Singlet oxygen promoted carbon-heteroatom bond cleavage in dibenzyl sulfides and tertiary dibenzylamines. Structural effects and the role of exciplexes

Article abstract of DOI:10.1021/jo701641b

(Chemical Equation Presented) The C-heteroatom cleavage reactions of substituted dibenzyl sulfides and substituted dibenzylcyclohexylamines promoted by singlet oxygen in MeCN have been investigated. In both systems, the cleavage reactions (leading to benzaldehyde and substituted benzaldehyde) were slightly favored by electron-withdrawing substituents with ρ values of +0.47 (sulfides) and +0.27 (amines). With dibenzyl sulfides, sulfones were also obtained whereas sulfoxide formation became negligible when the reactions were carried out in the presence of a base. Through a careful product study for the oxidation of dibenzyl sulfide, in the presence and in the absence of Ph ₂SO, it was established that sulfone and cleavage product (benzaldehyde) do not come by the same route (involving the persulfoxide and the hydroperoxysulfonium ylide) as required by the generally accepted mechanism (Scheme 1) for C-heteroatom cleavage reactions of sulfides promoted by singlet oxygen. On this basis and in light of the similar structural effects noted above it is suggested that dibenzyl sulfides and dibenzylamines form benzaldehydes by a very similar mechanism. The reaction with singlet oxygen leads to an exciplex that can undergo an intracomplex hydrogen atom transfer to produce a radical pair. With sulfides, collapse of the radical pair leads to an α-hydroperoxy sulfide than can give benzaldehyde by an intramolecular path as described in Scheme 3. With amines, the radical pair undergoes an electron-transfer reaction to form an iminium cation that hydrolyzes to benzaldehyde. From a kinetic study it has been established that the fraction of exciplex converted to aldehyde is ca. 20% with sulfides and ca. 7% with amines.

Full text of DOI:10.1021/jo701641b

Singlet Oxygen Promoted Carbon-Heteroatom Bond Cleavage in
Dibenzyl Sulfides and Tertiary Dibenzylamines. Structural Effects
and the Role of Exciplexes
Enrico Baciocchi,*,† Tiziana Del Giacco,‡ Osvaldo Lanzalunga,§,† and Andrea Lapi*,§,†
Dipartimento di Chimica and Istituto CNR di Metodologie Chimiche-IMC, Sezione Meccanismi di
Reazione c/o Dipartimento di Chimica, Sapienza UniVersita` di Roma, P.le A. Moro 5, 00185 Rome, Italy,
and Dipartimento di Chimica and Centro di Eccellenza Materiali InnoVatiVi Nanostrutturati, UniVersita` di
Perugia, Via Elce di sotto 8, 06123 Perugia, Italy
enrico.baciocchi@uniroma1.it; andrea.lapi@uniroma1.it
ReceiVed July 26, 2007
The C-heteroatom cleavage reactions of substituted dibenzyl sulfides and substituted dibenzylcyclo-
hexylamines promoted by singlet oxygen in MeCN have been investigated. In both systems, the cleavage
reactions (leading to benzaldehyde and substituted benzaldehyde) were slightly favored by electron-
withdrawing substituents with F values of +0.47 (sulfides) and +0.27 (amines). With dibenzyl sulfides,
sulfones were also obtained whereas sulfoxide formation became negligible when the reactions were
carried out in the presence of a base. Through a careful product study for the oxidation of dibenzyl
sulfide, in the presence and in the absence of Ph2SO, it was established that sulfone and cleavage product
(benzaldehyde) do not come by the same route (involving the persulfoxide and the hydroperoxysulfonium
ylide) as required by the generally accepted mechanism (Scheme 1) for C-heteroatom cleavage reactions
of sulfides promoted by singlet oxygen. On this basis and in light of the similar structural effects noted
above it is suggested that dibenzyl sulfides and dibenzylamines form benzaldehydes by a very similar
mechanism. The reaction with singlet oxygen leads to an exciplex that can undergo an intracomplex
hydrogen atom transfer to produce a radical pair. With sulfides, collapse of the radical pair leads to an
R-hydroperoxy sulfide than can give benzaldehyde by an intramolecular path as described in Scheme 3.
With amines, the radical pair undergoes an electron-transfer reaction to form an iminium cation that
hydrolyzes to benzaldehyde. From a kinetic study it has been established that the fraction of exciplex
converted to aldehyde is ca. 20% with sulfides and ca. 7% with amines.
Introduction
R-CH bond in the sulfide is relatively weak (e.g., when R )
Ph).1
It is well-known that in the reaction of sulfides with singlet
oxygen (1∆g), hereafter indicated as 1O2, products of C-S bond
cleavage can be formed in addition to the normal S-oxidation
products, viz., sulfoxides and sulfones.1 The C-S bond cleavage
reaction, eq 1, can become particularly important (and represent
the major reaction pathway) in aprotic solvents and when the
RCH2SR′ + 1O2 f RCHO + S-oxidation products (1)
Reaction 1 is receiving increasing attention and is generally
discussed in the framework of the widely accepted mechanism
of the 1O2-promoted sulfoxidation of sulfides. According to the
most recent mechanistic studies,1f-h the key step is the conver-
sion of the first formed persulfoxide 1 into the S-hydroperox-
ysulfonium ylide 2 (Scheme 1, path a). The latter compound
can lead to sulfoxide and sulfone but it may also undergo a
Pummerer rearrangement to form the R-hydroperoxy sulfide 3
* Address correspondence to this author. A.L.: phone +39-06-49913683;
fax +39-06-490421. E.B.: phone +39-06-49913711; fax +39-06-490421.
† Istituto CNR di Metodologie Chimiche-IMC, Sezione Meccanismi di
Reazione.
‡ Universita` di Perugia.
§ Dipartimento di Chimica, Sapienza Universita` di Roma.
