Role of sulfide radical cations in electron transfer promoted molecular oxygenations at sulfur

Article abstract of DOI:10.1021/ja710301s

The methylene blue, N-methylquinolinium tetrafluoroborate, and pyrylium-cation-sensitized photooxygenations of 5H, 7H-dibenzo[b,g] [1,5]dithiocin, 1, and 1,5-dithiacyclooctane, 2, have been investigated. The methylene blue sensitized reactions exhibit all of the characteristics of a singlet oxygen reaction including isotope effects for the formation of a hydroperoxysulfonium ylide and the ability of 1 and 2 to quench the time-resolved emission of singlet oxygen at 1270 nm. The product compositions in the N-methylquinolinium tetrafluoroborate and pyrylium-cation-sensitized reactions are dramatically different and are both different from that anticipated for the participation of singlet oxygen. This argues for different reaction mechanisms for all three sensitizers. However, both the quinolinium and pyrylium-cation-sensitized reactions display all of the characteristics of electron-transfer-initiated photooxygenations. Both sensitizers were quenched at nearly diffusion-limited rates by 1 and 2. Laser flash photolysis of mixtures of either sensitizer and 1 or 2 resulted in direct observation of the reduced sensitizer and the sulfide radical cation. In addition, electron-transfer reactions involving both sensitizers were shown to be exergonic. These results are consistent with the previously proposed outer sphere electron-transfer mechanism for N-methylquinolinium tetrafluoroborate and were used to argue for a new inner sphere mechanism for the pyrylium cation reactions.

Full text of DOI:10.1021/ja710301s

Published on Web 02/29/2008
Role of Sulfide Radical Cations in Electron Transfer Promoted
Molecular Oxygenations at Sulfur
Edward L. Clennan* and Chen Liao
Department of Chemistry, UniVersity of Wyoming, 1000 East UniVersity AVenue,
Laramie, Wyoming 82071
Received November 20, 2007; E-mail: clennane@uwyo.edu
Abstract: The methylene blue, N-methylquinolinium tetrafluoroborate, and pyrylium-cation-sensitized
photooxygenations of 5H, 7H-dibenzo[b,g] [1,5]dithiocin, 1, and 1,5-dithiacyclooctane, 2, have been
investigated. The methylene blue sensitized reactions exhibit all of the characteristics of a singlet oxygen
reaction including isotope effects for the formation of a hydroperoxysulfonium ylide and the ability of 1 and
2 to quench the time-resolved emission of singlet oxygen at 1270 nm. The product compositions in the
N-methylquinolinium tetrafluoroborate and pyrylium-cation-sensitized reactions are dramatically different
and are both different from that anticipated for the participation of singlet oxygen. This argues for different
reaction mechanisms for all three sensitizers. However, both the quinolinium and pyrylium-cation-sensitized
reactions display all of the characteristics of electron-transfer-initiated photooxygenations. Both sensitizers
were quenched at nearly diffusion-limited rates by 1 and 2. Laser flash photolysis of mixtures of either
sensitizer and 1 or 2 resulted in direct observation of the reduced sensitizer and the sulfide radical cation.
In addition, electron-transfer reactions involving both sensitizers were shown to be exergonic. These results
are consistent with the previously proposed outer sphere electron-transfer mechanism for N-methylquino-
linium tetrafluoroborate and were used to argue for a new inner sphere mechanism for the pyrylium cation
reactions.
Dioxygen is one of the most economical and environmentally
friendly reagents available for functionalization of organic
molecules to produce value-added products. These value-added
oxidized organic substrates are important feedstock for the
chemical and pharmaceutical industries. Unfortunately, many
organic substrates of interest are notoriously resistant to
molecular oxygenations with ground state triplet oxygen (3∑g)
under environmentally benign and easily controlled conditions.
This problem has been circumvented in many cases by either
activating oxygen or activating the substrate. Activation of
oxygen has been achieved by either energy transfer to form
singlet oxygen (1∆g) or by electron transfer to form superoxide
(O•2). Activation of organic substrates has been achieved by
either electron transfer to form radical cations or anions or by
hydrogen transfer to form radicals.
co-workers6-9 suggested that the first intermediate is a persul-
foxide (PS, Scheme 1) and the second intermediate a hydrop-
eroxysulfonium ylide (HPY, Scheme 1). The persulfoxide
suggestion was supported by trapping studies with substituted
sulfoxides that established it as a nucleophilic oxygen transfer
agent.3 The hydroperoxysulfonium ylide suggestion replaced a
previous assumption that the second intermediate was a thia-
dioxirane (TD, Scheme 1). It was demonstrated computationally
that the activation barrier separating the persulfoxide and
hydroperoxysulfonium ylide was far more consistent with
experimental results than the barrier between the persulfoxide
and thiadioxirane.9 The hydroperoxysulfonium ylide suggestion
was supported experimentally by trapping experiments that
established its electrophilic oxygen transfer character, by isotope
effect studies employing R-deuterium substituted sulfides,10,11
and in an examination of the reactions of â-chlorosulfides.12
In 1977, Foote and co-workers13 suggested that sulfide radical
cations14 were involved in the 9,10-dicyanoanthracene (DCA)
photooxygenations of diethyl- and diphenylsulfide (Scheme 2).
It has occasionally been assumed that the intermediate, R2SO2,
The reactions of organic sulfides with singlet oxygen have
been extensively studied since first reported by Schenck and
Krauch in the early 1960s.1 Experimental evidence that two
intermediates were involved on the potential energy surfaces
of these reactions, on the way to the sulfoxide major products,
was subsequently provided by Foote and co-workers in the
1970s and early 1980s.2-5 Computational studies by Jensen and
(6) Jensen, F. J. Org. Chem. 1992, 57, 6478-6487.
(7) Jensen, F. Theoretical Aspects of the Reactions of Organic Sulfur and
Phosphorus Compounds with Singlet Oxygen. In AdVances in Oxygenated
Processes; Baumstark, A. L., Ed.; JAI Press: Greenwich, CT, 1995; Vol.
4, pp 1-48.
(1) Schenck, G. O.; Krauch, C. H. Angew. Chem. 1962, 74, 510.
(2) Foote, C. S.; Peters, J. W. J. Am. Chem. Soc. 1971, 93, 3795-3796.
(3) Gu, C.; Foote, C. S.; Kacher, M. L. J. Am. Chem. Soc. 1981, 103, 5949-
5951.
(8) Jensen, F.; Foote, C. S. J. Am. Chem. Soc. 1988, 110, 2368-2375.
(9) Jensen, F.; Greer, A.; Clennan, E. L. J. Am. Chem. Soc. 1998, 120, 4439-
4449.
(4) Gu, C.-L.; Foote, C. S. J. Am. Chem. Soc. 1982, 104, 6060-6063.
(5) Liang, J.-J.; Gu, C.-L.; Kacher, M. L.; Foote, C. S. J. Am. Chem. Soc.
1983, 105, 4717-4721.
(10) Clennan, E. L.; Liao, C. Tetrahedron 2006, 10724-10728.
(11) Toutchkine, A.; Clennan, E. L. J. Org. Chem. 1999, 64, 5620-5625.
(12) Toutchkine, A.; Clennan, E. L. Tetrahedron Lett. 1999, 40, 6519-6522.
9
10.1021/ja710301s CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 4057-4068
4057

