The nickel-catalyzed carbonylative cycloaddition of allyl halides and acetylenes: An efficient tool for cyclopentane annelation

Article abstract of DOI:10.1002/chem.200903273

From a practical synthetic point of view, the nickel-mediated carbonylative cycloaddition of alkynes and allyl halides is a straightforward method for obtaining the cyclopentane skeleton in high yields and with controlled stereochemistry, especially when considering the efficiency of the intermolecular version of the reaction. The efforts to make the previously stoichiometric process catalytic in nickel, after experimental mechanistic observations, are reported herein. The unexpected intervention of iron as a reductant and the isolation of a final dimeric species that exhibits interesting tautomeric behavior are also presented. An extension of the reaction to new substrates has led to the conclusion that, although the steric and electronic effects of the alkyne substituents are generally irrelevant in relation to the adducts and their yields, those of the allylic counterpart may have a significant influence on the outcome of the reaction. However, the presence of the amine moiety in the alkyne completely inhibited the reaction. The feasibility of a multicentered reaction was verified with a triacetylene in which up to 12 bonds were created at once and in good yield.

Full text of DOI:10.1002/chem.200903273

FULL PAPER
DOI: 10.1002/chem.200903273
The Nickel-Catalyzed Carbonylative Cycloaddition of Allyl Halides and
Acetylenes: An Efficient Tool for Cyclopentane Annelation
Daniel del Moral, Susagna Ricart,* and Josep M. Moretꢀ*[a]
This paper is dedicated to Professor Francesc Camps and Professor Pedro Molina
Abstract: From a practical synthetic
point of view, the nickel-mediated car-
bonylative cycloaddition of alkynes and
vations, are reported herein. The unex-
pected intervention of iron as a reduc-
tant and the isolation of a final dimeric
species that exhibits interesting tauto-
meric behavior are also presented. An
extension of the reaction to new sub-
strates has led to the conclusion that,
although the steric and electronic ef-
fects of the alkyne substituents are gen-
erally irrelevant in relation to the ad-
ducts and their yields, those of the al-
lylic counterpart may have a significant
influence on the outcome of the reac-
tion. However, the presence of the
amine moiety in the alkyne completely
inhibited the reaction. The feasibility
of a multicentered reaction was veri-
fied with a triacetylene in which up to
12 bonds were created at once and in
good yield.
allyl halides is
a
straightforward
method for obtaining the cyclopentane
skeleton in high yields and with con-
trolled
stereochemistry,
especially
when considering the efficiency of the
intermolecular version of the reaction.
The efforts to make the previously stoi-
chiometric process catalytic in nickel,
after experimental mechanistic obser-
Keywords: carbonylation · cataly-
sis · cycloaddition · iron · nickel
Introduction
ic chemist with an efficient catalytic Pauson–Khand reaction
have been reflected in many recent contributions to the
chemical literature, although the most efficient procedures
involve the use of expensive metals and apply mainly to the
intramolecular version of the reaction.[2]
For two decades, we have studied the carbocyclization re-
actions of allyl halides and acetylenes mediated by nickel, a
reaction formally related to the aforementioned Pauson–
Khand process.[3] This reaction was originally reported by
Chiusoli and Cassar to give allylacrylic acids[4] and later ap-
plied also by Oppolzer et al. to synthesize cyclopente-
nones.[5] More recently, a similar palladium-catalyzed pro-
cess with allyl tosylates was reported by Tsukada et al.[6] We
initially found that, by using stoichiometric amounts of the
hazardous [Ni(CO)4] as a mediator and disubstituted al-
kynes as reagents, cyclopentenones were exclusively obtai-
ned[3a] (Scheme 1).