10.1021/jo701641b CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/15/2007
9582
J. Org. Chem. 2007, 72, 9582-9589

Singlet Oxygen Promoted C-Heteroatom Bond CleaVage
SCHEME 1
SCHEME 2
cleavage of the C-N bond to form an aldehyde and a secondary
amine. In the amines case, no evidence, either theoretical or
experimental, has been ever found for intermediates en route
from the charge-transfer complex to the radical couple. Thus,
two very similar processes, both induced by the same species,
appear to involve completely different mechanisms.3
In this context, it seems of interest to study structural effects
in the 1O2-promoted C-heteroatom bond cleavage reactions of
dibenzyl sulfides and dibenzylamines with the aim of investigat-
ing whether and how these effects are influenced by the
mechanistic differences noted above for the two systems. The
structural effects could be quite different in the hydrogen transfer
step that in both cases should be the key step leading to the
heteroatom-carbon cleavage. According to the mechanisms
displayed in Schemes 1 and 2, the hydrogen transfer step for
dibenzyl sulfides is the one where the persulfoxide intermediate
is converted into the hydroperoxysulfonium ylide (Scheme 1,
path a), whereas for dibenzylamines it is the one where the
exciplex is converted into the radical pair (Scheme 2, path a).
Thus, we have studied the reaction of a series of arylmethyl
benzyl sulfides (4) and arylmethyl(benzyl)cyclohexylamines4
(5) with 1O2 in MeCN. Structural effects on the C-heteroatom
bond cleavage can be easily evaluated by measuring the 4-X-
C6H4CHO/benzaldehyde ratio in the products mixture.
(path b) that can react with another molecule of sulfide forming
a sulfoxide and an R-hydroxy sulfide (path c). The hydrolysis
of the latter generates the cleavage products (path d) that,
however, can also be formed from 3 by a unimolecular
mechanism that does not produce sulfoxide (path e).
In many respects, the C-S bond cleavage of sulfides presents
a close analogy to the C-N bond cleavage (N-dealkylation)
that is the main outcome of the reaction of tertiary trialkylamines
with 1O2.2 In this case, however, the cleavage is thought to take
place by a simple mechanism involving a hydrogen abstraction
reaction occurring inside the charge-transfer complex (exciplex)
formed by interaction of the amine lone pair with 1O2 (Scheme
2).2j,k A radical couple is formed (path a) that can undergo an
•
electron transfer from the carbon radical to HO2 to form an
iminium cation (path b). The latter can easily hydrolyze with
(1) (a) Corey, E. J.; Ouannes, C. Tetrahedron Lett. 1976, 17, 4263. (b)
Takata, T.; Ishibashi, K.; Ando, W. Tetrahedron Lett. 1985, 26, 4609. (c)
Akasaka, T.; Sakurai, A.; Ando, W. J. Am. Chem. Soc. 1991, 113, 2696.
(d) Sheu, C.; Foote, C. S.; Gu, C.-L. J. Am. Chem. Soc. 1992, 114, 3015.
(e) Bonesi, S. M.; Mella, M.; d’Alessandro, N.; Aloisi, G. G.; Vanossi, M.;
Albini, A. J. Org. Chem. 1998, 63, 9946. (f) Toutchkine, A.; Clennan, E.
L. J. Am. Chem. Soc. 2000, 122, 1834. (g) Toutchkine, A.; Aebisher, D.;
Clennan, E. L. J. Am. Chem. Soc. 2001, 123, 4966. (h) Bonesi, S. M.;
Fagnoni, M.; Monti, S.; Albini, A. Tetrahedron 2006, 62, 10716.
(2) (a) Fisch, M. H.; Gramin, J. C.; Olesen, J. A. Chem. Commun. 1970,
13. (b) Fisch, M. H.; Gramin, J. C.; Olesen, J. A. Chem. Commun. 1971,
663. (c) Bellus, D.; Lind, H.; Wyatt, J. F. Chem. Commun. 1972, 1199. (d)
Smith, W. F. J. Am. Chem. Soc. 1972, 94, 186. (e) Tsubomura, H.; Yagishita,
T.; Toi, H. Bull. Chem. Soc. Jpn. 1973, 46, 3051. (f) Gollnick, K.; Lindner,
J. H. E. Tetrahedron Lett. 1973, 1903. (g) Inoue, K.; Saito, I.; Matsuura,
T. Chem. Lett. 1977, 607. (h) Haugen, C. M.; Bergmark, W. R.; Whitten,
D. G. J. Am. Chem. Soc. 1992, 114, 10293. (i) Cocquet, G.; Rool, P.;
Ferroud, C. J. Chem. Soc., Perkin Trans. 1 2000, 14, 2277. (j) Baciocchi,
E.; Del Giacco, T.; Lapi, A. Org. Lett. 2004, 6, 4791. (k) Baciocchi, E.;
Del Giacco, T.; Lapi, A. HelV. Chim. Acta 2006, 89, 2273. (l) Baciocchi,
E.; Del Giacco, T.; Lanzalunga, O.; Lapi, A.; Raponi, D. J. Org. Chem.,
2007, 72, 5912.
To gain a better mechanistic understanding of the sulfoxi-
dation reaction, a careful determination of the sulfoxidation
products in the reaction of dibenzyl sulfide with O2 was also
1
carried out. Moreover, the study was completed with kinetic
measurements for the determination of the rates of total and
1
chemical quenching of O2 by dibenzylamines and dibenzyl
sulfides.
(3) The similar outcome of the reactions of benzyl sulfides and tertiary
1
benzylamines with O2 has been already noted by Albini and co-workers,
who have considered whether the analogy could imply a similar structural
dependence.1h They have concluded negatively, but on the basis of a general
comparison between sulfides and amines, i.e., not strictly limited to those
systems that undergo carbon-heteroatom scission.
(4) (a) The cyclohexyl group was chosen as alkyl substituent since it
1
does not show any reactivity with O2.2k,4b (b) Baciocchi, E.; Del Giacco,
T.; Lapi, A. Org. Lett. 2006, 8, 1783.
J. Org. Chem, Vol. 72, No. 25, 2007 9583

Baciocchi et al.