A R T I C L E S
Scheme 1
Clennan and Liao
thiadioxiranes, TD.20 Consequently, they argued that superoxide
is a byproduct not involved in product formation, and they
suggested an alternative unitary chain-reaction mechanism
for all of the sensitizers, as depicted in Scheme 4. This
mechanism, as well as the mechanisms of Che and
Memarian (Scheme 3), invoke direct reaction between a sulfide
radical cation and oxygen, a reaction that has not been observed
on the pulse radiolysis time scale.21,22 In addition, a computa-
tional study of Rauk23 does not characterize dimethylpersul-
foxide radical cation, PS•, as a covalently bound species
but rather as a loose ion-induced dipole complex bound by
3.2 kcal/mol with a sulfur-oxygen distance only slightly
shorter than the sum of the van der Waals radii. Albini and
co-workers20 got around this latter problem by pointing out
in this reaction is the persulfoxide, PS, and that its subsequent
transformation to product occurs in an identical fashion to that
observed in the singlet oxygen reaction (Scheme 1).15,16
However, Baciocchi and co-workers17 discovered in 2003 that
diphenylsulfoxide does not trap any intermediate formed in the
N-methylquinolinium tetrafluoroborate (NMQ+)-sensitized pho-
tooxygenations of sulfides. This led to the suggestion that
NMQ+ sensitization, via the mechanism previously established
for DCA (Scheme 2), does not proceed via a persulfoxide, PS,
but that superoxide and the sulfide radical cation collapse
directly to a thiadioxirane, TD.
that addition of the ion-dipole complex to R2S (step kR S in
2
Scheme 4) is highly exergonic and drives the entire reaction
sequence.
Given the central role that oxidations of methionine residues
in proteins play in a variety of degenerate diseases including
Alzheimer’s disease, we have undertaken a careful study of the
photooxygenations of organic sulfides under a variety of
conditions. We report here our results with 5H, 7H-dibenzo-
[b,g] [1,5]dithiocin, 1, and 1,5-dithiacyclooctane, 2, with dif-
ferent sensitizers. These sensitizers react to provide unique
product compositions and leave very different footprints with
an arsenal of mechanistic tools. This pattern recognition
approach provides compelling evidence for different sensitizer-
dependent reactions and rules out the concept of a single unitary
electron-transfer photooxygenation mechanism.
Scheme 2
One-electron reduced DCA and NMQ+ are both thermody-
namically capable of reducing oxygen to superoxide as required
by the mechanism in Scheme 2. On the other hand, it has been
established that 2,4,6-triarylpyrylium tetrafluoroborates, TPP+s,
(Scheme 3), which form radicals incapable of producing
superoxide, also catalyze sulfide photooxygenation reactions.18,19
To rationalize these non-superoxide photooxygenations, Che18
and Memarian19 both suggested mechanisms initiated by direct
reaction of the sulfide radical cation with oxygen to give a
persulfoxide radical cation, PS• (Scheme 3). In the Che
mechanism, formation of the persulfoxide, PS, is circumvented
by direct formation of the thiadioxirane, TD, from pyrylium
radical reduction of the persulfoxide radical cation, PS•,
intermediate.
Results
The synthesis of 5H, 7H-dibenzo[b,g] [1,5]dithiocin, 1, is
shown in Scheme 5.24 The proton NMR spectrum of 1 exhibits
a large AB quartet and a smaller singlet (relative intensity
singlet/AB quartet ) 0.11) consistent with the previous
suggestion that it exists in two non-interconverting con-
formations on the NMR time scale.24 DFT calculations
(B3LYP/6-31G(d)) located both Cs symmetric boat-chair (BC)
and C2 symmetric twist-boat (TB) conformational minima
(Figure 1). The BC conformation is 2.23 kcal/mol more
stable than the TB conformation. The small singlet broadened
as the temperature of the NMR sample was lowered and
then sharpened to form an AB quartet consistent with slow
interconversion between the TB mirror images at low
temperature. A complete line shape fit gave ∆G* (228.16 K)
) 10.4 kcal/mol, ∆H* ) 7.1 kcal/mol, and ∆S* ) -14.4 cal/
mol‚K, in reasonable agreement with ∆G* (228.16 K) ) 10.9
kcal/mol that was determined for this dynamic process by
In 2006, Albini and co-workers pointed out that unproductive
electron transfers between superoxide and sulfur radical cations
are substantially more exothermic than reactions that form the
(13) Eriksen, J.; Foote, C. S.; Parker, T. L. J. Am. Chem. Soc. 1977, 99, 6455-
6456.
(14) Glass, R. S. Sulfur Radical Cations. In Organosulfur Chemistry II; Page,
P. C. B., Ed.; Springer-Verlag: Berlin, 1999; Vol. 205, pp 1-87.
(15) Baciocchi, E.; Crescenzi, C.; Lanzalunga, O. Tetrahedron 1997, 53, 4469-
4478.
(16) Akasaka, T.; Ando, W. Tetrahedron Lett. 1985, 26, 5049-5052.
(17) Baciocchi, E.; Del Giacco, T.; Elisei, F.; Gerini, M. F.; Guerra, M.; Lapi,
A.; Liberali, P. J. Am. Chem. Soc. 2003, 125, 16444-16454.
(18) Che, Y.; Ma, W.; Ji, H.; Zhao, J.; Zang, L. J. Phys. Chem. B 2006, 110,
2942-2948.
(19) Memarian, H. R.; Mohammadpoor-Baltork, I. Bahrami, K. Bull. Korean
Chem. Soc. 2006, 27, 106-110.
(20) Bonesi, S. M.; Manet, I.; Freccero, M.; Fagnoni, M.; Albini, A. Chem.s
Eur. J. 2006, 12, 4844-4857.
(21) Scha¨fer, K.; Bonifacic, M.; Bahnemann, D.; Asmus, K.-D. J. Phys. Chem.
1978, 82, 2777-2780.
(22) Scho¨neich, C.; Aced, A.; Asmus, K.-D. J. Am. Chem. Soc. 1993, 115,
11376-11383.
(23) Huang, M. L.; Rauk, A. J. Phys. Chem. A 2004, 108, 6222-6230.
(24) Gellatly, R. P.; Ollis, W. D.; Sutherland, I. O. J. Chem. Soc., Perkin Trans.
II 1976, 913-925.
9
4058 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008

Sulfide Radical Cations in Electron Transfer
A R T I C L E S
Scheme 3
Scheme 4
fitting a broadened spectrum at a single temperature.24 The
singlet and AB quartet observed at room temperature also
began to broaden as the temperature was raised consistently
with the onset of interconversion between the BC and TB
conformations on the NMR time scale. All of these dynamic
events were reversible, consistent with their conformational
origin.
Figure 2. ORTEP plot and three ball-and-stick projections of the X-ray
structure of bissulfoxide, 5.
Scheme 5
Methylene Blue Photosensitized Reaction of 1. An oxygen-
saturated CD3CN solution 1.1 × 10-2 M in 1 and 1 × 10-4
M
in methylene blue containing 8.75 × 10-3 M benzyl benzoate
as an internal standard was irradiated with a 600 W halogen-
tungsten lamp through 1 cm of a saturated NaNO2 filter solution.
The major product of the reaction is 5H, 7H-dibenzo[b,g] [1,5]-
dithiocin-6-oxide, 3, with a small amount of 5H, 7H-dibenzo-
[b,g] [1,5]dithiocin-6,6-dioxide, 4, and a trace of trans-5H, 7H-
dibenzo[b,g] [1,5]dithiocin-6,12-dioxide, 5, appearing after
extended irradiation (eq 1). The trans-configuration of dioxide,
5, was established by X-ray diffraction of an independently
synthesized sample that was shown to be identical to the material
formed in the photooxygenation (Figure 2). The dioxide adopts
a BC conformation in the crystal with two equatorial-like
oxygens. This isomer is the most stable of a set of four BC and
three TB 6,12-dioxides that were located by density functional
calculations (DFT) using the B3LYP/6-31G(d) model (Sup-
porting Information).
The proton NMR spectrum of sulfoxide 3 was more complex
than previously reported.25 Rather than two AB quartets, a set
of three AB quartets between 3.8 and 5.6 ppm were observed
in a 1.0:0.26:0.14 ratio at 400 MHz. This suggests that at
Figure 1. Boat-chair (BC) and C2 symmetric twist-boat (TB) confor-
mational minima of 5H, 7H-dibenzo[b,g] [1,5]dithiocin, 1.
(25) Ohkata, K.; Okada, K.; Akiba, K. Heteroat. Chem. 1995, 6, 145-153.
9
J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008 4059

