Electrophilic oxidant produced in the photodeoxygenation of 1,2-benzodiphenylene sulfoxide

Article abstract of DOI:10.1021/jo010009z

We report that the photodeoxygenation of 1,2-benzodiphenylene sulfoxide, 1, generates an intermediate capable of oxidizing the solvent benzene to phenol. The reactivity of the intermediate was probed with various substrates (2-methylbutane, chloride ion,

Full text of DOI:10.1021/jo010009z

4576
J . Org. Chem. 2001, 66, 4576-4579
Electr op h ilic Oxid a n t P r od u ced in th e P h otod eoxygen a tion of
1,2-Ben zod ip h en ylen e Su lfoxid e
Edlaine Lucien and Alexander Greer*
Department of Chemistry, Graduate School and University Center & The City University of New York
(CUNY), Brooklyn College, Brooklyn, New York 11210
agreer@brooklyn.cuny.edu
Received J anuary 3, 2001
We report that the photodeoxygenation of 1,2-benzodiphenylene sulfoxide, 1, generates an
intermediate capable of oxidizing the solvent benzene to phenol. The reactivity of the intermediate
was probed with various substrates (2-methylbutane, chloride ion, and para-substituted aryl
sulfides). The intermediate produced in the sulfoxide photodeoxygenation displays an electrophilic
oxidation chemistry. Our data on 1 contrast with the behavior of hydroxyl radical but resemble
the chemistry observed for gas-phase atomic oxygen [O(3P)] and for solution-phase photodeoxy-
genations of dibenzothiophene sulfoxide, 3, and pyridine N-oxide, 5. Correlations are made between
the ionization potential of the acceptor molecules and the logarithm of the relative rate constants
in order to advance the idea that the oxidizing agent of the title reaction may be solution-phase
O(3P).
Sch em e 1
In tr od u ction a n d Ba ck gr ou n d
The first reports of photoinduced deoxygenation of
aromatic sulfoxides came approximately 25 years ago.1,2
Since this time J enks and co-workers3,4 suggested a
mechanism for the photodeoxygenation of dibenzo-
thiophene sulfoxide (3), which suggested that along with
dibenzothiophene (4) an intermediate was produced
capable of oxidizing the solvent. The intermediate gener-
ated in the photodeoxygenation of 3 was tentatively
assigned as atomic oxygen O(3P) or a closely related
noncovalent complex. This assignment was based on the
oxidation chemistry, on comparison with known rate
constants of O(3P), and on the quantum yield data
(Scheme 1).
with gas-phase data on O(3P). Only minor amounts of
O(3P) appear to be formed (∼5%) in the photolysis of 5
compared to the photochemically rearranged product,
oxaziridine (6).
The gas-phase chemistry of singlet-excited (O(1D)) and
triplet-ground-state atomic oxygen (O(3P)) is known but
little is known of the condensed phase chemistry. The
gas-phase reactivity of O(3P) is established;5 however, the
potential value of O(3P) as a reagent in organic chemistry
and biochemistry remains unknown. The lack of informa-
tion partly reflects the scarcity of sources of O(3P) in
solution. Numerous studies have addressed metal-dioxo
species, superoxides, and other activated molecular oxy-
gen species, but none are known to produce O(3P).
Elegant work of Scaiano6 suggest that the formation of
O(3P) in acetonitrile solution occurs upon irradiation at
308 nm of pyridine N-oxide (5). The assignment of O(3P)
was based on the fact that the experimental method
allowed one to monitor the product of its reaction with
the solvent, acetonitrile oxide, and also by comparisons
Gas-phase methods that produce O(3P) are not readily
compatible with solution phase methods. Gas-phase
mercury triplet photosensitized decomposition of N2O
indicates formation of O(3P) based on energetics;5,7 how-
ever, a common method for spectroscopic detection of
O(3P) requires monitoring its vacuum UV laser-induced
fluorescence (LIF) signal at ∼130 nm.8 The disadvantage
of current methods that generate O(1D) or O(3P) in
solution, include (1) harsh photolysis conditions (e.g.,
[60Co] γ-ray irradiation of liquid water9 or alkaline H2O2
solutions,10 185 nm photolyses of N2O11), (2) the presence
of other reactive intermediates (e.g., oxoanions and
oxoacids,12 photodecomposition products of NO3-13), or (3)
(1) Gurria, G. M.; Posner, G. H. J . Org. Chem. 1973, 38, 2419-2420.
