C-S bond cleavage in the sensitized photooxygenation of tert-alkyl phenyl sulfides. The role of superoxide anion

Article abstract of DOI:10.1016/j.tet.2005.09.154

The N-methylquinolinium tetrafluoroborate (NMQ⁺)-photosensitized oxidation of tert-alkyl phenyl sulfides 1a-c (1a, tert-alkyl=tert-butyl; 1b, tert-alkyl=2-phenyl-2-propyl; 1c, tert-alkyl=1,1-diphenylethyl) and benzyl phenyl sulfide (2) were investigated in CH₃CN by nanosecond laser flash photolysis (LFP) and steady-state irradiation either under nitrogen or in the presence of O₂. By laser irradiation, the formation of sulfide radical cations 1a^{+{radical dot}}-c^{+{radical dot}} in the monomeric form (λ_max=520 nm) and of 2^{+{radical dot}} in both the monomeric (λ_max=520 nm) and dimeric form (λ_max=780 nm) were observed within the laser pulse. In both cases, the radical cations decayed by second-order kinetics without any apparent formation of transients attributable to C-S bond rupture. In line with these results, very small amounts of photoproducts were obtained under nitrogen thus suggesting that the sulfide radical cations mainly undergo a back electron transfer process with the reduced N-methylquinolinium (NMQ^{{radical dot}}). A different situation was found in the presence of O₂ since steady-state photolysis produced substantial amounts of C-S bond cleavage products (alcohols, alkenes, and ketones from 1a-c and benzaldehyde from 2), in contrast with LFP experiments. Formation of products was, however, significantly reduced in the presence of benzoquinone, a trap for O₂^{-{radical dot}} generated by NMQ^{{radical dot}} and O₂. For the tert-alkyl phenyl sulfides, 1a-c, these results have been interpreted by suggesting that C-S bond cleavage products in the presence of oxygen mostly derive from the decomposition of a thiadioxirane 6 formed by the reaction of the sulfide radical cation with O₂^{-{radical dot}}. In this cleavage a sulfinate and a carbocation formed. The former is oxidized to sulfonate, whereas the carbocation can react with adventitious water to form the alcohol (and the alkene therefrom) and with O₂^{-{radical dot}} to produce the ketone. For 2 (a sulfide with α-CH bonds) probably a different mechanism holds, benzaldehyde coming from the α-phenylthio carbon radical formed from deprotonation by O₂^{-{radical dot}} of 2^{n+{radical dot}}.

Full text of DOI:10.1016/j.tet.2005.09.154

Tetrahedron 62 (2006) 6566–6573
C–S bond cleavage in the sensitized photooxygenation of tert-alkyl
phenyl sulfides. The role of superoxide anion
b,
b
a,
Enrico Baciocchi,a, Tiziana Del Giacco, Paolo Giombolini and Osvaldo Lanzalunga
*
*
*
aDipartimento di Chimica, Universitaꢀ degli Studi di Roma ‘La Sapienza’ and Istituto CNR di Metodologie Chimiche (IMC-CNR),
Sezione Meccanismi di Reazione, P.le A. Moro, 5 I-00185 Rome, Italy
bDipartimento di Chimica and Centro di Eccellenza Materiali Innovativi Nanostrutturati (CEMIN), Universitaꢀ di Perugia,
Via Elce di Sotto 8, 06123 Perugia, Italy
Received 2 September 2005; accepted 26 September 2005
Available online 18 May 2006
Abstract—The N-methylquinolinium tetrafluoroborate (NMQ+)-photosensitized oxidation of tert-alkyl phenyl sulfides 1a–c (1a, tert-
alkyl¼tert-butyl; 1b, tert-alkyl¼2-phenyl-2-propyl; 1c, tert-alkyl¼1,1-diphenylethyl) and benzyl phenyl sulfide (2) were investigated in
CH3CN by nanosecond laser flash photolysis (LFP) and steady-state irradiation either under nitrogen or in the presence of O2. By laser irra-
ꢀ
ꢀ
ꢀ
diation, the formation of sulfide radical cations 1a+ –c+ in the monomeric form (lmax¼520 nm) and of 2+ in both the monomeric
(lmax¼520 nm) and dimeric form (lmax¼780 nm) were observed within the laser pulse. In both cases, the radical cations decayed by second-
order kinetics without any apparent formation of transients attributable to C–S bond rupture. In line with these results, very small amounts of
photoproducts were obtained under nitrogen thus suggesting that the sulfide radical cations mainly undergo a back electron transfer process
with the reduced N-methylquinolinium (NMQꢀ). A different situation was found in the presence of O2 since steady-state photolysis produced
substantial amounts of C–S bond cleavage products (alcohols, alkenes, and ketones from 1a–c and benzaldehyde from 2), in contrast with LFP
experiments. Formation of products was, however, significantly reduced in the presence of benzoquinone, a trap for Oꢁ2 ꢀ generated by NMQꢀ
and O2. For the tert-alkyl phenyl sulfides, 1a–c, these results have been interpreted by suggesting that C–S bond cleavage products in the
ꢀ
presence of oxygen mostly derive from the decomposition of a thiadioxirane 6 formed by the reaction of the sulfide radical cation with Oꢁ2 .
