Cobalt-catalyzed coupling of alkyl iodides with alkenes: Deprotonation of hydridocobalt enables turnover

Article abstract of DOI:10.1002/anie.201105235

Completing the cycle: Cobalt complexes I and II were found to catalyze intramolecular alkyl Heck-type coupling reactions of alkyl iodides to alkenes upon irradiation with visible light in the presence of a tertiary amine base. The use of base is key to the catalytic turnover. The compatibility of the method with a broad range of functional groups opens opportunities for synthesis. Copyright

Full text of DOI:10.1002/anie.201105235

DOI: 10.1002/anie.201105235
Cobalt-Catalyzed Cyclizations
Cobalt-Catalyzed Coupling of Alkyl Iodides with Alkenes:
Deprotonation of Hydridocobalt Enables Turnover**
Matthias E. Weiss, Lukas M. Kreis, Alex Lauber, and Erick M. Carreira*
The use of radical reactions to assemble complex molecular
architectures is acknowledged as a very powerful tool in
organic synthesis.[1] The fact that transition metals can
influence the reaction outcome has added further value to
this field.[2] Among the various metals used in radical
additions, cobalt has a unique position.[3] A particularly
useful application formally corresponds to a Heck-type
coupling, in which an alkyl group derived from an alkyl
halide undergoes addition to an olefin followed by unsatura-
tion to return a substituted olefin.[4] Herein, we report an
intramolecular, cobalt-catalyzed alkyl Heck-type reaction of
alkyl iodides to olefins (Scheme 1, 1!2). The transformation
proceeds under exceptionally mild reaction conditions upon
irradiation of the reaction vessel with visible light in the
presence of an amine base. Significantly, catalysis is enabled
by the presence of a base that deprotonates a putative
hydridocobalt intermediate.
Scheme 2. Previous use of organocobalt reagents/catalysts in alkyl
Heck-type couplings.
ates (i.e. radicals), which are known to participate in a host of
reactions.[10] In synthetic applications employing stoichiomet-
ric organocobalt reagents the reactive intermediates most
typically undergo addition to an olefin to give a new
organocobalt species. These then undergo disproportionation
(or formal b-hydride elimination) to give a new olefin and an
hydridocobalt(III) (step iv). The latter subsequently decom-
pose to H2 and ill-defined CoII complexes (step v).[11b]
A number of groups have recognized that the regener-
ation of a cobalt(I) species (step vi) would enable catalysis
(steps ii–vi). This has been most notably effected under
reductive conditions, namely electrolytically, by Zn or by the
stoichiometric use of Grignard reagents.[12]
Scheme 1. Cobalt-catalyzed Heck-type cyclization.
Careful analysis of these methods reveals potential
complications with respect to functional group compatibility
(e.g. aldehydes and ketones); indeed, it has been noted that
enoates are unsuitable substrates in such Heck-type cou-
plings.[12a] We became intrigued with the possibility of
generating a catalytic cycle under nonreductive conditions.
Consequently, we envisioned the conversion of the hydrido-
Alkyl- or stannyl cobaloximes (bis(dimethylglyoximato)-
cobalt), such as I and II and structurally related complexes,
have emerged as useful radical precursors[3,5b] and have been
intensively investigated as model systems for coenzyme B12.[6]
They are traditionally obtained in moderate yields by the
reaction of a preformed cobalt(I) complex with an electro-
phile (Scheme 2, step ii).[7,8] Once formed, these cobaloximes
are air-stable and can be stored for months on the bench if
kept in the dark.[9] The light emitted by an incandescent lamp
cobalt intermediate to the desired CoI species by the action of
I
base (CoIII H + R3N!Co + R3NH+). In this respect,
ꢀ
Schrauzer et al. have shown that hydridocobaloximes possess
pKa values of approximately 10 and are readily deprotonated
under alkaline conditions.[11]
ꢀ
is usually sufficient to induce homolytic Co R bond cleavage
(step iii) and generate a CoII complex along with intermedi-
In our working hypothesis for the Heck-type coupling of
alkyl iodides and olefins, bench-stable cobaloxime complexes
I and II were expected to provide convenient entry into a
catalytic cycle. For example the reaction of I[5b] and an alkyl
iodide under irradiation would produce Ph3SnI (3) and A
(Scheme 3), which in turn should add to an olefin to afford
adduct B. Following disproportionation, B would furnish a
new olefin and hydridocobalt C. If the deprotonation of C
could be effected under reaction conditions compatible with
the overall process then a CoI species D would be generated.
