Activation of molecular oxygen by S-radicals: Experimental and computational studies on a novel oxidation of alkynes to α-diketones

Article abstract of DOI:10.1039/b815358b

Thiylperoxyl radicals 14 are the suggested reactive key-intermediates in the oxidation of bis-aromatic alkynes to α-diketones using molecular oxygen "activated" by thiyl radicals. The Royal Society of Chemistry.

Full text of DOI:10.1039/b815358b

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Activation of molecular oxygen by S-radicals: experimental and
computational studies on a novel oxidation of alkynes to a-diketonesw
Kristine J. Tan and Uta Wille*
Received (in Cambridge, UK) 3rd September 2008, Accepted 30th September 2008
First published as an Advance Article on the web 16th October 2008
DOI: 10.1039/b815358b
Thiylperoxyl radicals 14 are the suggested reactive key-inter-
mediates in the oxidation of bis-aromatic alkynes to a-diketones
using molecular oxygen ‘‘activated’’ by thiyl radicals.
The development of new oxidation procedures that make use
of the environmentally most benign and cheapest oxidant, i.e.
molecular oxygen, O2, are in high demand. We wish to report
Scheme 1
Table 1 Experimental conditions and results for the PhSꢀ mediated
oxidation of 6 in the presence of O2
preliminary results on the mechanism of a newly discovered
reaction that enables oxidation of the CRC triple bond in bis-
aromatic alkynes to a-diketones using O2 ‘‘activated’’ by thiyl
radicals under metal-free conditions. This reaction could be a
useful addition to the toolbox of synthetic methodology, as it
provides a mild alternative to the existing procedures for
alkyne - a-diketone transformations, which generally require
transition metals and/or drastic reaction conditions, and are
often impaired by total cleavage of the former alkyne bond.1
During our studies on exploring S-centred radicals in self-
terminating radical cyclizations,2 we observed an unexpected
influence of O2 on the reaction outcome (Scheme 1).
Reaction of cyclodecyne (1) with phenylthiyl radicals, PhSꢀ,
under exclusion of O2 led to the bicyclic thioethers 2 and 3 in
70% combined yield through a radical addition/translocation
cascade (not shown). In the presence of O2 the bicyclic ketones
4 and 5 were obtained as major by-products.3 Since their
formation through further oxidation of 2/3 could be clearly
excluded, the finding that 4/5 must arise from an independent
pathway raised our curiosity whether S-centred radicals in
combination with O2 might promote oxidation of alkynes to
carbonyl compounds. While reactions of S-radicals in the
presence of O2 have been widely investigated with alkenes,4
only few studies involving alkynes are available.5
We explored the PhSꢀ/O2 mediated alkyne oxidation using
diphenylacetylene (6) as model substrate. PhSꢀ was generated
in the presence of O2 by various methods, namely (i) photo-
decomposition of diphenyl disulfide,4 (ii) autoxidation of
thiophenol, PhSH,5b (iii) reaction of PhSH with triethyl-
borane, Et3B,6 and (iv) anodic oxidation of PhSH5b (for details
see ESIw). A selection of relevant results is shown in Table 1.
We were pleased to find that photogenerated PhSꢀ leads to
formation of the a-diketone 7 (Scheme 2) in 66% yield in
benzene (data not shown), however, the significant amount of
a,b
Yieldc (%)
Entry
[6]/mM
[PhSH]/mM Solvent
7
8a
(a) PhSꢀ from autoxidation of PhSHd
1
2
100
100
100
100
100
100
100
100
100
100
400
400
400
400
400
400
400
400
400
400
C6H6
41 (49) 35 (41)
48 (52) 33 (36)
53 (55) 27 (28)
MeC6H5
i-C8H18
C6F6
MeCN
Acetone
CH2Cl2
CCl4
3
4e
5
22 (53)
9 (22)
23 (26) 61 (69)
36 (41) 42 (47)
19 (25) 36 (46)
19 (37) 18 (34)
32 (32) 47 (47)
6
7e
8e
9f
10f
MeOH
MeOH–C6H6 46 (46) 47 (47)
(1:4)
C6H6
11g
100
400
1
0
(n.d.)h (n.d.)
