap. c The
base peak at m/z 97 is likely to be deprotonated cyclohexanone, formed
via stepwise and/or concerted loss of water and carbon monoxide.
mechanisms to account for the observed rearrangement, namely,
a nucleophilic aromatic substitution at the ipso carbon (similar
to the Smiles rearrangement)26,30 or attack at the ortho position
followed by proton transfer. In the context of the perbenzoate
anion, these two mechanistic possibilities are outlined in Scheme
2.
Collision-Induced Dissociation of C6H5CO3-: The ESI-MS/
MS spectrum of the perbenzoate anion at m/z 137 is shown in
Figure 1b. The base peak in this spectrum appears at m/z 121
and corresponds to a loss of atomic oxygen (16 Da) and
formation of the benzoate anion from homolytic cleavage of
the O-O bond. The bond dissociation energy required for this
cleavage is calculated to be 84 kJ mol-1. Surprisingly, however,
an abundant fragment at m/z 93 is also observed that corresponds
to a loss of 44 Da. This fragmentation may be rationalized in
terms of the loss of carbon dioxide and formation of the
phenoxide anion: a pathway likely to be highly exothermic.
Furthermore, the fragment ion observed at m/z 65 in Figure 1b
is consistent with secondary fragmentation of phenoxide via loss
of carbon monoxide to form the aromatic cyclopentadienide
anion (Scheme 2). The latter fragmentation has previously been
reported for the collision-induced dissociation (CID) of the
phenoxide anion.25-27 The fragment ion at m/z 77 in Figure 1b
is most likely the phenyl anion resulting from the loss of 60
Da from the precursor. This neutral loss corresponds to a
concerted loss of CO3 or, alternatively, may arise from decar-
boxylation of the benzoate anion (m/z 121) in a stepwise process.
The former process seems unlikely on thermochemical grounds
given the relative instability of neutral CO3,28 while the
observation of facile decarboxylation of the benzoate anion in
a separate CID experiment (Table 1) supports the stepwise
mechanism.
Nucleophilic attack at the ortho position (Scheme 2a) should
result in a less strained five-centered cyclic intermediate;
however, the subsequent decarboxylation of this intermediate
requires a 1,2-proton transfer which is likely to represent the
rate-determining step. In contrast, the ipso mechanism does not
require an energetic proton transfer but occurs via the more
strained four-membered ring intermediate. These two possible
rearrangement mechanisms were investigated using isotopic
labeling and the examination of saturated homologues. The first
step of both mechanistic proposals involves intramolecular
nucleophilic attack on the aromatic ring. As such, the role of
the benzene ring was examined by obtaining the ESI-MS/MS
spectrum of the analogous saturated cyclic hydroperoxide, the
[M-H]- anion from cyclohexylpercarboxylic acid (Table 1).
This spectrum shows only a vanishingly small ion abundance
corresponding to the loss of carbon dioxide, suggesting that
indeed the aromatic ring plays a central role in the decarboxy-
lation of perbenzoate anions. The ESI-MS/MS spectrum of the
13C-isotopologue of perbenzoic acid labeled at the percarboxy-
late moiety (Table 1) shows exclusive formation of m/z 93,
confirming that the percarboxylate carbon is lost as carbon
dioxide. This observation is consistent with both the ipso and
ortho mechanisms shown in Scheme 2. The 13C-isotopologue
of perbenzoic acid with the label at the 1-position in the benzene
ring was also prepared. The ESI-MS/MS spectrum of the ring
labeled isotopologue (Table 1) shows the loss of unlabeled
carbon dioxide and formation of the 13C-phenoxide fragment
ion at m/z 94. This spectrum also shows that the cyclopenta-
dienide fragment ion at m/z 66 carries the 13C-label, indicating
that the 13C-labeled phenoxide anion has lost unlabeled carbon
monoxide. Previous CID studies of labeled phenoxide indicate
that the carbon bound to the oxygen is lost as carbon
monoxide.25-27 The asterisks in Scheme 2 indicate where the
13C-label would be expected following loss of carbon dioxide
via either the ortho or ipso pathway and thus where the label
would be expected in the resulting phenoxide. Clearly, the ipso
The carbon dioxide loss is competitive with fragmentation
via simple homolytic cleavage of the O-O bond and necessitates
an efficient rearrangement of the percarboxylate moiety. Such
a process has not previously been reported for peroxide anions
but, in some respects, is analogous to the expulsion of carbon
monoxide from deprotonated benzyl benzoates reported by
Bowie and co-workers.29 These authors proposed two possible
(24) de la Mora, J. F.; Van Berkel, G. J.; Enke, C. G.; Cole, R. B.;
Martinez-Sanchez, M.; Fenn, J. B. J. Mass Spectrom. 2000, 35, 939.
(25) Bowie, J. H.; Stringer, M. B. Org. Mass Spectrom. 1985, 20, 138.
(26) Eichinger, P. C. H.; Bowie, J. H.; Hayes, R. N. J. Am. Chem. Soc.
1989, 111, 4224.
(27) Binkley, R. W.; Tevesz, M. J. S.; Winnik, W. J. Org. Chem. 1992,
57, 5507.
(28) Pople, J. A.; Seeger, U.; Seeger, R.; Schleyer, P. v. R. J. Comput.
Chem. 1980, 1, 199.
(29) Chia, C. S. B.; Taylor, M. S.; Dua, S.; Blanksby, S. J.; Bowie, J.
H. J. Chem. Soc., Perkin Trans. 2 1998, 1435.
(30) Warren, L. A.; Smiles, S. J. Chem. Soc. 1930, 956.
