The Unprecedented cobalt-catalysed oxidative glaser coupling under reductive conditions

Article abstract of DOI:10.1055/s-0028-1083316

The cobalt-catalysed Glaser-type coupling of terminal alkynes was achieved utilising nitrobenzene as a stoichiometric oxidising agent under reductive conditions. The proposed electron transfer from zinc powder to a nitrobenzene coordinated to the cobalt centre initiates the coupling of the coordinated alkynes. Other aryl-, alkenyl-, alkyl-, and silylacetylenes besides phenylacetylene could also be coupled to generate the 1,3-diynes in moderate to very good yields. Georg Thieme Verlag Stuttgart.

Full text of DOI:10.1055/s-0028-1083316

PAPER
395
The Unprecedented Cobalt-Catalysed Oxidative Glaser Coupling under
Reductive Conditions
R
G
eductive Glaser
c
e
oupling rhard Hilt,* Christoph Hengst, Marion Arndt
Fachbereich Chemie, Philipps-Universität Marburg, Hans-Meerwein-Straße, 35043 Marburg, Germany
Fax +49(6421)2825677; E-mail: Hilt@chemie.uni-marburg.de
Received 18 July 2008; revised 6 October 2008
other and performed the cobalt-catalysed Glaser-type cou-
pling of phenylacetylene in the presence of nitrobenzene
Abstract: The cobalt-catalysed Glaser-type coupling of terminal
alkynes was achieved utilising nitrobenzene as a stoichiometric ox-
idising agent under reductive conditions. The proposed electron
(
Scheme 2). The reaction produced the diyne 3 (see ex-
transfer from zinc powder to a nitrobenzene coordinated to the co- perimental section) in excellent yield without any traces
balt centre initiates the coupling of the coordinated alkynes. Other of the previously reported cyclotrimerisation product of
3
aryl-, alkenyl-, alkyl-, and silylacetylenes besides phenylacetylene
type 1. Therefore, the nitrobenzene efficiently prevented
could also be coupled to generate the 1,3-diynes in moderate to very
the cobalt catalyst from initiating the trimerisation reac-
good yields.
tion pathway. This can be tentatively explained by assum-
ing that the nitro group occupies coordination sites at the
Key words: alkyne, cobalt, Glaser coupling, oxidation, diynes
cobalt centre. To further investigate this reaction, the
components were varied and the results are summarised in
The synthesis of symmetrical diynes by oxidative cou- Table 1.
pling of terminal alkynes has a long tradition in organic
synthesis starting with the first observation of a copper-
Ph
Ph
initiated process by Glaser in 1869. Stoichiometric as well
as catalytic couplings utilising copper as transition metal
are well established.
CoBr , Zn
2
Ph
+
1
ZnI , MeCN
2
Ph Ph
Ph
Over the course of our investigations concerning the cy-
clotrimerisation of alkynes and other [2+2+2] cyclo-
additions with low-valent cobalt catalysts we applied
PhNO2
2
Ph
Ph
9
9%
3
-ethynylnitrobenzene in the cyclotrimerisation reaction
3
(
Scheme 1).
Scheme 2 Cobalt-catalysed Glaser coupling utilising phenylacetylene
in the presence of nitrobenzene
NO2
Ar
Ar
CoBr2, Zn
Ar
Table 1 Cobalt-Catalysed Glaser Coupling of Phenylacetylene to
Generate 3
+
ZnI2, MeCN
Ar Ar
Ar
1
a
1b
Entry
CoBr2
Zn
+
ZnI2
+
PhNO2
Yield (%)a
Ar = 3-NO2C6H4
1
2
3
4
5
a
+
+
+
+
–
+
–
99
0
Ar
Ar
38%
+
+
2
Scheme 1 Cobalt-catalysed Glaser coupling utilising a nitro-func-
+
–
+
+
+
0
tionalised alkyne
–
+
0
+
+
0
To our surprise, no cyclotrimerisation product 1a or 1b
was obtained, but the 1,3-diyne 2 was isolated in moderate
yield (38%). This result is extremely unusual since the
Glaser-type coupling to afford product 2 is an overall ox-
idation process. In contrast to the classical Glaser cou-
Phenylacetylene (1.0 mmol), CoBr (0.1 mmol), Zn (1.0 mmol), ZnI
2
2
(
0.2 mmol), nitrobenzene (1 mmol), and MeCN (1 mL), 12 h, 80 °C.
pling reaction, the cobalt-catalysed reaction was Only when all components were present was the forma-
performed under reductive conditions (zinc powder). To tion of the diyne 3 observed, which can be rationalised in
investigate this coupling reaction in more detail, we sepa- the following way; the cobalt bromide is dissolved in ac-
rated the nitro- and the alkyne functionalities from each etonitrile and forms a complex with the solvent. The zinc
iodide acts as Lewis acid to abstract the bromide ligands
from the proposed CoBr (MeCN) to increase the Lewis
acidity of the cobalt centre. Nitrobenzene as well as the
alkyne displaces the acetonitrile as coordinating ligands.
2
n
SYNTHESIS 2009, No. 3, pp 0395–0398
x
x
.
x
x
.2
0
0
9
Advanced online publication: 09.01.2009
DOI: 10.1055/s-0028-1083316; Art ID: Z17008SS
Georg Thieme Verlag Stuttgart · New York
©

