In situ generation of the coates catalyst: A practical and versatile catalytic system for the carbonylation of meso-epoxides

Article abstract of DOI:10.1021/ol201043d

A highly active catalytic system for the carbonylation of meso- and terminal epoxides to β-lactones is described. The active catalyst, analogous to Coates' catalyst, is generated in situ from commercially available (TPP)CrCl and Co₂(CO)₈

Full text of DOI:10.1021/ol201043d

ORGANIC
LETTERS
2
011
Vol. 13, No. 12
142–3145
In Situ Generation of the Coates Catalyst:
A Practical and Versatile Catalytic System
for the Carbonylation of meso-Epoxides
3
Prasad Ganji, David J. Doyle, and Hasim Ibrahim*
Centre for Synthesis and Chemical Biology, School of Chemistry and Chemical Biology,
University College Dublin, Belfield, Dublin 4, Ireland
hasim.ibrahim@ucd.ie
Received April 19, 2011
ABSTRACT
A highly active catalytic system for the carbonylation of meso- and terminal epoxides to β-lactones is described. The active catalyst, analogous to
Coates’ catalyst, is generated in situ from commercially available (TPP)CrCl and Co (CO) . This practical system circumvents the preparation of
air sensitive cobaltate salts, operates at low catalyst loadings, and allows the carbonylation of functionalized, sterically demanding and
2
8
heterocyclic meso-epoxides.
1
The carbonylation of epoxides is an industrially rele-
vant process that provides direct access to versatile syn-
developed a system based on the cobaltate salt [PPN][Co-
(CO) ] (PPN = PPh dNdPPh ) which in combination
4
3
3
2
3
thons suchasβ-hydroxyesters, β-hydroxyamides, and β-
lactones. Well into the 1990s, the carbonylation of
with BF Et O activation selectively produced β-lactones
3 2
3
4
6
from a series of terminal epoxides. This system incorpo-
ratesa Lewisacid(LA) inducedactivationofthe epoxide in
conjunction with a ring-opening nucleophilic attack of
epoxides to β-lactones was the subject of limited studies due
to the lack of catalysts that selectively produce β-lactones
without competing polymerization to poly(β-hydroxyal-
kanoate)s or isomerization of the starting epoxides to
ketones. The disclosure of a patent by Drent in 1994 claim-
ing that a catalytic system comprised of Co (CO) and
ꢀ
7
[Co(CO) ] , a phenomenon first studied in detail by Heck.
4
Cobaltate salts as a concept were further elaborated by
the Coates group who designed well-defined catalysts of
þ
ꢀ
the general form [LA] [Co(CO) ] that incorporate Lewis
4
2
8
3
lactones to polyesters from terminal epoxides reinvigo-
-hydroxypyridine could give rise to high ratios of β-
acidic, oxophilic counterions such as the depicted
þ
þ
[(porphyrin)Cr] salts 1/2 and [(salph)Al/Cr] salts 3a/b
8
(Figure 1).
5
rated interest in this reaction. Insubsequent studies, Alper
In a different approach, Rieger reported on a catalytic
system comprising Co (CO) and AlMe that proved
highly active for terminal epoxides. Using detailed in situ
(
1) For reviews, see: (a) Church, T. L.; Getzler, Y. D. Y. L.; Byrne,
2
8
3
C. M.; Coates, G. W. Chem. Commun. 2007, 657–674. (b) Nakano, K.;
Nazoki, K. Top. Organomet. Chem. 2006, 18, 223–238.
(
2) (a) Deng, F.-G.; Hu, B.; Sun, W.; Chen, J.; Xia, C.-G. Dalton.
Trans. 2007, 4262–4267. (b) Denmark, S. E.; Ahmad, M. J. Org. Chem.