Multicomponent reactions are recognized as being of the
utmost interest in synthetic organic chemistry as a result of
their atom economy and the time-saving features in the re-
action processes. The vast majority of these reactions re-
quire a transition-metal center to activate the substrates to
allow unequivocally and sequentially the synthesis of the re-
quired linkages to yield the final product.[1] However, much
of their practical interest is lost when the metal mediator is
involved in the final product or spoiled during the reaction,
especially when an expensive metal center or a very sophis-
ticated ligand is involved. The efforts to provide the synthet-
[a] D. del Moral,+ Dr. S. Ricart, Prof. J. M. Moretꢀ
Homogeneous Catalysis Group
Institut de Ciꢁncia de Materials de Barcelona
Esfera, Universitat Autꢂnoma de Barcelona (UAB)
CSIC, 08193 Bellaterra (Spain)
Fax : (+34)935-805-729
E-mail: ricart@icmab.es
[+] D. Del Moral is enrolled in a Ph.D. program at the UAB.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200903273.
Scheme 1. General scheme for the carbonylative cycloaddition reaction
of allyl halides and disubstituted alkynes mediated by [Ni(CO)4].
Chem. Eur. J. 2010, 16, 9193 – 9202
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9193

The reaction, which provided adducts in acceptable-to-
high yields depending on the substrates, proved to be regio-
selective in both components and stereoselective when one
of them carried a stereogenic center. Moreover, and more
importantly when compared with other metal-mediated cy-
cloaddition reactions, the intermolecular version of the reac-
tion was also efficient. Thus, in the next step we tried to re-
place the dangerous mediator by the less hazardous [Ni-
Results and Discussion
When considering the mechanism proposed for the original
reaction, one could anticipate that it should display catalytic
character because the putative side-product released in the
last step, HNiX, would presumably suffer reductive elimina-
tion to the corresponding hydrogen halide to regenerate a
Ni0 species. This is not so, and products arising from solvo-
lytic cleavage of the p-allyl complex have previously been
found by Chiusoli et al.[15] and ourselves[3a] (reductive substi-
tution) with the concomitant formation of inactive NiII spe-
cies. More pernicious than that, however, is the known and
very efficient concurrent allyl halide self-coupling reaction,
which was the reason for the use of hyperstoichiometric
amounts of allyl halides and nickel(0) mediators in our earli-
er methodology (Scheme 2). Both processes yield inactive
nickel dihalide.
ACHTUNGTRENNUNG(cod)2] under a CO atmosphere. The reactions performed
with this mediator proved to be successful and we were able
to report a range of new applications.[3d] However, the labili-
ty and stoichiometric use of this mediator did not augur well
for its practical and large-scale application in organic syn-
thesis. Therefore we directed our efforts towards establish-
ing a process truly catalytic in nickel. Two facts lent support
to this proposal: Oppolzer et al.[5] had proved the substoichi-
ometry of the reaction (four cycles, intramolecularly) and
we had observed that, starting from 1-halo-1,4-dienes, the
carbocyclization proceeded catalytically.[7] This latter obser-
vation told us that the step (or steps) responsible for the
noncatalytic character of the global reaction was the allyl-
metallation of the alkyne because beyond this step the same
intermediates should be involved. Bearing these facts in
mind, we explored new conditions in an effort to make the
process catalytic and to find a cheap and stable source of
mediator that would allow its practical use. This was ach-
ieved by the use of iron in the form of a powder[8] and,
three years later, the reaction was applied to strained ole-
fins[9] to construct, in both cases, the cyclopentane skeleton.
In recent years the importance of environmentally benign
syntheses has been increasingly recognized and parameters
such as the E factor,[10] atom efficiency,[1a] and the “12 Prin-
ciples of Green Chemistry”[11] are often considered as essen-
tial driving forces in the quest for sustainable chemical pro-
cesses. The cycloaddition process reported herein deals with
three of the twelve green principles: The reaction is catalytic
in nickel, hazardous and cumbersome Ni0 species have been
replaced by available NiII salts, and up to four bonds are cre-
ated in one single synthetic operation. Furthermore, in
recent work we have shown that the process can take place
by using compressed fluids, with an important reduction in
the amount of organic solvent used and with an unexpected
chemoselectivity.[12] On the other hand, the reaction product,
the cyclopentane ring, is often present in natural products of
bioactive interest. Thus, aside from steroids and prostaglan-
dins, they are found in iridoids, which display different bio-
logical activities, for example, antibiotics (methylenomicin,
pentalenolactone), antitumorals (coriolin, quadrone),[13] crop
control agents (specionin, strygol), and perfumes (jasmine,
vetivone).[14]
Scheme 2. Solvolytic cleavage and allyl halide self-coupling leading to in-
active nickel dihalide.