(2) Shelton, J . R.; Davis, K. E. Int. J . Sulfur Chem. 1973, 8, 217-
228.
(7) Cvetanovic, R. J . J . Chem. Phys. 1955, 23, 1203-1207.
(8) Gas-phase studies of atomic oxygen often employ the VUV-LIF
method, for example: Taniguchi, N.; Takahashi, K.; Matsumi, Y. J .
Phys. Chem. A 2000, 104, 8936-8944.
(3) Wan, Z.; J enks, W. S. J . Am. Chem. Soc. 1995, 117, 2667-8.
(4) Gregory, D. D.; Wan, Z.; J enks, W. S. J . Am. Chem. Soc. 1997,
944-102.
(9) Brown, W. G.; Hart, E. J . Radiat. Res. 1972, 51, 249-253.
(10) Sauer, M. C., J r.; Brown, W. G.; Hart, E. J . J . Phys. Chem. 1984,
88, 1398-1400.
(5) For a review of kinetic data on reactions of O(3P) atoms with
hydrocarbons see: Cvetanovic, R. J . J . Phys. Chem. Ref. Dat. 1987,
16, 261-326.
(11) Varkony, T. H.; Pass, S.; Mazur, Y. 1975, J . Chem. Soc., Chem.
Commun. 457-458.
(6) Bucher, G.; Scaiano, J . C. J . Phys. Chem. 1994, 98, 12471-12473.
10.1021/jo010009z CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/25/2001

Photodeoxygenation of 1,2-Benzodiphenylene Sulfoxide
J . Org. Chem., Vol. 66, No. 13, 2001 4577
Since gas-phase studies have established the electrophi-
licity of O(3P) by correlating such rate dependencies on
acceptor molecule ionization potentials,17-20 we reasoned
that a similar correlation could add confidence to the
suggestion that 3 and 5 may lose oxygen to yield O(3P)
in solution. The quality of the data in Figure 1B,C does
not constitute a “proof” of the O(3P) intermediate in
solution; however, it does provide strong circumstantial
evidence.
Solven t Oxid a tion . In the present work, we focus on
1, which is similar to J enks’ compound 33,4 except for the
additional fused benzene ring. This change moves the
second UV band maximum from 321 to 343 nm, and
enables us to induce similar chemistry at lower energies.
Compound 1 is a candidate for unimolecular SdO cleav-
age because of a comparable excited state (∼75-78 kcal/
mol, from the UV-visible spectrum) and sulfur-oxygen
bond dissociation energy (76.8 kcal/mol, from UB3LYP/
6-31G(d)//B3LYP/3-21G calculations).21 In the presence
of 385 nm light in argon-saturated benzene at room
temperature, 1 (2-6 mM) deoxygenated to give phenol
and 1,2-benzodiphenylene sulfide (2). Deoxygenation of
1 also proceeded in the presence of 300, 320, 340, or 365
nm light. Products that would indicate disproportionation
(1,2-benzodiphenylene sulfone) or sulfur monoxide extru-
sion (2-phenylnaphthalene) were not detected. At 8%
conversion the ratio of phenol:2 is 0.81 by gas chromato-
graphic analysis. Identical results were obtained in
acetonitrile:benzene (2.9 M:9.5 M) mixtures at room
temperature. Photolyses were conducted with a 75-W
Xenon lamp focused on a monochromator (linear disper-
sion equal to ∼(12 nm). Control experiments demon-
strate that benzene is inert to 1 in the absence of light.
We next wanted to learn if this hydroxylation exhibits
any selectivity.
Alk a n e Hyd r oxyla tion . Photodeoxygenation of 1 (4
mM) in acetonitrile solution containing 0.687 M 2-me-
thylbutane demonstrated a regioselective hydroxylation
yielding 2-methyl-2-butanol (71%), 3-methyl-2-butanol
(19%), 2-methyl-1-butanol (4%), and 3-methyl-1-butanol
(5%). The ratios of alcohol products observed from the
photodeoxygenation of 1 are similar to those reported
from the photodeoxygenation of 3.3,4 One might suggest
that the selectivity for the formation of alcohol products
in the photodeoxygenation of 1 and 3 is derived from a
recombination of a caged hydroxyl radical-alkyl radical
pairs in a stepwise carbon-hydrogen bond insertion
process. Gas-phase alkane hydroxylations have shown
that O(3P) reacts selectively at tertiary C-H,22 whereas
O(1D)11 and OH radicals6,23 show little if any selectively.
