An alternative mechanism for the cobalt-catalyzed isomerization of terminal alkenes to (Z)-2-alkenes

Article abstract of DOI:10.1002/anie.201409902

The cobalt-catalyzed selective isomerization of terminal alkenes to the thermodynamically less-stable (Z)-2-alkenes at ambient temperatures takes place by a new mechanism involving the transfer of a hydrogen atom from a Ph₂PH ligand to the starting material and the formation of a phosphenium complex, which recycles the Ph₂PH complex through a 1,2-H shift.

Full text of DOI:10.1002/anie.201409902

Angewandte
Chemie
DOI: 10.1002/anie.201409902
Synthetic Methods
An Alternative Mechanism for the Cobalt-Catalyzed Isomerization of
Terminal Alkenes to (Z)-2-Alkenes**
Anastasia Schmidt, Alexander R. Nçdling, and Gerhard Hilt*
Abstract: The cobalt-catalyzed selective isomerization of
terminal alkenes to the thermodynamically less-stable (Z)-2-
alkenes at ambient temperatures takes place by a new mech-
anism involving the transfer of a hydrogen atom from a Ph2PH
ligand to the starting material and the formation of a phosphe-
nium complex, which recycles the Ph2PH complex through
a 1,2-H shift.
some alkenes, such as allylbenzene, the (E)-2-alkene was the
primary product in the first place.
During the course of our continuing investigations of
cobalt catalyst systems for the isomerization of 1,3-dienes,[5]
we found that the dpppBuEt ligand[6] allowed the desired
transformation of 1-hexadecene (1, R = C9H21; 95% conver-
sion, ambient temperature, 5 h) to E/Z-2 with only minor
amounts of the undesired higher homologues. Unfortunately,
the results obtained with a catalyst mixture consisting of
[CoBr2(dpppBuEt)], zinc powder, and zinc iodide were
difficult to reproduce and, therefore, we started a detailed
investigation into the reason for the inconsistent results.
Eventually, we realized that an impurity in the ligand,
a remnant of the ligands synthesis (Ph2PH), could be
responsible for the formation of the active cobalt catalyst.
Accordingly, when Ph2PH was added in catalytic amounts to
the precatalyst mixture, the reaction of 1-hexadecene pro-
ceeded within 1.5 h to 95% conversion and an E/Z ratio of
15:85. From this point onward the cobalt-catalyzed isomer-
ization was highly reproducible and could be performed
under very convenient reaction conditions on a millimolar
scale, often within 1–3 h. The transformations were conducted
in anhydrous dichloromethane at ambient temperature (with
few exceptions) under exclusion of oxygen and with com-
mercially available starting materials, without the need for
previous purification.
T
he isomerization of alkenes comprises two reaction modes.
The first is the inversion of the double-bond configuration.
This type of isomerization can be realized by various methods,
such as by photochemical isomerization[1] or by several
chemical transformations.[2] The second mode for the isomer-
ization of an alkene is the translocation of the double bond
along a carbon chain. When a terminal alkene, such as 1, is
subjected to the reaction, isomerization of the double bond to
position 2 leads to an E/Z mixture of the 2-alkene (2). With
transition-metal catalysts, the isomerization is often realized
via a transition-metal hydride (Met-H) as the active catalyst
species (Scheme 1).[3] According to the general mechanism
We were able to demonstrate in control experiments that
all the components of the catalyst mixture must be present to
obtain the desired transformation in a short reaction time at
ambient temperatures. We also ascertained that the amount
of Ph2PH has an influence on the reaction rate as well as on
the E/Z ratio. The reaction proceeds faster with a higher
Ph2PH content, but with a reduced excess of the Z isomer.
The isomerization reaction of 1-alkenes using the cobalt
precatalyst with dppp-type ligands (such as dppp, bdpp, or
dpppBuEt)[6] as well as Ph2PH as an additive under the
conditions exemplified in Scheme 2 could be applied to
a variety of other terminal alkenes. The results of these
transformations are summarized in Table 1.
Scheme 1. General mechanism for the isomerization of alkenes by
metal hydride species.
for such isomerization reactions via metal hydride catalysts,
the addition and b-hydride elimination by the Met-H result in
isomerization by a formal 1,3-H shift.