007, 72, 9630–9634. (c) Kim, H. S.; Bae, J. Y.; Lee, J. S.; Jeong., C., II;
Choi, D. K.; Kang, S. O.; Cheong, M. Appl. Catal., A: Gen. 2006, 301,
2
(6) Lee, J. T.; Thomas, P. J.; Alper, H. J. Org. Chem. 2001, 66, 5424–
5426.
7
5. (d) Hinterding, K.; Jacobsen J. Org. Chem. 1999, 64, 2164–2165.
3) (a) Goodman, S. N.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2002,
1, 4703–4705. (b) Watanabe, Y.; Nishiyama, K.; Zhang, K.; Okuda, F.;
(7) Heck, R. F. J. Am. Chem. Soc. 1963, 85, 1460–1463.
(8) (a) Kramer, J. W.; Lobkovsky, E. B.; Coates, G. W. Org. Lett.
2006, 8, 3709–3712. (b) Schmidt, J. A. R.; Lobkovsky, E. B.; Coates,
G. W. J. Am. Chem. Soc. 2005, 127, 11426–11435. (c) Schmidt, J. A. R.;
Mahadevan, V.; Getzler, Y. D. Y. L.; Coates, G. W. Org. Lett. 2004, 6,
373–376. (d) Mahadevan, V.; Getzler, Y. D. Y. L.; Coates, G. W. Angew.
Chem., Int. Ed. 2002, 41, 2781–2784. (f) Getzler, Y. D. Y. L.; Mahadevan,
V.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc. 2002, 124, 1174–
1175.
(
4
Kondo, T.; Tsuji, Y. Bull. Chem. Soc. Jpn. 1994, 67, 879–882.
4) For reviews on β-lactones, see: (a) Pommier, A.; Pons, J. M.
Synthesis 1993, 441–459. (b) Yang, H. W.; Romo, D. Tetrahedron 1999,
5, 6403–6434.
5) Drent, E.; Kragtwijk, E. Eur. Pat. Appl. 577206, 1994; Chem.
Abstr. 1994, 120, 191517c.
(
5
(
1
0.1021/ol201043d r 2011 American Chemical Society
Published on Web 05/25/2011

ꢀ
generation of the nucleophilic [Co(CO) ] ion. In order to
4
test our hypothesis, we examined metal halides of varying
Lewis acidity in the carbonylation of cyclooctene oxide.
a
Table 1. Screening of Lewis Acidic Metal Halides
b
L
n
MX
n
Co
2
(CO)
8
P(CO)
(psi)
conv
(%)
entry
L
n
MX
n
(mol %)
(mol %)
Figure 1. Coates’ catalysts for epoxide carbonylation.
1
2
3
4
5
6
7
8
9
ꢀ
ꢀ
5.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.1
900
900
900
900
900
900
900
900
900
900
900
900
500
400
100
500
0
AlCl
ZnCl
3
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.5
0.5
0.5
0.5
0.1
0
IR studies, the in situ generated cobaltate salt [Me Al-
2
0
þ
ꢀ
9
2
(
diglyme)] [Co(CO) ] was identified as the active catalyst.
4
ZrCl
Cp TiCl
( PrO) TiCl
4
0
Most of the catalytic systems described above show high
2
2
trace
trace
24
i
activity for terminal epoxides; however, only salts 1 and 2
display high activity with 1,2-disubstituted meso-epoxides.
Catalysts 1 and 2 are air sensitive and require the use of
glovebox and Schlenk-line techniques for their synthesis
and handling, which impedes their general use within
nonspecialized laboratories.
We report here on a catalytic system comprised of
commercially available components that circumvents the
preparation and handling of air sensitive cobaltate salts.
Moreover, it displays activities similar or higher than the
best catalysts describedtodate for both meso- and terminal
epoxides.
2
2
TiBr
4
(TPP)CrCl
(TPP)AlCl
(TPP)FeCl
(TPP)CoCl
(TPP)CrCl
(TPP)CrCl
(TPP)CrCl
(TPP)CrCl
(TPP)CrCl
g98
18
1
0
trace
0
11
1
2
3
g98
g98
91
1
14
15
59
1
6
22
a
Reaction conditions: epoxide (2 mmol), THF (3 mL, 0.67 M).