3
96
G. Hilt et al.
PAPER
The alkynes coordinate to the cobalt centre either as a bis-
h alkyne 4 or a cobaltacycle 5 which is generated via a
formal oxidative addition (Scheme 3).
2
e– (Zn)
Co
Ph
Ph
Co
Ph
Ph
Ph
Ph
O
N
O
(L)n
(L)n
4
–O N Ph
6
O
Ph
CoBr2(L)n
.
"
H "-shift
2
ZnI2
L = MeCN
PhNO2
Ph
2
ZnI2Br–
2
Ph
.
Co
"H "-shift
Co
Ph
O
N
O
N
Ph
Co
Ph
(
L)n
8
– H2O
(L)n
Co
Ph
Ph
7
O
N
(L)n
O
(L)n
Ph
HO
Ph
N
5
4
reductive
elimination
O
Ph
O
Ph
3
+ Co(L)n(ONPh)
Scheme 3 Proposed initial steps towards the cobalt-catalysed Glaser
coupling utilising phenylacetylene in the presence of nitrobenzene
Scheme 4 Proposed initial steps towards the cobalt-catalysed Glaser
coupling, utilising phenylacetylene in the presence of nitrobenzene
These assumptions become reasonable because the reac-
tion can also be conducted with additional bidentate di- tected. This leads to the interpretation that the zinc powder
imine and pyridine-imine ligands on the cobalt(II) centre undergoes a single electron reduction to initiate the pro-
(
see below) and an octahedral coordination sphere is a cess which the borohydride can not. The use of catalytic
prerequisite to accommodate these cobalt diimine type amounts of zinc powder (20 mol%) led only to a conver-
complexes. In a cobalt(I) complex, which adopts a trigo- sion of 23% indicating that at least one electron from zinc
nal bipyramidal coordination sphere, the starting materi- powder is needed to initiate the overall reaction. The re-
als and reagents can not be incorporated easily. duction of nitro- via nitrosobenzene led to aniline and di-
Nevertheless, a low-valent cobalt(I) complex generated azobenzene which were detected via GCMS. The
upon reduction with zinc powder might act as a reducing oxidising agent nitrobenzene can be altered and use of
agent towards a coordinated nitrobenzene ligand, result- other aromatic nitro compounds led to complete conver-
4
ing in the identical intermediate 6 (see Scheme 4).
sions such as C
4
6
D
5
NO
2
and 1-nitronaphthylene. When
-nitrobenzaldehyde, 4-cyanonitrobenzene and 1,3-dini-
The formation of 3 from the cobaltacycle 5 seems unlike-
ly, because the two hydrogen atoms, which must be re-
moved, are coplanar with the cobalt centre so that b-
hydride elimination can be excluded as the oxidation path-
way. Therefore, we further propose that the Glaser-type
oxidation process of the coordinated starting materials is
initiated by an electron transfer from zinc powder to the
activated nitrobenzene coordinated to the Lewis acidic co-
balt centre in 4 (Scheme 4). This one electron reduction
generates a radical type intermediate 6, which could act as
a strong oxidising agent to initiate a net H-shift to form 7.
Another net H-shift could result in the formation of water
and the cobalt intermediate 8 where the two alkynes are s-
bonded to the cobalt. A reductive elimination would then
generate the product 3 and a cobalt nitrosobenzene com-
plex.
trobenzene were used only small amounts of 3 were ob-
tained (10–30%) and the use of nitromethane did not
result in the formation of 3 at all. 4-Chloro- and 4-fluoro-
nitrobenzene are as effective as nitrobenzene itself result-
ing in 76%, 71% and 73% conversion after 3.5 hours of
reaction time, respectively. 4-Methyl- and 4-methoxyni-
trobenzene were inferior to nitrobenzene resulting only in
5
7% and 45% conversion after 3.5 hours of reaction time.
This indicates that the redox potentials seem to be critical
for the successful transformation and that electron-rich as
well as very strongly electron-deficient nitrobenzene de-
rivatives are less effective in this transformation. The
amount of converted nitrobenzene was determined by GC
analysis utilising decane as internal standard. Reproduc-
ibly, on a 2 mmol scale (alkyne and nitrobenzene) only
3
0% of the nitrobenzene was converted. The consumed
In a separate experiment nitrosobenzene also proved to be
active as an oxidising agent in this unprecedented Glaser-
type coupling reaction. Nevertheless, the replacement of
nitrobenzene by nitrosobenzene does give the coupling
product 3, but small amounts (5–10%) of the cyclotri-
merisation products of type 1 are also detectable.
amount of nitrobenzene is sufficient to account for all re-
dox equivalents required for the quantitative conversion
of the alkyne. When the total amount of nitrobenzene was
reduced to substoichiometric quantities (0.5 equiv with re-
spect to the alkyne) the conversion of the alkyne was in-
complete and the cyclotrimerisation product of the alkyne
The cobalt-catalysed Glaser-type coupling was also of type 1a (Ar = Ph) was observed.
achieved utilising a preformed cobalt complex with a
N,N¢-dicyclohexylethanediimine or triisopropyl phosphite
ligand. In both cases the conversion of phenylacetylene
with zinc powder led to the desired product 3 in 99%
yields, whereas when tetra-n-butylammonium borohy-
dride was used as the reducing agent no product was de-
The alternative oxidation pathway involving H genera-
2
tion can be excluded because MS measurements of the gas
atmosphere over the reaction mixture after complete con-
version of the alkyne in a sealed reaction vessel revealed
that no hydrogen gas was generated.
Synthesis 2009, No. 3, 395–398 © Thieme Stuttgart · New York