TABLE 1. 1O2 Total Quenching Rate Constants (kT) by Dibenzyl
Sulfides 4a-c and Dibenzylamines 5a-c in MeCN at 22 °C
TABLE 2. Benzaldehydes Yields in the Oxidation of
Dibenzylamines 5a-c by Photochemically and Thermally Generated
1O2
a
a
X
compd
kT (M-1 s-1
)
compd
kT (M-1 s-1
)
benzaldehydes yields (%)a
4-X-C6H4CHO/
H
OMe
Me
4
1.06 × 107
1.34 × 107
1.37 × 107
7.34 × 106
4a
4b
4c
5a
5b
5c
4.04 × 106
4.05 × 106
3.05 × 106
X
PhCHO
4-X-C6H4CHO
PhCHO
OMe
hνb
6.4 ( 0.3
3.7 ( 0.2
7.5 ( 0.4
3.9 ( 0.2
3.1 ( 0.2
5.7 ( 0.3
5.0 ( 0.3
3.0 ( 0.2
6.7 ( 0.3
3.5 ( 0.2
4.9 ( 0.3
8.0 ( 0.4
0.79
0.82
0.89
0.89
1.5
CN
c
∆
a Errors are (10%.
Me
CN
hνb
c
∆
hνd
e
∆
1.4
Results
a Referred to the initial amount of substrate and determined by GC
analysis. Average of at least four determinations. b 15 min irradiation.
c Endoperoxide/substrate ratio ) 2:1. d 30 min irradiation. e Endoperoxide/
substrate ratio ) 10:1.
Kinetic Study. The rate constants (kT) for the total quenching
(physical and chemical) of 1O2 by dibenzyl sulfides 4 and 4a-c
and dibenzylcyclohexylamines 5a-c were measured in MeCN
by laser flash photolysis experiments following the decay rate
of the singlet oxygen luminescence at 1270 nm. The results are
reported in Table 1 and it can be immediately noted that in
both systems the rate of total quenching is little influenced by
the substituent in the ring.
TABLE 3. Benzaldehydes Yields in the Oxidation of Dibenzyl
1
Sulfides 4a-c by Photochemically and Thermally Generated O2
benzaldehydes yields (%)a
4-X-C6H4CHO/
X
PhCHO
4-X-C6H4CHO
PhCHO
Benzylamines. Singlet oxygen was both photochemically and
thermally generated. In the former case tetraphenylporphyrin
(TPP) was used as the photosensitizer, whereas in the latter
process, singlet oxygen was generated by thermal decomposition
of 1,4-dimethylnaphthalene endoperoxide.5 In the photochemical
reactions, the substrate (10-2 M) was irradiated (400-600 nm)
in O2-saturated dry acetonitrile in the presence of TPP (10-4
M) at 25 °C with an irradiation time ranging from 15 to 30
min. During the reaction, further TPP was added at 5-min
intervals due to its degradation in the reaction medium. In all
cases, no product formation was observed when irradiations
were performed in the absence of TPP or O2. Thermal reactions
were carried out by heating (40 °C) a solution of the amine
(10-2 M) in the presence of 1,4-dimethylnaphthalene endoper-
oxide (2 × 10-2 or 0.1 M) in dry MeCN for 4 h in the dark.
Blank experiments carried out under the same conditions in O2-
saturated MeCN but in the absence of 1,4-dimethylnaphthalene
endoperoxide showed the absence of reaction products. Product
analysis was carried out by GC, GC-MS, and 1H NMR. In both
photochemical and thermal reactions, substrate conversion was
kept below 20% and the mass balance was always greater than
95%, with respect to benzaldehydes.
From the dibenzylamines 5a-c, 4-substituted and unsubsti-
tuted benzaldehydes were obtained together with the corre-
sponding secondary amines. Moreover, benzylidene cyclohex-
ylamines were also observed, which likely derive from the
further oxidation of the secondary amines.6 As previously found
in the oxidation of N,N-dimethylbenzylamine with 1O2,2j it was
observed that H2O2 is formed in a molar amount very close to
the sum of the substituted and unsubstituted benzaldehydes.7
The yields of substituted and unsubstituted benzaldehydes,
as well as the 4-X-C6H4CHO/PhCHO molar ratios, are reported
in Table 2. It can be noted that there is an excellent agreement
between the ratios obtained in the photochemical and thermal
experiments.
OMe
hνb
5.8 ( 0.4
3.3 ( 0.2
5.3 ( 0.4
2.9 ( 0.2
4.0 ( 0.3
1.8 ( 0.1
4.3 ( 0.3
2.1 ( 0.1
4.4 ( 0.3
2.2 ( 0.2
7.6 ( 0.5
3.6 ( 0.3
0.74
0.64
0.83
0.77
1.9
c
∆
Me
CN
hνb
c
∆
hνb
c
∆
2.0
a Referred to the initial amount of substrate and determined by 1H NMR
analysis. Average of at least four independent determinations. b 15 min
irradiation. c Endoperoxide/substrate ratio ) 1:3.
1O2. Since it is known that in MeCN the endoperoxide generates
1O2 in a 70% yield,4b from the O2 self-decay rate constant in
1
MeCN (1.4 × 104 s-1)8 and the kT values (Table 1) it is possible
1
to calculate the molar amount of O2 that actually reacts with
the dibenzylamine under our conditions (see the Supporting
Information). From this amount and the molar amount of
benzaldehydes formed in the reactions with thermally generated
1O2, the fraction of exciplex converted to benzaldehydes, in
competition with intersystem crossing (see Scheme 2), can be
calculated (e.g., 7% in the case of 5b). The kR(ArCHO) value can
then be obtained from the kT value reported in Table 1 (details
in the Supporting Information). This analysis, in the case of
5b, gave a kR(ArCHO) value of 2.8 × 105 M-1 s-1. In view of the
small substituent effects on the aldehyde yields as well as on
the kT values, similar values of kR(ArCHO) are obtained for the
other compounds (Supporting Information).
Benzyl Sulfides. Dibenzyl sulfides 5a-c were also reacted
with both photochemically (TPP) and thermally (1,4-dimeth-
ylnaphthalene endoperoxide) generated 1O2. In both cases, C-S
bond cleavage products (4-X-C6H4CHO and PhCHO) were the
main reaction products accompanied by S-oxidation products
and sulfur-containing cleavage products (see later). No formation
of H2O2 was observed.