F igu r e 1. Correlation of log krel for oxidation with adiabatic
ionization potentials of the acceptor molecules. Oxidation of
alkenes (A, slope ) -0.950 eV-1, r2 ) 0.898; ref 5; a ) 2,3-
dimethyl-2-butene; b ) 1-methyl-1-cyclohexene; c ) 2-methyl-
2-butene; d ) cyclohexene; e ) cyclopentene; f ) 2-pentene;
g ) trans-2-butene; h ) cis-2-butene; i ) 2-methyl-2-propene;
j ) 1-hexene; k ) 1-pentene; l ) 3-methyl-1-butene; m )
1-butene; n ) propene; o ) ethene) with gas-phase O(3P)
atoms. Oxidation of substrates from the photolysis of 5 (B,
slope ) -1.193 eV-1, r2 ) 0.845, ref 6; p ) trimethyl phosphite;
q ) cyclopentene; r ) pyridine; s ) 1-octyne; t ) cyclohexane;
u ) 1-octene; v ) benzene; w ) n-pentane; x ) cyclopentane;
y ) dichlromethane; z ) chloroform; a ′ ) acetonitrile; b′ )
methane) in acetonitrile solution and 3 (C (not labeled), slope
-1.377 eV-1, r2 ) 0.515, refs 3,4; c′ ) cyclohexane; d ′ )
cyclohexane-d12; e′ ) 2-methylbutane) in alkane solutions.
the need for high-pressure (e.g., SF6)14 or low temperature
(N2)15 solvents. We now report a much cleaner oxidation
method. Mildly energetic 385 nm light can initiate the
deoxygenation of 1,2-benzodiphenylene sulfoxide (1),
which generates an intermediate capable of oxidizing the
solvent benzene to phenol. The reactivity of the inter-
mediate was probed with various substrates (2-meth-
ylbutane, chloride ion, and para-substituted aryl sulfides)
and is found to have the electrophilic oxidation chemistry
that is expected of atomic oxygen [O(3P)] in solution.
Resu lts a n d Discu ssion
In our initial efforts, we sought a basis to compare
established gas-phase O(3P) reactions with solution phase
methods that potentially yield O(3P). A similarity emerges
when comparing molecular oxidations from (1) gas-phase
O(3P),5 and (2) solution phase 3- and 5-photodeoxygen-
ations3,4,6 (Figure 1). A linear relationship is found when
the logarithm of the relative rates, taken from refs 4-6,
are plotted versus ionization potentials of their respective
acceptor molecules.16 The data falls into two groups; one
for the gas phase (Figure 1A) and the other for the
solution phase (Figure 1B,C). Both rates correlate with
the ionization potential and have about the same slope.
Hyp och lor ite Ion F or m a tion . These studies are the
first to indicate that hypochlorite ion can form from a
sulfoxide photodeoxygenation reaction in the presence of
(17) Cvetanovic, R. J . Can. J . Chem. 1960, 38, 1678-1687
(18) Slagle, I. R.; Baiocchi, F.; Gutman, D. J . Phys. Chem. 1978,
82, 2, 1333-1336.
(19) Nip, W. S.; Singleton, D. A.; Cvetanovic, R. J .; J . Am. Chem.
Soc. 1981, 103, 3526-3530.
(12) Kla¨ning, U. K.; Sehested, K.; Wolff, T. J . Chem. Soc., Faraday
Trans. 1 1984, 80, 2969-79.
(20) Cvetanovic, R. J .; Singleton, D. A.; Irwin, R. S. J . Am. Chem.
Soc. 1981, 103, 3530-3539.
(13) Warneck, P.; Wurzinger, C. J . Phys. Chem. 1988, 92, 6278-
6283.
