Photosensitized reduction of sulfonium salts: Evidence for nondissociative electron transfer

Article abstract of DOI:10.1021/ja982932x

The photosensitized reductions of 4-cyanobenzylmethylphenyl sulfonium tetrafluoroborate (1-BF₄) by 9-phenylanthracene, 2-ethyl-9,10-dimethoxyanthracene, or perylene excited singlet states gives quantitative yields of phenyl methyl sulfide. Fluorescence quenching and the quantum yields for product formation, as functions of [1-BF₄], give bimolecular rate constants (0.58-1.6 × 10¹⁰ M^-1 s^-1) for reaction of the excited sensitizers with the sulfonium salt. The limiting quantum yields, corresponding to infinite [1-BF₄], are 0.65-0.77 for the three sensitizers, revealing significant inefficiencies in the photoreduction. These inefficiencies are assigned to the partitioning of a sulfuranyl radical intermediate in a two-step associative electron-transfer mechanism.

Full text of DOI:10.1021/ja982932x

4364
J. Am. Chem. Soc. 1999, 121, 4364-4368
Photosensitized Reduction of Sulfonium Salts: Evidence for
Nondissociative Electron Transfer
Xiuzhi Wang, Franklin D. Saeva,† and J. A. Kampmeier*
Contribution from the Center for Photoinduced Charge Transfer, Department of Chemistry, UniVersity of
Rochester, Rochester, New York 14627-0216
ReceiVed August 14, 1998
Abstract: The photosensitized reductions of 4-cyanobenzylmethylphenyl sulfonium tetrafluoroborate (1-BF4)
by 9-phenylanthracene, 2-ethyl-9,10-dimethoxyanthracene, or perylene excited singlet states gives quantitative
yields of phenyl methyl sulfide. Fluorescence quenching and the quantum yields for product formation, as
functions of [1-BF4], give bimolecular rate constants (0.58-1.6 × 1010 M-1 s-1) for reaction of the excited
sensitizers with the sulfonium salt. The limiting quantum yields, corresponding to infinite [1-BF4], are 0.65-
0.77 for the three sensitizers, revealing significant inefficiencies in the photoreduction. These inefficiencies
are assigned to the partitioning of a sulfuranyl radical intermediate in a two-step associative electron-transfer
mechanism.
Introduction
also implicated in the reduction of sulfonium salts. Chemical
experiments with potassium on graphite as the reductant showed
that the relative rates of cleavage of different groups from
unsymmetrical sulfonium salts depend on the specific structure
of the sulfonium salts and not just on the nature of the groups
being cleaved.6
There is a significant body of evidence that 9-S-3 sulfuranyl
radicals, [R3S]•, are formed as reactive intermediates in the
reactions of free radicals with sulfides and in the reductive
cleavage of sulfonium salts. Most directly, sulfuranyl radicals
with electronegative groups such as alkoxy have been observed
by ESR and absorption spectroscopy.1 In the absence of
stabilizing substituents, alkyl and aryl sulfuranyl radicals are
apparently short-lived and have not been observed.2 Even so,
indirect tests have disclosed the intermediacy of sulfuranyl
radicals in a variety of reactions. In one case, a labeling study
of an intramolecular free radical displacement at sulfur revealed
an intermediate that partitioned among competitive pathways.3
In another case, a study of the reactions of hydrogen atoms with
benzylalkyl sulfides gave different Hammett plots for the relative
rates of reaction of the sulfides and the relative rates of
competing benzyl versus alkyl cleavages. These results indicate
that the radical addition to the sulfide and the cleavage are
distinct steps, corresponding to the formation and decomposition
of an intermediate sulfuranyl radical.4 In an especially interesting
case, a [R3S]• intermediate was formed by free radical addition
to a sulfide and trapped by oxidation.5 Sulfuranyl radicals are
Photochemical reductions of sulfonium salts also point to
intermediate sulfuranyl radicals. For example, the quantum yield
for the destruction of sulfonium salt by the photolysis of the
triphenylsulfonium iodide charge-transfer complex is only 0.35.7
This result suggests a triphenylsulfuranyl radical intermediate
that fractionates between productive fragmentation and unpro-
ductive return electron transfer.2,7
The reductive cleavage of sulfonium salts can also be
accomplished electrochemically.8 Saeva suggested that some of
these reactions might proceed by a concerted mechanism.9 In
general, the detailed characteristics of the cyclic voltammetric
behavior of most sulfonium salts are in accord with a two-step
“EC” mechanism-fast associative electron transfer followed by
slower unimolecular fragmentation of an intermediate. In one
favorable case, the oxidation of the intermediate was actually
observed at high scan rates.10 On the other hand, Savea´nt and
Saeva concluded that some phenacyl- and 4-cyanobenzyl-
substituted sulfonium salts are cleaved in a dissociative mech-
anism; i.e., electron transfer and bond fragmentation are
concerted. The stimulating conclusion that the mechanism of
electrochemical cleavage can be associative (two-step) or
* Corresponding author: kamp@chem.rochester.edu.