7998 J. Org. Chem., Vol. 71, No. 21, 2006

Loss of CO2 from Perbenzoate Anions in the Gas Phase
SCHEME 3
position, as illustrated by the resonance structures in Scheme 3
and in apparent contradiction of labeling data from the perben-
zoate anion. This possibility was further investigated by
comparison of the CID spectrum of m/z 138 ions formed by
loss of CO2 in the ion source of the mass spectrometer with the
isobaric [M-H]- ions from ortho-, meta-, and para-nitrophe-
nol: the putative products of the decarboxylation. These data
are presented in Table 2 and show that fragmentation of the
source-formed [M-H-CO2]- ion from para-nitroperbenzoic
acid occurs predominantly via loss of NO and NO2 and
formation of the NO2 anion, which are analogous to the
fragments and relative abundances observed from the [M-H]-
anion of para-nitrophenol at the same collision energy. In
contrast, the relative abundances of these losses are different
for the meta- and ortho-nitrophenoxide anions (Table 2), and
furthermore, these species show significant fragment ions at m/z
91 and 80 not observed for the para-regioisomer. These data
suggest that the structure of the [M-H-CO2]- ion from para-
nitroperbenzoic acid is most closely related to the para-
nitrophenoxide anion and is thus consistent with loss of carbon
dioxide from the perbenzoate anion occurring via the ipso
mechanism (Scheme 3b).
FIGURE 2. ESI-MS/MS spectra of a 10 µM aqueous methanol
-
solution of (a) para-NO2-C6H4CO3- and (b) para-CH3O-C6H4CO3
.
Both spectra were obtained using a QuattroMicro (Waters, Manchester)
triple quadrupole mass spectrometer.
-
mechanism (Scheme 2b) would result in unlabeled C5H5 at
m/z 65, while the ortho mechanism (Scheme 2a) should produce
a
13C-labeled anion at m/z 66 as observed experimentally. To
confirm that m/z 66 is a secondary fragment arising from
decomposition of the labeled phenoxide, the CID of m/z 94 was
conducted in an MS3 experiment using an ion-trap mass
spectrometer (Table 1). This experiment showed exclusive loss
of unlabeled carbon monoxide, confirming that the rearrange-
ment of the perbenzoate anion occurs via the ortho mechanism
(Scheme 2a). This suggests that the proton transfer (rate-
determining) step in the ortho rearrangement requires less
activation energy than the four-centered intermediate in the ipso
rearrangement. This raises the intriguing question of whether
the activation energies and thus the reaction pathway are altered
by substitution of the aromatic ring.
Collision-Induced Dissociation of X-C6H4CO3-: Nucleo-
philic aromatic substitution reactions are generally favored in
aromatic compounds bearing electron-withdrawing groups due
to stabilization of negative charge in the Meisenheimer inter-
mediate31,32 and the consequent lowering of the activation
barrier. To further examine the generality and the mechanisms
of perbenzoate decarboxylation, we synthesized substituted
analogues and prepared the gas phase anions by ESI.
The CID spectrum of the [M-H]- anion from para-
nitroperbenzoic acid (m/z 182) is shown in Figure 2a and reveals
a significant abundance of the fragment ion at m/z 138
corresponding to loss of carbon dioxide and a base peak at m/z
108 that corresponds to the loss of CO2 and NO. Interestingly,
the loss of atomic oxygen by fission of the O-O bondsthe
base peak in the analogous spectrum of the perbenzoate anion
(Figure 1b)sis not observed in the spectrum of the para-nitro-
substituted homologue. Given that the O-O bond energy is
expected to be largely unperturbed by substitution of the
aromatic ring, this observation suggests that the activation
energy required for carbon dioxide expulsion has been signifi-
cantly reduced by nitro substitution. Surprisingly, however,
enhancement of the rearrangement channel by substitution of a
strong electron-withdrawing moiety at the para position is
indicative of nucleophilic attack at the ipso rather than the ortho
A similar trend was observed for ortho-nitroperbenzoic acid,
which also showed significant loss of carbon dioxide and a
product ion with strong structural similarities to ortho-nitro-
phenoxide (Table 2). In contrast, the meta-nitroperbenzoate
anion demonstrated less fragmentation with a normalized
precursor ion abundance of 40% compared to 28 and 17% for
ortho- and para-regioisomers, respectively. Furthermore, the
CID spectrum of source-formed [M-H-CO2]- product ions
(Table 2) did not match any of the isomeric nitrophenoxide
anions and indeed revealed abundant fragment ions not observed
for any of the authentic phenoxides, such as m/z 73, 63, and
41. These data indicate that the meta-isomer undergoes loss of
carbon dioxide via a mechanism different than that of the ortho-
and para-isomers. The alternative mechanism does not appear
to result in direct formation of pure nitrophenoxide anions as
predicted by Scheme 3, and thus an alternative explanation is
required.
For comparison, the effects of substitution with an electron-
donating group were also investigated, and the CID spectra of
meta- and para-methoxyperbenzoate anions are presented in
Table 3. Measured at the same collision energy as the nitro-
substituted homologues, the methoxyperbenzoate anions showed
significantly less fragmentation corresponding to the loss of
carbon dioxide and less fragmentation overall with normalized
precursor ion abundances of around 70% compared to 20-40%
for the nitroperbenzoates. These observations are consistent with
the destabilization of intermediates in the aromatic substitution
reaction by the electron-donating methoxy moiety. The CID
(31) Meisenheimer, J. Liebigs Ann. Chem. 1902, 323, 205.
(32) Artamkina, G. A.; Egorov, M. P.; Beletskaya, I. P. Chem. ReV. 1982,
82, 427.
J. Org. Chem, Vol. 71, No. 21, 2006 7999

Harman et al.