To suppress the reductive substitution of the allyl compo-
nent we added a compatible base (sodium acetate) to a new
system ([NiACHTNUGTRENUNG(cod)2], CO, 2-butyn-1-ol, allyl chloride, and
methanol). In this way we attained slightly more than two
cycles (Scheme 3). Although similar results had already
Scheme 3. Carbonylative cycloaddition reaction of 2-butyn-1-ol (3 equiv)
and allyl chloride (3 equiv) using [NiACHTNUGRTENUNG(cod)2] (0.5 equiv) at room tempera-
ture with methanol as solvent in the presence of NaAcO.
We report herein a full account of our research that final-
ly led to the nickel-catalyzed version of the carbonylative
cycloaddition reaction between allyl halides and acetylenes
and insights into the reaction mechanism. A study of the
scope and limitations of this reaction is also reported.
been reported by Oppolzer et al. by adding a chelating
phosphine to avoid catalyst degradation,[5a] our own inter-
molecular system was now operating catalytically thus prov-
ing its feasibility.[16]
9194
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9193 – 9202

Cycloaddition of Allyl Halides and Acetylenes
FULL PAPER
From this result it seemed that allyl self-coupling was the
main side-reaction responsible for catalyst degradation to in-
active NiII salts. This reaction had previously been studied
mechanistically and a NiI free radical seemed to be the key
for the reaction to proceed.[17] In our attempt to establish
conditions that are unfavorable to allyl self-coupling, we
changed the substrate model and solvent (Scheme 4). Unex-
pectedly, we concluded that the same NiI pseudoradical was
responsible for both processes and the result depended on
the relative kinetics of the reaction of a common NiI inter-
mediate towards allyl halide or alkyne.[8]
Although the results in entries 1–4 of Table 1 were very
disappointing, in all cases the major product was, by far, the
expected cycloadduct 3, which confirmed the mediation of
NiI radicals in the whole process. Despite that, the reduced
nickel species was easily reoxidized by the excess allyl
halide and the turnover was stopped. Tetrakis(diethylami-
no)ethylene (TDAE)[19] was the first reductant to confer a
true catalytic role on the nickel (entry 5). Unfortunately, we
could not attain more than four cycles with it.[20] Although
excess tin showed an acceptable efficiency, iron was even
better, attaining up to 16 cycles (entry 7). In all cases, the
CO uptake was about 2 mol of CO per mol of Ni,[21] which
approximately corresponds to the theoretical uptake for the
full conversion of the substrates into cycloadduct 3. Unlike
tin and iron, zinc and magnesium gave black metallic nickel
with no cycloadduct being obtained. Powdered manganese
and CrII salts were also used without success despite the fact
that, in the presence of a catalytic amount of quinoline,
Scheme 4. Model substrates phenylacetylene and allyl bromide used in
the cycloaddition reaction to give the cycloadduct 3a.
