Autoxidative carbon-carbon bond formation from carbonhydrogen bonds

Article abstract of DOI:10.1002/anie.201000711

(Figure Presented) Only oxygen and acid! The oxidative coupling of xanthene and other activated benzylic compounds with carbon nucleophiles such as ketones, can be performed under ambient conditions without solvent by simply using oxygen and catalytic amounts of methanesulfonic acid. The proposed reaction mechanism involves substrate activation by formation of hydroperoxides; the method can therefore be regarded as an "autoxidative coupling reaction".

Full text of DOI:10.1002/anie.201000711

Communications
DOI: 10.1002/anie.201000711
Oxidative Coupling
Autoxidative Carbon–Carbon Bond Formation from Carbon–
Hydrogen Bonds**
ꢀron Pintꢁr, Abhishek Sud, Devarajulu Sureshkumar, and Martin Klussmann*
Cross-coupling reactions are of paramount importance in the
synthesis of larger organic molecules. Many successful
methods have been developed, but most of them rely on
special activating or leaving groups which create unwanted
waste and extra synthetic steps for their introduction. The
principle of “green chemistry” raises awareness for more
Scheme 1. Metal-free, acid-catalyzed aerobic coupling of xanthene (1)
and cyclopentanone (2).
environmentally friendly reactions that should avoid unnec-
essary waste, cost, and energy.[1] In this context, the activation
of carbon–hydrogen bonds for coupling reactions is a
À
promising strategy, as C H bonds are ubiquitous in organic
molecules and thus the introduction of activating groups
would no longer be necessary.[2] Under oxidative conditions,
low levels of waste. To develop a general coupling method, we
investigated this reaction in more detail (Table 1). The
coupling could be performed under a high partial pressure
À
À
two C H bonds can be coupled to form a new C C bond;
hydrogen acts as a “leaving group” and in the best case water
as the only waste product.[3] These oxidative coupling
reactions can be performed catalytically using only simple
metal salts or organic redox-active molecules for the activa-
Table 1: Optimization of reaction conditions for the synthesis of 3a by
the aerobic coupling of xanthene (1) and cyclopentanone (2).[a]
Entry
O2 [bar]
Acid (mol%)
Solvent
Yield [%][b]
À
tion of C H bonds. However, often synthetic oxidants, harsh
1
2
3
4
5
6
7
8
8[c]
1
TfOH (5)
TfOH (5)
TfOH (5)
–
–
–
–
–
–
–
–
–
76
64
59
0
0
20
80
76
69
63
65
85
76
conditions, and expensive reagents are required, thus dimin-
ishing the overall sustainability of the process. Here, we
present an oxidative cross-coupling reaction for the formation
0.2[d]
1
À
À
of C C bonds from two C H bonds, requiring neither metal
catalyst nor synthetic oxidant, only catalytic amounts of a
simple acid and elemental oxygen.
1
1
1
1
1
1
1
1
AcOH (5)
CF3CO2H (5)
CH3SO3H (5)
pTsOH (5)
CH3SO3H (5)
CH3SO3H (5)
CH3SO3H (5)
CH3SO3H (7)
CH3SO3H (10)
CH3SO3H (20)
CH3SO3H (7)
During a study of metal-catalyzed aerobic oxidative
coupling reactions under acidic conditions, we performed a
control experiment without the metal catalyst, which unex-
pectedly resulted in similar yields of the desired product. With
xanthene (1) as the electrophilic substrate and ketones like
cyclopentanone (2), we could obtain high yields of the
coupling product 3a using only catalytic amounts of trifluoro-
methanesulfonic acid (TfOH) and elemental oxygen at
elevated pressure (Scheme 1).
This reaction is remarkable, as it does not involve any
redox-active catalyst or reagent commonly used for C H
bond activation and thus holds great potential for the design
of sustainable syntheses with simple and cheap reagents and
9
CH2Cl2
hexane
EtOAc
–
–
–
–
10
11
12
13
14
15[e]
1
1
1
70
94 (90)[f]
[a] Reaction conditions: 1 (0.5 mmol), 2 (2.5 mmol), and acid were
stirred under an atmosphere of oxygen at room temperature for 1 day.
[b] Yields were determined by GC analysis of the crude reaction mixture
unless indicated otherwise. [c] 80 bar of 10% O2 in N2. [d] Air rather than
oxygen. [e] 18 h, 408C. [f] Yield of isolated product. pTsOH=p-toluene
sulfonic acid.