A R T I C L E S
Clennan and Liao
Scheme 6
experiments with Ph2SO that generated Ph2SO2 consistent
with oxygen transfer from a persulfoxide intermediate. In the
latter trapping experiments 5H, 7H-dibenzo[b,g][1,5]dithiocin-
12-oxide, 6, was not formed, suggesting that the 12-persulfoxide
is not a transient intermediate that decomposes via physical
quenching, kq (Scheme 6). This is consistent with the
known lack of reactivity of diphenylsulfide with singlet oxygen.
The formation of the well-established persulfoxide singlet
oxygen adduct during the reaction of 1,5-dithiacyclooctane, 2,
was independently demonstrated by the photooxygenation of
2,2,8,8-tetradeuterio-1,5-dithiacyclooctane, 2d4.10 In this case
the product isotope effect ([2SOh4]/[2SOd4]) of 1.23 ( 0.08
reflects preferential formation of OOH in comparison to OOD
(Scheme 7) as a result of the increased difficulty of abstraction
of deuterium and concomitant more favorable partitioning along
the physical quenching pathway, kq, of persulfoxide PSHγ in
comparison to persulfoxide PSHR.
The total rate constants, kT, for reactions of 1 and 2 with
singlet oxygen were extracted from the observed pseudo first-
order rate constants for quenching of singlet oxygen phospho-
rescence at 1270 nm (Figure 4) by plotting kobs versus the
concentrations of the sulfide (Inset Figure 4; kobs ) kT[sulfide]
+ kd). The derived kT, which represents the rate constant of all
chemical, kr, and physical, kq, sulfide-induced pathways for the
deactivation of singlet oxygen, was more than 30 times smaller
for 1 (kT ) [6.7 ( 0.4] × 105M-1s-1) than for 2 (kT ) [2.1 (
0.2] × 107M-1s-1). The reduced rate of the reaction of 1 in
comparison to 2 is anticipated based upon the shielding of the
bottom face of the molecule by the phenyl rings in the most
stable butterfly-like BC conformation (Figure 1) and the known
lack of reactivity of Ph2S with singlet oxygen.
Figure 3. Two views of the twist-boat, TB, and boat-chair, BC,
conformations of 3.
least three conformational isomers of 3 are significantly
populated. This suggestion was verified by B3LYP/6-31G(d)
calculations that located two BCs, one with an axial-like
sulfoxide oxygen, BCAx, and a second with an equatorial-like
oxygen, BCEq, and a TB conformation. (Figure 3) A comparison
of the zero point corrected energies reveal a stability sequence
BCEq > TB > BCAx. Conversion of these energies, after
adjusting the energy of the chiral TB conformation for the
entropy of mixing, gives a population ratio TB/BCEq/BCAx of
1.0:0.55:0.06. The remarkable similarity of this calculated
population ratio to the experimental ratio, 1.0:0.26:0.14, provides
compelling confirmation of the assignment of the three AB
quartets to three conformational isomers. In addition, all three
AB quartets collapsed to a broad singlet when a DMSO-d6
sample of 3 was heated to 120 °C in the probe of a 400 MHz
NMR. These spectral changes were completely reversible,
consistent with a dynamic interconversion of three conforma-
tional isomers.
Singlet oxygen was identified as the key oxidant in this
reaction (eq 1) by DABCO quenching experiments that
completely suppressed product formation and by trapping
Figure 4. Time-resolved emission of singlet oxygen at 1270 nm as a function of the concentration of 1. Inset, plot of the pseudo-first-order rate constants
derived by fitting the decay curves versus the concentration of 1.
9
4060 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008

Sulfide Radical Cations in Electron Transfer
A R T I C L E S
Scheme 7. Reaction of 2d4 with Singlet Oxygen
N-Methylquinolinium Tetrafluoroborate (NMQ+BF4
)
the lifetime of singlet excited NMQ+ BF4- 32 is also similar to
,
-
Photosensitized Reaction of 1. Steady-state photolysis of a
other electron-transfer rate constants measured for this sensi-
tizer.17 Finally, a laser flash photolysis study generated a
transient absorption spectrum with two bands at 550 and 360
nm (Figure S17 in the Supporting Information). The 550 nm
peak can be confidently assigned to NMQ• based upon
comparison to the published spectrum of this radical and the
sensitivity of this peak to quenching by oxygen. However, the
transient at 360 nm, in contrast to NMQ•, was not quenched by
oxygen on the laser flash photolysis time scale, suggesting
assignment of this higher energy band to 1+•. Pyrylium salts
absorb strongly at 360 nm, and their bleaching in laser flash
photolysis studies obscure the absorbance of 1+• and preclude
its direct observation in these experiments.
To provide independent confirmation of the assignment of
the 360 nm peak to 1+•, we have used the B3LYP/6-31G(d)
model to locate stable conformations of 1+• and have calculated
their absorption spectra using the time-dependent DFT method.
Frequency calculations were used to verify the location of energy
minima and the absence of spin contamination was checked by
examination of S2 , which showed values acceptably close to
0.75 for these doublet cation radical species.33 Three confor-
mational minima were located and are depicted in Figure 5.
The most stable conformation of 1+• is a twist-boat-boat
conformation with C1 symmetry that is characterized by a short
sulfur-sulfur distance of 2.76 Å, well within the sum of two
sulfur van der Waals radii of 3.6 Å.34 The C2 TB conformation
is 2.91 kcal/mol above the global minimum and is characterized
by a significant distance of 4.2 Å between the two sulfurs. The
CD3CN solution of 1 in the presence of 5 mol % NMQ+BF4
-
was conducted under constant agitation with oxygen by irradia-
tion with a 450 W medium-pressure mercury lamp through a
pyrex filter. NMR analysis revealed the formation of both 3
and 5, which previously were observed in the methylene blue
reaction, and a new product that was identified as 5H, 7H-
dibenzo[b,g][1,5]dithiocin-12-oxide, 6, by comparison to an
authentic sample (eq 2). Product formation was quenched by
para-benzoquinone, the well-established superoxide trap,26 but
not by DABCO. The absence of quenching by DABCO is
especially significant because Baciocchi and co-workers17 have
-
previously demonstrated that NMQ+BF4 generates singlet
oxygen. The absence of a singlet-oxygen component to this
reaction is perhaps not surprising given the small rate constant
(vide supra) for the singlet oxygen reaction.
These results suggest the possibility that the NMQ+ sensitized
reaction of 1 occurs by an electron-transfer mechanism. The
oxidation potential of 1 (E°(D+/D), eq 3) has not been reported,
and a cyclic voltammetry study of 1 in CH3CN with a platinum
working electrode exhibited chemically irreversible behavior
precluding determination of a thermodynamically significant
oxidation potential. Nevertheless, using oxidation potentials of
thianthrene (1.23 V vs SCE)14,27 and 1,5-dithiacyclooctane (0.72
V vs SCE)28,29 as limiting values in conjunction with the Rehm
Weller equation30 (eq 3), an estimate of a very exergonic free
energy of electron transfer for 1 to the excited-state of
(26) Manring, L. E.; Kramer, M. K.; Foote, C. S. Tetrahedron Lett. 1984, 25,
2523-2526.
(27) Zhao, B. J.; Evans, D. H.; Macias-Ruvalcaba, N. A.; Shine, H. J. J. Org.
Chem. 2006, 71, 3737-3742.
(28) Wilson, G. S.; Swanson, D. D.; Klug, J. T.; Glass, R. S.; Ryan, M. D.;
Musker, W. K. J. Am. Chem. Soc. 1979, 101, 1040-1042.
(29) Goto, Y.; Matsui, T.; Ozaki, S.; Watanabe, Y.; Fukuzumi, S. J. Am. Chem.
Soc. 1999, 121, 9497-9502.
NMQ+BF4 of ∆Get ) -33 to -45 kcal/mol was obtained.31
-
∆Get(kcal/mol-1) ) 23.06{[E°(D+/D) - E°(A/A-)] -
Wp - ∆E00} (3)
(30) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259-271.
(31) The other values in eq 3 necessary to do this calculation are the reduction
potential of the NMQ+ (E°(A/A-) ) -0.85 V vs SCE), the singlet excited
state energy of NMQ+ (∆E00 ) 81.5 kcal/mol), and the energy necessary
to separate the ions, Wp, which is taken as zero in this case as a result of
the lack of Coulombic interaction between the NMQ• radical and the sulfide
radical cation.49
The ability of 1 to quench the fluorescence emission of
NMQ+ BF4 at 390 nm in acetonitrile is also consistent with
-
(32) Fukuzumi, S.; Fujita, M.; Noura, S.; Ohkubo, K.; Suenobu, T.; Araki, Y.;
Ito, O. J. Phys. Chem. A 2001, 105, 1857-1868.
an electron-transfer mechanism. The magnitude of the quenching
rate constant of 2.23 × 1010M-1s-1, which was derived from a
Stern-Volmer analysis using a published value of 20 ns for
(33) Cramer, C. J. Essentials of Computational Chemistry. Theories and Models;
John Wiley & Sons Ltd.: Chichester, England, 2002.
(34) Bondi, A. J. Phys. Chem. 1964, 68, 441-451.
9
J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008 4061

