Photosensitized oxygenation of benzyl ethyl sulfide

Article abstract of DOI:10.1021/jo9817927

Singlet oxygen adds to benzyl ethyl sulfide (5, total quenching rate ca. 1 x 10⁷ M^-1 s^-1, little dependent on the solvent) to ultimately give benzaldehyde (6) and a small amount of the sulfone (8) in aprotic media (rate of the chemical reaction in benzene 5.5 x 10⁶ M^-1 s^-1) and mainly the sulfoxide (7) in protic media (1.1 x 10⁷ M^-1 s^-1 in methanol). In the presence of small amounts (0.002-0.3 M) of protic additives (alcohols, phenol, carboxylic acids), the sulfoxide becomes the main product in benzene also. Various evidence support the formation of two intermediates in aprotic solvents. The first one is an exciplex or a syn persulfoxide. It undergoes intramolecular hydrogen abstraction to give a ylide and finally benzaldehyde. Such rearrangement is either a concerted or a radicalic process and not a proton transfer (as indicated by the deuterium effect observed with the α-d benzyl sulfide and the occurrence of the process with the p-nitro and p- methoxy derivatives, 5' and 5''). This intermediate is not quenched except under relatively strong acidic conditions. A second intermediate, arising either from the first one or through a parallel path, has the properties usually associated with the persulfoxide (possibly it is the anti rotamer). This species gives some sulfoxide but mainly decays to the unreacted sulfide; it can be trapped intermolecularly, however, both by acids and by diphenyl sulfoxide (in the latter case it gives more of sulfone 8 than of sulfoxide 7). The relative rates of protonation of both the first and the second intermediate (determined in benzene doped with protic additives) correlate with the gas-phase acidity of such additives. As for the reaction in neat alcohols and in benzene doped with acids, a single intermediate intervenes. This is better described as a S-hydroperoxy cation rather than a neutral hydroperoxysulfurane and is trapped by both diphenyl sulfoxide and sulfone at rates close to those measured for the photo-oxidation of diethyl sulfide.

Full text of DOI:10.1021/jo9817927

9946
J . Org. Chem. 1998, 63, 9946-9955
P h otosen sitized Oxygen a tion of Ben zyl Eth yl Su lfid e
S. M. Bonesi,† M. Mella,† N. d’Alessandro,‡ G. G. Aloisi,§ M. Vanossi,† and A. Albini†
Department of Organic Chemistry, University of Pavia, v. Taramelli 10, 27100 Pavia, Italy,
Department of General and Organic Chemistry, University of Torino, v. Giuria 7, 10125 Torino, Italy,
and Department of Chemistry, University of Perugia, v. Elce di Sotto 8, 06100 Perugia, Italy
Received September 3, 1998
Singlet oxygen adds to benzyl ethyl sulfide (5, total quenching rate ca. 1 × 107 M-1 s-1, little
dependent on the solvent) to ultimately give benzaldehyde (6) and a small amount of the sulfone
(8) in aprotic media (rate of the chemical reaction in benzene 5.5 × 106 M-1 s-1) and mainly the
sulfoxide (7) in protic media (1.1 × 107 M-1 s-1 in methanol). In the presence of small amounts
(0.002-0.3 M) of protic additives (alcohols, phenol, carboxylic acids), the sulfoxide becomes the
main product in benzene also. Various evidence support the formation of two intermediates in
aprotic solvents. The first one is an exciplex or a syn persulfoxide. It undergoes intramolecular
hydrogen abstraction to give a ylide and finally benzaldehyde. Such rearrangement is either a
concerted or a radicalic process and not a proton transfer (as indicated by the deuterium effect
observed with the R-d benzyl sulfide and the occurrence of the process with the p-nitro and
p-methoxy derivatives, 5′ and 5′′). This intermediate is not quenched except under relatively strong
acidic conditions. A second intermediate, arising either from the first one or through a parallel
path, has the properties usually associated with the persulfoxide (possibly it is the anti rotamer).
This species gives some sulfoxide but mainly decays to the unreacted sulfide; it can be trapped
intermolecularly, however, both by acids and by diphenyl sulfoxide (in the latter case it gives more
of sulfone 8 than of sulfoxide 7). The relative rates of protonation of both the first and the second
intermediate (determined in benzene doped with protic additives) correlate with the gas-phase
acidity of such additives. As for the reaction in neat alcohols and in benzene doped with acids, a
single intermediate intervenes. This is better described as a S-hydroperoxy cation rather than a
neutral hydroperoxysulfurane and is trapped by both diphenyl sulfoxide and sulfone at rates close
to those measured for the photo-oxidation of diethyl sulfide.
Sch em e 1
Dialkyl- and aryl alkyl- (but not diaryl-) sulfides give
the corresponding sulfoxides by reaction with singlet
oxygen.1,2 This process has attracted considerable interest
in view of its synthetic and biological implications. The
course of the reaction depends on the medium.3-7 In
aprotic solvents, oxygenation is inefficient, i.e., interac-
* Fax: 39-382-507323; E-mail albini@chifis.unipv.it.
† University of Pavia.
‡ University of Torino.
§ University of Perugia.
(1) (a) Clennan, E. L. Sulfur Rep. 1996, 19, 171. (b) Ando, W.;
Takada, T. In Singlet O2; Frimer, A. A., Ed.; CRC: Boca Raton, 1985;
Vol. 3, p 1. (c) Ando, W. Sulfur Rep. 1981, 1, 143. Foote, C. S. In Singlet
Oxygen; Wasserman, H. H., Murray, R. W., Eds.; Academic Press: New
York, 1979; p 139. (d) Gorman, A. A. Adv. Photochem. 1992, 17, 217.
(e) J ensen, F. In Advances in Oxygenated Processes; Baumstark, A.
L., Ed.; J AI Press: Greenwich, CT, 1995; Vol. 4, p 49. (f) Bellus, D.
Adv. Photochem. 1979, 11, 105.
(2) (a) Schenk, G. O.; Krauch, C. H. Angew. Chem. 1962, 74, 510.
(b) Foote, C. S.; Peters, J . W. J . Am. Chem. Soc. 1971, 93, 3795. (c)
Foote, C. S.; Denny, R. W.; Weaver, L.; Chang, Y.; Peters, J . W. Ann.
N.Y. Acad. Sci. 1970, 171, 139. (d) Ando, W.; Kabe, Y.; Miyazaki, H.
Photochem. Photobiol. 1980, 31, 191. (e) Wasserman, H. H.; Strehlow,
W. Tetrahedron Lett. 1970, 795. (f) Murray, R. W.; J indal, S. L.
Photochem. Photobiol. 1972, 16, 147. (g) Casagrande, M.; Gennari, G.;
Cauzzo, G. Gazz. Chim. Ital. 1974, 104, 1251. (h) Monroe, B. M.
Photochem. Photobiol. 1979, 29, 761. (i) Martin, C. D.; Martin, J . C. J .
Am. Chem. Soc. 1977, 99, 3511. (j) Sawaki, Y.; Ogata, Y. J . Am. Chem.
Soc. 1981, 103, 5947. (k) Gu, C. L.; Foote, C. S. J . Am. Chem. Soc.