TABLE 2. (a) The Tandem Mass Spectra of the [M-H]- Ions from Nitroperbenzoic Acids (collision energy ) 10 eV) and (b) The Tandem
Mass Spectra of Source-Formed [M-H-CO2]- Ions from Nitroperbenzoic Acids and the [M-H]- Ions Formed from Related Phenols (collision
energy ) 20 eV)a
(a) tandem mass spectra of the [M-H]- ions
m/z
182
[M-H]
152
-NO
138
-CO2
136
-NO2
135
-HNO2
120
-NO2
-O
108
-CO2
-NO
44(17)
6(3)
92
91
80
C5H4O
64
C5H4
46
NO2
loss or fragment
-CO2
-NO2
14(5)
-CO2
-HNO2
2(1)
100(41)
2(1)
ortho-NO2
meta-NO2
para-NO2
72b(28)c
96(40)
24(17)
9(4)
2(1)
100(39)
36(15)
17(12)
4(2)
1(1)
2(1)
1(0)
7(3)
2(1)
2(1)
100(69)
(b) tandem mass spectra of source-formed [M-H-CO2]- ions
m/z
138
108
94
-CO2
7
92
-NO2
21
13
14
91
-HNO2
7
70
2
13
7
80
C5H4O
16
19
2
73
C2H3NO2
63
C5H3
46
NO2
62
68
18
41
C2HO
loss or fragment
ortho-NO2 peracid
meta-NO2 peracid
para-NO2 peracid
ortho-NO2 phenol
meta-NO2 phenol
-NO
100
100
100
100
49
58
43
44
59
37
14
28
38
28
4
8
67
78
100
a Ion abundances are reported for each m/z as a percentage of the base peak, and for some spectra, normalized ion abundances are provided in parentheses.
All spectra reported in these tables were recorded on a Waters QuattroMicro triple quadrupole mass spectrometer. b Ion abundance as a percentage of the
base peak. c Normalized ion abundance as a percentage of total ion abundance calculated from In/∑Ii × 100.
TABLE 3. (a) The Tandem Mass Spectra of the [M-H]- Ions from Methoxyperbenzoic Acids (collision energy ) 10 eV) and (b) the Tandem
Mass Spectra of Source-Formed [M-H-CO2]- Ions from Methoxyperbenzoic Acids and the [M-H]- Ions Formed from Related Phenols
(collision energy ) 10 eV)a
(a) tandem mass spectra of the [M-H]- ions
m/z
167
[M-H]
152
-CH3
123
-CO2
108
-CH3
-CO2
15(11)
5(4)
loss or fragment
meta-OCH3
para- OCH3
100b(68)c
100(73)
26(18)
10(7)
5(3)
22(16)
(b) tandem mass spectra of source-formed [M-H-CO2]- ions
m/z
123
[M-H]
108
-CH3
105
-H2O
95
-CO
93
-CH2O
91
-CH3O
-H
80
-CH3
-CO
3
79
-CO2
77
C6H5
61
C6H
loss or fragment
meta-OCH3 peracid
para-OCH3 peracid
ortho-OCH3 phenol
meta-OCH3 phenol
para-OCH3 phenol
100
100
56
100
17
60
38
100
53
18
14
2
1
1
4
3
9
2
4
3
5
100
a Ion abundances are reported for each m/z as a percentage of the base peak, and for some spectra, normalized ion abundances are provided in parentheses.
All spectra reported in these tables were recorded on a Waters QuattroMicro triple quadrupole mass spectrometer. b Ion abundance as a percentage of the
base peak. c Normalized ion abundance as a percentage of total ion abundance calculated from In/∑Ii × 100.
TABLE 4. The Electronic Energy (Hartrees), Zero-Point
spectra obtained from source-formed [M-H-CO2]- ions from
(Hartrees) and Relative Energies (0 K, kJ mol-1) Calculated at the
both meta- and para-methoxyperbenzoic acid are presented in
B3LYP/6-311++G(d,p) for All Stationary Points Involved in the
Rearrangement of the Perbenzoate Anion as Illustrated in Figure 3a
Table 3 and are similar in fragment ions and abundances to
each other as well as to the meta-methoxyphenoxide anion. This
is consistent with minimization of the destabilizing effects of
the electron-donating substituent in the ortho mechanism for
para-methoxyperbenzoate (Scheme 3a) and in the ipso mech-
anism in the case of meta-methoxyperbenzoate (Scheme 3b).
Some variation was observed between the [M-H-CO2]- ions
from the methoxyperbenzoic acids and the authentic methoxy-
phenoxides, most notably in the relative abundance of ions at
m/z 95 corresponding to carbon monoxide loss. As was the case
for the meta-nitroperbenzoate anions, this indicates the presence
of isomeric product ions and thus alternative or competing
mechanisms for CO2 loss from the perbenzoate.
These data indicate that both rearrangement mechanisms
originally proposed in Scheme 2 are plausible and that the
energetic requirements are comparable and are influenced by
the nature and position of substituents on the benzene ring.
Furthermore, there is evidence that the loss of CO2 does not
result solely in pure phenoxide ions as suggested by Schemes
2 and 3 but that isomeric ions are also present.
molecular
anion
electronic
energy
zero-point
correction
relative
energy
-
C6H5-CO3
-495.52139
-495.48059
-495.57069
-495.56461
-495.57369
-306.85362
-495.49161
-495.49680
-495.48366
-495.57260
-306.84514
-495.45073
-306.99406
-188.64691
0.10407
0.10135
0.10438
0.10308
0.10388
0.08712
0.10317
0.10394
0.10239
0.10486
0.00944
0.10049
0.08945
0.01169
0
ipso-TS1 (-568 cm-1
ipso-IM1
)
100.0
-128.6
-116.1
-137.8
ipso-TS2 (-413 cm-1
ipso-IM2
)
ipso-P1
ortho-TS1 (-323 cm-1
ortho-IM1
)
75.9
64.2
94.7
ortho-TS2 (-503 cm-1
ortho-IM2
)
-132.4
ortho-P2
ortho-TS3 (-474 cm-1
)
176.1
C6H5O-
CO2
a Transition state imaginary frequencies are provided in parentheses, and
structures of all stationary points are provided as Supporting Information.