In this cleavage a sulfinate and a carbocation formed. The former is oxidized to sulfonate, whereas the carbocation can react with adventitious
water to form the alcohol (and the alkene therefrom) and with Oꢁ2 ꢀ to produce the ketone. For 2 (a sulfide with a-CH bonds) probably a different
ꢀ
ꢀ
mechanism holds, benzaldehyde coming from the a-phenylthio carbon radical formed from deprotonation by Oꢁ2 of 2n+
.
Ó 2006 Elsevier Ltd. All rights reserved.
hν
1. Introduction
S
S*
(a) Type I Mechanism
The sensitized photooxygenation of organic sulfides is re-
ceiving continuous attention for the important biological
implications as well as the possible practical applications
in organic synthesis.1–14 Two different mechanisms have
been suggested for this process.15,16 The Type I mechanism
involves the intervention of 1O2 generated by the reaction of
the excited sensitizer with 3O2 (henceforth simply indicated
as O2), whereas in the Type II mechanism the excited sensi-
tizer abstracts an electron from the substrate forming a radi-
1
3
O
+
S
O
+
+
S*
2
2
2
O
O
S
O
S
1
R
R
O
R S
2
R
R
persulfoxide
(b) Type II Mechanism
ꢀ
cal cation that then reacts with Oꢁ2 formed by O2 and the
R S
2
+
S
R S
+
S*
2
reduced form of the sensitizer. Recent work has presented
evidence indicating that the two mechanisms involve differ-
ent intermediates, namely a persulfoxide in the Type I mech-
anism and a thiadioxirane in the Type II mechanism
(Scheme 1).15,16
S
+
O
S
+
O
R
2
2
O
S
O
O
S
R
O
+
R S
2
R
R
2
thiadioxirane
Scheme 1.
In both mechanisms, sulfoxides are generally the main
products. However, variable amounts of C–S bond cleavage
products can also be obtained, depending on the substrate
* Corresponding authors. Tel.: +39 0649913711; fax: +39 06490421;
e-mail: enrico.baciocchi@uniroma1.it
0040–4020/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2005.09.154

E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
6567
structure. The formation of these products has been dis-
cussed in some detail for the Type I mechanism, where
they are suggested to be derived by additional pathways
available to the persulfoxide intermediate.10,11 In the elec-
tron transfer mechanism the fragmentation products are
attributed to possible reactions of the sulfide radical cation
involving the cleavage of a C–S or an a-C–H bond, in com-
from the alkyl moiety. Benzenesulfonate is the main product
derived from the sulfide moiety even though a small amount
of diphenyl disulfide was detected. No quantum yields were
determined for the sulfur containing products. Alcohols,
alkenes, and ketones were the products formed from 1b
and 1c as described in Scheme 3, where only the products
derived from the alkyl moiety are indicated. The correspond-
ing alcohol and ketone were also formed with 1a. In this case,
N-tert-butylacetamide was also produced, whereas the
alkene was not observed.
ꢀ
petition with the reaction with Oꢁ2 leading to sulfoxide
(Scheme 2).13,17
a
+
R S
+ O
2
R SO
2
2
R
CH
3
+
OH
NMQ+
1
R
R
R
1
1
b
+
R
1
S
C
R
CH
O
3
hν
R
2
2
R
2
2
+
+
RS
+
R
or R(-H)SR + H
3b-3c
4a-4c
5a-5c
1a R ,R = CH
3
1
2
Scheme 2.
1b
R = C H , R = CH
1 6 5 2 3
1c R ,R = C H
1
2
6
5
In view of the importance of sulfides photooxygenation, it is
certainly of interest to get more information on this problem.
Thus, we have carried out a detailed investigation of the
photochemical oxygenation of tert-alkyl phenyl sulfides
1a–c sensitized by N-methylquinolinium tetrafluoroborate
(NMQ+) in CH3CN. These reactions, that exclusively lead
to fragmentation products, have been studied by nanosecond
laser flash photolysis (LFP) and steady-state photolysis and
the transient species produced, their decay pathways, and
the final photoproducts have been identified. The results of
this study, reported herewith, have led to the important con-
clusion that the main pathway for the formation of fragmen-
tation products is not the direct C–S bond cleavage in the
radical cation, but the reaction of the sulfide radical cation
with the superoxide anion generated in the reaction of
NMQꢀ with oxygen. For comparison purposes, the photolysis
of benzyl phenyl sulfide (2) has also been investigated.
Fragmentation products are formed in this case too, but the
mechanism is probably different than that proposed for the
tert-alkyl phenyl sulfides.
Scheme 3.
A further notation, however, was that with 1b and 1c, the
alcohol is progressively converted into alkene as the reaction
proceeds, as shown, for example, from the data in Table 1 for
the photolysis of 1b.
Table 1. Quantum yields of the photoproducts formed in the oxidation of
2-phenyl-2-propyl phenyl sulfide (1b) photosensitized by NMQ+ in air-
saturated CH3CNa
T (min)
F(3b)
F(4b)
F(5b)
F(3b)+F(4b)
60
90
120
0.018
0.13
0.16
0.17
0.060
0.057
0.10
0.086
0.096
0.188
0.190
0.217
[sulfide]¼0.01 M. [NMQ+]¼3.4ꢂ10ꢁ3 M.