À
[*] Dr. ꢀ. Pintꢁr, Dr. A. Sud, Dr. D. Sureshkumar, Dr. M. Klussmann
Max-Planck-Institut fꢂr Kohlenforschung
Kaiser-Wilhelm-Platz 1
of oxygen (up to 8 bar with a mixture of 10% O2 and 90% N2)
but also at ambient pressure with only a slightly reduced
reaction rate by simply using a balloon filled with pure oxygen
or even air (Table 1, entries 1–3). Performing the reaction
under an atmosphere of argon resulted in only traces of the
desired product 3a (see the Supporting Information for
further details on the optimization study). To simplify the
procedure, we choose pure oxygen at ambient pressure as the
standard condition in further experiments. A catalytic amount
of a strong acid is needed: sulfonic acids (Table 1, entries 2, 7,
and 8) and mineral acids are effective, whereas a weaker acid
45470 Mꢂlheim an der Ruhr (Germany)
Fax: (+49)208-306-2980
E-mail: klusi@mpi-muelheim.mpg.de
Homepage: http://www.kofo.mpg.de/klussmann/
[**] We thank Prof. Benjamin List and the Alexander von Humboldt-
Foundation (scholarship to D.S.) for financial support and Esther
Bꢃß, Jan Kꢂmmel, and Tim Hillringhaus for the synthesis of some
starting materials.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201000711.
5004
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5004 –5007

Angewandte
Chemie
Table 2: Aerobic coupling of xanthene (1) with various C nucleophiles.[a]
like trifluoroacetic acid gave much lower reaction rates and
acetic acid gave no reaction at all (Table 1, entries 5 and 6).
Without acid, no reaction took place (Table 1, entry 4).
Among the sulfonic acids, methanesulfonic acid and p-
toluene sulfonic acid worked best and gave comparable
results (Table 1, entries 7 and 8). Trifluoromethanesulfonic
acid is considerably more expensive than the other sulfonic
acids and in a few cases gave irreproducible results. We also
tested the polymeric sulfonic acid Nafion as a catalyst; it gave
high conversion but a significantly reduced reaction rate. We
chose the liquid methanesulfonic acid for further studies
because it is easy to handle and gave the highest yields.
The coupling reaction can be performed in solvents: the
best results were obtained with dichloromethane, ethyl
acetate, and hexane (Table 1, entries 9–11). However, the
highest yields were obtained when the reaction was per-
formed without any solvent, provided the reaction mixture
was homogeneous (Table 1, entry 7). To ensure that xanthene
was completely in solution, we generally used five equivalents
of liquid ketone, but lower amounts could also be used. A
slightly higher acid loading (7 mol%) proved to be optimal,
while higher loadings resulted in decreased yields (Table 1,
entries 12–14). The reaction proceeds well at ambient temper-
ature, but a slightly elevated temperature of 408C improves
the rate and gives optimal results (Table 1, entry 15). At
higher temperatures the yields dropped again and the
increased formation of aldol condensation products from
cyclopentanone was observed.
Entry
Product
t [h]
Yield [%][b]
1
2
3
4
3a n=1
3b n=2
3c n=3
3d n=4
18
36
64
64
90
94
81[c]
71[c]
5
6
7
8
9
3e R’=H, R’’=Me
3 f R’=Me, R’’=Et
3g R’=H, R’’=Ph
3h R’=Ph, R’’=Me
3i R’=Ph, R’’=nBu
15
15
24
5
34
68
49
95
93
8
Using these optimized conditions, we tested a variety of
other carbon nucleophiles in the reaction with xanthene
(Table 2). In many cases, full conversion and high yields of
isolated products were achieved after reaction times of up to
two days. Xanthene (1) could be coupled with cyclic as well as
open-chain ketones, giving yields of up to 95% (Table 2,
entries 1–9). In some cases the reaction stopped before full
conversion was reached, but most of the leftover xanthene
could be recovered. (Table 2, entries 5 and 7). Further
additions of acid or different amounts of acetone did not
“restart” the reaction or change the yield.
10
11
12
13
3j R’=Me, R’’=OEt
3k R’=Ph, R’’=OEt
3l R’, R’’=OMe
64
96
128
128
85[c]
80[d]
42[c,e]
39[c,e]
3m R’, R’’=OEt
[a] Reactions were conducted at ambient pressure and 408C, unless
stated otherwise. [b] Yields of isolated products. [c] Reaction at 6 bar
partial pressure of O2 and room temperature. [d] 2.5 equiv nucleophile.
[e] CH2Cl2 as solvent and 7 mol% TfOH.