† Research Laboratories, Eastman Kodak Company, Rochester, New York
14650-2110.
(1) (a) Morton, J. R.; Preston, K. F. J. Phys. Chem. 1973, 77, 2645-
2648. (b) Gilbert, B. C.; Hodgeman, D. K. C.; Norman, R. O. C. J. Chem.
Soc., Perkin Trans. 2 1973, 1748-1752. (c) Chapman, J. S.; Cooper, J.
W.; Roberts, B. P. J. Chem. Soc., Chem. Commun. 1976, 835-836. (d)
Perkins, C. W.; Martin, J. C.; Arduengo, A. J.; Lau, W.; Alegria, A.; Kochi,
J. K. J. Am. Chem. Soc. 1980, 102, 7753-7759. Perkins, C. W.; Clarkson,
R. B.; Martin, J. C. J. Am. Chem. Soc. 1986, 108, 3206-3210. (e)
Chatgilialoglu, C.; Castelhano, A. L.; Griller, D. J. Org. Chem. 1985, 50,
2516-2518.
(6) Beak, P.; Sullivan, T. A. J. Am. Chem. Soc. 1982, 104, 4450-4457.
(7) Nickol, S. L.; Kampmeier, J. A. J. Am. Chem. Soc. 1973, 95, 1908-
1915.
(8) (a) Horner, L. In Organic Electrochemistry; Baizer, M. M., Lund,
H., Eds.; Marcel Dekker: New York, 1983. (b) Grimshaw, J. In The
Chemistry of the Sulfonium Group; Stirling, C. J. M., Patai, S., Eds.; Wiley-
Interscience: New York, 1981; Chapter 7. (c) Simonet, J. In Supplement
S, The Chemistry of Sulphur-Containing Functional Groups; Patai, S.,
Rappaport, Z., Eds.; John Wiley & Sons, Ltd.: West Sussex, England, 1993;
Chapter 1.
(9) Saeva, F. D.; Morgan, B. P. J. Am. Chem. Soc. 1984, 106, 4121-
4125.
(10) Andrieux, C. P.; Robert, M.; Saeva, F. D.; Save´ant, J.-M. J. Am.
Chem. Soc. 1994, 116, 7864-7871.
(2) (a) DeVoe, R. J.; Sahyun, M. R. V.; Schmidt, E.; Serpone, N.; Sharma,
D. K. Can. J. Chem. 1988, 66, 319-324. (b) Eckert, G.; Goez, M.; Maiwald,
B.; Mu¨ller, U. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 1191-1198.
(3) Fong, M. C.; Schiesser, C. H. Aust. J. Chem. 1992, 45, 475-477.
(4) Tanner, D. D.; Koppula, S.; Kandanarachchi, P. J. Org. Chem. 1997,
62, 4210-4215.