A subsequent reaction with the starting alkyl iodide would
[*] M. E. Weiss, L. M. Kreis, A. Lauber, Prof. Dr. E. M. Carreira
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich, HCI H335
Wolfgang-Pauli-Strasse 10, 8093 Zꢀrich (Switzerland)
E-mail: carreira@org.chem.ethz.ch
Homepage: https://www.carreira.ethz.ch
[**] We thank the ETH Zꢀrich for financial support and the Swiss
National Science Foundation.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201105235.
Angew. Chem. Int. Ed. 2011, 50, 11125 –11128
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11125

Communications
experiments. As irradiation with light proved to be crucial,
the use of another light source was investigated. Commercial
blue-light-emitting diodes[16] (LEDs; lmax 465 nm) were
selected on the basis of the absorption spectra of I and
II.[3f,5a] We were pleased to find that the reaction proceeded
smoothly upon irradiation with blue LEDs, and full con-
version was achieved after 16 h using only 10 mol% catalyst
(Table 1, entry 5). When using blue LEDs, base was also
necessary (Table 1, entry 6). The reaction rate depends
strongly on the light density, and the best results were
obtained if the flask was placed at a 5 mm distance from the
LEDs.[17] Tin-free catalyst II proved to be equally as
efficient[18] as I but afforded significantly higher amounts of
elimination product 6 for this specific substrate (Table 1,
entry 7). This is rather an unexpected outcome, which
suggests that the stannane coproduct 3 is not an innocent
bystander. Interestingly, temperature had no effect on the
reaction outcome and identical conversion and selectivity was
observed at 08C, 258C, and 658C.
With the optimized reaction conditions in hand, we
investigated the functional group tolerance and the scope of
the method (Table 2). The cobalt-catalyzed cyclization fol-
lows the same trend observed for free radical cyclizations, that
is, the formation of 5-membered rings is highly preferred and
products arising from 6-endo addition were not observed. In
contrast to the model substrate 4, elimination to furnish
dienes was not observed to be in competition with the
formation of 5-membered-ring products when either I or II
were examined as catalysts. Addition to terminal, di- and
trisubstituted olefins proceeds with high efficiency, and the
last of these olefins is particularly noteworthy as it leads to a
Scheme 3. Working hypothesis for base-mediated catalytic alkyl Heck-
type couplings.
regenerate A and thereby complete the catalytic cycle.
Despite the simplicity of the proposed turnover step there is
no literature report of its application to enable catalysis.
The hypothesis delineated above was investigated with
substrate 4 using the readily prepared, air- and moisture-
stable catalyst I.[5] Irradiation (500 W sunlamp) of a 0.05m
solution of 4 (1 equiv), I (0.2 equiv), and iPr2NEt (1.5 equiv)
in degassed[13] MeCN for 16 h resulted in nearly full con-
sumption (95%) of 4 and produced a 5:1 mixture of the
cyclization product 5 and the elimination product 6 (Table 1,
entry 1).[14] With the proof of concept validated, the effect of
varying the reaction parameters was investigated for the
purposes of optimization.[15] Using the initial reaction con-
ditions but with a lower catalyst loading (0.1 equiv) led to
incomplete conversion (48%; Table 1, entry 2). At higher
concentrations of the reactants (0.1m in 4) a significant drop in
reactivity (< 30% conversion) was observed. The use of
benzene as solvent[5b] did not prove advantageous. The
substitution of iPr2NEt for Et3N or 1,8-bis(dimethylamino)-
naphthalene did not significantly affect the conversion (see
the Supporting Information). The importance of the cobalt
catalyst (Table 1, entry 3) and irradiation (Table 1, entry 4)
was established by conducting the corresponding control
cyclization product incorporating
(Table 2, entry 4).