Electronic Structure Calculations. The experimental data
discussed above suggest that carbon dioxide loss from perben-
8000 J. Org. Chem., Vol. 71, No. 21, 2006

Loss of CO2 from Perbenzoate Anions in the Gas Phase
-
FIGURE 3. Reaction coordinate diagram for the unimolecular rearrangement of C6H5CO3 calculated at the B3LYP/6-311++G(d,p) level of
theory. All energies are given in kJ mol-1
.
SCHEME 4
zoate anions can proceed via either an ipso or an ortho
mechanism (Scheme 2) depending on the nature and position
of substitution of the aromatic ring. Electronic structure calcula-
tions have been performed to investigate these mechanisms in
detail. The calculated potential energy surface for rearrangement
-
of the C6H5-CO3 anion and subsequent decarboxylation is
shown in Figure 3 with energies given relative to the percar-
boxylate anion. The electronic energies of critical stationary
points are listed in Table 2 with molecular structures and full
Cartesian coordinates provided as Supporting Information.
The optimized structure of the anti-conformation of the
perbenzoate anion is provided as Supporting Information, and
energies of all stationary points are relative to this species.
Although the syn-conformation (not shown) is 0.8 kJ mol-1
lower in energy at the chosen level of theory, only the anti-
conformation is suitably oriented to undergo the nucleophilic
aromatic substitution reactions investigated here. Considering
the ortho mechanism first, attack of the peroxide oxygen at the
ortho carbon occurs via the transition state ortho-TS1. This
presents a barrier of 75.9 kJ mol-1 to rearrangement to the five-
centered endoperoxide, ortho-IM1, which lies some 64.2 kJ
mol-1 above the percarboxylate minimum. The structure of the
endoperoxide is representative of a Meisenheimer intermediate
with a near tetrahedral geometry about the ortho carbon with
the adjacent C-C bonds lengthened from a benzene-like 1.40
Å in the perbenzoate anion to an alkane-like 1.50 Å in the
intermediate structure. The endothermicity of this first step can
be attributed to the loss of aromaticity which is somewhat offset
by resonance stabilization of the negative charge over several
carbon centers and the carbonyl oxygen. A transition state,
ortho-TS3, was located for the 1,2-proton transfer step postu-
lated in Scheme 2a. Intrinsic reaction coordinate calculations
on this transition state show that it connects the endoperoxide
intermediate, ortho-IM1, with the minimum corresponding to
phenoxide anion and carbon dioxide. The structure for ortho-
TS3 shows a late transition state with the proton already bound
at the ipso position, reminiscent of the intermediate structure
proposed in Scheme 2a. The geometry of ortho-TS3 reveals a
lengthening of the O-O bond to 1.51 Å in preparation for
cleavage of peroxide bond and coincident lengthening of the
C-C(O)O bond to 1.51 Å for subsequent decarboxylation.
Although the phenoxide product channel accessed by this
transition state represents the global minimumssome 322 kJ
mol-1 lower in energy than the perbenzoate precursorsthe
barrier to this process is significant, Ea ) 176 kJ mol-1
.
Interestingly, a lower energy pathway to decarboxylation was
identified whereby the endoperoxide intermediate, ortho-IM1,
rearranges via transition state, ortho-TS2, to form the intermedi-
ate epoxide ortho-IM2. This reaction is driven by the cleavage
of the weak peroxide bond and formation of the stable
carboxylate anion as outlined in Scheme 4. The barrier to this
process is only 95 kJ mol-1, and the rearrangement provides
the isomerized intermediate with sufficient internal energy to
undergo decarboxylation resulting in the product ion, ortho-
P2. Although this product is more than 400 kJ mol-1 less stable
than the isomeric phenoxide anion, the overall activation energy
for this exit channel is some 81 kJ mol-1 lower than the
phenoxide forming mechanism (see Figure 3).
J. Org. Chem, Vol. 71, No. 21, 2006 8001

Harman et al.
13C-labeled cyclopentadienide anion would be expected in line
with experimental observations.34
SCHEME 5
The effects of substitution on the rearrangement of the
perbenzoate were also modeled by calculating the reaction
coordinate data for the para-methoxyperbenzoate (Figure 4a)
and para-nitroperbenzoate (Figure 4b). Full energy and geo-
metric data for each of the stationary points in these figures are
provided as Supporting Information. Considering the case of
nitro substitution first, the activation barrier to nucleophilic
aromatic substitution at the ipso position is lowered by 55 kJ
mol-1 with respect to the unsubstituted system, while the
activation energy for the rate-determining step along the ortho
pathway remains largely unperturbed by the nitro moiety. This
preferential stabilization by the electron-withdrawing nitro
substituent on the ipso-TS1 transition state may be understood
in terms of the delocalization of electron density onto the para-
carbon as outlined in Scheme 3b. Interestingly, the transition
state for 1,2-proton-transfer-coupled decarboxylation, ortho-TS3,
is also significantly lowered by the presence of the nitro moiety
and becomes more favorable than the competing epoxidation
channel. Despite the large decrease in activation energy (ca.
100 kJ mol-1) for this phenoxide-forming pathway, all ortho
mechanisms are well above the alternative ipso pathway that
requires only 45 kJ mol-1 of activation energy. Conversely, the
same rationale would predict an increase in the activation energy
required for the ipso rearrangement in the presence of the
electron-donating methoxy substituent. Indeed this trend is
observed, with the computational data presented in Figure 4a
showing the ipso pathway is destabilized by some 17 kJ mol-1
in the presence of the methoxy group, while the ortho coordinate
is largely unaffected by the substitution. The computational data
would thus predict a change in the mechanism leading to loss
of carbon dioxide with the change from electron-donating to
electron-withdrawing substituents at the para position. This
result is thus consistent with the experimental findings outlined
above. Furthermore, both potential energy surfaces reveal that
decarboxylation by either mechanism results, at least initially,
in a substituted oxepin or benzene oxide anions rather than
phenoxide ions as originally hypothesized. While subsequent
rearrangement to the global minimum seems likely, this may
occur to a greater or lesser extent depending on the internal
energy distribution of the nascent product ions. This may
account for some of the differences observed experimentally
between source-formed product ions and authentic phenoxide
anions (Tables 2 and 3).