In a few cases it was beneficial to run the reaction at an
elevated partial pressure of oxygen to increase the yield. For
example, the coupling product with cyclooctanone, 3d, was
formed in 71% yield at 6 bar (Table 2, entry 4) compared with
50% yield at ambient pressure. Similar trends were observed
for reactions with b-keto esters and 1,3-diesters as nucleo-
philes; usually high yields were achieved only at higher
pressure (Table 2, entries 10–13). In a few cases, the use of
triflic acid was beneficial (Table 2, entries 12 and 13). Also, in
these cases, a small amount of solvent was added to achieve a
homogeneous solution.
In previous studies of oxidative coupling reactions xan-
thene had been employed. In these studies an FeIII catalyst
together with tert-butyl peroxide,[4] stoichiometric amounts of
the quinone DDQ,[5] or equimolar amounts of a MnIII
compound at high temperature were used to synthesize
products like 3e and 3j–m.[6] Substituted xanthenes like 3e
and 3g have recently been synthesized via arynes.[7] The
present method offers an alternative with fewer steps, cheaper
reagents, and less waste.
Diphenylmethanes did not react under these conditions,
but acridanes 4 could be coupled with ketones (Figure 1). As
seen for products 3e and 3g, the reactions did not go to
completion but did not otherwise lead to side products. The
products were isolated with yields around 40% when oxygen
was used at ambient pressure (6b,d), while higher yields up to
77% were achieved at elevated oxygen pressure (6a,c,e).
Only tertiary N-substituted acridanes could be employed as
substrates; however, acridanes 6b and 6e should be depro-
tected easily to the secondary amines. N-phenyltetrahydro-
isoquinoline (5) could also be used successfully in the
coupling with ketones, provided elevated oxygen pressure
was used.[8] The products 7 and 8 were prepared in moderate
yields, but leftover starting material could be recovered.
Product 6c had been synthesized before from an acridi-
nium salt and a preformed enamine.[9] Product 7 had been
synthesized previously by oxidative copper catalysis using the
corresponding silyl enolate as the ketone equivalent,[10]
product 8 by metal catalysis at elevated temperature[11] or
with an amino acid co-catalyst and a peroxide oxidant.[12] The
Angew. Chem. Int. Ed. 2010, 49, 5004 –5007
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 5005

Communications
Scheme 2. a) Potential mechanism for the autoxidative coupling of
xanthene (1) and cyclopentanone (2) via hydroperoxide 10. b) Struc-
ture of the isolated side product 11.
In agreement with this mechanistic proposal, the addition
of catalytic amounts of a radical inhibitor (5 mol% 2,4,6-tris-
tert-butylphenol) completely blocked the reaction. We found
evidence for the formation of hydrogen peroxide in some
cases: addition of MnO2 at the end of a reaction resulted in
gas evolution, indicating catalytic decomposition. Since MnO2
was shown not to inhibit the reaction, its catalytic ability
offers an interesting improvement: it could be used to
continuously decompose any hydrogen peroxide formed to
oxygen, which can again enter the reaction, and water, which
then constitutes the final, truly environmentally benign waste
product.
Figure 1. Products from the aerobic coupling of acridanes 4 and
tetrahydroisoquinoline 5 with ketones. The reactions were performed
as described in Table 2. [a] Reaction at 6 bar partial pressure of O2 and
room temperature. [b] Solvent: CH2Cl2. [c] Solvent: MeOH. [d] 7 mol%
triflic acid. Cbz=benzyloxycarbonyl.
current method is clearly an improvement as the ketones can
be used directly without a metal catalyst.
Initially, we could not rule out that impurities of
transition-metal salts catalyze the reaction. Analysis of the
reaction mixture leading to the successful synthesis of 3a by
atomic adsorption spectroscopy did not reveal any common
redox-active transition metals like V, Mn, Fe, Cu, Ru, or Pd
above the detection limit of 0.5 ppm. Also, performing the
reaction in the presence of small amounts (1 mol%) of
various metal salts resulted at best in unchanged rates and
product yields, whereas CuII and FeII salts, likely impurities
and known redox catalysts, actually inhibited the reaction (see
the Supporting Information). Accordingly, we believe the
reaction to proceed without the involvement of metal
catalysts.
A potential mechanistic rationale involves autoxidative
formation of a xanthene hydroperoxide 10, which would react
in an acid-catalyzed SN1 type reaction with nucleophiles like
ketone 2 to form product 3a and hydrogen peroxide
(Scheme 2a). The oxidation of 1 to the hydroperoxide 10
could proceed by a radical pathway without the involvement
of a metal catalyst. Such autoxidation reactions are well
known for many organic compounds in the presence of
oxygen,[13] also, for example, for compound 5.[14] Hydro-
peroxides are not commonly used as electrophiles; instead
they are known to undergo acid-catalyzed rearrangement
reactions as in the synthesis of phenol from cumene hydro-
peroxide.[15] However, peroxides like 10 and 11 (Scheme 2b)
We could also isolate xanthene peroxide 11 (Scheme 2b)
as a side product of a reaction with trifluoroacetic acid as the
catalyst (Table 1, entry 6). This is strong support for the
occurrence of a hydroperoxide, as 11 likely forms from
reaction of hydroperoxide 10 by acid catalysis.[16] Strangely,
the addition of a radical initiator did not accelerate the
reaction. More detailed studies are needed to better under-
stand the mechanism of this reaction.