a quaternary center
As shown in Table 2, catalyst I outperforms II in the
cyclization reactions. Importantly, secondary and sterically
shielded primary iodides are only efficiently transformed to
the cyclized products by I. The stereochemical information of
secondary iodides is lost during the reaction (Table 2,
entry 5). The formation of 6-membered rings proceeds only
in synthetically useful yields if the olefin is activated with an
electron-withdrawing group. The reaction conditions proved
to be compatible with a broad range of functional groups
including aldehydes, ketones, aryl iodides and aryl bromides,
olefins, acetals, carbamates, pyridines, thiophenes, and unpro-
tected alcohols (Table 2; entries 9–15). We note that sub-
strates including nitroarenes, a-bromo- and a-chloro esters
are unsuitable for the reaction, as the starting material
remained unreacted. Additionally, mixtures of isomeric olefin
products are observed when there are multiple b-H elimi-
nation pathways for the intermediate organocobalt species, as
previously noted by Giese et al.[12b]
The utility of the method described is further highlighted
by a concise synthesis of (ꢁ)-samin (Scheme 4). Compound 7
was subjected to iodoetherification with NIS/allyl alcohol and
the product was then subjected to the optimal reaction
conditions, to afford known intermediate 8[19] as a single
diastereoisomer in excellent yield. Tetrahydrofurane 8 was
converted into (ꢁ)-samin following the sequence previously
described.
Table 1: Effect of reaction parameters on the cobalt-catalyzed alkyl Heck-
type reaction.
Entry
Cat. (equiv)
Light source
Conversion [%] (5/6)[a]
1
2
3
4
I (0.2)
I (0.1)
none
I (0.1)
I (0.1)
I (0.1)
II (0.1)
sunlamp[b]
sunlamp[b]
sunlamp
95 (5:1)
48 (5:1)
0
<5
>95 (5:1)
10
no light[c]
blue LEDs
blue LEDs
blue LEDs
5
6[d]
7
>95 (1.3:1)
Reaction conditions: 4 (0.05m) in degassed MeCN. The sealed reaction
vessel was placed in front of the light source (25 cm distance from
sunlamp, 5 mm from LEDs) and irradiated for 16 h. [a] Conversion based
on 1H NMR analysis of crude reaction mixture. [b] 500 W tungsten
halogen lamp. [c] Conducted at RT or 658C. [d] No iPr2NEt added.
Bn=benzyl.
11126 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11125 –11128

Table 2: Scope and functional group tolerance of the cobalt-catalyzed
alkyl Heck-type cyclization.
In summary, we have described a novel cobalt-catalyzed,
intramolecular Heck-type coupling of alkyl iodides with
olefins. The mild reaction conditions (iPr2NEt, cat. cobalox-
ime I and II, blue LEDs, RT) render the method compatible
with a wide range of functional groups, such as amides, esters,
ketones, and aldehydes. The development of the catalytic
method follows from the key insight that a mild base can
deprotonate a hydridocobalt intermediate and thereby regen-
erate a complex that is catalytically competent. In effect
Entry
1
Reactant
Product
Catalyst: yield [%]
I: 86
II: 91
ꢀ
proton abstraction from an intermediate Co H formally
corresponds to reduction at the metal center.[20] Future work
will be dedicated to explore the effects of changing the
catalyst structure, the axial ligand, and to expand this method
to substrates beyond alkyl iodides and intramolecular cou-
plings.
I: 91
II: 85
2
3
I: 83
Received: July 26, 2011
Published online: October 4, 2011
II: 47[a]
Keywords: blue LEDs · cobalt · cyclization ·
.
I: 83[b]
homogeneous catalysis · radical reactions
4[f]
II: 50[a]
I: 82
5[f]
II:<5[a]
[1] C. P. Jasperse, D. P. Curran, T. L. Fevig, Chem. Rev. 1991, 91,
1237 – 1286; “Radical Cyclization Reactions”: B. Giese, K.