In contrast, rearrangement of the perbenzoate anion via
nucleophilic aromatic substitution at the ipso carbon is calculated
to occur via the transition state ipso-TS1 that lies 100 kJ mol-1
above the entrance channel. The optimized structure of the
bicyclic transition state is reminiscent of the reactive intermedi-
ate proposed in Scheme 2b. The additional strain of the four-
membered ring in addition to the inability to delocalize the
charge onto oxygen in this structure may account for the greater
activation energy (24 kJ mol-1) required for ipso rearrangement.
No bicyclic minimum could be located, and intrinsic reaction
coordinate calculations on ipso-TS1 show that it is connected
to the substituted benzene oxide intermediate ipso-IM1 rather
than the phenoxide exit channel predicted in Scheme 2b. The
structure of ipso-IM1 is similar in geometry and energy to the
isomeric epoxide ortho-IM2 discussed above. The benzene
oxide-like intermediate, ipso-IM1, tautomerizes over a small
barrier (Ea ) 13 kJ mol-1) to form the more stable substituted
oxepin ipso-IM2. This pathway is analogous to the benzene
oxide-oxepin tautomerism calculated by Kassaee et al., who
found activation barriers of similar magnitude.33 The reaction
intermediate ipso-IM2 formed by rearrangement of the perben-
zoate anion has sufficient internal energy to undergo decar-
boxylation (Ea ) 213 kJ mol-1) to form the deprotonated oxepin
anion, ipso-P1. No transition state could be located for this
process, nor for direct decarboxylation of ipso-IM1. In both
instances, minimization of the carboxylate anions starting from
a range of C-C(O)O bond lengths resulted in either the starting
geometry or ipso-P1 and carbon dioxide: no stable minimum
corresponding to the deprotonated benzene oxide tautomer could
be located. Thus these calculations suggest a revision of the
mechanism for ipso rearrangement as outlined in Scheme 5.
To summarize the computed pathways for decarboxylation
of the perbenzoate anion, Figure 3 shows that the lowest energy
pathway is via initial nucleophilic aromatic substitution at the
ortho position as anticipated from experiment. Surprisingly,
however, the lowest energy pathways for decarboxylation via
either ortho or ipso mechanism do not result, at least initially,
in the thermodynamically favored phenoxide anion but rather
isobaric oxepin or benzene oxide anions. It seems highly
probable that subsequent rearrangement of these ions will result
in formation of the global minimum, but this may require
additional activation and is beyond the scope of this study. The
revised mechanisms shown in Schemes 4 and 5 appear
consistent with the results of the 13C-labeling studies discussed
earlier. The position of the label in each of these schemes is
indicated with an asterisk and shows that for the ipso mechanism
(Scheme 5) the label ends up adjacent to the oxygen in the
resulting oxepin anion, suggesting it would be lost in any
subsequent decarbonylation reaction. Conversely, for the ortho
mechanism (Scheme 4), the label is not bound to the oxygen
and is thus unlikely to be lost as carbon monoxide and thus a
Conclusions
The isotopic labeling data from the perbenzoate anion provide
strong evidence for unimolecular loss of carbon dioxide via the
ortho mechanism proposed in Scheme 2a. The combination of
substituent studies and electronic structure calculation, however,
reveals a somewhat more complex reaction manifold with
competing ortho and ipso mechanisms that have similar
activation energies directly influenced by the nature and position
of substituents on the aromatic ring. These data also indicate,
rather surprisingly, that the loss of carbon dioxide does not lead
directly to the phenoxide product ions but rather proceeds via
initial epoxidation of the benzene ring forming benzene oxide
(34) Hydrogen scrambling, which might confuse these experimental
results, has not been observed for deuterium-labeled C6H5O- anions even
under high energy collision conditions.27
(33) Kassaee, M. Z.; Arshadi, S.; Ahmadi-Taheri, N. J. Mol. Struct. 2005,
715, 107.
8002 J. Org. Chem., Vol. 71, No. 21, 2006

Loss of CO2 from Perbenzoate Anions in the Gas Phase
FIGURE 4. Reaction coordinate diagrams for the unimolecular rearrangement and decarboxylation of para-XC6H5CO3-, where (a) X ) OCH3
and (b) X ) NO2. All calculations were conducted at the B3LYP/6-311++G(d,p) level of theory with full geometric and energetic descriptions of
stationary points available as Supporting Information. Energies are given relative to the perbenzoate precursor ions in units of kJ mol-1. The dashed
lines indicate the possibility for further rearrangement of oxepin and benzene oxide isomers to phenoxide following decarboxylation that is outside
the scope of the present study.
or oxepin anions that may subsequently rearrange to the global
minimum (Schemes 4 and 5).
benzene ring during fragmentation of perbenzoate anions. The
decarboxylation reaction driving the epoxidation competes
favorably with homolytic processes, such as the loss of atomic
oxygen, nitric oxide, or methyl radicals. This intramolecular
epoxidation of the aromatic ring is analogous to the solution
phase intermolecular epoxidation of alkenes by percarboxylate
anions discussed earlier.11 It would therefore be of interest for
future studies to investigate whether perbenzoate anions can
initiate (i) analogous intramolecular aromatic epoxidation reac-
tions in solution36 and, conversely, (ii) intermolecular aromatic
epoxidation reactions in the gas phase.