(21) It is not known whether SdO cleavage can originate from the
singlet-excited state of 1. While the UB3LYP/6-31G(d)//B3LYP/3-21G
calculated energy split of the singlet-ground and triplet state (63.2 kcal/
mol) is below the BDE, this does not rule out possible SdO cleavage
from a vibrationally excited triplet state.
(14) Kajimoto, O.; Yamasaki, H.; Fueno, T. Chem. Lett. 1977, 349-
352
(15) Hirokami, S.; Cvetanovic, R. J . J . Am. Chem. Soc. 1974, 96,
3738-3746.
(16) Levin, R. D. Ionization potential and appearance potential
measurements; National Bureau of Standards: Washington, D. C.,
1982; Vol. IV.
(22) Cvetanovic, R. J . Adv. Photochem. 1963, 1, 115-182.
(23) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J .
Phys. Chem. Ref. Data 1988, 17, 513.

4578 J . Org. Chem., Vol. 66, No. 13, 2001
Lucien and Greer
chloride ion. Competitive kinetic studies were conducted
with the 365 nm photochemical decomposition of 1 (2
mM) with tetrabutylammonium chloride (38 mM) in
acetonitrile:benzene(3.8 M:8.4 M), which produced 2,
hypochlorite ion (OCl-), and phenol. After photolysis, the
mixture was stirred overnight, and hypochlorite ion was
detected as the chlorinated adduct of added 1,3,5-tri-
methoxybenzene (∼150 mM) using chemistry previously
established by Foote.24 The product ratio of OCl- (as
2-chloro-1,3,5-trimethoxybenzene) was found to be 191
( 30 times that of phenol, which is similar to the kinetic
ratio (200:1, chloride ion:benzene) derived from the
absolute rate constants for removal of O(3P) in acetoni-
trile.6,25 The above similarity represents a comparison
between product- and kinetic-based selectivities. Phenol
yields were determined in our product study but not in
Scaiano’s kinetic study.6 The reactivity of the intermedi-
ate in the photodeoxygenation of 1 is unlike that of OH
radical because OH radicals are known to react with
chloride ion and benzene with a similar preference
∼1:1.1.6,23 The preference for a reaction with an electron
rich substrate is indicated here for the first time in the
photodeoxygenation of an aromatic sulfoxide.
F igu r e 2. Correlation of log krel with the adiabatic ionization
potentials for the oxidation of sulfur compounds. Oxidation of
para-substituted aryl sulfides (X ) OMe, Me, H, Cl) in the
photodeoxygenation of 1 (A, open circles, slope ) -2.15, r2
)
0.998) and 3 (B, solid diamonds, slope ) -2.05, r2 ) 0.997) in
MeCN solution. Oxidation of MeSSMe, Me2S, EtSH, and MeSH
with gas-phase O(3P): C (open diamonds, slope ) -2.05, r2
)
0.996, ref 18) and D (solid triangles, slope ) -2.31, r2 ) 0.988,
ref 19,20).
Ha m m ett Stu d ies. We further demonstrate the pref-
erence of the intermediate for an electron rich substrate
via a Hammett trapping study of the sulfoxide photode-
oxygenation. Addition of electron-donating groups on the
substrate leads to enhanced relative reaction rates
toward the intermediate formed in the photodeoxygen-
ation of 1 and 3. Compound 1 was selectively irradiated
(hν ) 365 or 385 nm; 2-8 mM) in the presence of a series
of diaryl sulfides [(p-X-C6H4)2S, where X ) OMe, Me, H,
and Cl; 4-9 mM] in argon-saturated acetonitrile solu-
tions. Control reactions demonstrate that the aryl sul-
fides do not react with 1 in the absence of light. The
products formed are 2 and the corresponding aryl sul-
foxides, (p-X-C6H4)2SO. Maintaining <10% conversion
of the reaction components, the ratios kx/kH determined
from the slopes of plots of [(p-X-C6H4)2SO] versus [(C6H4)2-
SO] yielded evidence for an electrophilic oxidant (Ham-
mett F ) -1.57 ( 0.13, r2)0.975). Moreover, in an
analogous experiment, we demonstrated that the pho-
todeoxygenation of 3 with 350 nm light in the presence
of the same series of diaryl sulfides yielded a Hammett
plot with a F value ) -1.48 ( 0.11 (r2 ) 0.956). These
data suggest that the intermediate derived from the
photodeoxygenation of 1 and 3 are similar or identical.