(5) Leardini, R.; Pedulli, G. F.; Tundo, A.; Zanardi, G. J. Chem. Soc.,
Chem. Commun. 1985, 1390-1391.
10.1021/ja982932x CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/24/1999

Photosensitized Reduction of Sulfonium Salts
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4365
dissociative (one-step), depending on the structure of the
sulfonium salt, is not based on the direct observation of an
oxidation peak in the cyclic voltammogram. For example,
neither phenyldimethylsulfonium ion (“two-step”) nor
p-cyanobenzylmethylphenylsulfonium ion (1) (“one-step”) gives
reversible redox waves at scan rates up to 800 000 V/s. Rather,
the mechanistic distinction is based on detailed analyses of the
variations in peak potentials and peak widths as a function of
scan rates.10 Save´ant has argued the case for dissociative electron
transfers in a number of other reductive cleavages.11 There are,
of course, interesting conceptual11 and practical12 implications
of the idea of dissociative electron transfer.
As noted above, the reductive cleavage of sulfonium salts
can also be accomplished by photoinduced electron transfer.7
The quantum yields for destruction of triphenylsulfonium salts
with various sensitizers are routinely less than unity.2,7,13 The
forward electron transfer can be arranged so that all of the donor
excited states are captured by acceptor sulfonium salt to give
the ion-radical pair. Because return electron transfer can be a
very fast reaction, it can be a very demanding probe of the
structure and reactions of the primary ion-radical pair. A
quantum yield less than unity suggests an intermediate pair that
can partition between energy-wasting (back) and productive
(forward) steps14 (Scheme 1).
Scheme 2
Table 1. Physical Properties of Sensitizers and 1-BF4
1E, eV
E1/2, V (Ag/AgCl)
∆GET, eVd
PA
PE
EDA
1-BF4
3.1616
2.8416
3.01a
4.4b
1.30
1.05
-0.92
-0.85
-1.23
0.84
-0.94c
a Estimated from the midpoint of the absorption and emission spectra.
b Estimated from the maximum of the long wavelength absorption at
284 nm. c Irreversible reduction: recorded value is EP at 0.1 V/s.
d Calculated in acetonitrile as ∆GET ) E1/2(sens•+/sens) - EP(1+/1•) -
1E.
(EDA), and perylene (PE). The sensitizers were chosen to favor
electron transfer and disfavor energy transfer from the excited
singlet state of the sensitizer to the sulfonium salt. The
appropriate physical properties are given in Table 1.
Scheme 1
Solutions of sensitizer (∼0.01 M) and sulfonium salt (∼0.01
M) in acetonitrile were irradiated at wavelengths where the
sensitizer absorbed, but the sulfonium salt did not. The products
of the reaction are straightforward; phenyl methyl sulfide is the
only sulfide formed. The other possible cleavage products,
4-cyanobenzyl phenyl sulfide and 4-cyanobenzyl methyl sulfide,
could not be detected (<0.1%). Poly-4-cyanobenzylated
anthracenes were identified as products of the reactions using
the two anthracene sensitizers (PA and EDA). The radical
cleavage product (4-cyanobenzyl radical) couples with the
sensitizer cation radical to give a carbocation which is then
deprotonated to give the alkylated sensitizer. This chemistry is
strictly analogous to that first identified by Saeva,17 as the source
of the acid in the commercially useful reactions of sulfonium
salts. In addition to these important cross-coupling products,
trace amounts of other products characteristic of 4-cyanobenzyl
cleavage were identified in PA-sensitized reactions: 4-cyano-
toluene, 4-cyanobenzyl alcohol, and 4-cyanobenzaldehyde. The
dimer, 1,2-di(4-cyanophenyl)ethane, was not observed (<0.1%).
N-(4-Cyanobenzyl) acetamide was observed in 11% yield based
on phenyl methyl sulfide in reactions involving PA sensitizer,
but was not formed (<0.1%) when EDA or PE were used. The
formation of this interesting product will be discussed later.