Of the several transition-metal-catalyzed isomerization
reactions, Weix, Holland, and co-workers described a cobalt
catalyst which exhibits a high preference for the formation of
the (Z)-2-alkenes from 1-alkenes.[4] Unfortunately, this par-
ticular cobalt catalyst also caused further isomerization of the
double bond along the carbon chain so that higher isomers (3-
alkenes and so on) were also detected. Furthermore, the
primary (Z)-2-alkene product was isomerized to the thermo-
dynamically more-stable (E)-2-alkene over time, and for
The isomerization reactions with dppp are the fastest, but
the E/Z ratios are inferior to those when other ligands are
applied. The use of the bdpp ligand gave the best results for
[*] Dipl.-Chem. A. Schmidt, Dipl.-Chem. A. R. Nçdling, Prof. Dr. G. Hilt
Fachbereich Chemie, Philipps-Universitꢀt Marburg
Hans-Meerwein-Strasse, 35043 Marburg (Germany)
E-mail: Hilt@chemie.uni-marburg.de
[**] The donation of dpppBuEt from BASF SE is gratefully acknowl-
edged.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201409902.
Scheme 2. Cobalt-catalyzed isomerization of terminal alkenes.
Angew. Chem. Int. Ed. 2015, 54, 801 –804
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
801

.
Angewandte
Communications
Table 1: Results of the cobalt-catalyzed isomerization of alkenes.[a]
Entry
L
1-Alkene (3)
Main product (4)
Yield [%]
E/Z ratio
(conversion [%])
(sum of other isomers [%])
1
2
dpppBuEt
bdpp
99 (95)
92 (90)
15:85 (3)
14:86 (<1)
3
4
dppp
dppp
80[b] (51)
59 (100)
55:45 (<1)
100:0 (<1)
5
6
bdpp
84[c] (83)
86 (89)
8:92 (<1)
dpppBuEt
13:87 (<1)
7
8
9
bdpp
91[d] (100)
99 (92)
19:81 (<1)
16:84 (<1)
14:86 (<1)
dpppBuEt
dpppBuEt
95 (96)
10
11
12
dpppBuEt
bdpp
71 (93)
26:74 (<1)
4:96 (3)
96[e] (97)
94[d] (95)
bdpp
14:86 (5)
13
bdpp
80 (95)
16:84 (<1)
14
15
dppp
84[f] (93)
91 (88)
83:17 (9)
dpppBuEt
13:87 (<1)
16
17
18
19
dppp
bdpp
bdpp
dppp
99 (100)
84[f] (84)
98 (100)
72 (88)
100:0 (<1)
18:82 (<1)
87:13 (<1)
22:78 (8)
[a] Conditions: CoBr2(L) (5–10 mol%), Zn (10–20 mol%), ZnI2 (10–20 mol%), Ph2PH (2.5–5 mol%), and 1-alkene 3 (0.5 mmol, 1.0 equiv), CH2Cl2, 1–
24 h, rt (unless otherwise noted). Unreacted starting material and other isomers were not separated and are included in the yield (unless otherwise
noted). The conversion and E/Z ratio were determined by 1H NMR spectroscopy, GC, or GC-MS analysis. The amount of other isomers was
determined by GC or GC-MS analysis. [b] After 7 h, another 10 mol% CoBr2(L), 20 mol% Zn, 20 mol% ZnI2, and 5 mol% Ph2PH were added.
[c] 15 mol% CoBr2(L), 30 mol% Zn, 30 mol% ZnI2, and 7.5 mol% Ph2PH were used. [d] The reaction proceeded at 38C. [e] The reaction proceeded at
ꢀ38C. [f] The starting material was separated by flash chromatography. Bpin=(pinacolato)boron; TBS=tert-butyldimethylsilyl.