manganese efficiently reduces NiII to yield [Ni
easy well-known method.[22]
ACHTUNGRTENN(UNG cod)2] by an
We then directed our efforts towards the in situ genera-
tion of NiI species to promote the catalytic cycloaddition re-
action. We envisaged that an efficient generation of NiI spe-
cies, combined with the absence of any NiII in the medium,
would provide a long catalytic lifetime because the allyl self-
coupling reaction seems to require the presence of the two
valence states.[17c] In the absence of NiII species, the NiIII–p-
allyl complex generated after oxidative addition would coor-
dinate the acetylene and mediate the carbonylative cycload-
dition reaction.[18] The exclusive generation of NiI species
was attempted by the monoelectronic reduction of NiII salts
or by mixing equimolar amounts of Ni0 and NiII complexes
Role of iron in the reaction: Although iron was initially
thought to be only the reductant of the NiII halide, excess
iron was used to secure its effect under the heterogeneous
conditions of the reaction. The initial brick-red color of the
2À
NiI4 anion turned to grey in 30 min with a slight uptake of
CO. Then the substrate (phenylacetylene/allyl bromide, 1:1)
mixture was added dropwise to the rapidly stirred mixture
containing the catalyst. Soon gas absorption started and con-
tinued until the theoretical 2 mol of CO per mol of substrate
were reached. Then the reaction was stopped. By weighing
the remaining iron attached to the stirring bar we noted that
0.5 mol of iron per mol of substrate had disappeared, which
points to its active participation in the whole reaction.
Simple acidic work-up in air and removal of any small
amounts of iodine formed with Na2S2O3 solution afforded
the crude product, which consisted of 85% of the cycload-
duct 3a with the remaining 15% corresponding to two non-
polar adducts identified as 4a (minor) and 5a (1:1 mixture
of stereoisomers) derived from a further insertion of the
alkyne. Similar products have previously been reported in
related processes.[6]
under CO to stabilize the NiI species
ACTHNUTRGNEUNG(Table 1). Although
nickel iodide was not considered the best halide in terms of
preventing allyl self-coupling,[17c] it was chosen because of its
higher solubility.
Table 1. Catalytic efficiency of different reducing systems in the nickel-
mediated cyclocarbonylation of allyl halides and alkynes.
System
Subs./
Ni[a]
Conv.[b]
[%]
Yield[c]
[%]
Turnover[d]
1
2
Na
THF
[Ni(cod)2]/Na2ACHTUNGTRENNUNG[NiI4]/
G
2
25
15
<1
<1
2
32
30
THF
3
4
5
Et3SiH/Na
N
2
2
4
46
48
100
40
45
95
<1
<1
3.8
Et3SnH/Na CHTUNGTRENNUNG
1= TDAE/Na
2
We soon observed that the relative amounts of these side-
products decreased with the spontaneous oxidation of the
starting nickel iodide. We concluded that the iodine formed
in the oxidation of the nickel salt, as a Lewis acid, was bene-
ficial for the selectivity of the reaction and that we could re-
place the nickel iodide by a mixture of stable NiBr2, NaI
(6 mol), and AlBr3, with identical or better results than
those obtained with the nickel iodide (Scheme 5).
tone
6
7
Sn/Na
N
10
20
64
83
53
85
6.4
16
Fe/Na2ACHTUNGTRENNUNG[NiI4]/acetone
[a] Number of intended catalytic cycles. Substrate mixture: Phenylacety-
lene/allyl bromide 1:1. [b] Conversion, given as a percentage of the
crude product relative to the theoretical value, assuming total conversion
into the oxocyclopentenecarboxylic acid. [c] Total yield of 2-oxocyclo-
pent-3-eneacetic acid in the crude, as estimated from the corresponding
1H NMR spectrum. [d] Estimated turnover based on the total product.
Chem. Eur. J. 2010, 16, 9193 – 9202
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9195

S. Ricart, J. M. Moretꢀ et al.
and 5a. It was further found that the addition of small
amounts of aliphatic alcohols, either to the solvent or the
substrate mixture, had the same beneficial effect as water.
However, the effect of the added alcohols was limited com-
pared with the in situ generated water (ranging from 21%
of the tert-butyl ester when using t-butanol to 37% of the
methyl ester when using methanol).
Scheme 5. Carbonylative cycloaddition using NiBr2 (0.25 mmol) as the
catalyst, NaI (1.5 mmol), Fe (5 mmol), and AlBr3 (1 mmol) with phenyla-
cetylene (5 mmol) and allyl bromide (5.8 mmol) as the reagents.