The X-C6H4CHO/PhCHO molar ratios, determined as in the
benzylamines case, resulted very similar in photochemical and
thermal oxygenation and are displayed in Table 3. In this case
1
The results obtained with 1,4-dimethylnaphthalene endoper-
oxide allow us to calculate the rate constant for benzaldehydes
formation, kR(ArCHO), in the reaction of a dibenzylamine with
too, the results of the reaction with thermally generated O2
allow us to calculate the rate constant for benzaldehyde
formation following the procedure described above for the
benzylamines (details in the Supporting Information). For 4b,
(5) Turro, N. J.; Chow, M. F. J. Am. Chem. Soc. 1981, 103, 7218.
(6) After 5 min of irradiation under the same reaction conditions, it was
observed that N-benzylcyclohexylamine (10-2 M) was converted to ben-
zylidene cyclohexylamine in 65% yield.
k
R(ArCHO) is 3.0 × 106 M-1 s-1. Very similar values are obtained
for the other sulfides.
(7) For example, in the reaction of 5b, H2O2 was formed in a 13% yield,
very close to that of the two benzaldehydes (14%).
(8) Wilkinson, F.; Helman, W. P.; Ross, A. B. J. Phys. Chem. Ref. Data
1995, 24, 663.
9584 J. Org. Chem., Vol. 72, No. 25, 2007

Singlet Oxygen Promoted C-Heteroatom Bond CleaVage
TABLE 4. 1O2-Promoted Oxidation of 4 Photosensitized by Rose Bengal
a Referred to the initial amount of substrate and determined by HPLC analysis. Average of at least two determinations. b Rose bengal was added at 5-min
1
intervals due to its bleaching during the irradiation. The final RB concentration was 6 × 10-4 M. c 5 × 10-3 M. d Because of the O2 quenching property
of ethyl diisopropylamine, a longer irradiation time has been used to achieve the same benzaldehyde yield as in the reaction carried out in the absence of
the amine (see entry 2).
Product Study in the Reaction of Dibenzyl Sulfide. A
detailed product study was carried out for the photochemical
reaction of dibenzyl sulfide 4 with 1O2, with the aim of obtaining
further mechanistic information on the oxygenation of this
substrate.
in the presence of ethyl diisoproplyamine (5 × 10-3 M)12 no
bleaching of the rose bengal was observed and only a very small
amount of benzyl sulfoxide was found among the reaction
products (Table 4, entry 4; compare with entry 2). With use of
the substituted dibenzyl sulfides, it was verified that the 4-X-
C6H4CHO/PhCHO molar ratios (Table 3) are not influenced at
all by the presence or absence of the base in the reaction
medium.
The irradiation (400-600 nm) of a solution of 4 (10-2 M) in
the presence of rose bengal9 (10-4 M) as the photosensitizer in
O2-saturated dry MeCN afforded benzaldehyde, benzyl sulfox-
ide, benzyl sulfone, benzyl benzylthiosulfinate, and benzyl
benzylthiosulfonate. No dibenzyl disulfide was detected in the
reaction mixture. From the results reported in Table 4 (entries
1-3), it can be readily noted that the products distribution
changes with the reaction time. In particular, the amount of
dibenzyl sulfoxide formed, relative to that of benzaldehyde,
increases with the progress of photooxidation (the benzaldehyde/
benzyl sulfoxide ratio decreasing from 8.3 to 1.6 on going from
1 to 15 min of irradiation). An important observation, however,
was that under our reaction conditions, a rapid bleaching of
rose bengal took place, the solution appearing colorless after
only 2 min of irradiation. A further observation was that the
rose bengal color was not restored after addition of another
aliquot of the sensitizer even in the dark. We suspect that an
acidic compound is formed during the oxidation (presumably
benzylsulfenic acid, as reported in Scheme 1, or other acids
derived therefrom)1c,f-h and accordingly it was found that the
addition of an amine (N-methylpiperidine, N,N-dimethylben-
zylamines, or ethyl diisopropylamine) restored the original rose
color of the solution.10
The mass balance with respect to the starting material was
excellent (g95%) when benzaldehyde and sulfoxidation prod-
ucts (sulfone and sulfoxide) are considered. On the other hand,
a poor mass balance was observed for the sulfur-containing
products coming from the cleavage reaction (benzyl benzylth-
iosulfinate and benzyl benzylthiosulfonate). Probably under
these reaction conditions the main cleavage product, sulfenic
acid, is oxidized in large part to acid products that are difficult
to identify and isolate.13
Since the pioneering work by Foote and co-workers,14
1
experiments of O2-promoted sulfoxidation in the presence of
diphenyl sulfoxide have allowed considerable insight into the
reaction mechanism as this species functions as an excellent
trap for the persulfoxide, the key reaction intermediate. Thus,
it is of interest to look at the effect of adding diphenyl sulfoxide
during the oxidation of dibenzyl sulfide. In particular, we were
concerned with the effect of diphenyl sulfoxide on the benzal-
dehyde/dibenzyl sulfone ratio. According to the mechanism in
Scheme 1, trapping the persulfoxide should lower the yields of
aldehyde and sulfone, but leave their ratio unchanged. Substan-
tial formation of dibenzyl sulfoxide should also be observed.14
When the oxidation of dibenzyl sulfide was carried out under
the usual reaction conditions but in the presence of 0.5 M
diphenyl sulfoxide, a significant amount of benzyl sulfoxide
was formed, as expected (Table 5, entry 2). However, the
presence of diphenyl sulfoxide does not seem to significantly
affect the amount of benzaldehyde formed whereas a significant
decrease of benzyl sulfone is observed (the aldehyde/sulfone
Since it is well-known that the presence of acid species
catalyzes the sulfoxide formation in the 1O2-promoted oxidation
of sulfides,1e,11 the relative increase in the yield of benzyl
sulfoxide shown in entries 1-3 of Table 3 can be easily
explained. Indeed, when the oxygenation reaction was performed
(9) For this product study, rose Bengal was preferred to TPP since it
allowed us to use shorter irradiation time.