1982, 104, 49. (l) Clennan, E. L.; Oolman, K. A.; Yang, K.; Wang, D.
X. J . Am. Chem. Soc. 1992, 114, 3021.
tion with singlet oxygen mainly (>95%) leads to physical
quenching, although efficiency increases at low temper-
ature. Two discrete intermediates are involved under this
condition (Scheme 1). The first one has electrophilic
character and is quenched, e.g., by diphenyl sulfoxide,
and the latter one has nucleophilic character and is
quenched by diphenyl sulfide. In methanol, on the
contrary, chemical reaction predominates over physical
quenching and a single intermediate (quenched by both
Ph2S and Ph2SO) is involved. In Scheme 1, the interme-
diates have been identified as the persulfoxide 1, the
(3) Liang, J . J .; Gu, C. L.; Foote, C. S. J . Am. Chem. Soc. 1983, 105,
4717.
(4) J ensen, F.; Foote, C. S. J . Am. Chem. Soc. 1987, 109, 1478.
(5) Akasaka, T.; Kako, M.; Sonobe, H.; Ando, W. J . Am. Chem. Soc.
1988, 110, 494.
(6) Watanabe, Y.; Kuriki, N.; Ishiguro, K.; Sawaki, Y. J . Am. Chem.
Soc. 1991, 113, 2677.
(7) (a) J ensen, F. J . Org. Chem. 1992, 57, 6478. (b) J ensen, F.; Greer,
A.; Clennan, E. L. J . Am. Chem. Soc. 1998, 120, 4439.
10.1021/jo9817927 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/04/1998

Photo-Oxygenation of Benzyl Ethyl Sulfide
J . Org. Chem., Vol. 63, No. 26, 1998 9947
Sch em e 2
Sch em e 3
thiadioxirane 2 for aprotic media, and the adduct 3 for
the reaction in methanol. Several pieces of evidence
support this suggestion. However, other structures can
be (and have been) considered: a tight ion pair (1′) or a
diradical (1′′) rather than zwitterion 1; a sulfoxide-bound
persulfoxide (2′) or, very recently, a ylide (2′′) rather than
heterocycle 2; or a hydrogen-bonded persulfoxide (3′)
rather than the alkoxy adduct 3 (see Scheme 1, lower
part). Furthermore, the possibility that intermediates 1
and 2 are formed through parallel rather than sequential
paths has been envisaged.6 Direct identification has not
been possible except for a matrix isolation study,8 and
evidence for such structures has been sought through the
study of temperature, solvent, and substituent effect on
the singlet oxygen reaction (in the last case using aryl
alkyl sulfides),3-7 as well as through computational
studies.4,6,7 A general consensus on the details of the
mechanism seems not to have been reached as yet,
however.
As mentioned above, chemical reaction accounts for
only a small portion of the reaction of singlet oxygen with
sulfides in aprotic solvents. This leaves little room for
mechanistic studies under such conditions. Recently it
has been found that with sulfenates9 and sulfenamides10
(4, Scheme 2) chemical reaction better competes with
physical quenching, and this has allowed a detailed
study. However, with such substrates heteroatom sub-
stitution directly at the sulfur atom considerably alters
the electronic structure of the substrate and thus makes
extension of the conclusions to sulfides not straight-
forward.
Ta ble 1. Sen sitized P h oto-Oxygen a tion of Ben zyl Eth yl
Su lfid e (5)
product distributiona
kT
kr
b
c
solvent
benzene
acetonitrile
methanol
6
7
8
(M-1 s-1
)
(M-1 s-1
)
89
90
25
57
1
10
7
12
5
1.14 × 107 0.55 × 107
3
63
38
1.22
0.94
0.61
1.2
0.71
benzene,
MeOH 0.37 M
a
b
As determined at low (<20%) substrate conversion. Reaction
rate with singlet oxygen as determined by laser flash photolysis
through the quenching of the 1270 nm emission. c Rate of the
chemical reaction of 5 as measured in competition experiments
in the presence of octaline.
solutions of benzyl ethyl sulfide (5) in benzene in the
presence of tetraphenylporphine (TPP) gave benzalde-
hyde (6) as the major product (Scheme 3, Table 1), in
accord with the report by Corey and Ouanne`s.11 The
aldehyde could be obtained in a ca. 80% yield by distil-
lation at the end of the reaction. Minor products were
sulfoxide 7 and sulfone 8. Similar irradiation using Rose
Bengal (RB) as the sensitizer in acetonitrile also gave
mainly benzaldehyde. On the other hand, RB sensitized
photo-oxygenation in methanol, ethanol, tert-butyl alco-
hol, or trifluoroethanol gave the sulfoxide accompanied
by a lesser amount of benzaldehyde (Table 1). In all cases,
the oxygenation was quenched by DABCO, as previously
observed for dialkyl sulfides.2c
P h oto-Oxid a tion in Meth a n ol. As it appears from
the foregoing, the photo-oxygenation of benzylsulfide
followed two different paths, either S-oxidation or C-S
bond oxidative cleavage. The competition between the
two processes depended on the medium. To obtain
mechanistic information, the oxidation of 5 was examined
in more detail. All experiments for this purpose were
carried out at low conversion (<20%) to minimize second-
ary photoreactions.
As mentioned above, the major product from 5 in
methanol was sulfoxide 7. It has been demonstrated by
Foote3 that sensitized irradiation of diethyl sulfide in the
presence of both diphenyl sulfide and diphenyl sulfoxide
causes co-oxygenation of such substrates, although these
are both unreactive when irradiated in the absence of
Et2S. We thus carried out RB-sensitized oxidation of
various 5/Ph2S mixtures and observed oxidation of both
substrates to the corresponding sulfoxides. The ratio
[7]/[Ph2SO] obtained depended on the concentration of
both sulfides. Foote’s kinetic treatment was used; a plot
of the [7]/[Ph2SO] ratio vs [Ph2S]-1 at various concentra-
tions of 5 is shown in Figure 1. Likewise, Ph2SO was co-
oxidized to diphenyl sulfone, with the product distribu-
There is a class of sulfides which undergo a different
1
chemical reaction with O2, the benzyl sulfides. Twenty
years ago, Corey and Ouanne`s oxygenated some benzyl
sulfides and found that benzaldehyde was produced.11
This was rationalized as evidence in favor of the persul-
foxide 1. Partially related reactions of fluorenyl sulfides
and thiazolidines12 have been investigated, and further,
it has been found that co-oxidation of olefins take place
with such substrates. However, mechanistic data on the
oxidation of benzyl sulfides are limited.13 We thought that
a further investigation would be worthwhile; with these
substrates, the initial interaction with singlet oxygen
must be the same as with dialkyl sulfides, because there
is little difference in the electronic structure. Thus, any
change in the following chemical path may reveal the
chemical nature of the intermediate(s), and the informa-
tion obtained should be relevant for the general mecha-
nism of sulfide photo-oxygenation.