The magnitude of the reaction constants (F ) ∼-1.5)
implies a build-up of positive charge on the (p-X-C6H4)2S
sulfur in the transition state.
Me2S, EtSH, and MeSH (Figure 2C,D).18-20 The rate
constants are well correlated with the ionization poten-
tials for the solution and gas-phase data. The substantial
negative slopes in the IP plots provide evidence of partial
electron transfer or charge transfer between the electron
rich aryl sulfide additives and the electrophilic oxygenat-
ing intermediate. Our data on 1 is reminiscent of the
oxygenating character of high-valent iron porphyrins.27
The observation of similar slopes between Figure 2A,B
and 2C,D provides limited evidence for O(3P) as a
common intermediate because it is conceivable that the
solution-phase systems bear a rate dependence on the
ionization potential in a bimolecular reaction between the
excited state of 1 or 3 and the aryl sulfide acceptor. This
could especially be true because, by analogy, sulfoxides
are capable of forming dimeric structures, although
elevated concentrations are generally required.28-30 The
relative changes in the ionization potential may be
attributable to solvation given the abscissa of the plot is
the ionization potential of the molecule. The coincidence
of the gas and solution phase slopes32 (Figure 2A,B and
2C,D) may be fortuitous; however, we believe that the
data point to an identical reactive intermediate, O(3P).
Con clu sion
The properties of the photodeoxygenation of 1 have
been examined in order to investigate the nature of the
possible intermediate. An intermediate is generated
which is capable of oxidizing benzene, chloride ion,
2-methylbutane, and para-substituted aryl sulfides. With
the above data, it is not yet possible to prove absolutely
the existence of O(3P) in solution. However, the relative
Another measure of the character of the intermediate
comes from comparing the relationship between the
relative rate constants, and the (p-X-C6H4)2S ionization
potentials. A plot of the logarithm of the relative rates
of 1 and 3 versus the adiabatic ionization potentials of
the aryl sulfides16,26 revealed a linear dependence (Figure
2A,B). These observations are very similar to those
reported in the gas-phase O(3P) oxidations of MeSSMe,
(27) Traylor, T. G.; Xu, F. J . Am. Chem. Soc. 1988, 110, 1953-1958.
(28) Watson, R. F.; Eastham, J . F. J . Am. Chem. Soc. 1965, 87, 664-
665.
(29) Mislow, K.; Green, M. M.; Laur, P.; Chisolm, D. R. J . Am. Chem.
Soc. 1965, 87, 665-666.
(30) Mislow, K.; Green, M. M.; Laur, P.; Melillo, J . T.; Simmons, T.;
Ternay, A. L. J . Am. Chem. Soc. 1965, 87, 1958-1976.
(31) Takata, T.; Ando, W. Tetrahedron Lett. 1983, 24, 3631-3634.
(32) Solvation readily diminishes the energy required to form a
sulfide radical cation according to polarized continuum model UB3LYP
calculations. (Greer, A. Unpublished data)
(24) Foote, C. S.; Goyne, T. E.; Lehrer, R. I. Nature 1983, 301, 715-
716.
(25) Interestingly, the extent of electron transfer between chloride
ion and O(3P) may be significant since Scaiano’s pyridine N-oxide
system determined the rate of reaction to be above diffusion control.6
(26) The ionization potential of bis(p-methoxyphenyl)sulfide was
estimated as 7.79 eV from the half-wave potentials of a series of para-
substituted thiobenzoic acid S-n-butyl esters: Kunz, D. S.; Mayer, R.
Z. Chem. 1967, 194-194.

Photodeoxygenation of 1,2-Benzodiphenylene Sulfoxide
J . Org. Chem., Vol. 66, No. 13, 2001 4579
reactivity and similarity to gas-phase O(3P) and solution
systems 3 and 5 suggests O(3P) as a potential short-lived
intermediate formed in the UV deoxygenation of 1.