Finally, 2-ethyl-9,10-anthraquinone was observed in reactions
using EDA as a sensitizer. A related dealkylation path was
identified previously and attributed to reactions of the dialkoxy-
anthracene cation radical.18 The photosensitized reactions were
analyzed by 1H NMR as a function of photolysis time. As shown
in Figure 1a,b, there is a 1:1 correspondence between the
destruction of the sulfonium salt and the formation of phenyl
methyl sulfide. Accordingly, the formation of phenyl methyl
sulfide was used to monitor the photoreduction reaction.
Flash photolysis of solutions of PA in acetonitrile gives
readily detectable yields of 1PA (590 nm) and 3PA (430 nm).19
We have used this logic15 to probe the mechanism of the
reductive cleavage of 4-cyanobenzylmethylphenylsulfonium ion
(1), the best example of a dissociative electron-transfer mech-
anism in the electrochemical studies.10 Our predictive distinction
is that a one-step dissociative electron transfer will be revealed
by a 1:1 correspondence of the destruction of excited states and
the formation of cleavage products; i.e., Φ ) 1. A two-step
associative mechanism with an intermediate that partitions will
give Φ < 1 (see Scheme 2). The choice of substrate is
particularly important because the mechanism of electron
transfer might vary with the structure of the substrate.
Results and Discussion
We studied the photoinduced reduction of 4-cyanobenzyl-
methylphenyl sulfonium tetrafluoroborate [1-BF4] or hexafluo-
rophosphate [1-PF6] in acetonitrile using three different sensitiz-
ers, 9-phenylanthracene (PA), 2-ethyl-9,10-dimethoxyanthracene
(11) Save´ant, J.-M. Dissociative Electron Transfer. In AdVances in
Electron-Transfer Chemistry; Mariano, P. S., Ed.; JAI Press: Greenwich,
CT, 1994; Vol. 4, pp 53-116.
(12) (a) Saeva, F. D. In AdVances in Electron-Transfer Chemistry;
Mariano, P. S., Ed.; JAI Press: Greenwich, CT, 1994; Vol. 4, pp 1-25.
(b) DeVoe, R. J.; Olofson, R. M.; Sahyun, M. R. V. AdVances in
Photochemistry; Volman, D. H., Hammond, G. S., Neckers, D. C., Eds.;
John Wiley & Sons: New York, 1992, Vol. 17, pp 314-355. (c) Radiation
Curing in Polymer Science and Technology; Fouassier, J. P., Rabek, J. F.,
Eds.; Elsevier: London, 1993; Vol. 2. Crivello, J. V., Chapter 8; Hacker,
N. P., Chapter 9; Sahyun, M. R. V., DeVoe, R. J., Olofson, P. M., Chapter
10. (d) DeVoe, R. J.; Sahyun, M. R. V.; Schmidt, E. J. Imaging Sci. 1989,
33, 39-43.
(13) (a) Pappas, S. P.; Gatechair, L. R.; Jilek, J. H. J. Polym. Sci. Polym.
Chem. 1984, 22, 77-84. (b) Timpe, H.-J.; Bah, A. Makromol. Chem. Rapid
Commun. 1987, 8, 353-358. Timpe, H.-J. Pure Appl. Chem. 1988, 60,
1033-1038.
(14) Gould, I. R.; Ege, D.; Moser, J. E.; Farid, S. J. Am. Chem. Soc.
1990, 112, 4290-4301.
(15) Gan, H.; Kellett, M. A.; Leon, J. W.; Kloeppner, L.; Leinhos, U.;
Gould, I. R.; Farid, S.; Whitten, D. G. J. Photochem. Photobiol. A: Chem.
1994, 82, 211-218.
(16) Murov, S. V.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry; Marcel Dekker: New York, 1993.