57 (74) 13 (17)
43 (50) 22 (26)
12i
100
100
400
400
C6H6
C6H6
13e,j
(b) PhSꢀ from anodic oxidation of PhSHk,l
14
15
50
400
200
100
MeCN
MeCN–C6H6 18 (32)
(1:4)
7 (27)
0 (0)
5 (8)
16
50
500
MeCN–C6H6 35 (41) 18 (21)
(1:4)
a
b
O2-stream, unless otherwise stated. Yields determined by quantita-
c
tive GC using n-hexadecane as internal standard. Data in
d
parentheses are yields based on consumed 6. Reaction time 16–23 h.
e
f
Formation of significant amounts of 9. Formation of 8b as by-
g
product. In the presence of Et3N (100 mM). n.d = not determined.
i
h
j
80 psi. Reaction mixture pre-saturated with O2 (see
p(O2)
k
=
l
text). Reaction time 5–6 h. Et4NOTs (0.15–0.2 M) as supporting
electrolyte.
benzoic acid formed as by-product from photo-degradation of
77 rendered this radical source not suitable for our purposes.
ARC Centre of Excellence for Free Radical Chemistry and
Biotechnology, School of Chemistry and BIO21 Institute, The
University of Melbourne, Victoria, 3052, Australia
w Electronic supplementary information (ESI) available: Typical ex-
perimental conditions, spectroscopic data for 8b, Gaussian archive
entries for relevant optimized ground and transition state structures.
See DOI: 10.1039/b815358b
Scheme 2
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 6239–6241 | 6239

View Article Online
Reaction of 6 with PhSꢀ generated by slow autoxidation of a
fourfold excess of PhSH in an O2-stream gave diketone 7,
together with the a-carbonyl thioether 8a8 and minor amounts
of alkene 9. In all experiments formation of diphenyl disulfide,
resulting from PhSꢀ recombination, was observed. 7 and 8a did
not interconvert under the experimental conditions, suggesting
independent pathways for their formation (similar to the findings
outlined in Scheme 1). In nonpolar solvents such as benzene,
toluene and isooctane (entries 1–3) good yields of 7 were
obtained with reduced formation of 8a. Essentially similar yields
were obtained in hexafluorobenzene (entry 4), but the conversion
of 6 was significantly slower. Reactions performed in more polar
solvents or in haloalkanes (entries 5–8), resulted in lower yields of
7, with increased formation of the thioether 8a and alkene 9. In
methanol or methanol–benzene mixtures conversion of 6 was
faster, but the a,a-methoxythioether 8b was formed as additional
by-product through an unknown pathway (entries 9, 10). Neither
elevated temperatures, nor a larger excess of PhSH gave a higher
7/8a % yield ratio (not shown). Remarkably, a-diketone 7 and/
or the thioether 8a were not formed in the presence of base in
order to facilitate autoxidation of PhSH through deprotonation
(entry 11). Also, very low yields of diketone 7 (r3%) were
obtained, when alkyne 6 was used in excess (not shown).
The reaction of 6 with PhSꢀ generated using the Et3B/O2
system was fast, but resulted in predominant formation of the
unwanted by-product 8a (data not shown). On the other hand,
the a-diketone 7 was generally the major product in the
reactions involving electrogenerated PhSꢀ (entries 14–16).
We used the autoxidation system (method (ii)) to explore
the reaction mechanism. Interestingly, a time-dependent study
of the consumption of alkyne 6 using PhSH in at least 10-fold
excess in benzene revealed a pseudo-second order behavior.
Considering that [O2] in O2-saturated benzene solutions is ca.
9 mM,9 which was in the same order of magnitude as [6]t=0 in
these experiments (50 mM), we have: rate = k[6][PhSH][O2].