802
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 801 –804

Angewandte
Chemie
the branched and sterically more-demanding substrates,
whereas the dpppBuEt ligand was the ligand of choice for
the simple unbranched substrates. Simple alkenes without or
with functional groups at least two carbon–carbon bonds away
from the terminal alkene subunit, such as 3a, 3b, 3h–j, and 3l,
react smoothly and give the desired (Z)-2-alkenes in high
yields and good stereoselectivities. The amount of E isomer
increased significantly when an aryl substituent was placed in
the homoallylic position (3c), and the E isomer was obtained
as a single isomer when allylbenzene 3d or allyl ether 3p were
applied. Interestingly, additional methyl substituents in the
homoallylic position (3e, 3 f, 3q) were tolerated and, espe-
cially for 3e (entry 5), significantly better results were
obtained (compared to 3c, entry 3), with the more complex
Scheme 3. Proposed mechanism of the cobalt-catalyzed isomerization
of alkenes in the presence of a Ph2PH additive.
starting
material
increasing
the
selectivity
to
E/Z = 8:92. Fortunately, the highest Z selectivity was observed
for homoallylboronic pinacol ester 3k. In this case, the cobalt
catalyst led to the Z-configured allylboron building block 4k
in excellent yield and selectivity, with only minor amounts of
other isomers. In some cases the reactivity was so high that
lower reaction temperatures were necessary to obtain good
results (4g, 4k, 4l). An interesting observation is that
a carbonyl group in the vicinity of the alkene, such as in 3n,
leads to an inversion of the E/Z ratio (entries 14 and 15),
which can be rationalized by additional coordination of the
carbonyl moiety. Furthermore, a kinetic discrimination was
observed for 3r, where only the allylic ether double bond was
isomerized predominantly to the E isomer. It is noteworthy
that prolonged reaction times led to undesired isomerization
of the (Z)-2-alkene to the (E)-2-alkene (see the Supporting
Information). This process starts as soon as around 90% of the
starting material has been consumed. The most outstanding
feature of the cobalt-catalyzed isomerization reaction is the
fact that higher homologues are only found in trace amounts.
Increased amounts are found only after prolonged reaction
times (up to 9% after 24 h reaction for 1-hexadecene (3a)).
This finding led us to the conclusion that the isomerization
does not follow the generally accepted mechanism of addition/
elimination of a cobalt hydride species to the double bond.[3]
Instead, the role of the Ph2PH additive in the reaction
mechanism must be taken into consideration, which prompted
us to the following mechanistic proposal (Scheme 3).
In the low-valent cobalt(I) complex 5, generated by
reduction of the cobalt(II) precatalyst, the terminal alkene,
the dppp ligand, as well as the Ph2PH additive are coordinated.
This arrangement results in a vacant coordination site on the
cobalt center.[7] The outstanding activity of the cobalt complex
can be explained by the transfer of the hydrogen atom from the
P-H subunit to the terminal carbon atom of the double bond
and the formation of an alkyl cobalt species (6). The vacant
coordination site on the cobalt center allows b-hydride
elimination from the allylic position, which leads to migration
of the double bond. This results in an overall 1,3-hydrogen shift
in the substrate and formation of a cobalt hydride species 7.
Accordingly, the Ph2PH ligand in 5 is transformed into
a phosphenium-type ligand in structures 6 and 7.[8] If inter-
mediate 7 is short-lived and a 1,2-hydrogen shift from the
cobalt to the phosphorus center (7!8) takes place, the lifetime
of the cobalt-hydride intermediate 7 is reduced, which prohibits
the formation of higher homologues through re-addition of the
cobalt hydride to the 2-alkene and so on. The steric bulk of the
dppp ligand must be responsible for the kinetically controlled
formation of the Z isomer. Furthermore, the lower binding
affinity of the cobalt catalyst to the 2-alkene compared to
terminal alkenes then causes the fast exchange of the product
toward the starting material, and the active cobalt species 5 is
formed to complete the catalytic cycle of the reaction.
The formation of the alkyl-cobalt species 9 through
a nonproductive side reaction must also be taken into
consideration. Although this process does not lead to the
desired 2-alkene, the process is important to explain the
incorporation of deuterium in the product when deuterium-
labeled starting materials were used.
First, stoichiometric amounts of Ph2PD (96% D incorpo-
ration) were applied to verify that the deuterium atom is
transferred to the terminal carbon atom (5 to 6) and to the
carbon atom at position-2 via 5 to 9 of the 1-hexadecene in the
first step. The deuterium content of the isolated (E/Z)-2-
hexadecene in position-1 was found to be high (51%) and
a considerable amount of deuterium was detected at the
olefinic carbon atom at position-2 (18%, see the Supporting
Information). As an alternative mechanistic proposal, we
considered a reversible cobalt-catalyzed hydrophosphina-
tion,[9] the anti-Markovnikov or Markovnikov addition of
Ph2PH to 1-octene (10, Scheme 4), for the formation of 11 and
12. The 2-deuterated starting material can be generated from
11, whereas the 1-deuterium-labeled starting material or the
desired Z-configured product 13 would be accessible from 12.