Because all the observations indicated that iron was pres-
ent in the final product as a complexed iron salt, we directed
our efforts towards the separation of the metal salts under
nonoxidative conditions to allow the characterization of the
organic adduct by spectroscopic means. This was achieved
by the addition 6 mol of pyridine, which led to the removal
of all the iron as the corresponding tetracoordinated com-
plex, which could be readily filtered. The residual iron-free
crude oil could now be analyzed spectroscopically. The com-
We then directed our efforts towards characterizing the
primary reaction adduct and its side-products to determine
the role of iron. With this aim, the final mixture was dried
under vacuum. From the distillate of the final reaction mix-
ture we isolated significant amounts of mesityl oxide and an
insignificant amount of hydrogen bromide with a complete
absence of isopropyl alcohol. The IR spectrum of the resi-
1
bination of MS, IR, and H and 13C NMR data evidenced
À
due (Figure 1, spectrum a) showed no metal carbonyl band
the dimeric nature of the product obtained, m/z=410 being
the highest significative peak in the MS. In such a case,
owing to the stereogenic center of the a-keto carbon in the
cycle, duplicity of signals in the corresponding 13C NMR
spectrum was to be expected. However, the spectrum corre-
sponded to a single isomer.
The absence of signals at around 175–180 ppm in the
13C NMR spectrum (Figure 2, spectrum a) and beyond
1700 cmÀ1 in the corresponding IR spectrum (Figure 1, spec-
Figure 1. FTIR spectra of a) the crude residue, b) the residue after treat-
ment with pyridine (compound 6), and c) the crude product after oxida-
tive hydrolysis (compound 3a).
but strong bands at 3439, 1708, 1607, and 1466 cmÀ1, absent
in the IR spectrum of 3a, which indicates the presence of
hydroxy and extra double bond functionalities. The amount
of iron consumed was in agreement with a total abstraction
of the halide as FeBr2, which precludes an iron enolate as
the final reaction product. The product could not be a car-
boxylic acid nor an aldehyde nor a mixture of the two be-
cause there are no bands assignable to carbonyl functions in
the IR spectrum of the crude products other than those of
cyclopentenone.
Figure 2. 13C NMR spectra of a) the residue after treatment with pyridine
(compound 6) and b) compound 3a after oxidative hydrolysis.
trum b) excluded the presence of any a-diketo product[25] or
carboxylic functionality. The strong IR band at 1597 cmÀ1
(Figure 1, spectrum b) and the absence of the exocyclic
methylene and carboxylate signals and the presence of a-
diketo methine as a rudimentary signal in the 13C NMR
spectrum (Figure 2, spectrum a) together with a lack of reso-
All attempts to derivatize the crude product failed.[23] Be-
cause the ketone in the cycle in the final adduct was not
enolized (strong band at 1708 cmÀ1), we concluded that the
hydroxy group present in the final mixture must have arisen
from self-condensation of the acetone and addition of the
resulting water to the final adduct. This was apparently the
beneficial role of the Lewis acid (iodine[24] or AlBr3). Our
suspicion was confirmed when the addition of small amounts
of water to the solvent had the same effect as the Lewis
acid, that is, preventing the formation of side-products 4a
1
lution in the corresponding H NMR spectrum only for the
protons in the side-chain indicate rapid motion in the
system.
Because the crude product was able to undergo an aldolic
reaction with paraldehyde[8] we tentatively propose structure
6 as the primary product.[26] This structure contains the origi-
nally dehalogenated adducts plus one water moiety in agree-
9196
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9193 – 9202

Cycloaddition of Allyl Halides and Acetylenes
FULL PAPER
If water is scarce, a portion of the acyl radical might be
oxidized by the ferric ion to the corresponding cation to
give the Friedel–Crafts-type products 4a and 5a.[29]
It is plausible that some or all of the electron-transfer re-
actions can occur on a polynuclear iron–nickel species with
the metal centers in different oxidation states.[30] Thus, in the
light of our last results, we propose the mechanistic scheme
in Scheme 7 as the most plausible with only the iron atom
ment with the methodological requirement of a small
amount of water for the reaction to proceed. Because the
two enolized forms are degenerate in energy, they should
rapidly tautomerize through the dienol form 6’.[27]
But why do hydroxy groups appear in the IR spectrum of
the iron-containing crude? Our explanation is based on the
unique and fundamental role of iron. Like NiII, iron very
readily forms hydroxy complexes. Its fundamental role in
radical hydroxylation reactions of unsaturated organic sub-
strates, such as the Fenton reaction, is widely recognized
and applied. It is based on the easy single-electron oxidation
of FeII to FeIII. In the present case, despite the Mçssbauer
and EPR spectra of the crude indicating that most of the
iron is in a FeII oxidation state, about 15% of it is present as
FeIII (Figure 3).