A R T I C L E S
Clennan and Liao
Figure 5. Two views of the B3LYP/6-31G(d) conformations of the radical cation of 1.
CS BC conformation is 9.16 kcal/mol above the global minimum
and like the TB has a sulfur-sulfur distance (3.76 Å) greater
than the sum of their van der Waals radii.
The time-dependent density functional theory (TD-DFT)
calculations of the electronic spectra35 of each of the radical
cations were conducted both in the absence and presence of
acetonitrile, using the polarization continuum model. The results
are depicted in Figure 6 for the most stable twist-boat-boat
conformation (Supporting Information for results on other
conformations). The peak with the largest oscillator strength
(f) in both the presence (378 nm, f ) 0.118) and absence (375
nm, f ) 0.0774) of acetonitrile is very close to the peak observed
in the laser flash photolysis studies at 360 nm (vide supra). The
long wavelength bands with low oscillator strength were not
observed because of their low intensities and their wavelengths
that were outside of our observation window.
To establish the reliability of this time-dependent DFT
B3LYP/6-31G(d) model to calculate the absorption maximum
we have also computationally examined 1,5-dithiacyclooctane
radical cation, 2+•, because it has been directly observed and
has an absorption maximum at 405 nm in acetonitrile.36 This
study led to the location of the BC, chair-chair, and boat-
boat conformations that had been previously identified37 and a
fourth much higher energy TB conformation (Figure 7). TD-
DFT calculations on the BC global minimum faithfully repro-
duced the experimentally observed absorption peak (404 nm, f
) 0.1089; 414 nm (CH3CN), f ) 0.1543), providing compelling
verification of the veracity of this method for the study of 5H,
7H-dibenzo[b,g][1,5]dithiocin radical cation, 1+•
Figure 6. TD-DFT calculations on the B3LYP/6-31G(d)-minimized
Pyrylium Salt Photosensitized Reactions of 1. Steady-state
photolysis of 12.5 mM solutions of 1 in the presence of 1 mM
triarylpyrylium tetrafluoroborates, TPP, 2MeO-TPP, or 2MeO-
twist-boat-boat conformation of 1+• in the (a) gas phase and (b) in CH3-
CN.
STPP and 10 mM of benzylbenzoate as an internal standard in
oxygen-saturated CHCl3 led predominately to the formation of
sulfoxide 3 (eq 4a). After extended irradiation, a trace (<1%)
of an unknown product is produced, which we suspect is a ring
cleavage product; however, its quantity precluded isolation and
conclusive identification. Reaction of 1,5-dithiacyclooctane, 2,
(35) Petersilka, M.; Grossmann, U. J.; Gross, E. K. U. Phys. ReV. Lett. 1996,
76, 1212-1215.
(36) Tamaoki, M.; Serita, M.; Shiratori, Y.; Itoh, K. J. Phys. Chem. 1989, 93,
6052-6058.
(37) Stowasser, R.; Glass, R. S.; Hoffmann, R. J. Chem. Soc., Perkin Trans. 2
1999, 1559-1561.
9
4062 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008