(10) The same behavior was exhibited by TPP when it was used as the
photosensitizer in the photooxidation of 4. A rapid change in color of the
solution (from purple to green) was observed and the original color was
restored after addition of a base.
(11) (a) Clennan, E. L.; Greer, A. J. Org. Chem. 1996, 61, 4793. (b)
Bonesi, S. M.; Albini, A. J. Org. Chem. 2000, 65, 4532. (c) Bonesi, S. M.;
Fagnoni, M.; Monti, S.; Albini, A. Photochem. Photobiol. Sci. 2004, 3,
489.
(12) The sterically encumbered ethyl diisopropylamine was used to
minimize the 1O2 quenching by the amine (kT ) 7.56 × 107 M-1 s-1 in
MeCN at 22 °C; lower than the values generally shown by tertiary aliphatic
amines).8
(13) Davis, F. A.; Billmers, R. L. J. Am. Chem. Soc. 1981, 103, 7016.
(14) (a) Gu, C.-L.; Foote, C. S.; Kacher, M. L. J. Am. Chem. Soc. 1981,
103, 5949. (b) Liang, J.-J.; Gu, C.-L.; Kacher, M. L.; Foote, C. S. J. Am.
Chem. Soc. 1983, 105, 4717.
J. Org. Chem, Vol. 72, No. 25, 2007 9585

Baciocchi et al.
TABLE 5. 1O2-Promoted Oxidation of Dibenzyl Sulfide in the Presence of Ethyl Diisopropylamine
a Referred to the initial amount of substrate and determined by HPLC analysis. Average of seven independent determinations. b Determined by GC
analysis due to peak overlapping (Ph2SO) in HPLC analysis.
by electron-withdrawing substituents and from the Hammett plot
a positive F value (+0.47) can be calculated, slightly larger than
that found for tertiary benzylamines.15
According to the mechanism in Scheme 1, the key step in
the oxidation leading to the C-S bond cleavage products should
be the one involving the conversion of the persulfoxide into
the hydroperoxysulfonium ylide (path a). If this hypothesis is
correct (see later, however), it would seem that steps a in
Schemes 1 and 2 are characterized by quite similar transition
state structures.
Whereas this situation is certainly possible (for the persul-
foxide 1 a diradical structure has also been considered1e,16,17
and path a of Scheme 1 might have substantial hydrogen atom
transfer character and exhibit a small sensitivity to electronic
effects), it does not seem, however, probable on the basis of
the product study carried out with dibenzyl sulfide, whose results
are reported in Tables 4 and 5. Accordingly, this study provides
us with two pieces of important mechanistic information. The
first information is that the R-hydroperoxy sulfide (3), in
whatever way formed, must produce the aldehyde exclusively
by the intramolecular mechanism (Scheme 1, paths e-g).
Accordingly, the other mechanism leading to the aldehyde from
3 (paths c and d) appears not to be operating to a significant
extent since dibenzyl sulfoxide and dibenzyl disulfide, which
should accompany the formation of the aldehyde,1f,g are not
formed (dibenzyl disulfide) or are formed in negligible amounts
(dibenzyl sulfoxide).
The second, more crucial, information is that, in the presence
of 0.5 M Ph2SO, benzaldehyde yield remains substantially
constant, whereas the aldehyde/sulfone ratio increases signifi-
cantly from 4.0 to 4.9, largely outside experimental error. This
result is not in line with the mechanism described in Scheme
1. Accordingly, since Ph2SO should trap the intermediate
persulfoxide,14 we would expect, as already mentioned, that the
presence of Ph2SO should lead to lower yields of both aldehyde
and sulfone, whereas the aldehyde/sulfone ratio is unchanged,
which is not observed. Clearly, the effect of Ph2SO indicates
that persulfoxide is an intermediate for the formation of sulfone,
but not for the formation of the aldehyde, the C-S bond
FIGURE 1. Plot of the logarithms of the molar ratios of 4-X-C6H4-
CHO/PhCHO (average of data from both thermal and photochemical
reactions) vs Hammett σp in the oxidation of benzyl sulfides 4a-c
1
(squares) and dibenzylcyclohexylamines 5a-c (circles) with O2.
ratio increases from 4.0 to 4.9, largely outside experimental
error; compare entries 1 and 2 in Table 5).
Discussion
For the reaction of dibenzylamines with 1O2, the 4-X-C6H4-
CHO/PhCHO ratios shown in Table 2 indicate that the C-N
bond cleavage reaction is favored, even though to a very small
extent, by electron-withdrawing group on the aromatic ring.
When the log of these ratios (average values of photochemical
and thermal experiments) are plotted against the substituent
Hammett σp values, a nice plot is obtained (Figure 1) from which
a positive F value (+0.27) can be calculated. These results are
in line with the mechanism suggested in Scheme 2. The key
step in the formation of PhCHO or substituted PhCHO is the
hydrogen atom transfer inside the exciplex leading to the radical
couple. Since in the exciplex there is a partial transfer of charge,
the hydrogen atom transfer might have some character of proton
transfer, which may account for the observed positive F value.
Collapse of the radical pair by an intracomplex electron-transfer
reaction is consistent with the observation of almost identical
yields of benzaldehydes and H2O2.
(15) In principle, also the sulfone yield should be influenced by the
substituents. However, the substituent effect on the yield of sulfone can be
reasonably neglected given the small yield of this product with respect to
that of the aldehyde. Thus, the effects are probably within experimental
error.
The 4-X-C6H4CHO/PhCHO ratios for the substituted dibenzyl
sulfides are displayed in Table 3. Clearly, the C-S bond
cleavage reactions of dibenzyl sulfides exhibit a structure
dependence that is very close to that observed for the tertiary
benzylamines. In this case too, the cleavage reaction is favored
(16) Bonesi, S. M.; Torriani. R.; Mella, M.; Albini, A. Eur. J. Org. Chem.
1999, 1723.
(17) (a) Calculations, however, favor the zwitterionic structure.17b,c (b)
Greer, A.; Jensen, F.; Clennan, E. L. J. Org. Chem. 1996, 61, 4107. (c)
Clennan, E. L. Acc. Chem. Res. 2001, 34, 875.