Resu lts
Ma in F ea tu r es of Ben zyl Eth yl Su lfid e P h oto-
Oxygen a tion . Visible light irradiation of 0.01-0.1 M
(8) Akasaka, T.; Yabe, A.; Ando, W. J . Am. Chem. Soc. 1987, 109,
8085.
(9) Clennan, E. L.; Zhang, H. J . Am. Chem. Soc. 1994, 116, 809.
Clennan, E. L.; Chen, M. F. J . Org. Chem. 1995, 60, 6444.
(10) Clennan, E. L.; Zhang, H. J . Am. Chem. Soc. 1995, 117, 4218.
(11) Corey, E. J .; Ouanne`s, C. Tetrahedron Lett. 1976, 4263.
(12) Ando, W.; Nagashima, T.; Saito, K.; Kohmoto, S. J . Chem. Soc.,
Chem. Commun. 1979, 154.
(13) Akasaka, T.; Sakurai, A.; Ando, W. J . Am. Chem. Soc. 1991,
113, 2696.

9948 J . Org. Chem., Vol. 63, No. 26, 1998
Bonesi et al.
F igu r e 1. Ph2S trapping of the PhCH2SEt (5) photo-oxidation
intermediate as a function of both concentrations in MeOH;
slope ) 0.238, 0.56, and 0.87 for 0.02, 0.052, and 0.082 M 5,
respectively.
F igu r e 3. Co-oxidation of Ph2SO in the presence of 5 as a
function of both concentrations in benzene: (() 0.02 M 5, (9)
0.05 M 5, and (2) 0.1 M 5.
F igu r e 2. Ph2SO trapping of the PhCH2SEt (5) photo-
oxidation intermediate as a function of both concentrations
in MeOH; slope ) 0.128, 0.299, and 0.502 for 0.02, 0.052, and
0.082 M 5, respectively.
F igu r e 4. Ph2SO trapping of the PhCH2SEt (5) photo-
oxidation intermediate as a function of both concentrations
in benzene: (() slope ) 0.22, 0.02 M 5; (9) slope ) 0.23, 0.05
M 5; (b) slope ) 0.25, 0.1 M.
tion depending on the starting concentration of both
substrates (see the [7]/[Ph2SO2] vs [Ph2SO]-1 plot in
Figure 2).
and diethyl sulfide (both 0.05 M) in benzene caused an
increase in the oxidation of the latter substrate to Et2-
SO by a factor of 2.3 (with respect to the yield in the
absence of 5) and slowed down the oxidation of the former
with no change in the product distribution.
P h oto-Oxid a tion in Ben zen e. Under this condition,
the major product was benzaldehyde. One of the minor
products, sulfone 8, was present from the start and was
formed at a constant rate throughout the process. On the
contrary, practically no sulfoxide (7) was formed at low
conversion. The amount of this compound became sig-
nificant as long as the reaction proceeded, although it
remained a minor product under this condition (9% at
complete conversion of 5). Addition of Ph2S up to 0.5 M
caused neither co-oxidation of the additive nor a change
in the yield of the products arising from 5. On the other
hand, Ph2SO was co-oxidized when added to 5; in the
presence of this additive, the yield of benzaldehyde
remained constant, some sulfoxide 7 was formed, and the
yield of sulfone 8 grew conspicuously. The extent to which
the co-oxidation occurred did not depend on the concen-
tration of sulfide 5, as shown in Figure 3 (again, this held
for low-conversion experiments). Likewise, the ratio
[8]/[Ph2SO2] was not a function of [5] although it de-
pended on [Ph2SO]. A plot of this ratio vs [Ph2SO]-1 is
shown in Figure 4. Photo-oxygenation of a mixture of 5
P h oto-Oxid a tion in Ben zen e Dop ed w ith P r otic
Ad d itives. As shown above, the change from benzene
to methanol brought about a complete change in the type
of oxidative process occurring with benzyl ethyl sulfide.
Recent experiments by Clennan14 showed that it is
sufficient to add a few percent of methanol or of other
alcohols to make efficient the sluggish photo-oxidation
of diethyl sulfide in neat benzene. We looked at whether
a low percentage of alcohols could affect the oxidation of
5 in terms of efficiency or of product distribution. When
the photo-oxidation was carried out in benzene containing
some methanol, the yield of sulfoxide increased somewhat
with additions in the range of 0.37-1.85 M. tert-Butanol
had little effect up to 1.5 M (although the reaction in neat
t-BuOH gave 7 as the main product). Addition of trifluo-
roethanol was more effective, and the yield of sulfoxide
(14) Clennan, E. L.; Greer, A. J . Org. Chem. 1996, 61, 4793.

Photo-Oxygenation of Benzyl Ethyl Sulfide
J . Org. Chem., Vol. 63, No. 26, 1998 9949
Ta ble 2. TP P -Sen sitized Oxygen a tion of pa r a
Su bstitu ted Ben zyl Eth yl Su lfid esa
product distribution
substituent
6
7
8
kT (M-1 s-1
)
kr (M-1 s-1
)
4-OMe
4-NO2
82
70
8
5
5
1.66 × 107
0.52 × 107
0.17
0.10
a
Data obtained as in Table 1.
Ta ble 3. Deu ter a tion Exp er im en ts
substrate
conditions
PhCDO/PhCHOa
PhCHDSEt
PhCHDSEt
PhCHDSEt
PhCH2SEt
benzene, TPP
acetonitrile
methanol
3.63
2.95
2.70
0
methanol O-d
a
In all cases there was no measurable (gas/mass experiments)
change in the H/D ratio in either unreacted 5, sulfoxide 7, or
sulfone 8.
F igu r e 5. Yield of the photo-oxidation products from 5 (0.05
M) in benzene as a function of the concentration of added
trifluoroethanol: (9) benzaldehyde (6), (() sulfoxide (7).
4-methoxy- and 4-nitrobenzyl ethyl sulfides (5′ and 5′′
in Scheme 3) gave the corresponding benzaldehydes by
TPP-sensitized oxygenation in benzene. Both the rate of
total quenching (kT) and the rate for chemical reaction
(kr) were affected by the substituent, with a somewhat
lower effect in the latter case (Table 2).
Deu ter iu m La belin g. The photo-oxidation was simi-
larly carried out using 5-R-d. The D/H content in the
benzaldehyde (6) formed under this condition was >1,
as determined by GC/MS experiments. The results are
reported in Table 3. On the other hand, the D/H ratio
was unitary (in the limits of the experiment) in sulfone
7 and sulfoxide 8. Furthermore, no deuterium incorpora-
tion was measured in the recovered starting material or
in any of the products when the photooxidation of 5 was
carried out in methanol O-d.
F igu r e 6. Yield of the photo-oxidation products from 5 (0.05
M) in benzene as a function of the concentration of added
chloroacetic acid: (9) benzaldehyde (6), (() sulfoxide (7).
Discu ssion
Comparison with the thoroughly investigated photo-
oxygenation of diethyl sulfide1,3,6,14 will show in which
respect the benzyl sulfide (5) oxygenation is different and
thus where a new mechanistic insight can come from.
All of the observed reactions are quenched by DABCO,
just as with dialkyl sulfides,2c and involve singlet oxygen.