Scheme 7. Proposed mechanism for the carbonylative cycloaddition reac-
tion.
adjacent to the active nickel depicted although an undefined
number of them may be integrated into the reacting species.
Unlike the reaction with nonfunctionalized allyl halides,
when using methyl 4-bromocrotonate as reagent, all the the-
oretical iron was consumed
(iron/substrate, 1:1) and the re-
action, although with an inferi-
or yield, could be performed in
THF in the complete absence
of water. In this case, the for-
mation of the halo-enolate 8 in
complex form as the final reac-
Figure 3. EPR spectrum of the reaction crude.
tion product is plausibly as-
sumed.
This was unexpected under the strict reductive conditions
of the reaction. However, iron has been reported, under
noncarbonylative conditions, to be capable of coupling allyl
halides to give nonconjugated dienes.[28] Therefore allyl hal-
ides are efficient oxidants for iron. From a formal point of
view, product 6 can be regarded as the enolized ester of its
own reduced (aldehyde) and oxidized counterpart (the car-
boxylic acid 3a) arising from single-electron oxidation and
reduction of common radical intermediate 7 (Scheme 6).
Study of the reaction scope: Having ascertained the catalytic
nature of the reaction and taken into account the good effi-
ciency of the synthetic process, we then directed our efforts
towards extending its scope to more sophisticated substrates,
occasionally with the presence of different functionalities
and/or heteroatoms, either in the acetylenic or allylic com-
ponent, to encourage the synthetic chemist to apply the re-
action to the synthesis of more elaborate targets. We present
here the extension of the catalytic reaction to acetylenes
with ether, amino, alkene, and alkyne (conjugated and non-
conjugated) functionalities and to substituted linear and
cyclic allyl halides (Scheme 8).
For this study, we chose 10 different acetylenes (both in-
ternal and terminal, including a good variety of functional
groups in the structure) and also seven allyl derivatives
(mono-, di-, and trisusbtituted and cyclic). The results of
Scheme 6. Generation of product 6.
Chem. Eur. J. 2010, 16, 9193 – 9202
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9197

S. Ricart, J. M. Moretꢀ et al.
ducts in high yields (entries 14–16). Particularly remarkable
is the complete and opposite stereoselectivity in the result-
ing ring fusions. All these results parallel those found in the
stoichiometric version of the reaction, for which we have
given a plausible explanation.[3c]
Scheme 8. General scheme for the carbonylative cycloaddition reaction
between acetylenes (5 mmol) and allyl halides (5.8 mmol) catalyzed by
NiBr2 (0.25 mmol).
Because, in a few very slow reactions, we isolated the car-
bonylative cycloadduct from the reaction of the allyl bro-
mide with itself, we performed the reaction in the absence
of acetylene and a good yield of the cycloadduct 9 was ob-
tained (Table 2, entry 17). Although activated olefins like
norbornene or norbonadiene are also suitable substrates for
the reaction,[9] this is the first case of the reaction of a plain
olefin. Presumably, the 1,5-hexadiene is generated independ-
ently from the allyl bromide with an excess of iron[32] and
subsequently treated with the nickel–allyl active species to
give the final compound 9 with high stereoselectivity (only
one of the two possible stereoisomers is isolated after the
final work-up). Unfortunately, similar reactions did not take
place under the standard conditions with more hindered
allyl halides and when we performed the cross-reactions
with different allyl derivatives only undefined mixtures were
obtained. This result again indicates the importance of steric
hindrance in the allylic moiety.