Sulfide Radical Cations in Electron Transfer
A R T I C L E S
electron transfer to be in the Marcus inverted region, which is
often the case and has recently been experimentally verified
for the charge-shift reactions between tetralkyltins and acri-
dinium ions.40
Rehm Weller analyses (eq 3) of the electron-transfer reactions
suggested by these solvent effects for TPP (ES ) 65 kcal/mol,
ET ) 53 kcal/mol, E0(reduction) ) -0.30V vs SCE in CH3-
CN) and 2MeOTPP (ES ) 55 kcal/mol18, E0(reduction) ) -0.36
vs SCE in CH3CN) indicate that the reaction of 1 is substantially
exergonic with the singlet excited states of both sensitizers (TPP
∆G° ) -30 to -41 kcal/mol; 2MeOTPP ∆G° ) -20 to -31
kcal/mol) and with the triplet excited-state of TPP (∆G° ) -17
to -28 kcal/mol). In addition, Stern-Volmer quenching of TPP
(τS ) 4.2 ns)41 and of 2MeOTPP (τS ) 4.8 ns)18 by both 1 (kq
) (2.14 ( 0.13) × 1010 M-1s-1 and kq ) (1.10 ( 0.01) × 1010
M-1s-1, respectively) and by 2 (kq ) (2.39 ( 0.12) ×
1010M-1s-1 and (1.28 ( 0.01) × 1010M-1s-1, respectively) are
all diffusion-controlled processes.
Figure 7. B3LYP/6-31G(d) energies and conformations of 1,5-dithiacy-
clooctane radical cation.
Table 1. Extent of Product Formation during Photooxygenation of
a
1 with Several Pyrylium Salts in CD3CN and CDCl3
Nanosecond laser flash photolysis (LFP) studies of both TPP
and 2MeOTPP in an argon atmosphere in the presence of 1
also support an electron-transfer mechanism (Figure 8). In the
TPP experiment, a transient species appears at 550 nm that can
be safely assigned to the pyrylium radical (TPP•) based on
literature precedent42 (part a in Figure 8) We also assign the
transient observed in the laser flash photolysis experiment of a
mixture of 1 and 2MeOTPP to the corresponding pyrylium
radical by comparison to an electrochemically generated sample
(Supporting Information for details).
sensitizer
solvent
% product formation
CDCl3
CD3CN
CDCl3
CD3CN
CDCl3
CD3CN
CDCl3
CD3CN
36.5
0
28.9
14.0
58.6
6.8
TPP
2MeOTPP
2MeOSTPP
STPP
39.5
5.3
a Measured on a merry-go-round at a fixed irradiation time to allow
comparison of percent conversions (percent product formation).
Irradiations of 2MeOTPP+BF4 and 2MeOSTPP+BF4 did
not produce an emission signal at 1270 nm attributable to
sensitized formation of singlet oxygen (1∆g). The absence of a
singlet oxygen component to this reaction was also verified by
the absence of an isotope effect (kH/kD ) 0.99 ( 0.03) in the
-
-
in CH3CN produced a more complicated reaction mixture
consisting of 1,5-dithiacyclooctane sulfoxide, 2SO, and 1,5-
dithiacyclooctane bissulfoxide, 2SOSO (eq 4b). However,
examination of this reaction as a function of time clearly shows
that the bissulfoxide is a secondary product formed by over-
oxidation of 2SO.
2MeOTPP+BF4 sensitized reaction of 2d4 (Scheme 7).
-
The 2MeOTPP+BF4--sensitized photooxygenation of tert-
butyl phenylsulfide (eq 5) generated a preponderance of cleavage
products. Product formation was suppressed by 95 ( 5% in a
nitrogen atmosphere in the presence of 1.2 × 10-3 M benzo-
quinone. In contrast, under identical reaction conditions only
50 ( 3% of the photoxygenation of 1 was quenched by 1.2 ×
10-3M benzoquinone. Baciocchi and co-workers43 observed the
same oxygen dependent cleavage of this sulfide using NMQ+
and attributed it to rearrangement of a thiadioxirane intermediate.
The photooxygenations of 1 with all of the pyrylium salt
sensitizers were more rapid in the less polar CDCl3 (ET(30) )
39.1 kcal/mol) than in the more polar CD3CN (ET(30) ) 46.0
kcal/mol)38 (Table 1). Similar unusual solvent dependencies have
previously been observed with electron-transfer reactions that
generate radical/radical cation pairs that lack a Coulombic barrier
to separation.39 It has been suggested that back electron transfers,
which decrease the efficiencies of these charge-shift reactions,
are suppressed in the lower polarity solvents with smaller solvent
reorganization energies (λ). This suggestion requires the back
(40) Fukuzumi, S.; Ohkubo, K.; Suenobu, T.; Kato, K.; Fujitsuka, M.; Ito, O.
J. Am. Chem. Soc. 2001, 123, 8459-8467.
(41) Marin, M. L.; Miguel, A.; Santos-Juanes, L.; Arques, A.; Amat, A. M.;
Miranda, M. A. Photochem. Photobiol. Sci. 2007, 6, 848-852.
(42) Akaba, R.; Niimura, Y.; Fukushima, T.; Kawai, Y.; Tajima, T.; Kuragami,
T.; Negishi, A.; Kamata, M.; Sakuragi, H.; Tokumaru, K. J. Am. Chem.
Soc. 1992, 114, 4460-4464.
(38) Reichardt, C. Angew. Chem., Int. Ed. Engl. 1979, 18, 98-110.
(39) Todd, W. P.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould, I. R. J.
Am. Chem. Soc. 1991, 113, 3601-3602.
(43) Baciocchi, E.; Del Giacco, T.; Giombolini, P.; Lanzalunga, O. Tetrahedron
2006, 62, 6566-6573.
9
J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008 4063

A R T I C L E S
Clennan and Liao
Figure 8. Laser flash photolysis of (a) TPP and (b) 2MeOTPP at 355 nm in the absence and presence of 1.
Discussion
transfer pathway for the MB-sensitized reactions of 1 and 2
can be ruled out and the well-established singlet oxygen pathway
(Scheme 1) supported by several experimental observations. (1)
Both 1 and 2 quench the 1270 nm emission of singlet oxygen
with rate constants (1-[6.7 ( 0.4] × 105M-1s-1; 2-[2.1 ( 0.2]
× 107M-1s-1) much larger than that observed for cis-stilbene.
(2) Product formation in the reactions of 1 and 2 are both
quenched by the specific singlet oxygen quencher, DABCO.
(3) The kinetic isotope effect observed in the MB-sensitized
reaction of 2d4 is consistent with intermediacy of a persulfoxide,
a well-established intermediate in singlet-oxygen reactions of
sulfides. (4) Oxidation is not observed (eq 1) at S12, the sulfur
directly attached to the two phenyl rings in 1, consistent with
the lack of reactivity of diphenylsulfide with singlet oxygen.
(5) Diphenylsulfoxide, which is itself unreactive to singlet
oxygen, is oxidized to diphenylsulfone in the presence of 1,
The reactions of two 1,5-dithiacyclooctanes, 1 and 2, have
been examined with three different families of sensitizers,
thiazines (Methylene Blue), quinolinium salts, and pyrylium
cations.
Methylene Blue. Methylene Blue (MB) is a well-established
singlet sensitizer44,45 that promotes singlet oxygen reactions of
sulfides by the mechanism shown in Scheme 1. However, the
possibility of an electron-transfer pathway for the MB-sensitized
reaction of easily oxidized 1 (eq 1) cannot be summarily
discarded because Manring, Ericksen, and Foote46 have provided
convincing evidence for electron transfer in the MB photooxy-
genation of trans-stilbene. In this case, electron transfer occurs
because of the well-recognized lack of reactivity of trans-
stilbene toward singlet oxygen.47,48 Nevertheless, an electron-
(44) Stracke, F.; Heupel, M.; Thiel, E. J. Photochem. Photobiol., A 1999, 126,
51-58.
(48) Wilkinson, F.; Helman, W. P.; Ross, A. B. J. Phys. Chem. Ref. Data 1995,
24, 663-1021. The rate constant for the reaction of trans-stilbene with
singlet oxygen has not been reported; however, the rate constant for the
reaction of cis-stilbene (3 × 103M-1s-1) is one of the smallest reported in
the Wilkinson compilation of rate constants for over 1900 compounds.
(45) Tanielian, C.; Golder, L.; Wolff, C. J. Photochem. 1984, 25, 117-125.
(46) Manring, L. E.; Eriksen, J.; Foote, C. S. J. Am. Chem. Soc. 1980, 102,
4275-4277.
(47) Eriksen, J.; Foote, C. S. J. Am. Chem. Soc. 1980, 102, 6083-6088.
9
4064 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008