a
Accordingly, it can be seen that whereas the overall quantum
yield is substantially independent of irradiation time, those
of the alkene 3b and the alcohol 4b increase and decrease,
respectively, by increasing the irradiation time. However,
the sum F(3b)+F(4b) remains almost constant. Most likely,
alcohol is the primary product and there is a substantial and
increasing conversion of the alcohol into the alkene, proba-
bly promoted by H+ formed during the photolysis. Thus, in
Table 2, where the results for all tert-alkyl phenyl sulfides
are depicted, we report the F value for the ketone and the
overall quantum yield, F(3)+F(4), of alkene and alcohol
for 1b and 1c. In the same Table the results of the steady-
state photolysis of benzyl phenyl sulfide that leads to benzyl
phenyl sulfoxide and a fragmentation product, benzalde-
R
C
R
1
S
CH
2
S
CH
3
2
1a R ,R = CH
3
2
1
2
1b
R = C H , R = CH
1
6 5 2 3
1c R ,R = C H
1
2
6
5
2. Results
2.1. Steady-state photolysis
ꢀ
hyde, are also reported. The possible role of Oꢁ2 , that forms
The steady-state photooxidations sensitized by NMQ+ were
carried out in air-saturated CH3CN or for comparison pur-
pose under N2. No reaction was observed in the absence of
the sensitizer. The reaction products were identified by
in the reaction with NMQꢀ (Scheme 1b, S¼NMQ+), was
investigated by performing the photolysis in the presence
ꢀ
of benzoquinone, a well known trap for Oꢁ2 .18 A very sig-
nificant drop in the product quantum yields was observed,
particularly with 1a and 1b (Table 2).
1
GC–MS and H NMR (comparison with authentic speci-
mens, see Section 5).
When the above reactions were carried out in the absence of
oxygen (flushing the solutions with N2 before irradiation) no
products were observed with the NMQ+/1a and NMQ+/2
systems and a substantial decrease (about one order of mag-
nitude) in quantum yields was observed with the sulfides 1b
and 1c. Under these conditions, diphenyl disulfide was also
Photolysis in the presence of O2. Under air, quite efficient
photochemical processes were observed with 1a–c.
Fragmentation products were obtained exclusively in the
photoirradiation of tert-alkyl phenyl sulfides 1a–c with quan-
tum yields (F) between 0.3 and 0.5 for the products coming

6568
E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
Table 2. Quantum yields of the photoproducts formed in the oxidation of
tert-alkyl phenyl sulfides (1a–c) and benzyl phenyl sulfide (2) photosensi-
tized by NMQ+ in CH3CNa
observed. Two absorption bands were detected just after
the laser pulse in the 400 and 520 nm regions (time-resolved
absorption spectra for the NMQ+/toluene/1a system are
reported in Fig. 1a, time-resolved absorption spectra for
the NMQ+/toluene/1b and NMQ+/toluene/1c systems are
reported in Figs. S2 and S3 in the Supplementary data). These
bands were assigned to NMQꢀ (lmax¼400 and 550 nm)23 and
Substrate
Quantum yieldsb
F(3)+F(4)
FTOT
F(5)
1a
Air
Aird
N2
0.36
—
—
0.085c
—
—
0.21
—
—
e
e
e
e
e
e
the sulfide radical cations in the monomeric form (lmax
¼
520 nm).24
1b
1c
Air
Aird
N2
0.29
0.015
0.038
0.19
0.007
0.038
0.10
0.008
—
a
Air
Aird
N2
0.44
0.13
0.058
0.31
0.087
0.058
0.13
0.043
—
0,08
e
F(PhCH2SOPh)
0.03
F(PhCHO)
0.11
2
Air
Aird
N2
0.14
—
—
e
e
e
—
—
—
—
25 µs
0,04
e
e
e
[sulfide]¼1.0ꢂ10ꢁ2 M, [NMQ+]¼1.0ꢂ10ꢁ2 M under nitrogen and
[NMQ+]¼3.4ꢂ10ꢁ3 M in air-saturated condition. Quantum yields are
determined after 1 h irradiation.
a
∆A
b
Quantum yields of all the products from the alkyl moiety. The error is
ꢄ10%.
0,00
c
d
e
b
N-tert-butylacetamide (F¼0.065) is also formed.
In the presence of 1.2ꢂ10ꢁ3 M p-benzoquinone.
Under the detection limit (F<0.001).
0,08
formed, whereas no formation of the ketone 5 was observed.
These results are also displayed in Table 2.
0,04
0,00
25 µs
Finally, since excited NMQ+ can also react with O2 to form
1
1O2,15 we tested the reactivity of our sulfides toward O2.
Thus, the photolysis of 1b was carried out in the presence
of Rose Bengal in O2-saturated CH3CN.19 No detectable
amounts of photoproducts were obtained after 30 min of
irradiation.
400
600
800
λ (nm)
Figure 1. Time-resolved absorption spectra of the NMQ+ (3.4ꢂ10ꢁ3 M)/
toluene (1 M)/(CH3)3CSPh (1.0ꢂ10ꢁ2 M) system in CH3CN: (a) N2-
saturated, recorded 0.2 (6), 4.8 (:), 10 (B) and 16 (C) ms after the laser
pulse; inset: decay kinetics recorded at 520 nm. lexc¼355 nm. (b) O2-
saturated, recorded 0.08 (6), 2.0 (:), 4.0 (B), 6.1 (C) ms after the laser
pulse; inset: decay kinetics recorded at 520 nm. lexc¼355 nm.