9586 J. Org. Chem., Vol. 72, No. 25, 2007

Singlet Oxygen Promoted C-Heteroatom Bond CleaVage
SCHEME 3
intersystem crossing) is that of being converted into the
persulfoxide intermediate from which all sulfoxidation chemistry
generally observed ensues. It is, however, possible that with
systems where a relatively weak C-H bond is adjacent to sulfur
as in benzyl sulfides, an additional route (namely, path b in
Scheme 3) becomes available to the exciplex that can favorably
compete with conversion to persulfoxide (Scheme 3, path a).22
The formation of dibenzyl sulfone in the reaction of dibenzyl
sulfide with 1O2 shows that the exciplex, as already mentioned,
can also be converted to the persulfoxide from which the sulfone
is formed through the hydroperoxysulfonium ylide, as shown
in Scheme 1.23 The hydroperoxysulfonium ylide might also
undergo the Pummerer rearrangement, but this process seems
to be, at most, a very minor route. Accordingly, as already noted,
dibenzyl sulfoxide is formed in very small amounts and the yield
of aldehyde is almost insensitive to the presence of Ph2SO.
The rate constant values for formation of benzaldehydes also
deserve some comment. Considering the p-methyl derivatives
4b and 5b, kR(ArCHO) turns out to be about 10 times larger for
the sulfide than for the amine (3.0 × 106 vs 2.8 × 105 M-1
s-1). When the rates for total quenching of 1O2 (kT) of the two
cleavage product. In other words, sulfone and the aldehyde must
come from different routes.18
This conclusion together with the already noted strong
similarity of dibenzyl sulfide structural effects with those
observed for tertiary benzylamines19 led us to suggest that the
two systems react with 1O2 to form benzaldehydes by a similar
mechanism. Thus, as shown in Scheme 3, with sulfides too an
exciplex might be formed that, in addition to undergoing
intersystem crossing and being converted to persulfoxide, can
also undergo an intracomplex hydrogen transfer to form the
radical couple 6, analogous to that formed from tertiary
alkylamines (Scheme 2).20
substrates are considered, 1.37 × 107 and 4.05 × 106 M-1 s-1
,
respectively, it turns out that the fraction of exciplex leading to
benzaldehyde is 22% for dibenzyl sulfide 4b and 7% for
dibenzylcyclohexylamine 5b. The difference is not very large
and a role might be played by steric and stereoelectronic effects
that are probably different in the two systems.
The absence of H2O2 formation in the reaction of sulfides
(see Results), however, suggests that the collapse of 6 does not
occur by an electron-transfer reaction as proposed in the case
of amines (Scheme 2, path b).2j Since R-alkylthio carbon radicals
are much less easily oxidizable than R-dialkylamino carbon
radicals, a reasonable hypothesis is that this radical is not
Summary and Conclusions
The reactions of 1O2 with dibenzyl sulfides and dibenzylcy-
clohexylamines in MeCN have been investigated. With diben-
zylamines, the reaction leads to the formation of benzaldehyde
and the secondary amine as the exclusive reaction products.
Benzaldehyde is formed as the major product with dibenzyl
sulfides, accompanied by the formation of S-oxidation products.
Our investigation has led to the following main observations.
First, a careful product study of the reaction of dibenzyl sulfide
with 1O2 has shown that, when the reaction is carried out in the
presence of a base, dibenzyl sulfoxide is formed in almost
negligible amounts with respect to benzaldehyde and sulfone.
Most interestingly, addition of Ph2SO does not significantly
modify the benzaldehyde yield, whereas that of sulfone is
significantly lowered. Moreover, no dibenzyl disulfide is present
among the sulfur-containing cleavage products. Clearly, these
results do not fit the mechanism generally accepted (Scheme
1) to explain the formation of cleavage products in the oxidation
of sulfides by 1O2, but force us to suggest that dibenzyl sulfone
and benzaldehyde are formed by different routes. The second
observation comes from a study of the reactions of p-substituted
dibenzyl sulfides (4a-c) and dibenzylcyclohexylamines (5a-
c), which shows that the cleavage reactions in the two systems
respond in a similar way to electronic effects, the cleavage being
favored by electron-withdrawing substituents in both cases. The
effects are, however, very small as indicated by the F values
•
oxidized by HO2 , but undergoes a coupling reaction to form
the R-hydroperoxy sulfide 7 (Scheme 3, path c). The latter may
then be converted to PhCHO by the intramolecular path shown
in Scheme 1 (path e).1f,g
The possible involvement of exciplexes in the oxygenation
1
of sulfides by O2 has received very scarce attention, even
though these species have been called into play on several
occasions, particularly by Gorman.21,1e The probable reason is
that the most common situation in these reactions is the one
where the only fate of the exciplex (in addition to undergoing
(18) (a) A rewiever has suggested that the presence of 0.5 M diphenyl
sulfoxide (molar fraction <0.05) might significantly vary the polarity of
the medium, thus influencing the partition of the hydroperoxysulfonium
ylide 2 between benzaldehyde and benzyl sulfone (Scheme 1). This
possibility, however, appears very unlikely when it is considered that the
ET(30) polarity value of solutions of DMSO in MeCN remains practically
the same as that of pure MeCN up to 0.1 DMSO molar fraction.18b,c (b)
Bosch, E.; Rose´s, M. J. Chem. Soc., Faraday Trans. 1992, 88, 3541. (c)
Rose´s, M.; Ra`fols, C.; Ortega, J.; Bosch, E. J. Chem. Soc., Perkin Trans.
2 1995, 1607.
(19) In this respect, it can also be noted that very similar intramolecular
kinetic deuterium isotope effects have been found in the 1O2 promoted
benzaldehyde formation from benzyldimethylamine (3.5)2j and benzyl ethyl
sulfide (2.95)1e in MeCN.
(20) It must be acknowledged that the proposal that the C-S cleavage
products, in the oxidation of benzyl sulfides, might come from a route
parallel to that leading to sulfoxides and sulfones involving exciplexes
formation has been already put forward by the Albini group.1e However,
no consideration has been given later to this hypothesis in a work by the
same group on the same subject.1h Apart from this, the mechanism following
exciplex formation proposed by Albini and co-workers is different from
that presented here.