The rate constant of singlet oxygen quenching by 5 (kT
) 1.14 × 107 M-1 s-1 in benzene, see Table 1) is on the
same order as that obtained with Et2S (reported as 1.71,
already increased significantly at 0.02 M concentration;
this became the major product at 0.1 M (Figure 5). We
pursued the examination with other protic additives, viz.
m-nitrophenol, acetic acid, and chloroacetic acid. In these
cases, an even smaller addition was sufficient to obtain
a significant increase in the sulfoxide yield (e.g., with as
low as 0.002 M chloroacetic acid), and furthermore, a drop
in the benzaldehyde yield was observed, although this
effect was much less apparent (see Figure 6 for the case
of CH2ClCO2H).
Ra te of th e P h oto-Oxid a tion . The rate of the reac-
tion of 5 with singlet oxygen (kT) was determined by
quenching of the 1270 nm emission and was found to be
little affected by the nature of the solvent (Table 1). The
rate of chemical reaction of this sulfide (kr) was then
measured in competition experiments with octahydronaph-
thalene (supposed, as usual, to be photo-oxidized at a
solvent-independent rate), following the method used by
Foote.15 The data in Table 2 show that sulfide 5 was
oxidized more efficiently in methanol than in benzene,
although the increase was only by a factor of 2, not >20
as observed with diethyl sulfide.
2, or 3.04 × 107 M-1 s-1
)
10,14,16 but somewhat lower, a fact
that may be related to a slight steric hindering, as it is
known that this factor has a strong effect with more
bulky sulfides.2h This rate depends neither on the polarity
nor on the proticity of the medium (within 20%), as it is
most often the case for singlet oxygen reactions and has
been proved by Clennan for the case of alkyl sulfides.10,14
Electronic effects are significant, as shown by the 50%
increase of kT in the 4-methoxy substituted derivative 5′
(though, as expected, the effect is smaller than with
thioanisole, where a 4-OMe group increases kT by a factor
of 3)2h and the considerable drop of kT in the 4-nitro
derivative 5′′.
Summing up, the initial interaction of singlet oxygen
with 5 does not differ from the case of dialkyl sulfides.
Su bstitu en t Effect. The effect of a para substituent
on the reaction was tested. It turned out that both
(16) (a) Kacher, M. L.; Foote, C. S. Photochem. Photobiol. 1979, 29,
765. (b) Wilkinson, F. In Singlet Oxygen. Reaction with Organic
Compounds and Polymers; Ranby, B., Rabek, J . F., Eds.; Wiley:
London, 1978; p 27.
(15) Higgins, R.; Foote, C. S.; Cheng, H. Adv. Chem. Ser. 1968, 77,
102.

9950 J . Org. Chem., Vol. 63, No. 26, 1998
Bonesi et al.
(2)
Sch em e 4
2kSO [5]
[7]
) 1 +
[Ph2SO2]
kPhO [Ph2SO]
Figures 1 and 2 show that the expected linear dependence
of the ratios [7]/[Ph2SO] and [7]/[Ph2SO2], respectively,
on the reciprocal of the additive is followed. As expected,
the observed correlations depend on the concentration
of 5 and have a common intercept. Plots of the data at
three different concentrations are shown in Figures 1 and
2. These provide the ratios kSO/kPh (5.0) and kSO/kPhO (3.0),
respectively. These values are only slightly larger than
those obtained with diethyl sulfide (respectively, 4.0 and
2.77).3 Thus, the intermediate formed from 5 in protic
solvents has the same structure and the same reactivity
as that formed from other dialkyl sulfides.
However, this substrate is better suited for mechanistic
studies because the chemical reaction remains efficient
in aprotic media and product distribution is a sensitive
function of the medium characteristics (see e.g., Table 1
and Figures 5 and 6), whereas with dialkyl sulfides, a
single product, the sulfoxide, is obtained, efficiently in
alcohols and inefficiently in aprotic solvents. With 5,
cleavage of the C-S bond is dominant in aprotic solvents
and S-oxidation is dominant in protic solvents. It is
sufficient to add a small amount of a protic additive to
make S-oxidation occur also in a benzene solution (e.g.,
as low as 0.002 M with CH2ClCO2H). The most effective
among such additives also depress the yield of benz-
aldehyde, but this requires a larger addition than for
producing the sulfoxide form. Another characteristic is
that, in methanol, 5 co-oxidizes both Ph2S (an electro-
philic acceptor) and Ph2SO (a nucleophilic acceptor), just
as it occurs with Et2S, whereas in benzene, only the latter
additive is co-oxidized (and furthermore appears to
catalyze the oxidation of 5 to the sulfone). Clearly, a
variety of intermediates are involved. In the following,
we show that the previously proposed mechanism can be
conveniently adapted for accommodating the present
data in an economic way, and a new insight into the
nature of the intermediates is obtained.
In ter m ed ia tes in Ap r otic Med ia . Foote demon-
strated that, in contrast to the methanol case, the
intervention of two intermediates is required for ratio-
nalizing the photo-oxygenation of diethyl sulfide in
benzene (Scheme 1).3 These were identified as the nu-
cleophilic persulfoxide 1 and the electrophilic thiadiox-
irane 2. With 5, co-oxidation of Ph2SO but not of Ph2S is
observed. Furthermore, the experiments in benzene
doped with protic additives (Figure 5, Table 1) show that
an intermediate is intercepted and diverted to give the
sulfoxide 7 under low acidity conditions. It is reasonable
that both electrophiles, the proton and Ph2SO, trap the
same intermediate, i.e., the first one in Foote’s scheme.
Thus, the persulfoxide is also an intermediate in the
oxygenation of 5. However, under such conditions,
benzaldehyde is still formed in about the same yield as
in neat benzene, whereas its formation is quenched, with
a further increase in the formation of the sulfoxide, under
more acidic conditions (compare Figure 5, weak acid, with
Figure 6, stronger acid), though not in the presence of
Ph2SO. This demands a further intermediate leading to
benzaldehyde, which either is a precursor of the above
one or is formed through a parallel independent path and
which can be protonated as can the persulfoxide, though
less efficiently.
There are two viable possibilities for this pair of
intermediates. The first one is that one is a sulfide-
oxygen exciplex (10, not an ion pair because dialkyl
sulfides, including 5, are not sufficiently good donors for
reducing singlet oxygen) and the other one persulfoxide
11. The intermediacy of an exciplex has been previously
considered and fits well with the large increase of the
rate of photo-oxygenation for dialkyl sulfides observed
by lowering the temperature (although the total quench-
ing rate of singlet oxygen does not change).la,d,2k,l Exciplex
10 (see Scheme 5) has no chemical path available with
dialkyl sulfides and in that case simply decays to the
components (kd). With 5, however, intramolecular hydro-
gen transfer from the activated benzylic position (kr)
occurs rather efficiently and ultimately gives benzalde-
hyde. Persulfoxide 11 may be formed from it by S-O
bond formation (ki).