2.2. Fluorescence quenching
Fluorescence quenching experiments were carried out in air-
saturated CH3CN to determine the efficiency by which sub-
strates 1a–c and 2 quench the lowest excited singlet state of
NMQ+. The fluorescence intensity of NMQ+ was recorded
by steady-state experiments in the absence (Iꢃ) and in the
presence (I) of increasing concentrations of the substrates
1a–c and 2. The second-order rate constants for NMQ+ fluo-
rescence quenching (kq) were calculated from the slopes
of the linear Stern–Volmer plots (Iꢃ/I vs [1a–c, 2]) divided
Laser photolysis of the NMQ+/toluene/2 system under N2
showed an additional broad absorption band at 780 nm that
was assigned to the dimer sulfide radical cation (Fig. 2a).25
Very reasonably, the dimer was not observed with 1a–c
due to the steric hindrance of the tert-alkyl group.26
1
by the lifetime of NMQ+* (20 ns)20 (Fig. S1 and Table S1
in the Supplementary data). The Stern–Volmer plots show
that all the substrates quench the emission of NMQ+ with
rate constants, which are close to the diffusion limit (1.5–
1.8ꢂ1010 Mꢁ1 sꢁ1).21
The time-evolution of the absorption spectra shows that the
decay of the two signals recorded at ca. 400 and 520 nm is
not accompanied by the buildup of any transient, which
can be assigned to products of C–S bond cleavage, i.e., the
phenylthiyl radical PhSꢀ (lmax¼450–490 nm)27 and, in the
ꢀ
case of 1c+ , the 1,1-diphenylethyl cation (lmax¼420 nm).28
2.3. Laser flash photolysis studies
Moreover, in the LFP experiment with the NMQ+/toluene/2
system, analysis of the spectral evolution did not show the
growth of the absorption of PhSCHꢀPh (lmax¼350 nm),21
The laser photolysis experiments (lexc¼355 nm) were car-
ried out in CH3CN under N2- and O2-saturated conditions
in the presence of 1 M toluene, which was used as cosensi-
tizer to reduce the efficiency of the back electron transfer
process and, consequently, to increase the concentration of
the transient formed within the laser pulse.22 By laser photo-
lysis of the NMQ+/toluene/1a–c systems in N2-saturated
solutions, similar time-resolved absorption spectra were
ꢀ
i.e., the product of benzylic C–H deprotonation of 2+ . In
O2-saturated solutions, an identical situation was observed,
apart from the fact that the time-resolved absorption spectra
are modified by the fast decay of NMQꢀ, that is efficiently
quenched by molecular oxygen;15 thus, the shoulder at ca.
400 nm disappears and the only recorded transient concerns

E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
6569
3. Discussion
3.1. tert-Alkyl phenyl sulfides
0,12
0,08
0,04
a
The observation that sulfides 1a–c quench the fluorescence
emission of NMQ+ with rates close to the diffusion limit in-
dicates that NMQ+ acts as an electron transfer sensitizer. The
high efficiency is certainly related to the reduction potential
of 1NMQ+* (2.7 V vs SCE)22 that is much higher than those
25 µs
∆A
ꢀ
ꢀ
of the tert-alkyl phenyl sulfide radical cations 1a+ –c+ (ca.
1.6 V vs SCE).17 Any possible contribution of O2 to the
photooxidation has been shown to be negligible.
1
0,00
0,08
0,04
0,00
b
The ET mechanism from the sulfides to 1NMQ+* is also sup-
ported by the observation, in the laser flash photolysis exper-
iments of the NMQ+/toluene/1a–c systems in N2-saturated
solutions, of absorption bands that have been assigned to
NMQꢀ (lmax¼400 and 550 nm) and the sulfide radical cat-
ions (lmax¼520 nm). Under oxygen, the only recorded tran-
sient is the sulfide radical cation (lmax¼520 nm) since NMQꢀ
is converted to NMQ+.
25 µs
400
600
800
λ (nm)
However, both under oxygen and nitrogen the decay of the
radical cation is not accompanied by the buildup of any tran-
sient. Particularly, the transients expected for a direct C–S
bond cleavage in the radical cation (phenylthiyl radical in
Figure 2. Time-resolved absorption spectra of the NMQ+ (3.4ꢂ10ꢁ3 M)/
toluene (1 M)/PhCH2SPh (1.0ꢂ10ꢁ2 M) system in CH3CN: (a) N2-satu-
rated, recorded 0.08 (6), 0.32 (:), 1.0 (B) and 5.8 (C) ms after the laser
pulse; (b) O2-saturated, recorded 0.08 (6), 0.94 (:), 3.0 (B), 6.0 (C) ms
after the laser pulse. Inset: decay kinetics recorded at 530 nm. lexc¼355 nm.
ꢀ
all cases and the tertiary carbocation from 1c+ ) were not
observed. Since second-order kinetics were followed and
the rates are faster under oxygen, the decays are reasonably
attributed to back electron transfer, under nitrogen, and to
the sulfide radical cation (time-resolved absorption spectra
are reported in Fig. 1b and Fig. 2b for the NMQ+/toluene/
1a and NMQ+/toluene/2 systems and in Figures S2 and S3
in the Supplementary data for the NMQ+/toluene/1b and
NMQ+/toluene/1c systems). The decay rate of radical cat-
ꢀ
a reaction with Oꢁ2 , under oxygen.
The results of steady-state photoirradiations under nitrogen
(formation of small amounts of photoproducts with sulfides
1b and 1c, and no reaction at all with 1a) were in substantial
agreement with those of the laser photolysis study. The low
efficiency observed is fully consistent with a process domi-
nated by back electron transfer in the geminate radical/
radical cation pair (Scheme 4, path b).