(22) It cannot be excluded, however, that formation of the exciplex occurs
in a parallel path along with the formation of persulfoxide. Generally, the
exciplex undergoes intersystem crossing exclusively, but with substrates
with relatively weak R C-H bonds, it might lead to products.
(23) The fact that in the presence of diphenyl sulfoxide the yield of benzyl
sulfoxide becomes close to that of benzaldehyde (Table 5) shows that the
conversion of exciplex to persulfoxide is, at least, as important as that leading
to benzaldehyde.
(21) (a) Gorman, A. A. AdV. Photochem. 1992, 17, 218. (b) Clennan, E.
L.; Oolman, K.; Yang, K.; Wang, D.-X. J. Org. Chem. 1991, 56, 4286.
J. Org. Chem, Vol. 72, No. 25, 2007 9587

Baciocchi et al.
three fractions (50 mL each) of diethyl ether. The combined organic
layers were washed with water and brine and dried over anhydrous
MgSO4. After rotary evaporation, the residue was purified by flash
chromatography (silica gel; gradient from hexane to hexane/ethyl
acetate 4:1).
(+0.47 for sulfides and +0.27 for amines). On the basis of the
above results, it is suggested that the 1O2-promoted C-S bond
cleavage in benzyl sulfides occurs by the same mechanism as
that already proposed for the corresponding C-N bond cleavage
reactions in benzylamines: Formation of an exciplex between
the substrate and 1O2 which undergoes an intracomplex hydro-
gen transfer leading to a radical pair. With benzylamines the
radical pair undergoes an ET process leading to an iminium
cation from which benzaldehyde can be formed (Scheme 2).
With sulfides, the radical pair instead can collapse to the
R-hydroperoxy sulfide 3 (R-alkylthio carbon radicals are much
less easily oxidizable than R-dialkylamino carbon radicals) from
which benzaldehyde can be formed as described in Scheme 1,
path e. Part of the exciplex can also be converted into the
persulfoxide 1, which leads to the sulfone through the hydro-
peroxysulfonium ylide 2, as shown in Scheme 1. It also has
been determined that the fraction of exciplex that is converted
to aldehyde is quite close for the two systems (ca. 20% with
sulfides and ca. 7% with amines).
Determination of 1O2 Total Quenching Rate Constants (kT).
1O2 was produced in MeCN by energy transfer to O2 from the triplet
state of phenalenone (7.0 × 10-5 M), generated by excitation at
355 nm from a Nd:YAG laser (pulse width ca. 7 ns and energy <3
mJ per pulse). The phosphorescence emission of 1O2 emerging from
the cuvette passed through a cutoff filter at 1050 nm and three
pieces of gelatin cutoff filter at 870 nm and was detected by a
germanium diode detector (5 mm diameter). After amplification
with a two-stage homemade amplifier (ca. 100 MHz bandwidth,
14 dB), the output of the diode was fed into a digital signal analyzer
and computer stored and analyzed.27 Rate constants for the total
1
1
quenching of O2 (kT) were determined from the decrease of O2
emission lifetime in O2-saturated solvent, in the presence of various
amounts of substrates (0-8 × 10-3 M). All measurements were
carried out at 22 ( 2 °C (some examples are reported in the
Supporting Information).
Oxidation of Amines 5a-c by Photochemically Generated
Singlet Oxygen. Photooxygenation reactions were carried out in a
photoreactor equipped with ten 4500-6000 Å fluorescence lamps
(14 W each). A 4-mL sample of a solution containing the
dibenzylcyclohexylamine (10-2 M) and TPP (10-4 M, added as
0.2 mL solution in chloroform) in dry MeCN was irradiated at 25
°C in a thermostated jacketed tube for 15 (5a,b) or 30 min (5c)
under slight oxygen bubbling. Further aliquots of TPP (each equal
to the initial amount) were added in order to replace the bleached
photosensitizer. An internal standard (4-methylbenzophenone) was
added. Products analysis was carried out by GC and GC-MS
(comparison with authentic specimens). Blank experiments, carried
out under the same reaction conditions but in the absence of TPP
or O2, showed the absence of reaction products.
The amount of H2O2 was quantitatively determined by titration
with iodide ion; the solution was treated, after dilution, with an
excess of KI and a few drops of AcOH. The amount of I3- formed
was determined from UV spectra (ꢀ ) 2.50 × 104 M-1 cm-1 at
λmax ) 361 nm).28 Blank experiments, performed in the absence of
the substrate, showed no formation of H2O2.
Experimental Section
Materials. Commercially available benzyl sulfide (4) was
purified by silica gel chromatography (hexane/ethyl acetate 9:1)
and then recrystallized from hexane to eliminate traces of benzyl
disulfide. Phenyl sulfoxide was commercially available and purified
by recrystallization from hexane. Ethyl diisopropylamine was
commercially available and distilled before use. The photosensitizers
rose bengal disodium salt and tetraphenylporphyrin were com-
mercially available and were used as received. 1,4-Dimethylnaph-
thalene endoperoxide was prepared according to a literature
procedure.5 Benzaldehyde, 4-methoxybenzaldehyde, 4-methylben-
zaldehyde, 4-cyanobenzaldehyde, benzyl sulfoxide, and benzyl
sulfone were commercially available whereas benzyl benzylthio-
sulfinate24 and benzyl benzylthiosulfonate25 were prepared according
to literature procedures. Acetonitrile (HPLC plus grade) was
refluxed (3 h) and then distilled over calcium hydride. CD3CN
(99.8% D atom) was distilled over calcium hydride and stored over
molecular sieves.
General Procedure for the Synthesis of Benzyl Sulfides 4a-
c. Benzyl sulfides 4a-c were prepared by reaction of benzyl
mercaptane with 4-methoxybenzyl chloride, 4-methylbenzyl chlo-
ride, and 4-cyanobenzyl bromide, respectively. A 20 mL solution
of the 4-substituted benzyl halide (15 mmol) in EtOH was slowly
added to a stirred solution of benzylmercaptane (2.0 g, 16 mmol)
and KOH (1.0 g, 18 mmol) kept at 80 °C under Ar atmosphere.