(17) Saeva, F. D.; Morgan, B. P.; Luss, H. R. J. Org. Chem. 1985, 50,
4360-4362.
(18) DeVoe, R. J.; Sahyun, M. R. V.; Schmidt, E.; Sharma, D. K. Can.
J. Chem. 1980, 68, 612-619.

4366 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Wang et al.
Figure 2. Stern-Volmer analyses of fluorescence and lifetime
1
quenching of EDA by 1-BF4: 0 ) F0/F; ] ) τ0/τ.
Table 2. Results of Stern-Volmer Analyses of Fluorescence
Quenching by 1-BF4
10-10 kq,
sensitizer λex, nma λf, nmb kqτ0, M-1
rc
τ0, nsd M-1 s-1
PA
PE
EDA
362
380
383
380-500
400-600
390-550
49.66
50.83
87.36
0.9998 5.1
0.9997 4.6
0.9994 8.2
0.97
1.10
1.07
a Wavelength of exciting radiation. b Wavelength of emission band.
c Correlation coefficient. d Nondeoxygenated solutions.
concentration of sulfonium salt. This experiment, with EDA as
the sensitizer, is also shown in Figure 2. The good agreement
between these two different experiments shows that the domi-
nant quenching process is dynamic and that there is no
significant contribution from ground state complex formation.
Spectra of mixtures of each sensitizer with sulfonium salt in
acetonitrile showed no new absorptions in the UV region,
leading to the same conclusion. As part of the fluorescence
quenching studies, we also searched carefully for new fluores-
cence emissions; no new emissions were observed.
Since there is a 1:1 correspondence of the disappearance of
the sulfonium salt to the appearance of phenyl methyl sulfide,
the quantum yield for the two processes must be the same.
Because the reaction of the sensitizer singlet with sulfonium
salt is a dynamic process, the observed quantum yield for the
Figure 1. Concentrations of 4-cyanobenzylmethylphenyl sulfonium
salt [1-BF4] (0) and phenyl methyl sulfide (]) as a function of
photolysis time: (a) PA sensitizer; (b) EDA sensitizer.
The rate of disappearance of 1PA was markedly accelerated by
low concentrations of sulfonium salt 1-BF4, while the rate of
decay of 3PA was unchanged. Strong fluorescence in acetonitrile
was easily observed for all three sensitizers; the fluorescence
intensity decreased when the sulfonium salt 1-BF4 was added
to each solution of the sensitizers. These quenching processes
followed the Stern-Volmer equation, Fo/F ) 1 + kqτ0[1], where
F is the emission intensity integrated across the band. A typical
relationship is shown in Figure 2, and the results are summarized
in Table 2. Fluorescence lifetimes were measured by single-
photon counting under the same (nondeoxygenated) conditions
used for the Stern-Volmer analyses. These lifetimes are
reported in Table 2 also. The Stern-Volmer slopes and the
observed lifetimes give the quenching constants, kq. These kq
values are consistent with near-diffusion-controlled quenching
of the excited sensitizers by 1-BF4 (kdiff ) 1.9 × 1010 M-1 s-1
in CH3CN at 25 °C).16 Similar results were obtained in the
quenching of 1PA by 1-PF6 (kq ) 1.1 × 1010 M-1 s-1) and the
corresponding ethyl derivative, 2-PF6 (kq ) 1.2 × 1010 M-1
s-1).
formation of phenyl methyl sulfide is a function of the
PhSCH3
obs
concentration of sulfonium salt. Plots of 1/Φ
vs 1/[1-
BF4] were linear for all three sensitizers, Figure 3. The slopes,
in conjunction with the sensitizer singlet lifetimes measured
under the same conditions, give the bimolecular rate constants
for the reactions of the sensitizer singlets with sulfonium salt.
These rate constants are in acceptable agreement with those
derived from fluorescence quenching. The intercepts of the plots
give the limiting quantum yields, corresponding to the quantum
yield when 100% of the sensitizer singlets are captured by
reaction with the sulfonium salt. Values of τo, kq, and ΦLIM are
reported in Table 3.