In order to examine the role of [O2], the reaction was
performed at an O2 pressure, p(O2), of 80 psi, resulting in a
dramatic increase of the yield of diketone 7 to 74% at the
expense of thioether 8a (entry 12). This shows that formation
of diketone 7 is strongly affected by [O2]. On the other hand,
the reaction performed in benzene that has been saturated with
O2 prior to the experiment gave slightly smaller amounts of 8a
with the yield of 7 being unaffected, compared to the reaction
performed in an O2-stream (entry 13 vs. 1). Our finding that
no reaction occurred, when the electron-poor alkyne 4,40-
dinitrodiphenylacetylene was treated with PhSH/O2 under
autoxidation conditions, whereas the more electron-rich
4,40-dimethoxydiphenylacetylene reacted readily to yield the
respective a-diketone and a-carbonyl thioether (not shown)
suggests that addition of an electrophilic radical (i.e. S- or
O-centred) to the CRC triple bond is the initial reaction step.
A mechanism for formation of a-carbonyl thioether 8a (and
alkene 9) has been proposed in literature (Scheme 3).5b
Scheme 3 Proposed mechanism for formation of 8a and 9. BHLYP/
cc-pVTZ energies in kJ molꢂ1 (for R = Me; ZPC included).
alkene 9, which is attacked by a second PhSꢀ to form alkyl
radical 11. Trapping by O2, followed by hydrogen abstraction
and decomposition of hydroperoxide 13 gives thioether 8a.
In a strongly competing pathway, PhSꢀ can undergo rever-
sible oxidation to the thiylperoxyl radical 14 (Scheme 4).12
This is supported by computations, which show that with
Ez = 52.7 kJ molꢂ1 and DE = 33.2 kJ molꢂ1 the energetic
parameters for formation of peroxyl radical 14 are very similar
to those for formation of vinyl radical 10 shown in Scheme 3.
Thiylperoxyl radicals 14 can rearrange to sulfonyl radicals 15
in an exothermic process,4 for which a rate constant of 2 ꢃ
103 sꢂ1 was determined.13 Our BHLYP/cc-pVTZ energies of
Ez = 132.4 kJ molꢂ1 and DE = ꢂ232.9 kJ molꢂ1 in fact
support this trend. Although the spin density in sulfonyl
radicals of type 15 is delocalized over the entire sulfonyl moiety,
they add to p systems via sulfur.4 It is therefore unlikely that 15
is directly involved in the transformation 6 - 7.z However, 15
can react with O2 to give the sulfonylperoxyl radical 16 in an
exothermic (calculated DE = ꢂ23.3 kJ molꢂ1) and practically
diffusion controlled process (k ca. 109 Mꢂ1 ꢂ1),4,12a for which
s
we were not able to locate a transition state computationally.
Addition of the peroxyl radicals 14 or 16 to 6 leads to vinyl
radicals 17a,b, which possess a very labile O–O bond. Both
Initial PhSꢀ addition leads to vinyl radical 10. BHLYP/cc-
pVTZ computations10 using 2-butyne (R = Me) as a simplified
model for 6 (R = Ph) revealed that this step is associated with
an activation barrier, Ez, of 54.1 kJ molꢂ1 and is endothermic
by DE = 23.3 kJ molꢂ1 (zero-point energy correction, ZPC,
included).11 Hydrogen abstraction from PhSH leads to the
Scheme 4 Proposed mechanism for formation of 7. BHLYP/cc-pVTZ
energies in kJ molꢂ1 (for R = Me; ZPC included).
ꢁc
This journal is The Royal Society of Chemistry 2008
6240 | Chem. Commun., 2008, 6239–6241

View Article Online
radical additions are energetically feasible with a calculated Ez
dimer of the sulfinyl radical PhSꢀQO, i.e. the thiosulfonate
PhSSO2Ph.18 PhSꢀQO would be released in the conversion
17a - (3)20 via either 18a or (3)19 (see Scheme 4).8
of 72.1 or 50.5 kJ molꢂ1 and DE of ꢂ5.4 or ꢂ15.7 kJ molꢂ1
,
respectively. The vinyl radicals 17a,b could be converted into 7
through various routes. A pathway involving additional O2
could initially lead to peroxyl radicals 18a,b in a strongly
To conclude, we have discovered a new method for the
oxidative conversion of CRC triple bonds to a-diketones
using molecular oxygen activated by S-radicals. Mechanistic
studies indicate that this oxidation, which proceeds under very
mild conditions without cleavage of the former alkyne bond, is
initiated by intermediately formed thiylperoxyl radicals 14.