Determined by H NMR spectroscopy of the crude material.
b
1
It is well-documented that Lewis bases (LBs) dispropor-
tionate Co (CO) to form homonuclear ion pairs
2
8
For initial catalyst screening we employed 1 mol % of a
metal halide and equimolar amounts of Co (CO) under
1
0
(
by halide sources has been the subject of a relatively few
HNIPs). However, the disproportionation of Co (CO)8
2
2
8
900 psi of CO pressure, with reactions run at 70 °C and a
1
1
publications. Inparticular, a report by Braterman caught
our attention, where it was shown that free or complexed
halide sources could catalyze the disproportionation of
reaction time of 16 h. A control experiment with 5 mol %
of Co (CO) failed to give any product (Table 1, entry 1).
2
8
AlCl , ZnCl , and ZrCl asprecatalystsgaveno conversion
3
ꢀ
11a
2
4
Co (CO) to generate the nucleophilic [Co(CO) ] ion.
8
i
2
4
(
entries 2ꢀ4), while Cp TiCl and ( PrO) TiCl gave only
2
2
2
2
Our group is interested in utilizing Lewis acidic metal
halides as multifunctional reagents, for instance, simulta-
traces of the product (entries 5 and 6). We were pleased to
find that TiBr gave a 24% conversion to the product
1
2
4
neously as an LA and a halide source. We hypothesized
that the use of a Lewis acidic metal halide in conjunction
with Co (CO) could form in situ a catalyst that would
confirming our hypothesis that anin situ generated catalyst
from a Lewis acidic metal halide can catalyze the carbo-
nylation of epoxides (entry 7). We examined (TPP)CrCl
next, which has been reported to form the strongly Lewis
2
8
combine LA activation of the epoxide with halide-assisted
þ
13
acidic [(TPP)Cr(L)2] ion in the presence of epoxides,
(
9) (a) Molnar, F.; Luinstra, G. A.; Allmendinger, M.; Rieger, B.
Chem.;Eur. J. 2003, 9, 1273–1280. (b) Allmendinger, M.; Eberhardt,
R.; Luinstra, G. A.; Molnar, F.; Rieger, B. Z. Anorg. Allg. Chem. 2003,
and were delighted to find that cyclooctene oxide was
completely converted to the β-lactone (entry 8). In com-
parison, (TPP)AlCl gave a conversion of 18%, whereas
(TPP)FeCl produced only traces of the product, with
629, 1347–1352.
(
10) (a) Sisak, A.; Ungv ꢀa ry, F.; Mark oꢀ , L. Organometallics 1983, 2,
1
2
1
244–1246. (b) Sisak, A.; Mark oꢀ , L. J. Organomet. Chem. 1987, 330,
01–206. (c) Fachinetti, G.; Fochi, G.; Funaioli, T. J. Organomet. Chem.
986, 301, 91–97. (d) Tasi, M.; Sisak, A.; Ungv ꢀa ry, F.; P ꢀa lyi, G.
(
TPP)CoCl being inactive (entries 9ꢀ11). These findings
Monatsh. Chem. 1985, 116, 1103–1105. (e) Absi-Halabi, M.; Atwood,
J. D; Forbus, N. P.; Brown, T. L. J. Am. Chem. Soc. 1980, 102, 6248–
are in agreement with Coates’ work and underline the
Sub-
8b,c
uniqueness of the chromium porphyrin systems.
6254.
(
sequent reaction parameter optimization using (TPP)CrCl
showed that lowering the (TPP)CrCl and Co (CO) load-
ing to 0.5 mol % (entry 12) and the CO pressure to 500 psi
11) (a) Braterman, P. S.; Walker, B. S.; Robertson, T. H. J. Chem.