Th e Oxid a tion in P r otic Med ia . We begin with the
oxidation in methanol. Under this condition, the photo-
oxidation of 5 closely follows the well-known pattern of
other alkyl sulfides. Both Ph2S and Ph2SO are co-oxidized
under this condition, and the co-oxidation yield depends
on the concentration of 5, showing that both of these
additives compete with benzyl sulfide for the same
intermediate. Thus, the single-intermediate mechanism
demonstrated by Foote for Et2S operates also with 5, as
shown in Scheme 4.3 In this scheme, the quenchable
intermediate is indicated as the protonated persulfoxide,
viz., the S-hydroperoxy cation 9, rather than a methanol
adduct such as 3 (compare Scheme 1). This is because it
is the acidity of the medium, not its nucleophilicity, which
determines the course of the reaction, as will be discussed
later in more detail.
An alternative rationalization for the two intermedi-
ates would be a conformational equilibrium of the per-
sulfoxide. Calculations by J ensen and Foote17 and by
J ensen7a on the singlet oxygen/dimethyl sulfide addition
showed that the first intermediate is the persulfoxide
with CS symmetry and the oxygen-oxygen bond bisecting
If Scheme 4 holds, steady state eqs 1 and 2 can be
derived and are valid for low-conversion experiments,
where secondary oxygenation reactions are unimportant
(compare ref 3).
2kSO [5]
[7]
) 1 +
(1)
[Ph2SO]
kPh [Ph2S]
(17) J ensen, F.; Foote, C. S. J . Am. Chem. Soc. 1988, 110, 2368.

Photo-Oxygenation of Benzyl Ethyl Sulfide
J . Org. Chem., Vol. 63, No. 26, 1998 9951
Sch em e 5
F igu r e 7. Acid quenching of the photo-oxidation of 5 (0.05
M) in benzene as observed on the formation of sulfoxide 7 and
as a function of the additive concentration. For the sake of
clarity, only some of the points are shown: (×) methanol, slope
) 0.66; (O) trifluoroethanol, slope ) 0.117; (2) acetic acid, slope
) 0.0347; (0) m-nitrophenol, slope ) 0.0125; (() chloroacetic
acid, slope ) 0.0111.
Ta ble 4. Rela tive Ra te of P r oton a tion for Oxygen a tion
In ter m ed ia tesa
1st intermed, 10
2nd intermed, 11
methanol
1
trifluoroethanol
acetic acid
m-nitrophenol
chloroacetic acid
1
4.7
27.5
19.0
5.6
19.0
52.8
59.5
the Me-S-Me angle. However, a C1 rotational (by 120°)
isomer of the persulfoxide was also located as a minimum
and was only slightly more energetic (1.2 kcal/mol).
Clennan has suggested that activation by alcohols in-
volves concomitant rotation and complexation of the
persulfoxide.14 Thus, it may be that the syn persulfoxide
(12), better suited for intramolecular hydrogen abstrac-
tion, is formed first and then rearranges to the anti
rotamer 12′, better suited for intermolecular reactions,
e.g., protonation. In that case, formulae 12 and 12′
(bottom of Scheme 5) should be substituted for 10 and
11, respectively, in the scheme, and ki is the rotation rate.
The role of these intermediates in the reactions occurring
in benzene is discussed in the next sections.
P r oton a tion of th e In ter m ed ia tes. We propose that
both the first and the latter intermediates are protonated
in the presence of protic additives (rates kH and k′H,
respectively, see Scheme 5), giving the single intermedi-
ate characteristic of protic media (9, shown also in
Scheme 4). This step is required in order to generate
sulfoxide 7, which is not formed by irradiation in neat
benzene (at least at low conversion, see a further com-
ment below). In our scheme, the yield of sulfoxide 7
depends on the concentration of the protic additives in a
complex way, being the sum of two contributions (see eq
3, where m is the probability that the protonated
intermediate 9 is finally converted into sulfoxide 7).
a
Slopes of the correlation lines in Figures 8 and 7, respectively.
plateau value, [7]pl, is reached, with no significant
reduction in the benzaldehyde yield. Thus, the second
intermediate 11 is essentially completely protonated
before protonation of the first-formed complex 10 occurs
to a significant degree. Therefore, under such conditions,
the yield of 7 is given by the second term of eq 3, and eq
4 holds. Furthermore, with stronger acids (carboxylic
[7]pl kx + k′d + k′H[HA]
kx + k′d
k′H[HA]
)
) 1 +
(4)
[7]
k′H[HA]
acids or the nitrophenol) a sharp increase in the yield of
7 is obtained at a low concentration, where the yield of 5
undergoes only a marginal drop (see Figure 6 for the
example of chloroacetic acid 9). Therefore, eq 4 can also
be used for analyzing the protonation of the second
intermediate by strong acids, provided that the experi-
ments at low concentration are considered.
Figure 7 shows that the linear plots expected from eq
4 are obtained for all the protic additives we used. With
the reasonable assumption that unimolecular reactions
of the intermediate are not affected by these small
additions of protic agents, the slopes of such plots are
proportional to the rates of protonation of persulfoxide
11 by such additives. These values are reported in Table
4, taking methanol as the reference.
mkH[HA]
[7] )
+
kr + ki + kd + kH[HA]
mkik′H[HA]
On the other hand, at a higher acid concentration, the
yield of benzaldehyde decreases and a further increase
in the yield of sulfoxide 7 takes place. This effect is not
observed with methanol (up to 1.8 M), barely detectable
with trifluoroethanol, and clearly apparent with stronger
acids (see e.g., Figure 6 for chloroacetic acid). Apparently,
(3)
(kr + ki + kd + kH[HA])(kx + k′d + k′H[HA])
The rates of protonation for the two intermediates are
largely different. Figure 5 shows that when using weak
acids such as alcohols the yield of 7 increases until a

9952 J . Org. Chem., Vol. 63, No. 26, 1998
Bonesi et al.
F igu r e 8. Acid quenching of the photo-oxidation of 5 (0.05
M) in benzene as observed on the formation of benzaldehyde
(6) and as a function of the additive concentration. For the
sake of clarity, only some of the points are shown: (2) acetic
acid, slope ) 4.7; (0) m-nitrophenol, slope ) 27.5; (() chloro-
acetic acid, slope ) 18.6. Trifluoroethanol (not shown), slope
) 1.
F igu r e 9. Dependence of the relative rate of protonation of
the intermediates of the photo-oxidation of 5 in benzene (from
Figure 8) as a function of the acidity of the additive in the gas
phase: (b), first intermediate, 10; (9), second intermediate,
11.
Sch em e 6
under this condition, the first intermediate 10 is in turn
protonated. The yield of the sulfoxide now depends on
two contributions, but monitoring the drop in the alde-
hyde yield allows evaluation of the protonation of the first
intermediate alone. The ratio between the yield of the
aldehyde in the absence ([6]0) and in the presence of the
acid is given by eq 5. Figure 8 shows the expected linear
kr + ki + kd + kH[HA]
kr + ki + kd
kH[HA]
[6]°
[6]
)
) 1 +
(5)
kr + ki + kd
dependence of such ratio on the concentration of the acid
and allows the evaluation of the relative rate of proto-
nation of the complex 10 (Table 4, relative to trifluoro-
ethanol).
served a correlation of the diethyl sulfide photo-oxigena-
tion rate with the acidity of the additive in a study limited
to alcohols and concluded that a concerted protonation-
nucleophile addition to the persulfoxide took place to give
3 (Scheme 1).14 However, the present study shows that
the correlation with acidity still holds with much stronger
acids than the alcohols, and thus the determining factor
must be the acidity and not the nucleophilicity of the
additive.