Simple organic reactions that result in the oxidation of
benzene are relatively few due to the aromatic stability of the
benzene ring. This is one of the reasons that dioxin pollutants
are so persistent in the environment and so much effort is
currently being expended in developing enzymatic and metal-
based catalysts to facilitate oxidative degradation of these
aromatic compounds.35 It is intriguing therefore that in this study
we have observed facile unimolecular epoxidation of the
(35) Finocchio, E.; Busca, G.; Notaro, M. Appl. Catal., A 2006, 62, 12.
J. Org. Chem, Vol. 71, No. 21, 2006 8003

Harman et al.
containers must not be sealed.41,42 The highly concentrated hydrogen
peroxide prepared in this study was stored in the dark at 4 °C in a
clean Schott bottle with a vented cap.
Experimental Section
Mass Spectrometry. Standard solutions of peracids (10 µM)
were prepared in aqueous methanol (50% v/v) with the pH adjusted
to 9-10 using aqueous ammonia. Mass spectra were obtained using
a QuattroMicro triple quadrupole mass spectrometer (Waters,
Manchester, UK). Spectra were obtained by infusion of the standard
solution (10 µL/min), typical settings were cone voltage 60 V,
capillary voltage 3 kV, source temperature 80 °C. ESI-MS spectra
were obtained by scanning Q1 while operating Q3 in Rf-only mode.
Resolution for ESI-MS and ESI-MS/MS experiments was typically
0.5 Th across the entire mass range. ESI-MS/MS spectra were
obtained by mass-selecting the parent ion using Q1 and scanning
for product ions using Q3. Argon was used as the collision gas at
a pressure of 4 × 10-3 Torr. ESI-MS and ESI-MS/MS spectral
data presented in this paper result from the average of at least 50
scans. The data were baseline subtracted (40% background subtract
with a first-order polynomial) and smoothed (two mean smooths
using a peak width of 0.5 Th) using the MassLynx software (Waters,
Manchester UK). High-resolution ESI-MS spectra of deprotonated
peracids were obtained using a Q-ToF Ultima quadrupole-time-
of-flight instrument (Waters, Manchester, UK). MS3 spectra were
obtained in a LTQ quadrupole linear ion-trap mass spectrometer
(ThermoFinnigan, San Jose, USA).
Synthesis and Characterization of Perbenzoic acids. Most
peracids were prepared by procedures based on those of Swern
and co-workers, in which a carboxylic acid was suspended in
methanesulfonic acid and treated with concentrated aqueous
hydrogen peroxide.37,38 These preparations were carried out in open
beakers behind a blast shield as a precaution. The highly reactive
4-methoxyperbenzoic acid was prepared from the acid chloride by
adapting a procedure of Lefort and co-workers.39 Peracids were
stored in the dark at -18 °C, and purities were determined
iodometrically by treatment of a weighed sample in acetone/acetic
acid solution with excess saturated aqueous KI, then titration of
the liberated iodine with a standard aqueous Na2S2O3 solution.40
FT-IR spectra were obtained in chloroform solution. Additional
characterization is provided for previously reported compounds
where literature data are absent. Yields are adjusted for purity.
CAUTION: Peracids may explode when heated or detonated
by shock or sparks and are chemically incompatible with strong
reducing agents, transition metals, and easily oxidized organic
compounds. Open reaction vessels and nonflammable solvents were
used wherever possible as recommended by Swern.2 Excessive heat
and agitation of peracid solutions was avoided during rotary
evaporation, and samples were stored at -18 °C.
Concentration of Hydrogen Peroxide. A Dreschel bottle was
charged with aqueous H2O2 (500 mL, 44% m/m), and dry air was
bubbled slowly through the solution via a sintered glass frit at room
temperature until the volume reduced to ∼175 mL (6 weeks). The
concentration was found to be 84% by titration against an acidified
KMnO4 standard solution. This sample was purified by evaporation
and cold-trapping (liq. N2) almost to dryness under high vacuum
at room temperature. Concentration under vacuum at room tem-
perature yielded a 92% solution.
Cyclohexanepercarboxylic acid. Cyclohexanecarboxylic acid
(14 mg, 0.11 mmol) was added to a stirred solution of H2O2 (92%,
0.30 g, 8.0 mmol) in methanesulfonic acid (MsOH, 0.55 g, 5.7
mmol) at 0 °C. The mixture was heated gradually to 40 °C and
kept at that temperature for 2 h, before cooling to 0 °C and diluting
with cold H2O (5 mL). The resulting mixture was extracted with
CH2Cl2 (5 mL), the extract was washed with H2O (5 mL), dried
(MgSO4), and rotary evaporated at room temperature to yield a
colorless oil (16 mg, 97% yield): δH (CDCl3, 300 MHz) 1.20-
1.38 (3H, m), 1.40-1.55 (2H, m), 1.64-1.71 (1H, m), 1.75-1.78
(2H, m), 1.87-1.95 (2H, m), 2.31-2.39 (1H, m), 11.39 (1H, br);
δC (CDCl3, 75 MHz) δ 25.3, 25.7, 28.8, 42.8, 161.6; m/z (ESI)
found 143.0732, calcd for C7H11O3 (M - H)- 143.0708.
Perbenzoic acid.37 Hydrogen peroxide (3.2 g, 44% m/m, 41
mmol) was added dropwise to a stirred suspension of benzoic acid
(3.7 g, 30 mmol) in MsOH (14.4 g, 150 mmol) at 40 °C. After 4.5
h, the colorless solution was cooled, treated with ice (10 g), followed
by saturated aqueous (NH4)2SO4 solution (10 mL). The resulting
mixture was extracted with CH2Cl2 (3 × 10 mL), the combined
extracts were washed with saturated aqueous (NH4)2SO4 (3 × 10
mL), dried (MgSO4), and concentrated by rotary evaporation at
room temperature to afford a white crystalline solid (4.2 g, 83%
yield) that was analyzed as 83.0% peracid by iodometric titration.