Soc., Chem. Commun. 1977, 651–652. (b) Fachinetti, G.; Funaioli, T.;
Marcucci, M. J. Organomet. Chem. 1988, 353, 393–404. (c) Braterman,
P. S.; Leslie, A. E. J. Organomet. Chem. 1981, 214, C45–C49.
2
8
(12) (a) Akula, R.; Galligan, M. J.; Ibrahim, H. Chem. Commun.
2
2
009, 6991–6993. (b) Akula, R.; Galligan, M. J.; Ibrahim, H. Synthesis
011, 347–351.
(13) Chen, P.; Malcolm, H. C.; Gallucci, J. C.; Zhang, X.; Zhou, Z.
Inorg. Chem. 2005, 44, 2588–2595.
Org. Lett., Vol. 13, No. 12, 2011
3143

and or catalyst loading (entry 16) had a detrimental effect
on the conversion.
With the optimized conditions from Table 1, entry 13 in
hand, we set out to carbonylate a series of meso-epoxides
Table 2. (TPP)CrCl/Co (CO) Catalyzed Carbonylation of
2
meso-Epoxides (cat. = (TPP)CrCl)
8
a
(Table 2). Our system proved highly active in the carbo-
nylation of 12-, 8-, 7-, and 5-membered carbocyclic
epoxides with an epoxide/(TPP)CrCl loading of 200:1
(
Table 2, entries 2ꢀ5). Moreover, lowering the catalyst
loading to 1000:1 in the case of the 12-membered epoxide
still gave full conversion to the β-lactone (entry 2).
1
4
Dicarboxylate cyclopentene oxide was less reactive and
a loading of 100:1 was needed for a good conversion
(entry 6). We then examined carboacyclic meso-epoxides
and found that they displayed similar activities (entries
7
ꢀ11). However, 2,3-dibenzyl and 2,3-benzyloxymethyl
ethylene oxide proved less reactive and a higher loading
of 20:1 was required for a good conversion with the former
(entry 9).
We have also subjected epoxides derived from hetero-
cycles to our carbonylation conditions and were delighted
to find that they gave heterocyclic-fused β-lactone pro-
ducts in high yields. The epoxide derived from 2,5-dihy-
drofuran was carbonylated in 91% yield under the standard
loading of 200:1 (entry 12), whereas a higher loading of
5
0:1 was required to carbonylate the 1-tosyl-2,3,6,7-tetra-
hydro-1H-azepine derived epoxide, which was obtained in
7% yield (entry 13). An X-ray structure determination
8
confirmed the trans-stereochemistry of the tetrahydro-1H-
azepine derived β-lactone (Figure 2).
Figure 2. Tetrahydro-1H-azepine derived β-lactone with its
crystal structure drawn with 50% thermal ellipsoids.
In order to demonstrate the synthetic utility of these
heterocycle-fused β-lactones, we choose to convert the
tetrahydrofuran β-lactone into the literature known trans-
pentacin analogue trans-4-aminotetrahydrofuran-3-car-
1
5
boxylic acid. Thus, S 2 opening of the β-lactone with
N
NaN followed by azide reduction furnished the trans
3
1
6
β-amino acid in 56% overall yield (Scheme 1).
With our catalytic system displaying high activity with
meso-epoxides, we were interested in examining the
a
Reaction conditions: epoxide (2 mmol), THF (3 mL, 0.67 M),
b
(
2 8
TPP)CrCl/Co (CO) = 1:1.5, CO (500 psi), 70 °C, 16 h. Determined
1
by H NMR spectroscopy of the crude material; yield of isolated
β-lactone is given in parentheses.
(
14) Both, a synthesized and a purchased sample from Strem Che-
micals of (TPP)CrCl gave complete conversion to the product at a
loading of 1000:1.
(
15) Bunnage, M. E.; Davies, S. G.; Roberts, P. M.; Smith, A. D.;
Withey, J. M. Org. Biomol. Chem. 2004, 2, 2763–2776.