As outlined in the Introduction, our mechanistic distinction
is that a concerted, dissociative electron transfer should give a
correspondence of excited states captured to cleavage product
PhSCH3
LIM
The quenching process could also be monitored by studying
the lifetime of the fluorescent state as a function of the
formed, i.e., Φ
) 1. In contrast, electron transfer to give
an intermediate that can partition between paths that do and do
not give phenyl methyl sulfide will be inefficient, Φ
The observed values of Φ
PhSCH3
LIM
<1.
(19) Workentin, M. S.; Johnston, L. J.; Wayner, D. D. M.; Parker, V.
D. J. Am. Chem. Soc. 1994, 116, 8279-8287.
PhSCH3
LIM
for three different sensitizers

Photosensitized Reduction of Sulfonium Salts
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4367
Scheme 3
(forward reaction) are polarized to give emission. This result is
best interpreted2b in terms of successive caged radical pairs;
the partitioning of the primary intermediate [Ph2I•, sens•+] is
responsible for the polarization of Ph2I+ and the partitioning of
the secondary intermediate [PhI, Ph•, sens•+] is responsible for
the net emissive polarization in C6H5D. The photosensitized
CIDNP experiment with Ph3S+ still gives evidence for succes-
sive radical pairs, but fails to directly reveal the sulfuranyl
radical; C6H5D is polarized in emission, but polarized Ph3S+ is
not observed. The rate for diffusive separation in acetonitrile is
a reference clock for the return electron transfer and the
fragmentation rates. Both return and fragmentation are
sufficiently fast to compete successfully with diffusive separa-
tion,20 indicating rate constants >109 s-1. These rate constants
are more than an order of magnitude greater than the limit set
by the absence of an oxidation wave in the cyclic voltammogram
at high scan rates (>4 × 107 s-1).10
Figure 3. Stern-Volmer analysis of the yield of phenyl methyl sulfide
as a function of [1-BF4]: 0 ) 9-phenylanthracene; ] ) 2-eth-
yldimethoxyanthracene; O ) perylene.
Table 3. Results of Stern-Volmer Analyses of Product Formation
1
from the Reactions of Sensitizers with Sulfonium Salt 1-BF4
PhSCH3
LIM
sensitizer τ0,a ns 1010kq, M-1 s-1
Φ
∆GBET, eVb
PA
PE
EDA
7.2
5.9
16.1
0.58
1.2
1.6
0.65 ( 0.003
0.69 ( 0.01
0.77 ( 0.02
-2.2
-2.0
-1.8
a Deoxygenated solutions. b Calculated from ∆G ) EP(1+/1•) -
E
1/2(sens•+/sens).