ꢀ
ꢀ
ions 1a–c+ and 2+ was determined by following the absorp-
tion decay at 520 nm. Under nitrogen and oxygen the decay
kinetics followed second-order laws, (see insets in Figs. 1
ꢀ
ꢀ
and 2 for the decay of 1a+ and 2+ , respectively) and the de-
cay rate constants (k2) divided by the extinction coefficient
(3), for all the radical cations, measured at 25 ꢃC are reported
in Table 3.
The structure of photoproducts formed (alkenes 3b and 3c
and alcohols 4b and 4c, from 1b and 1c, respectively) sug-
ꢀ
gests that at least for 1b+ and 1c+ꢀ a pathway involving the
It can be observed that the decay rate constants are not
significantly influenced by the structure of the sulfide and
that the values under oxygen are higher than those observed
under nitrogen.
cleavage of the C–S bond, leading to the tertiary benzylic
cations and the phenylthiyl radical (Scheme 4, path c) can
also operate, albeit as a minor route (not revealable in LFP
experiments). The benzylic cations can react with adventi-
tious water to form the alcohols 4 or lose a proton to produce
the alkenes 3 (paths d and e in Scheme 4). However, alkenes
3 seem to be formed primarily by acid induced dehydration
of the alcohol (Scheme 4, path f, see Section 2). In both
cases, dimerization of the phenylthiyl radical leads to
Time-resolved spectra recorded by laser flash photolysis of
the NMQ+/1b system in the presence of BQ (9.7ꢂ10ꢁ4 M)
in O2-saturated CH3CN (see Fig. S4 in the Supplementary
ꢀ
data) showed the characteristic absorption band of BQꢁ
(lmax¼520 nm)29 thus confirming the electron transfer pro-
cess from Oꢁ2 ꢀ to BQ.
ꢀ
diphenyl disulfide. With 1a+ , C–S bond cleavage is probably
ꢀ
slower (a less stable carbocation is formed) than with 1b+
ꢀ
Table 3. Decay rate constants of 1a–c+ and 2+ꢀ formed by laser photolysis of
and 1c+ꢀ and cannot compete at all with back electron trans-
fer. Thus, no products are observed in the steady-state
photolysis.
the NMQ+ (3.4ꢂ10ꢁ3 M)/toluene (1 M)/1a–c, 2 (1.0ꢂ10ꢁ2 M) systems in
N2- and O2-saturated CH3CN (lecc¼355 nm)
Sulfide
k2/3 (106 sꢁ1 cm)
N2
O2
However, as shown by the data in Table 2, the efficiency of
photolysis of the sulfides 1a–c significantly increases in the
presence of oxygen (under air). From 1a, the alcohol 4a and
tert-butylacetamide are formed together with acetone (5a).
The ketones 5b and 5c, the alkenes 3b and 3c, and the
1a
1b
1c
2
4.0
2.7
5.0
4.0
12
12
16
15

6570
E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
BET
b
R
R
1
1
hν
+
+
NMQ
C
R
S
+ NMQ
C
R
S
a
2
2
2 R = R = CH
1
2
3
3 R1 = CH , R = Ph
3
2
c
R
1
C
+
S
R
2
S
+ H O
2
- H+
- H+
e
g
d
R
1
f
SS
C
R
C OH
CH
2
R
2
2
Scheme 4.
alcohols 4b and 4c, are formed from 1b and 1c, respectively.
In all cases, benzenesulfonate is the main product coming
from the phenyl sulfide moiety even though small amounts
of diphenyl disulfide are observed.
such a bimolecular reaction can become very difficult and
the thiadioxirane (that probably is in a vibrationally excited
state due to the very large exergonicity of its formation) may
instead undergo C–S bond cleavage, presumably concerted
with the opening of the three-member ring, leading to a quite
stable tertiary carbocation and phenylsulfinate (Scheme 5,
path b).31 The phenylsulfinate should be oxidized to the
sulfonate under the reaction conditions.32 The alcohols (and
alkenes) are formed from the carbocations as shown in
Scheme 4. With 1a, also N-tert-butylacetamide is formed,33
whereas the alkene is not observed.
Clearly, the presence of O2 promotes a fragmentation pro-
cess that is much more important than that coming from
the direct C–S bond cleavage in the radical cation mentioned
above. A possible proposal is that the fragmentation prod-
ucts in the presence of oxygen mostly derive from the de-
composition of an intermediate formed by the reaction of
the radical cation with the superoxide anion generated by
NMQꢀ and O2. As already mentioned, the reaction of sulfide
radical cations with Oꢁ2 ꢀ is the key step in the electron trans-
fer photoinduced sulfoxidations of sulfides and it is reason-
able to suggest that it can play a fundamental role also in the
fragmentation reactions investigated. This suggestion is
supported by the following. First, as already observed, the
presence of oxygen strongly increases the efficiency of the
photolysis, producing quantum yields of fragmentation
products much higher that those found in the presence of
R
C
R
R
C
R
1
O
O
1
a
+ O
CH
3
S
CH
S
2
3
2
2
6
b
O
S
R
C
R
1
ꢀ
nitrogen. Second, a reaction of Oꢁ2 with the radical cation
CH
O
3
is indicated by the results of time-resolved experiments.