16) Yokota, Y.; Cortez, G. S.; Romo, D. Tetrahedron 2002, 58,
7075–7080.
(entry 13) still gave complete conversion to the product.
However, a further decrease in CO pressure (entries 14 and 15)
(
3
144
Org. Lett., Vol. 13, No. 12, 2011

Scheme 1. Conversion of Tetrahydrofuran-Fused β-Lactone
into the trans-4-Aminotetrahydrofuran-3-carboxylic Acid
Scheme 3. Proposed Mechanism for the Carbonylation of Five-
Membered Ring Epoxides
catalytic activity with terminal epoxides. As expected, our
(
TPP)CrCl/Co (CO) system proved highly active with
2 8
1,2-epoxybutane (EB) as substrate, displaying unprece-
dented TONs (Scheme 2).
Scheme 2. Catalytic Carbonylation of 1,2-Epoxybutane with the
(
2 8
TPP)CrCl/Co (CO) System
In addition to the cationic mechanism proposed by
8b
On a 100 mmol scale and with an EB/(TPP)CrCl loading
of 10000:1, EB was fully converted to β-valerolactone, as
judged by H NMR spectroscopy (Scheme 1). Lowering
the EB/(TPP)CrCl loading even further to 20000:1 on the
same scale still gave a conversion of 80%.
Coates, we propose that a ring closure preceding epimer-
ization of the neutral chromium alkoxide/cobalt acyl inter-
mediate is feasible under the Lewis acidic reaction conditions.
The exact nature of this epimerization is not clear at present.
In conclusion, we have disclosed an in situ generated
catalytic system that allows the ring-expansive carbonyla-
tion of epoxides using low precatalyst loadings. Our
unique approach to catalyst generation permits the use
of commercially available air stable components thereby
eliminating the need for glovebox and Schlenk-line tech-
niques required for the preparation and handling of air
sensitive cobaltate salts. We anticipate that our new ap-
proach could greatly expand the general utility of this
important transformation.
1
We believe that, in the presented catalytic system, a
catalyst analogous to Coates’ chromium cobaltate com-
plex 1 is generated in situ by a chloride-assisted dispropor-
tionation of Co (CO) according to eq 1, where L = THF
2
8
1
1
or epoxide.
3
2
4 L
þ
f
s 2[(TPP)Cr(L) ] [Co(CO) ]
2 4
ꢀ
Co2(CO) þ 2(TPP)CrCl
8
þ CoCl2 þ 4CO
ð1Þ
However, as this reaction occurs in the presence of
epoxide and CO, we propose that a very fast ring opening
Acknowledgment. This work was supported by a UCD
Ad Astra Scholarship to P.G., the Centre for Synthesis and
Chemical Biology and in part by Science Foundation
Ireland (07/RFP/CHEF394). We thank Dr. Helge
M u€ ller-Bunz (UCD) for X-ray structure determination
and Drs Jimmy Muldoon (UCD) and Yannick Ortin
ꢀ
of the epoxide by [Co(CO) ] occurs within the ion pair
4
þ ꢀ
(TPP)Cr(L)(epoxide)] [Co(CO) ] , followed by an im-
4
[
mediate CO insertion, leading to a neutral chromium
7
1
alkoxide/cobalt acyl intermediate.
Finally, the observed cis stereochemistry in the case of
five-membered ring β-lactones (Table 2, entries 5, 6, and
(
UCD) for NMR spectra.
12) raises the possibility of an alternative mechanistic
pathway in operation to that of larger rings (Scheme 3).
Supporting Information Available. Experimental car-
bonylation procedures and analytical data of all β-lac-
tones. This material is available free of charge via the
Internet at http://pubs.acs.org.
(
17) For detailed mechanistic studies with aluminum complex 3a, see:
Church, T. L.; Getzler, Y. D. Y. L.; Coates, G. W. J. Am. Chem. Soc.
006, 128, 10125–10133.
2
Org. Lett., Vol. 13, No. 12, 2011
3145