M:9.5 M) mixtures. Argon was flushed through the reaction
mixtures for 15-30 min prior to irradiation. Phenol was
detected in the benzene and acetonitrile:benzene solutions by
gas chromatography when 1 was irradiated with 300, 320, 340,
365, or 385 nm light. Alkane hydroxylations were conducted
in Ar-saturated acetonitrile solutions containing 4 mM 1, 5 ×
10-4 M biphenyl as an internal standard, and 0.687 M
2-methylbutane using 365 or 385 nm light. Chloride ion
trapping was conducted by irradiating Ar-saturated acetoni-
trile:benzene (3.8 M:8.4 M) solutions with 365 nm light
containing 4 mM 1, 5 × 10-4 M biphenyl as an internal
standard, and 38 mM tetrabutylammonium chloride. After
photolysis the reaction mixture was stirred overnight and
hypochlorite ion detected as the chlorinated adduct of added
1,3,5-trimethoxybenzene (∼150 mM).24 The aryl sulfide Ham-
mett studies were conducted in 1-mL GC vials in Ar-saturated
acetonitrile solutions containing 4 mM 1, 5 × 10-4 M biphenyl
as an internal standard, and 4 mM Ph2S and 4 mM (p-X-
C6H4)2S and irradiating with 365 or 385 nm light. Hammett
studies were conducted in an identical fashion with 3 except
that 350 nm light was used. In all experiments the irradiation
times were selected (∼2-3 h) in order to ensure less than 10%
of any reaction component. The concentrations of products
were determined in the reaction mixtures immediately after
photolysis by gas chromatography.
Exp er im en ta l Section
Gen er a l Asp ects. 1,2-Benzodiphenylene sulfide (Aldrich
99%), dibenzothiophene sulfide (Aldrich 99%), phenol (Aldrich
99%), tetrabutylammonium chloride (Fluka 99%), 2-methylbu-
tane (anhydrous, Aldrich 99%), 2-methyl-2-butanol (Aldrich
99%), 3-methyl-2-butanol (Aldrich 98%), 2-methyl-1-butanol
(Aldrich 99%), 3-methyl-1-butanol (Aldrich 99%), sodium hy-
pochlorite (Aldrich 4% aqueous solution), 1,3,5-trimethoxy-
benzene (Aldrich 99%), biphenyl (Aldrich 99%), diphenyl
sulfoxide (Aldrich 96%), and acetonitrile (anhydrous, Aldrich
99.8%) were used as received. The purity of the reagents was
checked by GC or GC/MS prior to use. Diphenyl sulfide
(Aldrich 99%) was distilled under reduced pressure (bp 124
°C/6 mmHg). Benzene (anhydrous, Aldrich 99.8%) was distilled
over P2O5. Compounds 1 and 3 were synthesized and purified
using a literature method.31 Relative concentrations of 1-4,
phenol, 2-methyl-2-butanol, 3-methyl-2-butanol, 2-methyl-1-
butanol, 3-methyl-1-butanol, 2-chloro-1,3,5-trimethoxyben-
zene, and the para-substituted aryl sulfoxides (p-X-C6H4)2SO
were determined by reference to calibration curves constructed
from authentic samples. Gas chromatographic data were
collected on one of two gas chromatographs, a Hewlett-Packard
GC/MS instrument consisting of a 5890 series GC and a 5988A
series mass selective detector, or on a Shimadzu-17A auto-
sampler capillary gas chromatograph.
Ack n ow led gm en t. This work was supported by the
Petroleum Research Fund, administered by the Ameri-
can Chemical Society, a start-up grant from the City
University of New York, and by a research award from
PSC-CUNY. Computer time was provided by the Na-
tional Partnership for Advanced Computational Infra-
structure and the San Diego Super Computer Center.
E.L. is a Pfizer PREPARE award recipient. Special
thanks to the reviewers for their comments and to
Professors H. Gafney and P. Haberfield for several
fascinating discussions.
Su lfoxid e p h otod eoxygen a tion r ea ction s were carried
out using 1.5 mL airtight GC vials or 5-mm NMR tubes at
room temperature and irradiation with a 75-W Xenon Model
L-201 Arc lamp (Photon Technology International) focused on
a tunable monochromator to obtain monochromatic light over
the range 280-400 nm. A typical experiment was conducted
in a 1-mL solution of spectral grade benzene, which contained
3 mM 1 and 5 × 10-4 M biphenyl as an internal standard.
Often cosolvents were used, such as acetonitrile:benzene (2.9
J O010009Z