Sulfide Radical Cations in Electron Transfer
A R T I C L E S
Scheme 8
consistent with trapping a persulfoxide, a well-established
intermediate, in the reactions of sulfides with singlet oxygen.
N-Methylquinolinium Tetrafluoroborate. The N-meth-
ylquinolinium salt is a well-established potent electron-transfer
sensitizer with a high singlet excited-state energy of 81.5 kcal/
mol, a triplet excited-state energy of approximately 60.5 kcal/
mol, and a reduction potential of -0.85 V versus SCE.49 The
triplet excited-state is not populated significantly by intersystem
crossing from the singlet as revealed by a very high fluorescence
quantum yield but is formed in the presence of oxygen in a
process that produces singlet oxygen with a quantum yield of
2.0.17 Baciocchi and co-workers,17 however, provided convinc-
ing evidence that the NMQ+BF4--sensitized reaction of di-n-
butylsulfide occurs by the electron-transfer mechanism depicted
in Scheme 8. The key feature of this mechanism is the reaction
between superoxide and the di-n-butylsulfide radical cation that
generates the thiadioxirane (step 3 Scheme 8). This is also the
step that distinguishes it from singlet-oxygen reactions that
produce persulfoxide and hydroperoxysulfonium ylide inter-
mediates (vide supra). The absence of a persulfoxide in this
reaction was inferred from the inability to trap a nucleophilic
intermediate with diphenylsulfoxide.
kr, which must be no larger than the total rate constant (kT ) kr
+ kq ) [6.7 ( 0.4] × 105M-1s-1) for the removal of singlet
oxygen from solution by 1 is clearly too small to compete with
the electron-transfer pathway. Even di-n-butylsulfide, which
reacts more rapidly with singlet oxygen (kT ) 2.9 × 107M-1 -1
s )
than 1, is preferentially oxidized by an electron-transfer mech-
-
anism in the presence of NMQ+BF4 (vide supra).
Triarylpyrylium Salts. The product distributions for the
sensitized photooxygenations of 1 are very different for pyrylium
salts and NMQ+BF4-, which rule out the operation of any
unitary mechanism for these sensitizers such as depicted in
Scheme 4. In addition, the triarylpyrylium salts used in this study
do not sensitize the formation of singlet oxygen. Consequently,
we suggest the unique photooxygenation mechanism for these
triarylpyrylium cations depicted in Scheme 10. The formation
of the triplet excited-state of pyrylium cations (step 1, Scheme
10) is well established50 and in many cases the triplet-triplet
absorption can be observed by laser flash photolysis. The
arguments for the charge-shift electron transfer (step 2, Scheme
10) are also compelling and include, (1) the fluorescence of
TPP, and 2MeO-TPP are quenched at diffusion-controlled rates
by 1, (2) electron transfers from 1 to the excited pyrylium salts
-
The experimental results obtained in the NMQ+BF4 sensi-
Scheme 10
tized reaction of 1 are also more consistent with a reaction via
electron transfer than via singlet oxygen. For example, 5H, 7H-
dibenzo[b,g][1,5]dithiocin-12-oxide, 6, formed in the reaction
of 1 with NMQ+BF4 (eq 2), is clearly not a singlet-oxygen
-
product as established by the reaction with methylene blue (vide
supra). This conclusion is confirmed by the inability of DABCO
to quench the formation of 6. The formations of 3 and 5 are
also unaffected by the addition of DABCO to the NMQ+BF4
/
-
1 reaction, suggesting that singlet oxygen is not involved in
the formation of any of the reaction products. On the other hand,
1 quenches the fluorescence of NMQ+BF4- at a nearly diffusion
controlled rate and laser flash photolysis unambiguously identi-
fies the formation of NMQ•. These results coupled with a
favorable calculated exergonicity suggest that reaction of 1 with
NMQ+BF4 is initiated by electron transfer (Scheme 9). The
-
rate constant for the chemical reaction of 1 with singlet oxygen,
Scheme 9
(49) Yoon, U. C.; Quillen, S. L.; Mariano, P. S.; Swanson, R.; Stavinoha, J. L.;
Bay, E. J. Am. Chem. Soc. 1983, 105, 1204-1218.
(50) Miranda, M. A. Chem. ReV. 1994, 94, 1063-1069.
9
J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008 4065

A R T I C L E S
Scheme 11
Clennan and Liao
Table 2. Energy Differences between the Anti and Bisected
Conformations of Methyl Dimethylpersulfoxide
are exergonic, (3) the formation of the pyrylium radicals have
been observed in laser flash photolysis experiments in the
presence of 1, and (4) the observed increases in reaction
efficiencies with decreasing solvent polarity are diagnostic of
charge-shift reactions.
computational level
∆
E (anti-bisected), kcal/mol
MP2/6-31G(d)
MP2/6-311+G(d,p)
B3LYP/6-31G(d)
B3LYP/6-311+G(d,p)
B3LYP/6-311++G(3df,2dp)
1.94
2.04
1.73
1.89
1.83
Step 3 in Scheme 10 involves the addition of oxygen directly
to the pyrylium radicals to form covalently bound peroxy
radicals (Scheme 11). Pyrylium radicals TPP•, 2MeOTPP•, and
2MeOSTPP do not appear to react with oxygen on the LFP
time scale; however, they do react on the cyclic voltammetry
time scale (Figure S15 in the Supporting Information). It is the
oxidations of these peroxy radicals that we suggest are
responsible for the quenching of pyrylium-catalyzed photooxy-
genations by 1,4-benzoquinone. The trapping by benzoquinone
leads directly to the pyrylium cation, oxygen, and the benzo-
quinone radical that subsequently reduces the sulfide radical
cation. This oxidation competes with step 4, which is sterically
slower for tert-butylphenyl sulfide than for 1, which provides
an explanation for the less effective quenching of the photo-
oxygenation of 1.
The reaction of 2MeOTPP• with oxygen is a complicated
process as revealed by the absence of an isosbestic point in the
UV spectrum when electrochemically generated 2MeOTPP• is
exposed to oxygen (Figure S11 in the Supporting Information).
The initially formed product is likely a peroxy radical formed
by the addition of oxygen to either the 2 or 4 position of the
radical (Scheme 11). It is well established that pyrylium radicals
dimerize at the 4-position (Scheme 11);52 however, B3LYP/6-
31G(d) calculations (Supporting Information) identified three
rotomeric minima for the 2-peroxy radical, 7OO•, which are
all 1 to 2 kcal/mol more stable than the 4-peroxy radical
rotomers, 8OO• (Figure S16 in the Supporting Information).
In step 4 (Scheme 10), these peroxy pyrylium radicals, 7OO•
and 8OO•, add to the sulfide radical cation, located in close
proximity within the same solvent cage, to form the alkylated
persulfoxides shown in the box in Scheme 10. The other two
possible alkylated persulfoxide isomers are not formed because
the ortho hydrogens on the phenyl rings in the twist-boat-
boat conformation of the radical cation of 1 effectively block
the approach to S12. Computational evidence for these alkylated
persulfoxides was obtained by examining O-methyl dimethylp-
ersulfoxide. Using this simple model, two alkylated persulfoxide
rotomers, the bisected and anti conformations, have been located
as energy minima at a variety of different computational levels
(Table 2).
restoration of the pyrylium cation resonance energy. Nucleo-
philic oxygen transfer to the sterically unencumbered S6 sulfur
in 1 subsequently generates two molecules of 3 as the exclusive
product (step 6, Scheme 10). The oxygen-dependent pyrylium-
sensitized cleavage of tert-butylphenyl sulfide (eq 5) provides
the experimental evidence for the thiadioxirane as the key
intermediate in this pyrylium catalyzed process. However, the
thiadioxirane must be formed without the intermediacy of a
persulfoxide intermediate because a product isotope effect
([2SOh4]/[2SOd4]) of 0.99 ( 0.03 was observed in the 2MeOT-
PP+-catalyzed photooxygenation of 2d4 (Scheme 7).
The direct formation of a thiadioxirane bypassing the forma-
tion of a persulfoxide intermediate suggests that the anti rotomer
of the alkylated persulfoxide is the crucial reactive intermediate.
The MP2/6-31G(d) bisected conformation of the alkylated
persulfoxide has a core geometry nearly identical to the MP2/
6-31G(d) geometry of the stable bisected persulfoxide (Figure
9) and would be expected to collapse to this product by a least-
motion argument. On the other hand, the anti alkylated
persulfoxide has a core geometry nearly identical to that of
thiadioxirane (Figure 9) and very similar to the transition state
for the formation of the thiadioxirane and is its likely precursor.9
Indeed, removal of the methyl group from the MP2/6-31G(d)
structure of anti-methyl dimethylpersulfoxide followed by
geometry minimization leads directly to thiadioxirane. This is
consistent with previous computational studies that failed to
locate a stable anti persulfoxide.9
These unusual alkylated persulfoxides decompose to generate
the thiadioxirane in step 5 (Scheme 10) concomitantly with
(51) Sawyer, D. T. Oxygen Chemistry; Oxford University Press: New York,
Figure 9. Reaction of the rapidly inteconverting bisected and anti alkylated
persulfoxides to give the bisected persulfoxide and thiadioxirane, respec-
tively.
1991; p 223.
(52) Wintgens, V.; Pouliquen, J.; Kossanyi, J.; Heintz, M. NouV. J. Chim. 1986,
10, 345-350.
9
4066 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008