Accordingly, as already noted, the second-order decay of
the radical cation in the presence of O2 is ca. 3–4 times faster
than in the presence of nitrogen, being probably determined
2
Scheme 5.
ꢀ
by the reaction of the radical cation with Oꢁ2 , and there is no
The formation of ketones 5, that are a significant fraction of
the products mixture obtained under air, may be justified by
the possible reaction of the carbocation with Oꢀ2ꢁ. An alkyl-
peroxyl radical is formed that can be converted into ketone
as shown in Scheme 6. Indeed, by generating the cumyl-
peroxyl radical in the photolysis of cumene in the presence
effect of the sulfide structure on the reactivity. Third and still
more significant, in the presence of benzoquinone that can
ꢀ
trap Oꢁ2 , a substantial drop of the quantum yield in the
photolysis of 1a–c under air was observed.30
With convincing evidence in hand that the reaction of the
ꢀ
sulfide radical cations of 1a–c with Oꢁ2 is en route to frag-
mentation products, we can try to envisage a reasonable
mechanistic scheme. In previous work15 we have reported
evidence indicating that in electron transfer photooxygena-
tions, the reaction of the sulfide radical cation with Oꢁ2 ꢀ leads
to a thiadioxirane 6 (Scheme 5, path a). Usually, the thia-
dioxirane is expected to react with another sulfide molecule
to form the sulfoxide. However, it is not unreasonable to
suggest that with the sterically crowded sulfides 1a–c,
R
C
R
R
1
1
+
O
2
H C
H C C OO
3
3
R
2
2
R
1
R
1
+
CH
3
H C C O
O
3
R
2
R
2
5a-5c
Scheme 6.

E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
6571
ꢀ
of di-tert-butyl peroxide and oxygen, benzophenone 5b was
obtained as the main product accompanied by small amounts
of 3b and 4b (Section 5).34
the reaction of the sulfide radical cation with Oꢁ2 . The meso-
lytic rupture of the C–S bond in the intermediate radical
cation plays a minor role. A different situation holds with
benzyl phenyl sulfide 2 (a primary alkyl sulfide with a-CH
bonds) where the fragmentation product, benzaldehyde, is
suggested to come from deprotonation of 2ꢀ+ by O2ꢁꢀ in com-
petition with the formation of thiadioxirane. The latter spe-
cies is much less congested than that formed from tert-alkyl
phenyl sulfides and can therefore react, as expected, with
another sulfide molecule to give benzyl phenyl sulfoxide
that accordingly is among the reaction products.
3.2. Benzyl phenyl sulfide
The behaviors of sulfide (2) were similar to that of the tert-
alkyl phenyl sulfides, discussed above. Thus, laser photoly-
sis experiments showed the formation of the sulfide radical
cation (in this case the dimeric form was also observed)
that, however, decayed without the apparent formation of
any transient. The decay rate followed second-order kinetics
with a k23 value very close to that found with the radical cat-
ions of tert-alkyl phenyl sulfides. In this case too the decay
rate was faster under oxygen than under nitrogen (Table
3). As with tert-alkyl phenyl sulfides, steady-state photolysis
of 2 sensitized by NMQ+ led to significant amounts of
products (benzyl phenyl sulfoxide and benzaldehyde) only
in the presence of oxygen. However, when the photolysis
was run in the presence of benzoquinone the yield of
products dropped almost to zero in the presence of oxygen.
5. Experimental
5.1. Starting materials
Commercial benzyl phenyl sulfide (2) was further purified by
recrystallization from EtOH/H2O. tert-Butyl phenyl sulfide
(1a) was prepared by acid catalyzed reaction of thiophenol
with tert-butanol.36 2-Phenyl-2-propyl phenyl sulfide (1b)
and 1,1-diphenylethyl phenyl sulfide (1c) were prepared by
acid catalyzed addition of thiophenol on a-methylstyrene
and 1,1-diphenylethylene, respectively.37 CH3CN (spectro-
photometric grade) and toluene were used as received.
N-Methylquinolinium tetrafluoroborate was prepared
according to a literature procedure.38
It seems clear that with 2, the reaction of the sulfide radical
ꢀ
cation with Oꢁ2 also plays a key role in product formation.
When Oꢁ2 ꢀ attacks sulfur, thiadioxirane forms leading to sulf-
oxide by reaction with another molecule of sulfide, as previ-
ously suggested for electron transfer sulfoxygenations.15 In
this case, the sulfide is not sterically congested, thus this re-
action predominates with respect to the C–S bond cleavage
reactions observed with 1a–c.35 However, since a-CH bonds
5.2. Quantum yields
ꢀ
are present in 2+ , O2ꢁꢀ can also perform a deprotonation pro-
A 2 ml solution of NMQ+ (4ꢂ10ꢁ4 M) in CH3CN and 1a–c
and 2 (1.0ꢂ10ꢁ2 M) in CH3CN (CD3CN in the case of 1a)
was placed in a quartz cuvette and irradiated at 313 nm,
selected with a Balzer interference filter by a high pressure
Hg lamp. The substrate conversion to photoproduct was
held below 10% to avoid secondary reactions. The photo-
cess forming an a-phenylthio carbon radical that can lead to
benzaldehyde and diphenyl disulfides by reaction with O2, as
shown in Scheme 7. The possibility that in electron transfer
sulfoxygenations of sulfides with a-CH bonds, C–S bond
cleavage products would derive from deprotonation of the
sulfide radical cation had already been proposed,13 but the
key role played by Oꢁ2 ꢀ in this respect had not been hitherto
demonstrated.