Gçbel, J. Dickhaut, K. J. Kulicke, F. Trach in Organic Reactions,
Vol. 48 (Ed.: L. Paquette), Wiley, New York, 1996, pp. 301 – 348.
[2] a) J. Iqbal, B. Bhatia, N. K. Nayyar, Chem. Rev. 1994, 94, 519 –
564 and references therein; for a recent example of cobalt
catalysis, see: b) B. Gaspar, E. M. Carreira, J. Am. Chem. Soc.
2009, 131, 13214 – 13215.
I: 76[c]
6
II: 52[c]
I: 89
7
II:38[a]
[3] a) G. Pattenden, Chem. Soc. Rev. 1988, 17, 361 – 382, and
references therein; b) M. Okabe, M. Tada, Bull. Chem. Soc.
Jpn. 1982, 55, 1498 – 1503; c) B. P. Branchaud, M. S. Meier, Y.
Choi, Tetrahedron Lett. 1988, 29, 167 – 170; d) B. Giese, J.
Hartung, J. He, O. Hꢀter, A. Koch, Angew. Chem. 1989, 101,
334 – 336; Angew. Chem. Int. Ed. 1989, 28, 325 – 327; e) G. N.
Schrauzer, R. J. Windgassen, J. Am. Chem. Soc. 1966, 88, 3738 –
3743; f) G. N. Schrauzer, L. P. Lee, J. W. Sibert, J. Am. Chem.
Soc. 1970, 92, 2997 – 3005; g) M. Tada, Rev. Heteroat. Chem.
1999, 20, 97 – 144.
[4] For examples of Pd-catalyzed alkyl Heck reactions see: a) L.
Firmansjah, G. C. Fu, J. Am. Chem. Soc. 2007, 129, 11340 –
11341, and references therein; for a carbonylative alkyl Heck
reaction, see: b) K. S. Bloome, E. J. Alexanian, J. Am. Chem.
Soc. 2010, 132, 12823 – 12825.
8
9
X=4-MeC6H4
X=4-CNC6H4
X=4-IC6H4
II: 91
II: 81
10
11
12
13
14
15
II: 84
X=2-BrC6H4
II: 71[d]
II: 68[e]
II: 74
X=4-(CH2OH)C6H4
X=4-(CHO)C6H4
X=thiophen-2-yl
X=pyridin-2-yl
II: 78
II: 83[e]
All experiments were run in degassed MeCN (0.05m in substrate) in
sealed Schlenk tubes. [a] Conversion based on 1H NMR analysis of
unpurified reaction mixture. [b] Yield after Jones oxidation of unpurified
cyclization product. [c] Elimination accounts for the remaining material.
[d] Catalyst II (0.3 equiv) used. [e] 48 h reaction time. [f] The relative
configuration of the cyclized products was assigned in analogy to
literature precedence for the reductive cyclization of these substrates.
See Ref. [21]. BOC=tert-butyloxycarbonyl.
[5] a) G. N. Schrauzer, G. Kratel, Chem. Ber. 1969, 102, 2392 – 2407;
for an application in free radical cyclization, see: b) M. Tada, K.
Kaneko, J. Org. Chem. 1995, 60, 6635 – 6636.
[6] “Chemistry and Significance of Vitamin B12 model systems”: J.
Halpern in B12, Vol. 1. (Ed.: D. Dolphin), Wiley, New York,
1982, pp. 501 – 543.
[7] a) “12. Bis(dimethylglyoximato)cobalt complexes”: G. N.
Schrauzer in Inorganic Syntheses, Vol. XI (Ed.: W. L. Jolly),
McGraw-Hill Book Company, New York, 1968, pp. 61 – 70;
b) A. Ali, B. G. Gill, G. Pattenden, G. A. Roan, T.-S. Kam, J.
Chem. Soc. Perkin Trans. 1 1996, 1081 – 1093.
[8] For an alternative approach involving hydrocobaltation of
olefins: See Ref [3e].
[9] The thermal stability of alkyl cobaloximes ([Co]–R): R = Me @
Et ꢂ iPr@ tBu (cobaloximes with tertiary alkyl groups were not
isolable), see Ref. [8].