In preliminary experiments we have observed that tertiary
À
C H bonds in 9-methyl-substituted xanthenes are very
unreactive under our reaction conditions. Also, the addition
of substoichiometric amounts of product 3e to a reaction of
xanthene and cyclohexanone considerably slowed down the
reaction (Scheme 3).
Scheme 3. The rate of the formation of 3b is suppressed by the
addition of 3e.
À
are known to react by acid-catalyzed C O bond cleavage with
These results could indicate why 1) the reaction products
carbon nucleophiles.[16] We could successfully convert hydro-
peroxide 10 and cyclopentanone to 3a under our reaction
conditions without added oxygen. Similarly, other peroxides
have been employed as alkylating agents mediated by Lewis
acids.[17]
bearing tertiary C H groups are stable under the reaction
À
conditions and do not undergo a second coupling step or other
oxidative degradations and 2) why 3e is only formed in low
À
yields. The C H bond cleavage in the products might be
hindered for steric reasons, as has been observed for 9-
5006 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5004 –5007

Angewandte
Chemie
phenylxanthene.[18] Product 3e might suppress the reaction
rate because it acts as an antioxidant by formation of highly
stabilized radicals, slowing down and finally stopping the
autoxidation.[19]
The proposed reaction mechanism presents a possibility
for the further development of “green” oxidative coupling
reactions using oxygen, drawing on the rich experience in
radical chemistry and peroxide formation.[13b] Substrates not
as easily autoxidizable as xanthene could be employed at
higher temperatures, under irradiation with light or by using
radical initiators or singlet oxygen, for example. Thereby, a
Keywords: autoxidation · C H activation · cross-coupling ·
green chemistry · homogeneous catalysis
À
.
[1] P. Anastas, N. Eghbali, Chem. Soc. Rev. 2010, 39, 301 – 312.
[2] R. G. Bergman, Nature 2007, 446, 391 – 393.
[3] a) C. J. Scheuermann, Chem. Asian J. 2010, 5, 436 – 451; b) C.-J.
Li, Acc. Chem. Res. 2009, 42, 335 – 344.
[4] D. H. R. Barton, T.-L. Wang, Tetrahedron 1994, 50, 1011 – 1032.
[5] F. Benfatti, M. G. Capdevila, L. Zoli, E. Benedetto, P. G. Cozzi,
Chem. Commun. 2009, 5919 – 5921.
[6] a) H. Nishino, Bull. Chem. Soc. Jpn. 1986, 59, 1733 – 1739; b) H.
Nishino, H. Kamachi, H. Baba, K. Kurosawa, J. Org. Chem.
1992, 57, 3551 – 3557.
[7] X. Huang, T. Zhang, J. Org. Chem. 2010, 75, 506 – 509.
[8] For a potentially related example of an oxidative coupling of 5
without a redox catalyst, see: A. G. Condie, J. C. Gonzꢀlez-
Gꢁmez, C. R. J. Stephenson, J. Am. Chem. Soc. 2010, 132, 1464 –
1465.
À
wide range of organic compounds bearing C H bonds would
be accessible to autoxidative coupling reactions using only
oxygen, an acid catalyst, and maybe an initiator. One possible
À
À
difficulty may be the issue of acid-catalyzed O O versus C O
bond cleavage of the hydroperoxides, but the appropriate
choice of acid can steer the reaction course towards the
desired product.[17]
[9] O. N. Chupakhin, V. N. Charushin, E. O. Sidorov, Chem. Heter-
ocycl. Compd. 1979, 15, 541 – 544.
In summary, we have developed an aerobic oxidative
coupling method for activated benzylic CH2 groups like that
in xanthene with ketones and 1,3-dicarbonyl compounds; the
reaction proceeds smoothly under mild conditions without a
metal reagent and requires simply cheap methane sulfonic
acid and oxygen. The reaction fulfills several criteria of
“green” chemistry:[1] 1) It does not require intensive heating,
a solvent, or extractive workup, 2) it is catalytic and uses a
cheap acid, 3) it is very atom-economic as two hydrogen
atoms serve as “leaving groups”, and 4) it produces only
hydrogen peroxide as the waste product, which can be
decomposed to water in situ. Xanthene and acridane deriv-
atives are pharmaceutically active compounds and are used as
dyes and fluorescent materials and chiroptical molecular
switches; they also occur in natural products.[7,20] Accordingly,
we anticipate further elaborations of this autoxidative cou-
pling principle and applications towards the synthesis of
complex products.