PAPER
Reductive Glaser Coupling
397
All attempts to intercept possible alkyne radical type in-
termediates in a radical-based reaction pathway upon ad-
dition of further substrates such as terminal alkenes,
malonates, cinnamyl chloride, diazo esters, bromoform,
propargyl bromide, and acyl amides resulted either in the
unchanged formation of the diyne 3 or in no conversion at
all. Therefore, we came to the conclusion that no radical
type alkyne intermediates are present free in solution, but
are rather coordinated to the cobalt centre and lead to the
product 3 by rapid reductive elimination from 8.
Ph
Ph
Bu
Bu
3
0
CoBr2(Ligand)
Ph
Bu
+
Ph
Bu
1
Zn, ZnI2, PhNO2
MeCN, 80 °C
11
Scheme 6 Attempted synthesis of unsymmetrical 1,3-diynes
Table 3 Results of the Cobalt-Catalysed Glaser-Type Coupling
Utilising Two Different Alkynes
Entry
Ligand
Product ratioa
/10/11
The preparative application of the cobalt-catalysed
Glaser-type coupling was investigated on some selected
examples utilising representative terminal alkynes. The
results of these transformations (Scheme 5) are sum-
marised in Table 2.
3
1
2
3
none
1.00:0.47:0.08
1.00:0.72:0.38
1.00:0.33:0.42
2 P(Oi-Pr)3
dppe
CoBr2, Zn, ZnI2
R
R
R
4
1.00:0.50:0.11
N
N
PhNO2, MeCN, 80 °C
9
Scheme 5 Attempted synthesis of symmetrical diynes
5
1.00:0.54:0.41
N
N
Table 2 Results of the Cobalt-Catalysed Glaser-Type Coupling of
Various Alkynes
Entry Alkyne
1
Product 9
Yield (%)a
93
6
1.00:0.50:0.65
1.00:0.48:0.13
N
N
N
N
S
S
S
7
2
3
87
45
a
Average of GC and NMR integration after >95% of the phenylace-
tylene had converted.
was obtained for the phosphite ligand system, whereas the
diimine- and the pyridine imine-type ligands gave an al-
most constant ratio of 3 to 10, while the amount of 11
varied significantly. Nevertheless, complete chemoselec-
tivity was not expected in these reactions, but the influ-
ence of ligands on the ratios of products, especially of the
phosphite ligands, which gave the highest ratio of the un-
symmetrical product 10, might be worth investigating in
the future in more detail.
4
5
52
56
SiMe3
Me3Si
SiMe3
a
Alkyne (1.0 mmol), CoBr (0.05 mmol), Zn (1.0 mmol), ZnI (0.1
2
2
mmol), nitrobenzene (1 mmol), and MeCN (1 mL), 12 h, 80 °C.
While the cobalt-catalysed Glaser-type coupling of phe-
nylacetylene as well as of 2-ethynylthiophene proceeded
with excellent isolated yields, the results for alkenyl-,
alkyl- and silyl-substituted alkynes are moderate to good
utilising cobalt bromide in the absence of any additional
ligands.
In summary we were able to demonstrate that nitroben-
zene can be used as a stoichiometric oxidant in the cobalt-
catalysed Glaser-type coupling of aromatic alkynes
whereas acceptable yields could be obtained with alkenyl-,
aliphatic, and silyl-substituted alkynes. The unprecedent-
ed reaction mechanism of an oxidative coupling reaction
under formally reductive conditions (stoichiometric
amounts of zinc powder) could be useful in future appli-
cations where oxidation sensitive functional groups might
interfere with traditional reaction protocols.
Finally, we briefly investigated the application of cobalt
bromide and selected cobalt complexes in a coupling re-
action utilising 5 mol% of the cobalt catalyst to afford
symmetrical as well as unsymmetrical diynes (Scheme 6)
utilising equimolar quantities of phenylacetylene and hex-
1
-yne. The results are summarised in Table 3.5
The ratios of the desired products were detected by GC
and NMR integration of suitable signals. Surprisingly, in-
1
,4-Diphenylbuta-1,3-diyne (3); Typical Procedure
stead of a statistical mixture of products the phenylacety- To a suspension of CoBr (22 mg, 0.1 mmol, 5 mol%), Zn (131 mg,
2
lene was coupled more efficiently than hex-1-yne in all 2.0 mmol, 1.0 equiv) and ZnI (64 mg, 0.2 mmol, 10 mol%) in
2
MeCN (1.0 mL) were added nitrobenzene (246 mg, 206 mL, 2.0
mmol, 1.0 equiv) and phenylacetylene (204 mg, 220 mL, 2.0 mmol)
entries. The highest ratio of the unsymmetrical product 10
Synthesis 2009, No. 3, 395–398 © Thieme Stuttgart · New York