Scheme 4. Synthesis of (ꢁ)-samin.
Angew. Chem. Int. Ed. 2011, 50, 11125 –11128
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11127

Communications
[10] a) D. N. Ramakrishna, R. Symons, M. Symons, J. Chem. Soc.
[15] See the Supporting Information for a detailed table.
[16] For blue LEDs in the context of Ru(bpy)32+, see a) D. A.
Nicewicz, D. W. C. MacMillan, Science 2008, 322, 77 – 80; for
recent developments employing blue LEDs in displacement
reactions, see: b) L. Furst, B. S. Matsuura, J. M. R. Narayanam,
J. W. Tuckler, C. R. J Stephenson, Org. Lett. 2010, 12, 3104 –
3107.
[17] The cyclization is well scalable if the light density (= number of
LEDs) is increased proportional to the scale. Otherwise, longer
reaction times are necessary to achieve full conversion of the
Faraday Trans. 1 1984, 80, 423 – 434; b) B. D. Gupta, S. Roy,
Inorg. Chim. Acta 1988, 146, 209 – 221.
[11] a) G. N. Schrauzer, R. J. Windgassen, J. Kohnle, Chem. Ber. 1965,
98, 3324 – 3333; b) G. N. Schrauzer, R. J. Holland, J. Am. Chem.
Soc. 1971, 93, 1505 – 1506; c) G. N. Schrauzer, R. J. Holland, J.
Am. Chem. Soc. 1971, 93, 4060 – 4062; d) H. L. Stokes, M. E.
Welker, Organometallics 1996, 15, 2624 – 2632; e) H. Bhandal, G.
Pattenden, J. Chem. Soc. Chem. Commun. 1988, 1110 – 1112.
[12] a) W. Affo, H. Ohmiya, T. Fujioka, Y. Ikeda, T. Nakamura, H.
Yorimitsu, K. Oshima, Y. Imamura, T. Mizuta, K. Miyoshi, J.
Am. Chem. Soc. 2006, 128, 8068 – 8077; b) B. Giese, P. Erdmann,
T. Gobel, R. Springer, Tetrahedron Lett. 1992, 33, 4545 – 4548;
c) S. Torii, T. Inokuchi, T. Yukawa, J. Org. Chem. 1985, 50, 5875 –
5877; for catalytic, intermolecular additions to styrene, see:
d) B. P. Branchaud, W. D. Detlefsen, Tetrahedron Lett. 1991, 32,
6273 – 6276; for a reductive example, see: e) M. Okabe, M. Abe,
M. Tada, J. Org. Chem. 1982, 47, 1775 – 1777.
substrate. See the Supporting Information for
a detailed
experimental setup and for the cyclization of the substrate
shown in Table 2, entry 1 on a 1 mmol scale.
[18] Experiments in which the initiation was monitored by 1H NMR
spectroscopy suggest that II enters the catalytic cycle by forming
C and propene.
[19] G. Maiti, S. Adhikari, S. Chandra Roy, Tetrahedron 1995, 51,
8389 – 8396.
[13] Solutions of I and II are air stable in the dark. Exposure to
intensive light sources requires rigorous exclusion of oxygen. For
the reaction of alkyl cobaloximes with oxygen, see: C. Giannotti,
A. Gaudemer, C. Fontaine, Tetrahedron Lett. 1970, 11, 3209 –
3212.
[20] The fact that deprotonation of a hydridometal species results in
formal reduction follows from the ambiguity of assigning
oxidation states and is discussed in Ref. [11c]. Yet, this is not
uncommon, for example cobalt tetracarbonyl hydride is formally
Co+I but is known to have a pKa of 8.5.
[14] Products arising from reductive cyclization or deiodination were
not observed. The formation of 6 can be rationalized by
competing disproportionation before cyclization.
[21] a) X. J. Salom-Roig, F. Dꢁnꢂs, P. Renaud, Synthesis 2004, 1903 –
1928; b) A. Vaupel, P. Knochel, J. Org. Chem. 1996, 61, 5743 –
5753.
11128 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 11125 –11128