The mixture was then refluxed for 2 h, after which it was cooled
to room temperature, poured into 200 mL of water, and extracted
with three fractions (100 mL each) of chloroform. The combined
organic layers were washed with 25% aqueous KOH, water, and
brine and dried over anhydrous MgSO4. After rotary evaporation,
the residue was purified by flash chromatography.
General Procedure for the Synthesis of Dibenzylcyclohexy-
lamines 5a-c. Dibenzylcyclohexylamines 5a-c were prepared by
reaction of benzylcyclohexylamine26 with 4-methoxybenzyl chlo-
ride, 4-methylbenzyl chloride, and 4-cyanobenzyl bromide, respec-
tively. The benzyl halide (8.3 mmol) in MeCN (5 mL) was slowly
added to a stirred solution of benzylcyclohexylamine (1.0 g, 5.3
mmol) and K2CO3 in MeCN (20 mL) at room temperature. The
mixture was then refluxed for 6 h, after which it was cooled to
room temperature, poured into 40 mL of water, and extracted with
Oxidation of amines 5a-c by Thermally Generated Singlet
Oxygen. A 2-mL sample of amine (10-2 M) and 1,4-dimethyl-
naphthalene endoperoxide (2 × 10-2 M for 5a,b and 0.1 M for 5c)
in dry MeCN was heated at 40 °C (water bath) in the dark for 4 h.
An internal standard (4-methylbenzophenone) was added, and the
mixture was analyzed by GC and GC-MS. No product formation
was observed in blank experiments, carried out in the absence of
1,4-dimethylnaphthalene endoperoxide and in oxygen-saturated
solutions.
Oxidation of Sulfides 4a-c by Photochemically Generated
Singlet Oxygen. Photooxygenation reactions were carried out in a
photoreactor equipped with ten 4500-6000 Å lamps (14 W each).
A 2-mL sample of a solution containing the sulfide (10-2 M) and
TPP (10-4 M, added as 0.1 mL solution in CDCl3) in dry
O2-saturated CD3CN was irradiated at 25 °C in a thermostated
jacketed tube for 15 min. An internal standard (bibenzyl) was added.
Product analysis was carried out by 1H NMR spectroscopy (by GC
analysis, the products benzyl benzylthiosulfinate and benzyl ben-
zylthiosulfonate undergo thermal degradation forming a substantial
amount of benzaldehyde making this analytical method inappropri-
ate). When benzyl 4-methylbenzyl sulfide (4b) was irradiated for
30 min under the same reaction conditions but in the presence of
ethyl diisopropylamine (5 × 10-3 M), the 4-methylbenzaldehyde/
(24) Furukawa, N.; Morishita, T.; Akasaka, T.; Oae, S. J. Chem. Soc.,
Perkin Trans. 2 1980, 432.
(25) Boekman, K.; Voegtle, F. Chem. Ber. 1981, 114, 1048.
(26) Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G. Tetrahedron
Lett. 1990, 31, 5595.
(27) Elisei, F.; Aloisi, G. G.; Lattarini, C.; Latterini, L.; Dall’Acqua, F.;
Guiotto, A. Photochem. Photobiol. 1996, 64, 67.
(28) Fukuzumi, S.; Kuroda, S.; Tanaka, T. J. Am. Chem. Soc. 1985, 107,
3020.
9588 J. Org. Chem., Vol. 72, No. 25, 2007

Singlet Oxygen Promoted C-Heteroatom Bond CleaVage
benzaldehyde molar ratio appeared to be unaffected by the presence
of the amine. Blank experiments, carried out under the same reaction
conditions but in the absence of TPP or O2, showed the absence of
detectable products.
reported above by irradiating the solution with six lamps for 5 min.
In the reactions with diphenyl sulfoxide the intense peak of this
compound superimposed with that of dibenzyl sulfone, therefore
the quantitative analysis of the latter was carried out by GC. Blank
experiments, carried out under the same reaction conditions but in
the absence of RB or O2, showed the absence of detectable products.
Irradiation of ethyldiisopropylamine under the same reaction
conditions but in the absence of 5 showed the absence of any
detectable reaction product.
Oxidation of Sulfides 4a-c by Thermally Generated Singlet
Oxygen. A 1-mL sample of sulfide (10-2 M) and 1,4-dimethyl-
naphthalene endoperoxide (6.7 × 10-3 M) in CD3CN was heated
at 40 °C (water bath) in the dark for 4 h. An internal standard
1
(bibenzyl) was added, and the mixture was analyzed by H NMR
spectroscopy. Blank experiments, carried out in the absence of 1,4-
dimethylnaphthalene endoperoxide and in oxygen-saturated solu-
tions, showed the absence of reactivity.
Acknowledgment. MUR, University “La Sapienza” of
Rome, and University of Perugia are thanked for financial
support.
Oxidation of Benzyl Sulfide (4) by Photochemically Gener-
ated Singlet Oxygen. Photooxygenation reactions were carried out
in a photoreactor equipped with four 4500-6000 Å lamps (14 W
each). A 4-mL sample of a solution containing the sulfide (10-2
M) and rose bengal (10-4 M) in dry O2-saturated MeCN was
irradiated at 25 °C in a thermostated jacketed tube for 1, 5, and 15
min. An internal standard (4-methylbenzophenone) was added.
Product analysis was carried out by HPLC (C18 column, solvent
gradient from MeOH/H2O 75:25 to MeOH). Photoreactions in the
presence of ethyl diisopropylamine (5 × 10-3 M) alone or with
phenyl sulfoxide (0.5 M) were carried out under the conditions
Supporting Information Available: Characterization and spec-
tral properties of compounds 4a-c and 5a-c, determination of
k
r(ArCHO) rate constants, and dependence of observed rate constants
1
of O2 quenching on concentration of 4, 4a, 5a, and i-Pr2NEt in
MeCN. This material is available free of charge via the Internet at
http://pubs.acs.org.
JO701641B
J. Org. Chem, Vol. 72, No. 25, 2007 9589