Sulfide Radical Cations in Electron Transfer
A R T I C L E S
Patterson methods and refined by least-squares techniques adapting the
An alternative mechanism to that shown in Scheme 10
involves thermodynamically unfavorable formation of super-
oxide (E°(O2/O2-•) ) -0.87 V vs SCE in CH3CN)51 by
pyrylium radical (TPP•, 2MeOTPP•, and 2MeOSTPP•) reduction
of oxygen that is driven by its rapid reaction with sulfide radical
cation, or its formation by decomposition of 7SOO• and 8SOO•,
and its subsequent reaction via steps 3 and 4 in the Baciocchi
mechanism2 (Scheme 8). This would provide a rationale for the
observed quenching by benzoquinone. However, the absence
of the diagnostic superoxide product 6 formed in the NMQ+-
catalyzed reaction of 1 (eq 2) precludes this alternative pos-
sibility.
full-matrix weighted least-squares scheme, w-1 ) σ2Fo2 + (0.0596P)2
+ 0.17P, where P ) (Fo + 2Fc2)/3, on F2 using the SHELXTL
2
program.57,58
C14H12S2O2 crystallizes in the centrosymmetric triclinic space group
P1h. The asymmetric unit consists of a compound molecule. The
molecules are well ordered and well separated. All non-hydrogen atoms
were located in the difference maps during successive cycles of least-
squares and refined anisotropically. All of the hydrogen atoms except
for one were located in the Fourier maps and refined isotropically. One
of the hydrogen atoms was placed in a calculated position and refined
isotropically using a riding model. The final refinement parameters were
R1 ) 0.0364 and wR2 ) 0.1015 for data with F > 4σ(F), giving a
data-to-parameter ratio of 12. The refinement data for all of the data
were R1 ) 0.0403 and wR2 ) 0.1050.
Conclusion
We have demonstrated that electron-transfer photooxygen-
ations of sulfides are far more complex than previously assumed.
Although the experimental data suggests that NMQ+ functions
as a simple outer sphere electron shuttle between the sulfide
and oxygen, and pyrylium cations function by a less passive
inner sphere mechanism. We also introduced a new technique,
the photooxygenation of 2d4, to detect the absence of a
persulfoxide in electron-transfer-catalyzed oxygenations and
illustrated its use with the pyrylium cation sensitizer, 2MeOT-
PP+.
Laser Flash Photolysis. Nanosecond laser flash photolysis experi-
ments were conducted using an excitation wavelength of 355 nm (5-
20 mJ/pulse) from a Nd:YAG laser. The transient spectra were recorded
with a point-point technique with 5-10 nm intervals from 350 to 700
nm. The kinetic decays were the average of at least 10 shots. Samples
were degassed by either three freeze-pump-thaw cycles or by bubbling
with argon. The optical densities were adjusted to between 0.3 and 0.6
at the excitation wavelength. When oxygen was needed, the sample
was saturated with oxygen for 20 min. Several solvents were used
including methylene chloride, 1,2-dichloroethane, and acetonitrile. In
general, the most intense peaks were observed in 1,2-dichloroethane.
The singlet-oxygen experiments were done with oxygen-saturated
solutions using either the 355 or 532 nm output of a 10 Hz Nd:YAG
laser. The kinetic apparatus consisted of a germanium (Judson 5 mm
φ) diode detector/customized preamplifier and various optics: A 10
nm narrow band-pass non-fluoresceing filter centered at 1.27 µm placed
just ahead of the detector, a 500 MHz LeCroy transient digitizer/signal
averager interfaced to a computer, and an energy meter. Because the
experimental decay is a convolution of the detector response (fwhm of
approximately 10 ps for the 5 mm φ detector) and the sample decay,
it was necessary to implement a numerical deconvolution analysis to
accurately extract measured lifetimes 100 µs g τ g 2 ms from the
recorded data. The numerical deconvolution analysis of Demas59 for
exponential decays corrects artifacts caused by scattered excitation light
that may reach the detector that was used. This analysis was
implemented on our laboratory computer using a general scientific/
engineering analysis/data acquisition program called Lifetime. Signal
averaging 100 experiments, each with laser pump energies e5 mJ gave
8192-point decay curves, each of which produced pseudo first-order
rate constants with correlation coefficients (square root) better than 0.99.
Experimental Section
Sulfides 1, 2,10 2SO,10 2SOd4,10 2d4,10 3,25 4,25 5,25 6,25 and
NMQ+BF4
were all synthesized as reported in the literature.
-53
Commercial samples of TPP, and 2MeOTPP were used as received
for the photochemical reactions and purified by recrystallization from
ethanol for the electrochemical studies.
2-(p-Methoxyphenyl)-4,6-diphenyl-thiapyrylium Tetrafluorobo-
rate. 2-MeO-STPP was synthesized using a modified procedure of
Suld54 and Wizinger.55 A 2.5 mL H2O solution of sodium sulfide
nonahydrate (250 mg, 1.04 mmol) was added to a solution of 2MeOTPP
(0.21 g, 0.5 mmol) in 10 mL of acetone. After the addition was
complete, the mixture was stirred for 0.5 h, then acidified with 2.5 mL
of 48% HBF4, diluted with 10 mL of H2O, stirred for 2 h, and then
filtered. The crude product was recrystallized twice from ethanol. Yield
11%. 1H NMR (CD3CN) δ 3.96 (s, 3H), 7.27(d, J ) 8.7 Hz, 2H), 7.70-
7.81(m, 6H), 8.07-8.17(m, 6H), 8.80(d, J ) 14.7 Hz, 2H).
Radical Cation Preparation. Solutions of 1 and 2 were prepared
in a specially designed two-compartment cell with one compartment
filled with 0.1 M substrate and the second compartment with 1 equiv
of sublimed (220 °C) NOBF4 in acetonitrile. Both compartments were
subjected to a series of three freeze-pump-thaw cycles and then mixed
at room temperature under vacuum. The UV-vis spectra of 2+• was
collected immediately. The yellow color of 1+• disappeared too rapidly
to allow measurement of its UV-vis spectrum at room temperature.
Crystallographic Data for 5. X-ray diffraction data were collected
for C14H12S2O2 on a Bruker P4 Diffractometer equipped with a
molybdenum tube and a graphite monochromator at 25 °C. A colorless
rectangular prismatic crystal of approximate dimensions 0.52 × 0.36
× 0.24 mm3 glued to a glass fiber was used for data collection. A total
of 2812 (Rint ) 0.0226) independent reflections were gathered in the
2θ range of 4.04 to 55° with the data collected having -1 e h e 9,
-10 e k e 10, -13 e l e 13 using the XSCANS program.56 Three
standard reflections measured after every 97 reflections exhibited no
significant loss of intensity. The data were corrected for Lorentz-
polarization effects and absorption. The structure was solved by
Stern-Volmer Studies. All Stern-Volmer studies were done using
acetonitrile as the solvent and sensitizer concentrations of [NMQ+] )
1.5 × 10-5 M, [TPP+] ) 4 × 10-6 M, and [2MeOTPP+] ) 4 × 10-6
M. In the TPP+ quenching studies, the concentrations of 1 were varied
over the range of 5.16 × 10-4 to 9.82 × 10-3 M, the concentrations of
2 over the range of 5.7 × 10-4 to 5.7 × 10-3 M, and utilized an
excitation wavelength of 405 nm and slit widths of 2.5 nm on the
excitation side and 5 nm on the emission side. In the 2MeOTPP+
quenching studies, the concentrations of 1 were varied over the range
of 2.0 × 10-3 to 8.7 × 10-3 M, the concentrations of 2 over the range
of 6.6 × 10-4 to 7.9 × 10-3 M, and utilized an excitation wavelength
of 447 nm and slit widths of 5.0 nm on both the excitation side and
emission side. In the NMQ+ quenching studies, the 1 concentrations
were varied over the range of 2.5 × 10-4 to 1.5 × 10-3 M, and utilized
an excitation wavelength of 317 nm and slit widths of 2.5 nm on the
excitation side and 5 nm on the emission side. All slopes of the Stern-
(53) Donovan, P. F.; Conley, D. A. J. Chem. Eng. Data 1966, 11, 614-615.
(54) Suld, G.; Price, C. C. J. Am. Chem. Soc. 1962, 84, 2090-2094.
(55) Wizinger, R.; Ulrich, P HelV. Chim. Acta 1956, 39, 207-216.
(56) Bruker XSCANS, ver. 2.31; Bruker AXS: Madison, WI, 1993.
(57) Bruker SHELXTL, ver. 5.10; Bruker AXS: Madison, WI, 1997.
(58) Sheldrick, G. M. SHELXS97 and SHELXL97, University of Go¨ttingen:
Go¨ttingen, Germany, 1997.
(59) Love, J. C.; Demas, J. N. Anal. Chem. 1984, 56, 82.
9
J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008 4067