are all less than one, providing prima facie evidence for a two-
step mechanism involving an intermediate. Stated differently,
the mechanism of the photosensitized reductive cleavage of the
sulfonium salt provides an opportunity for the energy of the
incident light to be wasted. The rates, energetics, and products
of the reactions reported here and those of closely related
reactions2,12-14 are all in accord with a photoinduced electron
transfer from sensitizer donor to sulfonium acceptor to give the
primary intermediate, 3, in Scheme 2. Since the system is
arranged so that all of the light is absorbed by the sensitizer
and all of the sensitizer excited singlets are captured by the
sulfonium salt, and since there is no path for destruction of
sulfonium salt that does not give phenyl methyl sulfide, the
trivial sources of inefficiency are ruled out. Accordingly, the
source of the inefficiency must be in the partitioning of the
primary intermediate. The absence of any new emission
indicates the absence of a radiative pathway for 3 to return to
sulfonium salt. On the other hand, nonradiative return electron
transfer would waste the incident energy and regenerate the
starting material. As recorded in Table 3, the return electron
transfer is strongly exoergic in all three cases. Accordingly,
return electron transfer is expected to be a fast, sensitive probe
of the structure of the primary intermediate. The quantum yields
reveal a competition between reversion to starting sulfonium
ion and formation of product; i.e., the rate of the forward
reaction of the primary intermediate 3 must be comparable to
the rate of the return electron-transfer reactions. From our data,
it is not possible to determine whether this forward rate is for
fragmentation of the sulfuranyl radical within the primary cage
to give a secondary intermediate, 4, or for diffusive separation
of the components of the primary cage. However, the photo-
CIDNP results on the sensitized photoreductions of diphenyl-
iodonium and triphenylsulfonium salts in CD3CN do permit a
distinction.2 Both Ph2I+ (return electron transfer) and C6H5D
The formation of some 4-cyanobenzylacetamide (10% yield)
in the phenylanthracene-sensitized reaction of the sulfonium salt
stimulated us to think of a roundabout mechanism for wasting
the photochemical energy. The formation of the amide signals
the formation of 4-cyanobenzyl cation, which is captured by
acetonitrile and water to give the observed amide. One might
imagine that the initial electron transfer is actually concerted
to give directly the caged partners 4. Electron transfer from the
ox
4-cyanobenzyl radical (E ) 1.08 V, SCE)21 to the 9-phenyl-
1/2
ox
anthracene cation radical (E ) -1.13 V SCE)22 within the
1/2
cage would give the 4-cyanobenzyl cation, which could then
be captured by phenyl methyl sulfide to regenerate the starting
sulfonium ion. According to this hypothetical mechanism
(Scheme 3), the source of the inefficiency is a “three-step”
process in which phenyl methyl sulfide is the species that
partitions. However intriguing this mechanism might be, it
cannot be a generalized source of the photochemical inefficien-
cies because no 4-cyanobenzylacetamide can be detected
(<0.1%) in the reactions sensitized by EDA or PE. In fact, these
other sensitizers were chosen because the corresponding cation
radicals are less oxidizing (PE, -0.85 V, SCE;16 EDA, -0.66
V, SCE) than the cation radical derived from phenylanthracene.
A related “reassembly” mechanism for energy wasting could
involve initial dissociative electron transfer to give the caged
partners, 4. “Reaction” of these partners with concomitant
electron transfer would give back sulfonium ion 1 and the
sensitizer ground state. The implication of the forward dissocia-
tive electron transfer is that the sulfuranyl radical is not an
(20) Peters, K. S. In AdVances in Electron-Transfer Chemistry; Mariano,
P. S., Ed.; JAI Press: Greenwich, CT, 1994; Vol. 4, pp 27-52.
(21) Sim, B. A.; Milne, P. H.; Griller, D.; Wayner, D. D. M. J. Am.
Chem. Soc. 1990, 112, 6635-6638.
(22) Wilkinson, F.; McGarvey, D. J.; Olea, A. F. J. Am. Chem. Soc.
1993, 115, 12144-12151.

4368 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Wang et al.
energy minimum.23,24 Accordingly, the sulfuranyl radical cannot
be involved in the return process; i.e., the return reaction would
have to be an “associative electron transfer”. Although such a
process is conceptually valid,25 it is unprecedented.
transfer in the electrochemical studies. The most consistent
interpretation of the observed inefficiencies is that the electron
transfer is not concerted and that short-lived 9-S-3 sulfuranyl
radicals are involved as intermediates in the photosensitized
reductive cleavage reactions of sulfonium salts.27 The important
implication of this conclusion is that considerations of the
characteristics of these reductive cleavages must, necessarily,
involve considerations of the structures, stabilities, and relative
rates of formation and fragmentation of the sulfuranyl radical
intermediates.
One might imagine yet another source of inefficiency.