Further exploration of the synthetic potential of thiylperoxyl
radicals and this new methodology is currently underway.
This work was supported by the Australian Research
Council under the Centres of Excellence Program and the
Australian Partnership for Advanced Computing. Technical
support by Associate Professors Peter Tregloan, Steven Bottle
and Dr Kathryn Fairfull-Smith is gratefully acknowledged.
exothermic reaction (DE = ꢂ152.3 or ꢂ135.7 kJ molꢂ1
,
respectively), however, we could so far not locate a transition
state for this addition. Fragmentation of the peroxylic O–R00
bond in 18a,b leads to biradical (3)20, which could isomerize
(through intersystem crossing, ISC) to carbonyl oxide 21. The
fragmentation is not only exothermic, but shows also a very
low Ez for both 18a and 18b. Carbonyl oxides of type 21 are
suggested intermediates in the ozonolysis of alkynes, leading
to a-diketones.14 Alternatively, (3)20 could be formed via
biradical (3)19 (which is in resonance with an a-oxo carbene,
not shown) that results from exothermic g-cleavage of the
weak O–R00 bond in vinyl radical 17a,b,10 followed by trapping
with O2. Pathways that do not require additional O2 could
proceed through oxetene 22 resulting from homolytic cleavage
of the O–S bond in 17a,b in an exothermic reaction (according
to computations). Although we were not able to locate the
respective transition states for the reaction 17a,b - 22 yet, we
do not believe that the latter could compete with the (calcu-
lated) fast O–O bond scission in 17a,b - (3)19. Thus, a
concerted or stepwise rearrangement of 17a,b to radical inter-
mediate 23a,b, which undergoes homolytic O–S bond cleavage
in a subsequent step, is a more likely scenario in a mechanism
that would not require additional O2.
Notes and references
z The addition of 15 to CRC triple bonds via O is energetically unfav-
ourable (Ez = 104.4 kJ molꢂ1, DE = 22.9 kJ molꢂ1; BHLYP/cc-pVTZ).
y Conditions: Mercury lamp (Heraeus TQ150); l 4 500 nm (cut-off filter).
z According to BHLYP/cc-VTZ calculations, 14 is an electrophilic
radical, which interacts with the alkyne p orbitals in the transition state.
8 Sulfinyl radicals are also suggested products of the reaction between
thiylperoxyl radicals and thiols; see ref. 4.
1 Selected examples: C. Sheu, A. Sobkowiak, S. Jeon and
D. T. Sawyer, J. Am. Chem. Soc., 1990, 112, 879; P. Li,
F. H. Cheong, L. C. F. Chao, Y. H. Lin and I. D. Williams,
J. Mol. Catal. A: Chem., 1999, 145, 111; Y. H. Hwang and
C. P. Cheng, J. Chem. Soc., Chem. Commun., 1992, 317.
2 Selected examples: U. Wille, G. Heuger and C. Jargstorff, J. Org.
Chem., 2008, 73, 1413; U. Wille, J. Am. Chem. Soc., 2002, 124, 14.
3 K. J. Tan and U. Wille, to be published.
4 Overview: in S-Centered Radicals, ed. Z. B. Alfassi, John Wiley &
Sons, Chichester, 1999.
5 K. Griesbaum, A. A. Oswald and B. R. Hudson, Jr, J. Am. Chem.
Soc., 1963, 85, 1969; J. Yoshida, S. Nakatani and S. Isoe, J. Org.
Chem., 1993, 58, 4855.