The fact that the rate of quenching depends on the
acidity of quencher HA is per se no evidence about the
structure of the resulting species, viz., the single inter-
mediate characteristic of protic media. We indicated it
as cation 9 in our schemes, but it may be a hydrogen-
bonded complex (13, see Scheme 6) or an adduct (14t3
of Scheme 1 in the case of methanol). Recent calculations
failed to locate 3 as a minimum,7b and at any rate, if an
adduct is involved, its chemistry is not influenced by
group A (alkoxy, phenoxy, or carboxy). For the sake of
simplicity, therefore, we referred to cation 9 in the
present discussion. We notice that such species reason-
ably explain the observed oxidation of both sulfides
(electrophilic OH transfer) and sulfoxides (because initial
coulombic attraction between positively charged sulfur
and the sulfoxide oxygen may facilitate ensuing nucleo-
philic OH transfer, see Scheme 6, reactions indicated as
a and b) and is probably a better candidate than 14 (or
3 for Et2S).
If quenching of the intermediates indeed involves
protonation, a correlation of the observed effect with the
strength of the acids is expected. Figure 9 shows the
dependence of log kH(rel) (both series in Table 4) on the
gas-phase acidities.18 This parameter is chosen because
it is thought to better match the results in benzene than
would the pKa values in water or DMSO. In fact, this
fits the data satisfactorily, e.g., m-nitrophenol causes a
larger effect than acetic acid, corresponding to the acidity
order in the gas phase, not in water. Such a dependence
exists for the protonation of both 10 and 11 and is steeper
in the former case. The rate of protonation increases over
almost 2 orders of magnitude from methanol to chloro-
acetic acid. Less acidic alcohols, such as tert-butanol,
cause practically no effect unless present in a massive
amount.
We also studied the increased efficiency of the diethyl
sulfide S-oxidation in benzene doped with the same
additives and found that the protonation rate increases
in an even more conspicuous way (by a factor of 450 from
MeOH to CH2ClCO2H).19 Clennan had previously ob-
(18) Cummings, J . B.; Kebarle, P. Can. J . Chem. 1978, 56, 1. (b)
Batrmess, J . E.; Scott, J . A.; McIver, R. T. J . Am. Chem. Soc. 1979,
101, 6056. (c) Fujio, M.; McIver, R. T.; Taft, R. W. J . Am. Chem. Soc.
1981, 103, 4017. (d) Fujio, M.; McIver, R. T.; Taft, R. W.; Bordwell, F.
G.; Olmstead, W. N. J . Am. Chem. Soc. 1984, 106, 2717.
Tr a p p in g of th e In ter m ed ia tes by Electr op h iles.
The first intermediate (10) is protonated only under
(19) Unpublished data from the authors’ laboratory.

Photo-Oxygenation of Benzyl Ethyl Sulfide
J . Org. Chem., Vol. 63, No. 26, 1998 9953
Sch em e 7
reciprocal plot in Figure 3 shows that the linear behavior
predicted by eq 9 is followed.
akPhO[Ph2SO]
[7] ) [Ph2SO2] )
[7]-1 ) [Ph2SO2]-1
(8)
(9)
kPhO[Ph2SO] + kx + k′d
kx + k′d
1
a
)
+
akPhO[Ph2SO]
Benzyl sulfone 8, on the other hand, is formed through
two paths, both directly in neat benzene and via the Ph2-
SO-catalyzed path (eq 10). The ratio [8]/[Ph2SO2] is
expected to be unaffected by the concentration of 5 and
to show a reciprocal dependence on [Ph2SO], as indicated
in eq 11. This is verified in Figure 4. According to eqs 9
markedly acidic conditions since the competing uni-
molecular reaction (see further below) make its lifetime
too short. Therefore, it does not react significantly with
other electrophiles. The latter one (11), however, is
trapped by diphenyl sulfoxide (Scheme 7). Kinetically,
the observed trapping parallels that observed by Foote
in the case of diethyl sulfide3 in the sense that quenching
by Ph2SO does not depend on the concentration of 5 (see
Figure 3). However, the chemical reaction is different.
In the case of diethyl sulfide, the persulfoxide rearranges
to the thiadioxirane (2 in Scheme 1 or ylide 2′′ according
to a recent computation)7b and this gives diethyl sulfoxide
by reaction with Et2S or Ph2S. Quenching of the persul-
foxide by diphenyl sulfone also gives diethyl sulfoxide
according to eq 6 (R ) Et). On the other hand, in the
kx + (1 - a) kPhO[Ph2SO]
[8] )
(10)
kPhO[Ph2SO] + kx + k′d
kx + (1 - a) kPhO[Ph2SO]
akPhO[Ph2SO]
[8]
)
)
[Ph2SO2]
kx
akPhO[Ph2SO]
1 - a
a
+
(11)
and 11, the intercepts of the plots in Figures 3 and 4
allow evaluation of the partitioning from adduct 15′, and
the value of a = 0.2 results. From the slope of the line in
Figure 4 and by taking a ) 0.2, the ratio kx/kPhO is
calculated as 0.046. This is twice as much as that
determined by Foote for the analogous nucleophilic attack
of Ph2SO on the diethyl sulfide persulfoxide in benzene.3
The present evaluation should not be over-stressed
because it involves approximations from two different
measurements, but it is remarkable that the two evalu-
ations are quite close. Thus, unimolecular reactions and
reaction with Ph2SO competes in the same way in the
case of Et2S+OO- and in the case of the benzyl persul-
foxide (or at least of the anti rotamer 12′ if such an
equilibrium is involved), even if the chemical products
are in part different.
C-S Bon d Clea va ge. C-S bond cleavage to give
benzaldehyde is the main reaction in aprotic media and
is reasonably formulated as involving intramolecular
hydrogen transfer from the activated R position, as shown
in Scheme 5. The intermediacy of a S-hydroperoxysul-
fonium ylide (16 in the present case) has been initially
proposed by Ando for the photo-oxygenation of thiazoli-
dine and benzyl sulfides12 and later shown to be the
active species in the co-oxidation of alkenes observed in
the presence of such sulfides13 and in the oxidation of
thioanisole.21 This undergoes a Pummerer-type rear-
rangement to the R-hydroperoxysulfide 17 (Scheme 5),
isolated in analogous cases,13 and then cleaves to the
aldehyde.
RSEt + Ph2SO + O2 f RSOEt + Ph2SO2 (6)
case of benzyl sulfide 5, no sulfoxide is formed nor is there
quenching by Ph2S, and thus there is no rearrangement
to a thiadioxirane (see a further comment below). More-
over, addition of Ph2SO leads in part to the sulfoxide
according to eq 6 (R ) PhCH2) and in part appears to
catalyze formation of benzyl sulfone 8, according to eq 7.