To verify that the alternative mechanism was relevant in
the cobalt-catalyzed isomerization, the non-deuterated
isomer of 11, 1-octyldiphenylphosphine, was applied as the
additive instead of Ph2PH for the isomerization of 1-
hexadecene (3a). The reaction proceeded very slowly (49%
conversion after 25 h), but (Z)-2-hexadecene (4a) was
generated as with Ph2PH as additive. 1-Octene or (Z)-2-
octene should have been detected as side products if the alkyl
phosphine was an intermediate in the reaction, which was not
the case. Furthermore, no isomerization product (4a) was
detected when the Markovnikov-type product of the hydro-
phosphination, 2-octyldiphenylphosphine (12), was used as
the additive. Thus, the cobalt-catalyzed addition/elimination
of Ph2PH can be excluded as an alternative mechanism
Angew. Chem. Int. Ed. 2015, 54, 801 –804
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
803

.
Angewandte
Communications
suspended in anhydrous dichloromethane (0.5 mL) under argon. After
stirring the mixture for 10 min, Ph2PH (2.5–5 mol%; 0.125m in
CH2Cl2) was added and after additional stirring for 10 min, 1-alkene 3
(0.50 mmol; 1 equiv) was added. The reaction mixture was stirred at
the required temperature and the progress of the reaction was
monitored by GC or GC-MS analysis. The reaction mixture was
filtered over a short pad of silica gel with pentane or pentane/diethyl
ether as eluent. The solvent was removed under reduced pressure and
the residue was purified by flash column chromatography on silica gel.
Scheme 4. Deuterium-labeling experiment to verify the addition/elimi-
nation by hydrophosphination.
Received: October 9, 2014
Published online: November 27, 2014
because the 1- and 2-octyldiphenylphosphine were not found
to be intermediates in the isomerization reaction.
Keywords: alkenes · cobalt · double-bond migration ·
isomerization · stereoselectivity
.
We then treated 3,3-dideutero-1-hexadecene (14) with the
catalyst mixture in the presence of non-deuterated alkene 3i
(Scheme 5) to prove that the isomerization does not proceed
through an intramolecular 1,3-hydrogen shift reaction.
The 2-alkenes obtained from this scrambling experiment
were separated and analyzed individually. The isomerization
[1] a) K. Singh, S. J. Straig, J. D. Weaver, J. Am. Chem. Soc. 2014, 136,
5275 – 5278; b) Y. Ito, Y. Uozu, T. Dote, M. Ueda, T. Matsuura, J.
Am. Chem. Soc. 1988, 110, 189 – 198.
[2] a) N. F. Jain, P. F. Cirillo, J. V. Schaus, J. S. Panek, Tetrahedron
Lett. 1995, 36, 8723 – 8726; b) J. Q. Yu, M. J. Gaunt, J. B. Spencer,
J. Org. Chem. 2002, 67, 4627 – 4629; c) U. H. Andrews, J. E.
Baldwin, M. W. Grayston, J. Org. Chem. 1982, 47, 287 – 292;
d) M. T. Reetz, M. Plachky, Synthesis 1976, 199 – 200.
[3] a) R. H. Crabtree, The Organometallic Chemistry of the Transi-
tion Metals, 3rd ed., Wiley, New York, 2001, pp. 226 – 228;
b) G. W. Parshall, Homogeneous Catalysis, Wiley, New York,
1980, pp. 31 – 35.
[4] a) C. Chen, T. R. Dugan, W. W. Brennessel, D. J. Weix, P. L.
Holland, J. Am. Chem. Soc. 2014, 136, 945 – 955; b) G. Hilt,
ChemCatChem 2014, 6, 2484 – 2485; c) G. Erdogan, D. B. Grot-
jahn, Org. Lett. 2014, 16, 2818 – 2821; d) C. R. Larsen, G. Erdogan,
D. B. Grotjahn, J. Am. Chem. Soc. 2014, 136, 1226 – 1229;
e) S. M. M. Knapp, S. E. Shaner, D. Kim, D. Y. Shopov, J. A.