A R T I C L E S
Clennan and Liao
Volmer plots were extracted by a least-squares treatment of the relative
intensity data.
600 or 250 W halogen-tungsten lamp. The reaction mixtures were
analyzed by either proton NMR, GC, or GCMS.
Computational Studies. All of the calculations were done using
Gaussian 0360 and the B3LYP/6-31G(d) model except those calculations
appearing in Table 2. All of the stationary points were checked with
frequency calculations and the absence of spin contamination was
verified by examination of S2 .
Photooxygenation of tert-Butylphenylsulfide. The photooxygen-
ations of 0.1 M tert-butylphenylsulfide with 4 × 10-4 M NMQ+ and
4 × 10-4 M 2MeOTPP+ were conducted in acetonitrile using 3 × 10-4
M hexamethyldisiloxane as an internal standard. The NMQ+ reaction
was irradiated with a medium-pressure Hanovia lamp for 10 min and
the 2MeOTPP+ sensitized reaction was irradiated for 1 h with a 600
W halogen-tungsten lamp. Both reaction mixtures were examined by
proton NMR, clearly demonstrating the formation of identical products
in very similar ratios. Under these reaction conditions, 53% of the tert-
butylphenylsulfide had reacted in the NMQ+ reaction and 59% in the
2MeOTPP+ reaction. The major product was the tert-butylacetamide
in both cases (17% NMQ+, 26% 2MeOTPP+). Substantial amounts of
tert-butanol (13% NMQ+, 21% 2MeOTPP+) and acetone (9% NMQ+,
7% 2MeOTPP+) were also found in the reaction mixtures. Phenyl
sulfonic acid and diphenyldisulfide were partially obscured by overlap
with the starting material and were not quantified. When these reactions
were run under identical conditions, but, in the presence of 1.2 × 10-3
M benzoquinone, less than 8% of the products were observed in the
NMQ+ reaction and no products were observed in the 2MeOTPP+
reaction.
Cyclic Voltammetry. Electrochemical samples were prepared in
acetonitrile and were 2 × 10-3 M in substrate, 0.1 M in supporting
electrolyte (nBu4NClO4), and 2 × 10-3 M in ferrocene as an internal
standard. Cyclic voltammograms were collected at a variety of scan
rates, utilizing a three-electrode configuration including a glassy carbon
working electrode, a platinum wire as a counter electrode, and a Ag/
AgCl (NaCl-satd) reference electrode.
General NMQ+BF4 Photooxygenation Conditions. The photo-
-
oxygenations of 1 (1.25 × 10-2 M) and 2 sensitized by 5 mol %
NMQ+BF4- were carried out in dry-oxygen-agitated/saturated deuter-
ated acetonitrile solutions in an NMR tube. A 450 W medium-pressure
mercury lamp was used as the light source, and a Pyrex test tube was
used as a filter (310 nm cutoff). Oxidation products were detected and
monitored by either proton NMR or GCMS.
General Singlet Oxygen Photooxygenation Conditions. The
photooxygenations of 0.1 M 1 sensitized by methylene blue (10-4 M)
were carried out in oxygen-saturated deuterated acetonitrile solutions
in an NMR tube. The samples were irradiated through a 1 cm saturated
NaNO2 filter solution with a halogen-tungsten lamp set at 250 W.
The reaction mixtures were analyzed by proton NMR.
General Pyrylium Cation Photooxygenation Conditions. The
photooxygenations of 1-2 × 10-2 M of 1 or 2 with 8 mol % TPP+,
2MeOTPP+, and 2MeOSTPP+ were carried out in oxygen-agitated/
saturated deuterated acetonitrile solutions in an NMR tube. Irradiations
were carried out through 1 cm of a saturated NaNO2 solution with a
Solvent Effects (Table 1). A 1 mL CD3CN stock solution of the
sensitizer was used to prepare a 0.5 mL CD3CN sample containing 2
mg of 1 and 3 mol % (5 × 10-3 M) of the sensitizer. The CDCl3
samples were prepared by first removing the CD3CN under a vacuum
from a 100 µL aliquot of the stock solution before adding 0.5 mL of
CDCl3 containing 2 mg of 1. Both solutions were put into pyrex NMR
tubes, saturated with oxygen for 3 min, and then irradiated with UVA
lamps in a merry-go-round.
Acknowledgment. We thank the National Science Foundation
for their generous support of this research. We also thank Dr.
Doug Wheeler for development of the program Lifetime used
to extract lifetimes from the singlet-oxygen emission data.
(60) 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.
Supporting Information Available: Stern-Volmer quenching
data of sensitizers by 1 and 2, electrochemical data for
triarylpyrylium cations in argon and oxygen, TD-DFT results
for 1 and 2 radical cations, B3LYP/6-31G(d) structure data for
1 and its derivatives and for O-methyl dimethylpersulfoxide,
and a CIF file for the X-ray analysis of 5. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA710301S
9
4068 J. AM. CHEM. SOC. VOL. 130, NO. 12, 2008