RSEt + O2 Ph SO8 RSO2Et
(7)
2
The simplest rationalization of this result is that this had
to do with the adduct between persulfoxide and Ph2SO
(see Scheme 7). Ando demonstrated on the basis of
O-tracing experiments that it has the zwitterionic open-
chain structure corresponding to 15 for the oxidation of
dimethyl sulfide in the presence of dimethyl sulfoxide.20
In the present case, however, the steric effect by the
benzyl group may favor a cyclic structure such as 15′
which partitions between two alternative cleavage modes.
Path (a) gives 7 and Ph2SO2, and path (1-a) gives 8 and
Ph2SO (Scheme 7).
Because, as seen above, sulfoxide 7 is not formed in
benzene in the absence of Ph2SO, this compound arises
only via eq 7. The amount obtained is the same as that
of diphenyl sulfone and depends on the concentration of
added Ph2SO (see eqs 8 and 9). In our experiments, the
yield of 7 is somewhat larger than that of Ph2SO2;
presumably because of the formation of a little 7 from
another path (see next section). Therefore, the existence
of a correlation was tested with Ph2SO2, and the double
Experimental support for the intermediacy of ylide 16
is offered by the secondary oxidation we observe. Ando
has shown that such ylides (and not the quite stable
hydroperoxide analogous to 17) are oxidizing species.
Thus, these epoxidize alkenes (although only when
present in a large excess and high concentration, typically
7 M) and oxidize dimethyl sulfide.13 This explains why
sulfoxide 7, not a primary product and not detected at
(20) Akasaka, T.; Ando, W. J . Chem. Soc., Chem. Commun. 1983,
1203.
(21) Ishiguro, K.; Hayashi, M.; Sawaki, Y. J . Am. Chem. Soc. 1996,
118, 7265.

9954 J . Org. Chem., Vol. 63, No. 26, 1998
Bonesi et al.
the beginning of the reaction, is formed at a later stage
of the reaction. This reasonably occurs by oxygen transfer
from 16, which accumulates to a certain extent, to 5.
Likewise, co-oxidation of equimolecular amounts of 5 and
diethyl sulfide leads, as one may expect, to a decreased
reaction of 5 (this is proportional to the part of singlet
oxygen competitively quenched by Et2S and involves no
change in product distribution) but gives an enhanced
yield of Et2SO (2.3 times with respect to irradiation of
Et2S alone). Thus, here Et2S is oxidized by ylide 16.
As for the mechanism by which 16 is formed, very
recent calculations by J ensen et al7b suggest that hydro-
gen transfer to give a ylide is a viable path from the
persulfoxide in the case of dimethyl sulfide. However, the
above kinetic analysis shows that benzaldehyde is formed
from 5 via an intermediate which precedes the persul-
foxide, or at least a second intermediate that has the
chemical characteristics generally attributed to the per-
sulfoxide. Thus, as shown in Scheme 5, the possibilities
are either that the ylide is formed directly from complex
10 or that two persulfoxide conformations are formed
sequentially and only for the first one is hydrogen
transfer significant. Whichever is the exact mechanism,
experimental evidence suggests that the O-H bond is
formed to a considerable degree in the transition state,
as shown by the marked selectivity in the case of R-d 5
(ratio 3.63 for PhCDO in benzene), and that the reaction
is no proton transfer, because the rate of formation of
the benzaldehyde (kr) is not determined by charge
stabilization (it remains the same with the 4-methoxy
derivative of 5, and although it decreases with the 4-nitro
derivatives, the ratio kr/kT increases in that case, and the
effect is small, see Table 1). Thus, a concerted or radical
pathway is indicated. In the first case, the reaction would
proceed via exciplex 10 and follow a concerted path
analogous to the alkene ene addition, and in the latter,
a diradicalic character of the first-formed persulfoxide 12
would be indicated22 (also notice that a little bibenzyl is
formed during the photo-oxidation, again pointing to a
radical course). Calculation on dimethyl sulfide did not
support a concerted path,7b,21 but this may not be relevant
because C-S bond cleavage occurs only with activated
C-H, as with the present benzyl derivatives. Notice also
that the intermediate is somewhat polarized, being
quenched by acids, albeit only under relatively severe
conditions.
fides.24 Recent computational (on dimethyl sulfide)7b and
experimental (on thioanisole)21 evidence discard the
intermediacy of a thiadioxirane and suggest that in
apolar media the sulfone is formed via a hydroxypersul-
fonium ylide, so that formation of the sulfone or C-S
bond cleavage depend on the partitioning of such an
intermediate. As it appears from the previous section,
ylide 16 is an intermediate for benzaldehyde formation
in the present case and arises from the first intermediate,
an exciplex or a diradicalic persulfoxide. One may
consider whether sulfoxide may be formed through the
same path or via the second intermediate. Contrary to
the thioanisole case, the proportion of sulfoxide is low
here and it is difficult to arrive at an unambiguous
conclusion. However, the fact that acids quench the
formation of sulfone 8 while promoting that of sulfoxide
7 and the form of the observed [8]/[Ph2SO2] vs [Ph2SO]
relation (Figure 4) suggest that 8 results from the latter
one (kx in Scheme 5), although it gives no indication of
how such a conversion occurs. Certainly, there is no
evidence for a thiadioxirane in this case, nor for any
intermediate quenched by nucleophiles. As discussed in
a previous section, an enhanced yield of 8 is obtained in
the presence of Ph2SO, apparently via an unprecedented,
but not unreasonable, fragmentation of adduct 15′ (see
Scheme 7). This path does not affect formation of benz-
aldehyde but rather avoids unproductive decay (k′d) of
the second intermediate (persulfoxide 11 or 12′) to the
components.
Over a ll Mech a n ism . The photochemistry of benzyl
sulfide 5 is more complex than that of dialkyl sulfides.
The competition of various paths in aprotic solvents, with
relatively low amounts of some of the products, intro-
duces into some of the present evaluations a certain
degree of uncertainty. However, this varied chemistry
offers a wealth of indications and justifies choosing this
model for better understanding the nature of the inter-
mediates involved in the photo-oxygenation of sulfides.
Many of these results are easily reconciled with the
known scheme, but some revision is appropriate. Since
the electronic structure of 5 is little different from that
of other dialkyl sulfides, most indications (except of
course those based on steric factors) should have a
general bearing for sulfide oxygenation.
The most significant information from the present
results concerns the very first step of chemical reaction
and suggests that an exciplex (10) or a diradicalic syn
persulfoxide (12) is formed before the “classical” persul-
foxide or in parallel to it. Such an intermediate has no
path for chemical reaction with dialkyl sulfides, but an
activated R-hydrogen reveals its presence and makes the
reaction efficient also in aprotic solvents. The chemical
quantum yield for intramolecular hydrogen transfer is
ca. 0.5 in benzene, and thus, if the intermediates are
formed sequentially, kr = ki (see Scheme 5) and if they
are formed in parallel, their yield is about the same. Since
Φ , 0.05 with dialkyl sulfides, no hint can be obtained
for the elusive first intermediate in that case.
F or m a tion of Su lfoxid e. Finally, some notes are in
order with regard to the formation of sulfoxide 8.