Tendler, D. M. Pudalov, A. R. Chianese, Organometallics 2014,
33, 473 – 484; f) R. Jennerjahn, R. Jackstell, I. Piras, R. Franke, H.
Jiao, M. Bauer, M. Beller, ChemSusChem 2012, 5, 734 – 739; g) L.-
G. Zhuo, Z.-K. Yao, Z.-X. Yu, Org. Lett. 2013, 15, 4634 – 4637;
h) M. Mayer, A. Welther, A. J. von Wangelin, ChemCatChem
2011, 3, 1567 – 1571; i) T. Kobayashi, H. Yorimitsu, K. Oshima,
Chem. Asian J. 2009, 4, 1078 – 1083; j) H. J. Lim, C. R. Smith, T. V.
RajanBabu, J. Org. Chem. 2009, 74, 4565 – 4572; k) D. B. Grot-
jahn, C. R. Larsen, J. L. Gustafson, R. Nair, A. Sharma, J. Am.
Chem. Soc. 2007, 129, 9592 – 9593.
[5] a) A. Schmidt, G. Hilt, Chem. Asian J. 2014, 9, 2407 – 2410;
b) Y. N. Timsina, S. Biswas, T. V. RajanBabu, J. Am. Chem. Soc.
2014, 136, 6215 – 6218; c) F. Pꢀnner, A. Schmidt, G. Hilt, Angew.
Chem. Int. Ed. 2012, 51, 1270 – 1273; Angew. Chem. 2012, 124,
1296 – 1299; d) M. Sodeoka, H. Yamada, M. Shibasaki, J. Am.
Chem. Soc. 1990, 112, 4906 – 4911; e) M. Sodeoka, S. Satoh, M.
Shibasaki, J. Am. Chem. Soc. 1988, 110, 4823 – 4824.
Scheme 5. Deuterium-labeling experiment for the verification of inter-
molecular hydrogen transfer.
products derived from 14 (> 95% D) exhibited a decreased
deuterium incorporation, while the products derived from 3i
(natural abundance) showed an increased deuterium incor-
poration. The scrambling of the deuterium/hydrogen atoms
among the starting materials was proven by the identification
of about 25% of product 15 in the mixture and about 20% of
16 (see the Supporting Information). From these results we
conclude that either the deuterium atom from 14 or the
hydrogen atom from 3i was transferred to the cobalt atom, an
alkene exchange occurred, and that the isomerization is then
catalyzed by the intermediate cobalt complexes 6, 7, and 8.
In conclusion, we were able to show that a cobalt-
catalyzed isomerization of terminal alkenes to (Z)-2-alkenes
can be realized with Ph2PH as the decisive additive to
promote the isomerization. The desired Z isomers were
generated predominantly, a number of functional groups
were tolerated, and additional methyl groups in homoallylic
positions were not only tolerated but led in one case to even
better results. We are convinced that the new mechanism and
the mild reaction conditions will be a boost for further
developments in this field.
[6] dppp = 1,3-bis(diphenylphosphino)propane,
bdpp = 2,4-bis(di-
phenylphosphino)pentane, dpppBuEt = 2-butyl-2-ethyl-1,3-bis
(diphenylphosphino)propane.
[7] L. Fiebig, J. Kuttner, G. Hilt, M. Schwarzer, G. Frenking, H.-G.
Schmalz, M. Schꢁfer, J. Org. Chem. 2013, 78, 10485 – 10493.
[8] Leading references: a) L. Rosenberg, Coord. Chem. Rev. 2012,
256, 606 – 626; b) J.-P. Bezombes, P. B. Hitchcock, M. F. Lappert,
J. E. Nycz, Dalton Trans. 2004, 499 – 501; c) W. Malisch, K. Grꢀn,
O. Fey, C. Abd El Baky, J. Organomet. Chem. 2000, 595, 285 – 291.
[9] Leading reference: W. Malisch, B. Klꢀpfel, D. Schumacher, M.
Nieger, J. Organomet. Chem. 2002, 661, 95 – 110.
Experimental Section
Representative procedure: Anhydrous zinc iodide (10–20 mol%),
zinc powder (10–20 mol%), and cobalt complex (5–10 mol%) were
804
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 801 –804