1
products were quantified by GC and H NMR. Bibenzyl
was used as an internal standard. The products from sulfide
1a were analyzed exclusively by 1H NMR, because of their
high volatility. Water soluble photoproducts were identified
1
by H NMR after evaporation of the solvent CH3CN and
+
O2
Ph
S
CH2 Ph
Ph
Ph
S
CH
O2
O
addition of D2O. All products formed were identified by
comparison with authentic specimens. tert-Butyl acetamide,
acetophenone, benzophenone, 2-phenyl-2-propanol, a-
methylstyrene, 1,1-diphenylethanol, 1,1-diphenylethylene,
tert-butanol, acetone, benzaldehyde, diphenyl disulfide,
sodium benzenesulfonate were commercial products.
Benzyl phenyl sulfoxide was prepared by oxidation of benzyl
phenyl sulfide with NaIO4.39 The light intensity (ca.
1ꢂ1015 photons sꢁ1) was measured by potassium ferric
oxalate actinometry.40
HOO
O
Ph
S
CH Ph
PhCHO + PhSSPh
5.3. Steady-state photooxidation of cumene
Scheme 7.
By irradiation of tert-butyl peroxide (0.16 M) at 313 nm in
the presence of cumene (1.0ꢂ10ꢁ2 M) in O2-saturated
CH3CN, acetophenone (3.4%), a-methylstyrene (1.0%),
and 1,1-diphenylethanol (0.8%) formed after 7 h.
4. Conclusions
The results of LFP and steady-state photolysis experiments
have allowed us to reach the important conclusion that the
C–S bond cleavage products formed in the photooxygena-
tion of tert-alkyl phenyl sulfides sensitized by NMQ+ mainly
derive from the sterically congested thiadioxirane formed by
5.4. Fluorescence quenching
Measurements were carried out on a Spleg Fluorolog
F112AF spectrofluorometer. Relative emission intensities

6572
E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
at 390 nm (NMQ+ emission maximum) were measured by
irradiating at 315 nm (NMQ+ absorption maximum) a solu-
tion containing NMQ+ (1.0ꢂ10ꢁ5 M) and the substrates
1a–c and 2 at different concentrations (from 0 to 7.0ꢂ
10ꢁ3 M) in CH3CN at 22 ꢃC. The error estimated on the
Stern–Volmer constants (KSV) was ꢄ5%.
12. Clennan, E. L.; Zhou, W.; Chan, J. J. Org. Chem. 2002, 67,
9368–9378.
13. Lacombe, S.; Cardy, H.; Simon, M.; Khoukh, A.; Soumillion,
J. P.; Ayadim, M. Photochem. Photobiol. Sci. 2002, 1, 347–
354.
14. Zen, J. M.; Liou, S. L.; Kumar, A. S.; Hsia, M. S. Angew.
Chem., Int. Ed. 2003, 42, 577–579.
15. Baciocchi, E.; Del Giacco, T.; Elisei, F.; Gerini, M. F.; Guerra,
M.; Lapi, A.; Liberali, P. J. Am. Chem. Soc. 2003, 125, 16444–
16454.
16. Che, Y.; Ma, W.; Ren, Y.; Chen, C.; Zhang, X.; Zhao, J.; Zang,
L. J. Phys. Chem. B 2005, 109, 8270–8276.
17. Baciocchi, E.; Crescenzi, C.; Lanzalunga, O. Tetrahedron
1997, 53, 4469–4478.
18. Manring, L. E.; Kramer, M. K.; Foote, C. S. Tetrahedron Lett.
1984, 25, 2523–2526.
19. Bonesi, S. M.; Fagnoni, M.; Albini, A. J. Org. Chem. 2004, 69,
928–935.
20. Fukuzumi, S.; Fujita, M.; Noura, S.; Ohkubo, K.; Suenobu,
T.; Araki, Y.; Ito, O. J. Phys. Chem. A 2001, 105, 1857–
1868.
21. Baciocchi, E.; Del Giacco, T.; Elisei, F.; Gerini, M. F.; Lapi, A.;
Liberali, P.; Uzzoli, P. J. Org. Chem. 2004, 69, 8323–8330.
22. Dockery, K. P.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.;
Gould, I. R.; Todd, W. P. J. Am. Chem. Soc. 1997, 119,
1876–1883.
5.5. Laser flash photolysis
The excitation wavelength of 355 nm from a Nd:YAG laser
(Continuum, third harmonic) was used in nanosecond flash
photolysis experiments (pulse width ca. 7 ns and energy
<3 mJ per pulse). The transient spectra were obtained by
a point-to-point technique, monitoring the absorbance
changes (DA) after the flash at intervals of 5–10 nm over
the spectral range 350–850 nm, averaging at least 10 decays
at each wavelength. The lifetimevalues (the time at which the
initial signal is reduced to 1/e, experimental error ꢄ10%) are
reported for transients showing first-order decay kinetics.
The k2/3 values are reported for the transients showing sec-
ond-order kinetics. A 2 ml solution containing the substrate
(1.0ꢂ10ꢁ2 M), NMQ+ (4ꢂ10ꢁ4 M) and the cosensitizer
(1 M toluene) was flashed in a quartz photolysis cell while
nitrogen or oxygen was bubbling through them. All measure-
ments were carried out at 22ꢄ2 ꢃC. The experimental error
was ꢄ10%.