A small amount of the crude perbenzoic acid was purified further
by flash chromatography (Rf ) 0.3, CH2Cl2 on silica) for charac-
terization: νmax/cm-1 3514 br, 3265 br, 1728 vs, 935 w; δH (CDCl3,
500 MHz) 7.50 (2H, dd, J ) 7.8 Hz), 7.65 (1H, dd, J ) 7.8, 0.98
Hz), 8.00 (2H, m, J ) ∼8, 0.98 Hz), 11.64 (1H, s); δC (CDCl3,
125 MHz) 125.2, 128.9, 129.3, 134.4, 168.1; m/z (ESI) found
137.0270, calcd for C7H5O3 (M - H)- 137.0239.
Perbenzoic acid-Ring-1-13C and Perbenzoic acid-Carboxy-
13C. The 13C-isotopologues of perbenzoic acid, perbenzoic acid-
ring-1-13C (34%), and perbenzoic acid-carboxy-13C (33%) were
prepared and purified similarly on a 0.4 mmol scale in the yields
indicated, from benzoic acid-ring-1-13C (Icon Isotopes, 99 atom
%) and benzoic acid-carboxy-1-13C (Sigma-Aldrich 99 atom %),
respectively. The only differences were (a) the hydrogen peroxide
addition was done at 0 °C and the reaction was immediately allowed
to warm to room temperature; (b) the CH2Cl2 extracts were washed
with cold water instead of aqueous (NH4)2SO4.
2-Nitroperbenzoic acid.37 To a stirred suspension of 2-nitroben-
zoic acid (5.0 g, 30 mmol) in MsOH (14.4 g, 150 mmol) at 35 °C
was added aqueous H2O2 (92% m/m, 3.6 g, 98 mmol) dropwise, at
a rate which kept the temperature below 40 °C. After 4 h, the
mixture was cooled to 10 °C and treated with ice (10 g) followed
by ice water (20 mL). The resulting solid was filtered, washed twice
with ice water, and dried in vacuo at room temperature to yield
pale yellow crystals (4.5 g, yield 64%) that were analyzed as 79.0%
peracid by iodometric titration. The crude peracid was suspended
in water (30 mL) at 10 °C, and aqueous NaOH (10 mL, 0.2% w/v,
6 mmol) was added to extract the residual carboxylic acid. The
suspension was filtered, and the precipitate was washed with ice
water (20 mL) and dried in vacuo. Recrystallization from Et2O/
pentane (twice) yielded pale yellow crystals that were analyzed as
100% peracid: mp 95.5-96.5 °C (lit. 95-96 °C in a sealed
CAUTION: Decomposition of hydrogen peroxide to oxygen and
water can occur rapidly under high pH conditions and by exposure
to transition metals and their ions. Contact with combustibles may
result in a fire or explosion. Build-up of oxygen can also occur
due to gradual decomposition of high purity H2O2, and thus storage
capillary);37 max/cm-1 3502 br, 3304 br, 1755 vs, 962 w; δH (CD3-
ν
OD (19 mol %) in CDCl3, 500 MHz) 2.85 (1H, br), 7.69-7.79
(3H, m), 8.10 (1H, dd, J ) 8.0, 1.0 Hz); δC (CDCl3, 75 MHz)
124.4, 124.9, 129.8, 132.3, 133.8, 147.0, 165.7; m/z (ESI) found
182.0116, calcd for C7H4NO5 (M - H)- 182.0089.
(36) Preliminary investigations in our laboratory find that both perbenzoic
acid and 4-nitroperbenzoic acid undergo minimal reaction in aqueous
alkaline solution even when heated at 120 °C in a sealed vessel for 3 days.
(37) Silbert, L. S.; Siegel, E.; Swern, D. J. Org. Chem. 1962, 27, 1336.
(38) Silbert, L. S.; Siegel, E.; Swern, D. Org. Synth. 1963, 43, 93.
(39) Nedelec, J. Y.; Sorba, J.; Lefort, D. Synthesis 1976, 12, 821.
(40) Swern, D. Organic Reactions; Wiley: New York, 1953; Vol. 7, p
378.
(41) Schumb, W. C.; Satterfield, C. N.; Wentworth, R. L. Hydrogen
Peroxide; Reinhold Publishing: New York, 1955; p 152.
(42) Shanley, E. S. In Organic Peroxides; Swern, D., Ed.; Wiley-
Interscience: New York, 1972; Vol. 3, p 341.
8004 J. Org. Chem., Vol. 71, No. 21, 2006

Loss of CO2 from Perbenzoate Anions in the Gas Phase
3-Nitroperbenzoic acid.43 Procedure as for 2-nitroperbenzoic
acid. 3-Nitrobenzoic acid (5.0 g, 30 mmol), MsOH (14.4 g, 150
mmol), and H2O2 (3.6 g of 88%, 93 mmol) afforded the product as
pale yellow crystals (4.9 g, 89%), which was analyzed as 99.7%
peracid by iodometric titration: mp 80-82 °C; νmax/cm-1 3508
br, 3286 br, 1743 vs, 922; δH (CDCl3, 300 MHz) 7.77 (1H, t, J )
8.1 Hz), 8.34 (1H, ddd, J ) 7.6, 1.8, 1.2 Hz), 8.52 (1H, ddd, J )
8.1, 2.3, 1.2 Hz), 8.85 (1H, m), 11.55 (1H, br); δC (CDCl3, 75 MHz)
124.4, 127.1, 128.7, 130.3, 134.8, 148.3, 165.8; m/z (ESI) found
182.0128, calcd for C7H4NO5 (M - H)- 182.0089.
was analyzed as 68.2% peracid by iodometric titration. Further
purification by flash chromatography (Rf ) 0.3, cold CH2Cl2 on
silica) achieved a purity of 84.1%: δH (CDCl3, 500 MHz) 3.88
(3H, s), 6.97 (2H, d, J ) 8.8 Hz), 7.95 (2H, d, J ) 8.8 Hz), 11.62
(1H, br); δC (CDCl3, 125 MHz) 55.5, 114.2, 121.9, 131.5, 164.4,
167.9; m/z (ESI) found 167.0368, calcd for C8H7O4 (M - H)-
167.0344.