Although our computational findings would suggest that
the pathways involving the sulfonylperoxyl radical 16 are
energetically slightly less favourable than those involving the
thiylperoxyl radical 14, a clear conclusion which of these two
radicals is responsible for the conversion 6 - 7 cannot be
drawn. Therefore, 16 was independently generated through
reaction of phenylsulfonyl chloride with superoxide15 in the
presence of alkyne 6 (not shown). Diketone 7 was indeed
formed, but the very low yield of 3% (GC) clearly cannot
account for the observed yields shown in Table 1.
6 M. Ueda, H. Miyabe, H. Shimizu, H. Sugino, O. Miyata and
T. Naito, Angew. Chem., Int. Ed., 2008, 47, 5600.
7 D. L. Bunbury and C. T. Wang, Can. J. Chem., 1968, 46, 1473.
8 A. R. Katrizky, L. Xie and L. Serdyuk, J. Org. Chem., 1996, 61,
7564.
The ‘‘bottleneck’’ for formation of sulfonylperoxyl radical
16 is the rearrangement 14 - 15 (see Scheme 4). This process
proceeds with high efficiency (F = 0.8) using light with l =
546 nm.16 However, when the autoxidation reaction (method
(ii)) was performed under irradiation,y similar yields of dike-
tone 7 were obtained as in control experiments performed
under strict exclusion of light. Therefore, we conclude that 16
is not involved in the conversion of alkyne 6 into diketone 7.
Finally, since peroxyl radicals are known to react with
organic compounds via H-abstraction to give hydroperoxides,
we explored whether the transformation 6 - 7 might be
simply the result of a non-radical epoxidation of the CRC
bond in 6 by a peroxide, followed by isomerization to an a-oxo
carbene (see above).17 However, reaction of alkyne 6 with
hydroperoxides, for example H2O2 or t-BuOOH, both in the
presence and absence of O2 did not lead to any formation of
diketone 7 after 24 h reaction time. We therefore suggest that
PhSOOꢀ (14) must be the reactive key-intermediate in the
oxidation of 6 to the a-diketone 7.z This is supported by the
GC/MS detection of a by-product, which can be assigned to a
9 S. L. Murov, I Carmichael and G. L. Hug, Handbook of Photo-
chemistry, Marcel Dekker, Inc., New York, Basel, Hong Kong,
2nd edn, 1993.
10 U. Wille and J. Andropof, Aust. J. Chem., 2007, 60, 420.
11 Rate constants between 0.83 ꢃ 104 Mꢂ1 sꢂ1 and 79 ꢃ 104 Mꢂ1 sꢂ1
were determined for the addition of PhSꢀ to terminal alkynes: O. Ito,
R. Omori and M. Matsuda, J. Am. Chem. Soc., 1982, 104, 3934.
12 Rate constants for this reaction are between 104–109 Mꢂ1 sꢂ1, with
PhSꢀ being on the lowest side of the reactivity range: D. Becker,
S. Summerfield, S. Gillich and M. D. Sevilla, Int. J. Radiat. Biol.,
1994, 65, 537; K. Schaefer, M. Bonifacic, D. Bahnemann and
K. D. Asmus, J. Phys. Chem., 1978, 82, 2777; O. Ito and
M. Matsuda, J. Am. Chem. Soc., 1979, 101, 1815.
13 X. Zhang, N. Zhang, H.-P. Schuchmann and C. von Sonntag,
J. Phys. Chem., 1994, 98, 6541.
14 Selected examples: N. C. Yang and J. Libman, J. Org. Chem.,
1974, 39, 1782; K. Griesbaum, Y. Dong and K. J. McCullough,
J. Org. Chem., 1997, 62, 6129.
15 M. Y. Park, S. G. Yang and Y. H. Kim, Heteroat. Chem., 2002, 13,
431.
16 Y. Razskazovskii, A.-O. Colson and M. D. Sevilla, J. Phys. Chem.,
1995, 99, 7993.
17 See for example J. Ciabattoni, R. A. Campbell, C. A. Renner and
P. W. Concannon, J. Am. Chem. Soc., 1970, 92, 3826.
18 M. Iino and M. Matsuda, J. Org. Chem., 1983, 48, 3108.
ꢁc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 6239–6241 | 6241