1
Suppose that the sulfonium salt reacts with sens not only by
electron transfer, but also catalyzes some intersystem crossing
of 1sens to 3sens. The energetics of electron transfer from 3sens
to sulfonium salt are unfavorable26 and 3sens would, therefore,
simply decay nonradiatively to the ground state. In this
hypothetical scenario, it is the excited singlet state of the
sensitizer that fractionates. This intersystem crossing was
previously excluded for PA by spectroscopic experiments.2a
Clearly, the best way to settle the question of the 9-S-3
intermediate is to observe it. We attempted to detect the primary
intermediates, 3, for each sensitizer by picosecond absorption
spectroscopy, but failed to observe any new well-defined
absorptions that could be assigned to a sulfuranyl radical. The
“intersystem crossing” mechanism described above requires a
significant yield (Φ ≈ 0.2-0.3) of sensitizer triplets, followed
by slow unimolecular decay of the triplets. This process would
have been easily detected in nano- and picosecond absorption
experiments with perylene since the intersystem crossing yield
in the absence of sulfonium salt is only 1%. In the presence of
sulfonium salt 1-BF4, the absorption of 1PE (710 nm) was
decreased with no concomitant increase in the absorption of
[3PE] at 480 nm.
Acknowledgment. This research was supported by the
National Science Foundation Center for Photoinduced Charge
Transfer, CHE-9120001. We are especially grateful to Professor
Joshua L. Goodman for an extended series of instructive
discussions
Supporting Information Available: Experimental details
for the preparation and characterization of sulfonium salts 1-BF4
and 1-PF6, 2-BF4, and 2-PF6, for the characterization of
photolysis products, and for photochemical measurements,
including a table of quantum yields for the formation of PhSCH3
as a function of [1-BF4] (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
JA982932X
(27) Whether there is a contradiction between the electrochemical and
the photochemical observations remains to be seen. Save´ant has discussed
other cases (e.g., CF3Br, ArCH2Br) where different modes of reduction have
led to different conclusions about the mechanism of the reductive cleav-
age.10,11 Some of the differences were attributed to environmental effects
(solvation, ion pairing) that stabilize an anion radical intermediate and
thereby favor a stepwise path. Those factors are presumably not important
in the case of a neutral sulfuranyl radical intermediate. Alternatively, Save´ant
has argued that the mechanism can change as a function of the driving
force. Just such a crossover was reported for benzylphenylmethylsulfonium
ion; increasing driving force (scan rate) led to a decrease in the cyclic
voltammetric peak width that was attributed to a change in mechanism from
concerted to stepwise.10 In the present case, the argument would be that
the lower driving force in the electrochemical reduction of 4-cyanoben-
zylmethylphenyl sulfonium salt favors the concerted mechanism and the
larger driving force in the photochemical reduction favors the stepwise
mechanism. More recently, Roberts and Save´ant have invoked an avoided
crossing model to rationalize the different mechanistic conclusions derived
from photochemical and electrochemical experiments.28
Conclusion
The experiments and the logic outlined here rule out
alternative sources of inefficiency in the reactions of sensitizer
singlet excited states with 4-cyanobenzylmethylphenyl sulfo-
nium salt, the “best-case” candidate for dissociative electron
(23) Eldin, S.; Jencks, W. P. J. Am. Chem. Soc. 1995, 117, 9415-9418.
(24) Save´ant has a different view. For a recent discussion, see: Andrieux,
C. P.; Save´ant, J.-M.; Tallec, A.; Tardivel, R.; Tardy, C. J. Am. Chem. Soc.
1997, 119, 2420-2429.
(25) Karki, S. B.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould,
I. R.; Zona, T. A. J. Am. Chem. Soc. 1997, 119, 431-432.
(26) The energies16 of 3PA (1.82 eV) and 3PE (1.57 eV), in conjunction
with the redox potentials in Table l, lead to the conclusion that ∆GET is
(28) Robert, M.; Save´ant, J.-M., 1999, submitted for publication. We
are grateful to Professor Save´ant for sharing a preprint with us.
3
3
+0.42 eV for both PA and PE.