Electronic Structure Calculations. Geometry optimizations
were carried out with the B3LYP method46,47 using the 6-311++G-
(d,p) basis set within the GAUSSIAN03 suite of programs.48 All
stationary points on the potential energy surface were characterized
as either minima (no imaginary frequencies) or transition states (one
imaginary frequency) by calculation of the frequencies using
analytical gradient procedures. Frequency calculations provided
zero-point energies, which were added to the calculated electronic
energy. The minima connected by a given transition state were
confirmed by inspection of the animated imaginary frequency using
the MOLDEN package49 and by intrinsic reaction coordinate
calculation.50,51
4-Nitroperbenzoic acid.37,43 Preparation as for 2-nitroperbenzoic
acid. 4-Nitrobenzoic acid (5.0 g, 30 mmol), MsOH (14.4 g, 150
mmol), and H2O2 (3.6 g of 88%, 93 mmol) afforded the product as
yellow crystals (5.4 g, 93%) that were analyzed as 95.8% peracid
by iodometric titration. Further purification was achieved by
washing a CH2Cl2 solution (300 mL) of the crude product (5.3 g)
with aqueous NaOH (0.038 M, 50 mL) then water (50 mL). Drying
(MgSO4) and solvent removal at room temperature afforded material
of 98.8% purity: mp 139 °C (vigorous decomposition) (lit. 138
°C decomposition);37 max/cm-1 3503 br, 3283 br, 1741 vs; δH
ν
(CDCl3, 300 MHz) 8.20 (2H, dm, J ) 9.0 Hz), 8.37 (2H, dm, J )
9.0 Hz), 11.56 (1H, s); δC (CDCl3, 75 MHz) 123.7, 124.0, 130.6,
131.3, 166.1; m/z (ESI) found 182.0123, calcd for C7H4NO5 (M -
H)- 182.0089.
Acknowledgment. S.J.B. acknowledges the financial support
of the University of Wollongong (URC New Researcher Grant),
the Institute for Biomolecular Science, and the Australian
Research Council (DP0452849). We also acknowledge the
support of the Australian Partnership for Advanced Computing
(ANU, Canberra) for a generous allocation of supercomputing
time.
3-Methoxyperbenzoic acid.44 Although Swern and co-workers37
report that their procedure was unsuccessful in the synthesis of
3-methoxyperbenzoic acid, in our hands, slight changes to the
method permitted its preparation. Stirred 3-methoxybenzoic acid
(5.0 g, 33 mmol) in MsOH (14.7 g, 153 mmol) was cooled to 0
°C, and H2O2 (88%, 3.5 g, 91 mmol) was added dropwise at such
a rate to keep the temperature below 50 °C. After 4 h, ice (50 g)
was added, the mixture extracted with CHCl3 (3 × 50 mL), and
the combined organic extracts were washed with H2O (50 mL) and
dried (MgSO4). Removal of solvent in vacuo at room temperature
afforded a brown/yellow solid (3.7 g, 26% yield), 38% peracid by
iodometric titration. A small quantity was purified for characteriza-
tion by flash chromatography (Rf ) 0.34, CH2Cl2 on silica), and
upon cooling, the resulting colorless oil crystallized: mp 32-34
°C; νmax/cm-1 3516 br, 3265 br, 1732 vs, 925 w; δH (CDCl3, 300
MHz) 3.85 (3H, s), 7.18 (1H, ddd, J ) 8.3, 2.6, 1.0 Hz), 7.41 (1H,
t, J ) 7.9 Hz), 7.49 (1H, dd, J ) 2.6, 1.8 Hz), 7.59 (1H, dt, J )
7.6, ∼1.2 Hz), 11.62 (1H, s); δC (CDCl3, 75 MHz) 55.5, 113.6,
121.0, 121.6, 126.3, 130.0, 159.8, 168.0; m/z (ESI) found 167.0371,
calcd for C8H7O4 (M - H)- 167.0344.
Supporting Information Available: Electronic energies, zero-
point energies, and molecular geometries (as Cartesian coordinates)
for all stationary points discussed in the text. This material is
available free of charge via the Internet at http://pubs.acs.org.
JO060730A
(46) Becke, A. D. J. Chem. Phys. 1993, 98, 1372.
(47) Lee, C. T.; Yang, W. T.; Parr, R. G. Phys. ReV. B: Condens. Matter
1988, 37, 785.
(48) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A.
G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, revision C.02; Gaussian Inc.: Wallingford, CT, 2004.
4-Methoxyperbenzoic acid.45 A solution of freshly prepared
4-methoxybenzoyl chloride (3.0 g, 18 mmol) in CH2Cl2 (6 mL)
was added dropwise to a stirred suspension of aqueous H2O2 (93%,
3.0 g, 82 mmol) in CH2Cl2 (6 mL) at 0 °C. After 2.5 h at 0 °C, the
mixture was poured into ice water (15 mL), extracted with CH2Cl2
(3 × 15 mL), and the combined extracts were washed with cold
water (10 mL) and dried (MgSO4). Rotary evaporation of solvent
at room temperature yielded a white solid (3.1 g, yield 72%) that
(49) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000,
14, 123.
(50) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154.
(51) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523.
(43) Pande, C. S.; Jain, N. Synth. Commun. 1988, 18, 2123.
(44) Blake, R. C., II; Coon, M. J. J. Biol. Chem. 1980, 255, 4100.
(45) Konen, D. A.; Silbert, L. S. J. Org. Chem. 1971, 36, 2162.
J. Org. Chem, Vol. 71, No. 21, 2006 8005