3
98
G. Hilt et al.
PAPER
under argon. The reaction mixture was stirred at 80 °C for 12 h. Af-
(2) For recent examples, see: (a) Hilt, G.; Hengst, C.; Hess, W.
Eur. J. Org. Chem. 2008, 2293. (b) Hilt, G.; Hess, W.;
Harms, K. Synthesis 2008, 75. (c) Hilt, G.; Paul, A.; Harms,
K. J. Org. Chem. 2008, 73, 5187.
ter the addition of Et O (5.0 mL), the mixture was filtered over silica
2
gel (eluent: pentane–Et O, 2:1). The volatiles were removed and the
2
product was purified by column chromatography (SiO , pentane)
2
affording the product 3 (200 mg, 0.99 mmol, 99%) as a colourless
solid. The analytical data are consistent with those reported in the
literature.1b
(3) (a) Hilt, G.; Hess, W.; Vogler, T.; Hengst, C. J. Organomet.
Chem. 2005, 690, 5170. (b) Hilt, G.; Vogler, T.; Hess, W.;
Galbiati, F. Chem. Commun. 2005, 1474.
(
4) The use of deuterated nitrobenzene led to the corresponding
aniline and diazobenzene side products without loss of any
deuterium labeling.
References
(
5) For the Cadiot–Chodkiewicz coupling, see: (a) Cadiot, P.;
Chodkiewicz, W. In Chemistry of Acetylenes; Viehe, H. G.,
Ed.; Marcel Dekker: New York, 1969, 597. (b) Siemsen, P.;
Felber, B. In Handbook of C-H Transformations, Vol. 1;
Dyker, G., Ed.; 2005, 53–62. (c) Siemsen, P.; Felber, B. In
Handbook of C-H Transformations, Vol. 1; Dyker, G., Ed.;
(1) (a) Glaser, C. Ber. Dtsch Chem. Ges. 1869, 2, 422.
(
(
(
b) Eglinton, G.; Galbraith, A. R. J. Chem. Soc. 1959, 889.
c) Eglinton, G.; McCrae, W. Adv. Org. Chem. 1963, 4, 225.
d) Hay, A. S. J. Org. Chem. 1962, 27, 3320. (e) Boldi, A.
M.; Anthony, J.; Knobler, C. B.; Diederich, F. Angew.
Chem., Int. Ed. Engl 1992, 31, 1240; Angew. Chem. 1992,
2005, 83–84. (d) Marino, J. P.; Nguyen, H. N. J. Org. Chem.
002, 67, 6841.
1
04, 1270. (f) Siemsen, P.; Livingston, R. C.; Diederich, F.
Angew. Chem. Int. Ed. 2000, 39, 2632; Angew. Chem. 2000,
12, 2740.
2
1
Synthesis 2009, No. 3, 395–398 © Thieme Stuttgart · New York