Whether the intermediate leading to the sulfone is the
same as that intercepted by sulfides has been the subject
of some controversy.1,6 In particular, the work by Sawaki
suggests that the thiadioxirane is formed competitively
with, not sequentially to, the persulfoxide and then opens
up to the sulfone.6 The unimolecular mechanism for the
formation of the sulfone was supported by 18O2 experi-
ments. Sulfones are formed in high yield from Et2S at
low temperature in an apolar medium23 and from disul-
(22) (a) The diradialic character of 1,3-dipoles has been demon-
strated in a number of cases; see, e.g., ref 22b. Notice also that the
overall deuterium effect on a multistep reaction such as that leading
to benzaldehyde does not necessarily represent that involved in the
first step, as assumed in the present discussion for the sake of
The second intermediate has all the characteristics
attributed to the persulfoxide in the previous work.
Indeed, such a classical persulfoxide (11 or, specifically,
the anti rotamer 12′) mainly decays to unreacted sulfide
simplicity. Calculations by J ensen et al. show that
a significant
deuterium effect may also be involved in the second step, the shift of
the OOH group; see ref 7b. (b) Hiberty, P. C. Isr. J . Chem. 1983, 23,
10.
(24) Clennan, E. L.; Zhang, H. J . Org. Chem. 1994, 7952. Clennan,
E. L.; Wang, D.; Zhang, H.; Clifton, C. H. Tetrahedron Lett. 1994, 35,
4723.
(23) Foote, C. S.; Peters, J . W. J . Am. Chem. Soc. 1971, 93, 3795.

Photo-Oxygenation of Benzyl Ethyl Sulfide
J . Org. Chem., Vol. 63, No. 26, 1998 9955
purity was checked by NMR and mass spectroscopy. The
sulfoxide (7)29 and the sulfone (8),30 as well as the correspond-
ing methoxy- and nitro-substituted derivatives, were prepared
by literature methods. Diphenyl sulfoxide and diphenyl sulfone
(commercial samples from Aldrich) were purified by chroma-
tography and recrystallization of the central fraction. Rose
Bengal, tetraphenylporphine, and octahydronaphthalene were
commercial products (Aldrich).
P h otoch em ica l Rea ction s. Irradiations were carried out
in rubber-stoppered Pyrex tubes while passing a stream of
oxygen saturated with the solvent. Four phosphor-coated 15
W lamps (Applied Photophysics) emitting from 350 to 700 nm
were used as the light source. The course of the reaction was
followed by VPC and/or HPLC and all reactions were discon-
tinued at <20% conversion. The reported data are the average
of at least three measurements within 5%. In the experiments
with octaline, the solution was treated with triphenylphos-
phine before analysis to reduce the hydroperoxide to the allylic
alcohol and then analyzed as above.
independent of the presence of a benzylic hydrogen, as
one would expect because here only polar (and not radical
or concerted) chemical paths are available. This inter-
mediate can be made to react efficiently by protonation
or trapping by Ph2SO according to the well-known
general pattern1,3,10,25 (e.g., the ratio kx/kPhO is within a
factor of 2 with Et2S and with 5). Indeed, both intermedi-
ates are protonated, though with different efficiency, and
certainly not at a diffusion-controlled rate, consistent
with the idea that charge localization is not too high and
these transients are moderate bases (and nucleophiles).
As an example, with the strongest acid we used, chloro-
acetic acid, the second intermediate is wholly protonated
at ca. 0.02 M, and three-fourths of the first one is
protonated at 0.1 M; furthermore, only the latter one is
trapped by Ph2SO.
In turn, the protonated intermediate behaves exactly
as the corresponding transient from Et2S, giving the
sulfoxide and co-oxidizing both Ph2S and Ph2SO. How-
ever, we noticed that this chemistry occurs also in the
presence of phenol and carboxylic acids, not only alcohols
as previously known,14 and its occurrence is proportional
to the acidity and not the nucleophilicity of the additive.
The effect of tiny additions of acids may be useful for
directing the reaction and further suggests that this
intermediate is better represented as a cation (9) than
as a neutral adduct. Further subtle, probably steric,
effects influence the evolution of some intermediates and,
e.g., make the formation of the sulfone in aprotic solvents
more important for 5 than from Et2S, both directly and
in the presence of Ph2SO. Formation of the sulfone seems
not to involve the ylide, in contrast to what was recently
found for dialkyl sulfides7b and thioanisole.21,26
An a lytica l Meth od s. Gas chromatographic determination
were carried out by means of a Hewlett Packard mode 5890
with a Innowax column (30 m × 0.25 mm, 0.5 mm film
thickness) and He as the carrier gas. The injector pressure
was constant (12 psi). The temperature program was 50-140
°C (5 °C/min), 140-250 °C/min, 7 min final time). For mass
determinations, this was equipped with a model 5971 mass
detector. Under this condition, PhCHO and PhCDO were
completely separated.
P h otop h ysica l Mea su r em en ts. Rate constants for the
quenching of singlet oxygen were measured by laser pulse
spectroscopy (Spectra Physics Nd:YAG laser) by measuring the
emission lifetime at 1270 nm after oxygen sensitization by
phenalenone.31
Ack n ow led gm en t . S.M.B. (permanent address:
Dept. of Organic Chemistry, University of Buenos Aires,
Ciudad Universitaria, CP1428 Buenos Aires, Agentina)
thanks FOMEC for a fellowship. M.V. thanks Prochim-
ica for a fellowship. Dr. P. P. Rossi (Prochimica) is
thanked for discussing this project. Financial support
by Consiglio Nazionale delle Ricerche, Rome, is grate-
fully acknowledged.
Exp er im en ta l Section
Ma ter ia ls. Spectroscopic grade solvents were used. Benzene
and acetonitrile were distilled from calcium hydride before use.
Benzyl ethyl sulfide (5) and the substituted derivatives 5′ and
5′′ were prepared by the reported method.27 R-d 5 was
analogously prepared from R-d benzyl chloride,28 and the
J O9817927
(27) Modena, G. Gazz. Chim. Ital. 1960, 89, 834.
(28) Meyerson, S.; Rylander, P. N.; Eliel, E. L.; McCollum, J . D. J .
Am. Chem. Soc. 1959, 81, 2606.
(29) Cerniani, A.; Modena, G.; Todesco, P. E. Gazz. Chim. Ital. 1960,
90, 3.
(30) Buechi, J .; Prost, M.; Eichenberger, H.; Lieberherr, R. Helv.
Chim. Acta 1952, 35, 1527.
(31) Schmidt, R. J . Photochem. Photobiol., A 1994, 79, 11.
(25) Adam, W.; Golsch, D.; Goerth, F. C. Chem. Eur. J . 1996, 2, 255.
(26) The mechanism of formation of sulfones has not been fully
clarified even with dialkyl sulfides. As an example, at low temperature,
the sulfoxide-sulfone ratio depends in a complex way on the methanol
content in methanol-apolar solvent mixtures; see ref 26b. (b). Clennan,
E. L.; Yang, K. Tetrahedron Lett. 1993, 34, 1697.