[10] D. Sureshkumar, A. Sud, M. Klussmann, Synlett 2009, 1558 –
1561.
[11] Y. Shen, M. Li, S. Wang, T. Zhan, Z. Tan, C.-C. Guo, Chem.
Commun. 2009, 953 – 955.
[12] A. Sud, D. Sureshkumar, M. Klussmann, Chem. Commun. 2009,
3169 – 3171.
[13] a) N. A. Milas, Chem. Rev. 1932, 10, 295 – 364; b) Science of
Synthesis. Compounds with One Saturated Carbon–Heteroatom
Bond. Peroxides, Vol. 38 (Ed.: A. Berkessel), Thieme, Stuttgart,
2009, pp. 9 – 141.
[14] A. Rieche, E. Hoeft, H. Schultze, Chem. Ber. 1964, 97, 195 – 201.
[15] V. Zakoshansky, Russ. J. Gen. Chem. 2009, 79, 2244 – 2266.
[16] A. G. Davies, R. V. Foster, R. Nery, J. Chem. Soc. 1954, 2204 –
2209.
[17] a) L. Liguori, H.-R. Bjorsvik, F. Fontana, D. Bosco, L. Galim-
berti, F. Minisci, J. Org. Chem. 1999, 64, 8812 – 8815; b) P. H.
Dussault, H.-J. Lee, X. Liu, J. Chem. Soc. Perkin Trans. 1 2000,
3006 – 3013.
[18] F. G. Bordwell, J. Cheng, G. Z. Ji, A. V. Satish, X. Zhang, J. Am.
Chem. Soc. 1991, 113, 9790 – 9795.
[19] We thank a referee for pointing out the following reference,
describing a diphenylmethyl antioxidant: J. C. Scaiano, A.
Martin, G. P. A. Yap, K. U. Ingold, Org. Lett. 2000, 2, 899 – 901.
[20] For representative examples, see: a) W. Tang, H. Hioki, K.
Harada, M. Kubo, Y. Fukuyama, J. Nat. Prod. 2007, 70, 2010 –
2013; b) M. S. T. Gonꢂalves, Chem. Rev. 2009, 109, 190 – 212;
c) B. L. Feringa, J. Org. Chem. 2007, 72, 6635 – 6652; d) T.
Troxler, K. Hurth, H. Mattes, M. Prashad, P. Schoeffter, D.
Langenegger, A. Enz, D. Hoyer, Bioorg. Med. Chem. Lett. 2009,
19, 1305 – 1309; e) K. Chibale, M. Visser, D. van Schalkwyk, P. J.
Smith, A. Saravanamuthu, A. H. Fairlamb, Tetrahedron 2003, 59,
2289 – 2296; f) R. Pellicciari, G. Costantino, M. Marinozzi, A.
Macchiarulo, L. Amori, P. Josef Flor, F. Gasparini, R. Kuhn, S.
Urwyler, Bioorg. Med. Chem. Lett. 2001, 11, 3179 – 3182; g) M.
Nꢁgrꢀdi in Science of Synthesis, Vol. 14 (Ed.: E. J. Thomas),
Thieme, Stuttgart, 2003, pp. 201 – 273; h) R. H. Prager, C. M.
Williams in Science of Synthesis, Vol. 15 (Ed.: D. S. Black),
Thieme, Stuttgart, 2004, pp. 987 – 1028.
Experimental Section
Representative procedure: Xanthene (1) (91 mg, 0.5 mmol), cyclo-
pentanone (2) (0.22 mL; 2.5 mmol) and methanesulfonic acid (2.3 mL,
7 mol%) were mixed in a glass vial or round-bottom flask equipped
with a magnetic stirring bar. The vial was flushed with oxygen and
then connected to a balloon filled with oxygen. After rapid stirring at
408C for 10–18 h, the reaction mixture was purified directly by
column chromatography (silica gel, toluene) without further workup,
giving the product 3a as colorless solid (119 mg, 90% yield).
Reactions at high pressure were performed with 10% oxygen in
nitrogen for reasons of safety, as most flammable organic compounds
have a limiting oxygen concentration around 10%.
Received: February 5, 2010
Published online: June 11, 2010
Angew. Chem. Int. Ed. 2010, 49, 5004 –5007
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 5007