23. Bockman, T. M.; Kochi, J. K. J. Am. Chem. Soc. 1989, 111,
4669–4683.
Acknowledgements
´
24. Jonsson, M.; Lind, J.; Merenyi, G.; Eriksen, T. E. J. Chem. Soc.,
Perkin Trans. 2 1995, 67–70; Ioele, M.; Steenken, S.;
Baciocchi, E. J. Phys. Chem. A 1997, 101, 2979–2987;
Mohan, H.; Mittal, J. P. J. Phys. Chem. A 1997, 101, 10012–
10017.
MIUR, University ‘la Sapienza’ of Rome and University of
Perugia are thanked for the financial support.
25. Yokoi, H.; Hatta, A.; Ishiguro, K.; Sawaki, Y. J. Am. Chem. Soc.
1998, 120, 12728–12733.
Supplementary data
26. Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436–442.
27. Armstrong, D. A.; Sun, Q.; Schuler, R. H. J. Phys. Chem. 1996,
100, 9892–9899.
Supplementary data associated with this article can be found
in the online version, at doi:10.1016/j.tet.2005.09.154.
28. Cozens, F. L.; Kanagasabapathy, V. M.; McClelland, R. A.;
Steenken, S. Can. J. Chem. 1999, 77, 2069–2082.
29. Fukuzumi, S.; Nakanishi, J.; Maruta, J.; Yorisue, J.; Suenobu,
T.; Itoh, S.; Arakawa, R.; Kadish, K. M. J. Am. Chem. Soc.
1998, 120, 6673–6680.
References and notes
1. Eriksen, J.; Foote, C. S.; Parker, T. L. J. Am. Chem. Soc. 1977,
99, 6455–6456.
2. Ando, W.; Kabe, Y.; Shohei, K.; Takyu, C.; Yamagishi, A.;
Inaba, H. J. Am. Chem. Soc. 1980, 102, 4526–4528.
3. Marciniak, B.; Hug, G. L.; Rozwadowski, J.; Bobrowski, K.
J. Am. Chem. Soc. 1995, 117, 127–134.
30. The effect of benzoquinone also shows that the higher quantum
yields in the presence of oxygen are not due to a suppression of
the BET process by the reaction of NMQꢀ with O2. This possi-
bility is also made unlikely by the fact that no transients were
observed in the laser photolysis experiments in the presence
of O2.
4. Miller, B.; Williams, T. D.; Schoneich, C. J. Am. Chem. Soc.
1996, 118, 11014–11025.
5. Bobrowski, K.; Hug, G. L.; Marciniak, B.; Miller, B.;
31. Theoretical calculations (B3LYP/6-31G* level) have shown
that such a C–S bond cleavage, in CH3CN as solvent, is exo-
thermic by 11 kcal molꢁ1. We are very grateful to Dr. M. F.
Gerini for performing this calculation.
¨
Schoneich, C. J. Am. Chem. Soc. 1997, 119, 8000–8011.
¨
6. Tung, C.-H.; Guan, J.-Q. J. Am. Chem. Soc. 1998, 120, 11874–
11879.
7. Bonesi, S. M.; Mella, M.; d’Alessandro, N.; Aloisi, G. G.;
Vanossi, M.; Albini, A. J. Org. Chem. 1998, 63, 9946–9955.
8. Toutchkine, A.; Clennan, E. L. J. Org. Chem. 1999, 64, 5620–
5625.
9. Bonesi, S. M.; Albini, A. J. Org. Chem. 2000, 65, 4532–4536.
10. Toutchkine, A.; Aebisher, D.; Clennan, E. L. J. Am. Chem. Soc.
2001, 123, 4966–4973.
32. Oae, S.; Takata, T.; Kim, H. K. Tetrahedron 1981, 37, 37–44.
33. tert-Butyl cation is less stable and less selective than the cations
from 1a and 1b, therefore it can also react with the solvent that
is a weaker nucleophile than water.
34. It might be suggested that the C–S bond breaking in the thia-
dioxirane is homolytic forming a benzenesulfonyl radical and
a carbon radical. Reaction of the carbon radical with oxygen
should lead to the observed reaction products. However, the
formation of N-tert-butylacetamide is a clear indication that
a carbocation is also formed and supports the heterolytic
11. Clennan, E. L. Acc. Chem. Res. 2001, 34, 875–884.

E. Baciocchi et al. / Tetrahedron 62 (2006) 6566–6573
6573
cleavage. Moreover, in this case the ketone would be the main
reaction product, which is not observed.
37. Screttas, C. G.; Micha-Screttas, M. J. Org. Chem. 1979, 44,
713–719.
35. In the photooxygenation of 2 sensitized by 9,10-dicyano-
anthracene we have also observed the formation of benzyl
derivatives.17 However, those experiments were carried out
at a significantly higher temperature, ꢅ40 ꢃC.
38. Donovan, P. F.; Conley, D. A. J. Chem. Eng. Data 1966, 11,
614–615.
39. Johnson, C. R.; Leonard, N. J. J. Org. Chem. 1962, 27,
282–284.
40. Bunce, N. J. Handbook of Organic Photochemistry; Scaiano,
J. C., Ed.; Academic: New York, NY, 1999; Vol. 1, p 243.
36. Cutress, N. C.; Grindley, T. B.; Katritzky, A. R.; Topsom, R. D.
J. Chem. Soc., Perkin Trans. 2 1974, 263–268.