Again, to account for this result we feel obliged to pro-
pose the polynuclear model. Because, most probably, the
formation of a coordinated 1,5-hexadiene requires two
metal centers[32] in close proximity, and because iron does
not perform any cyclocarbonylation under these conditions,
the collaborative presence of three metal centers turns out
to be very plausible. Such a mediator would impose severe
steric restrictions on the substrates and products and this
would be reflected in the strong stereoselectivity of the reac-
tion. Almost with no exception, only one single diastereoiso-
mer was obtained in each case.
these reactions, performed under standard conditions, are
shown in Table 2.
The reactions of alkynes 1a and 1b have already been re-
ported to give good yields in the presence of an added
Lewis acid.[8] The substitution of the Lewis acid by a small
amount of water improved the yields (entries 1 and 2). As a
general trend, we observe that the reactions under these
new conditions take place in good yields with most of the
acetylenes.
The presence of an alcohol or ether linkage in the alkyne
also provided the expected cycloadduct in good yields (en-
tries 3–5). It is surprising that the use of alkyne 1c led to the
exclusive formation of ester 3c in view of the low yields of
ester obtained when noncoordinating alcohols replace the
added water, as seen above. The ability of triple bonds to
coordinate the active species may account for this result and
others reported below.[31] Although oxygen functionalities
are completely compatible with this reaction, amines are not
(entries 9 and 10). The complete inhibition of the reaction
brought about by amines has already been observed in the
stoichiometric version of the reaction and also when amines
were used as proton scavengers to turn the reaction into a
catalytic process. The acetylenic amine was recovered un-
changed in these cases. Competition with the alkyne for an
active coordination site may explain this result.
Unlike the results obtained in the stoichiometric reaction
with conjugated allylic systems,[3b] the enyne 1 f reacted with
allyl bromide to give the cycloadduct 3 f in good yield
(entry 6). The extraordinary affinity of triple bonds for the
activated allyl ligand is evidenced in entries 7 and 8 in which
diacetylenes, either conjugated or not, gave excellent yields
of dicycloadducts (using 2 equiv of the allyl derivative).
These new results, together with those previously reported,
led us to the conclusion that only very severe steric hin-
drance on the alkyne substituents can preclude the reaction
taking place in a satisfactory way. In contrast, the reaction is
very sensitive to the steric restrictions imposed by the sub-
stituents on the allyl system. Although monosubstitution of
the allyl functionality still provides a good yield of the cyclo-
adduct (entries 11 and 12), gem disubstitution slows down
the reaction considerably thereby allowing a further inser-
tion of the alkyne into the unfinished cycloadduct “tail” in a
pseudo-Friedel–Crafts coupling reaction (entry 13). Al-
though 1,3-disubstituted 3-bromocyclohexene displays con-
formational steric constraints leading to results similar to
those obtained with gem-disubstituted homologues, the
more flexible seven- and eight-membered cyclic allyl sys-
tems afford the corresponding oxocyclopenteneacetic ad-
Finally, as an example of the feasibility of the present re-
action, we report the reaction of 1,3,5-trialkynylbenzene 1k
in which 12 bonds were formed at once in good yield
(Scheme 9), again with a single isomer being observed.
However, in this case the tri-acid obtained after the usual
work-up was insoluble in the solvents used in the oxidative
treatment. After esterification of the crude product in acidic
media, the corresponding final product 10 was isolated in a
yield of 51%.
Scheme 9. Triple carbonylative cycloaddition between 1,3,5-trialkynylben-
zene (5 mmol) and allyl bromide (17.4 mmol) using NiBr2 (0.25 mmol).
9198
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9193 – 9202

Cycloaddition of Allyl Halides and Acetylenes
FULL PAPER
Table 2. Products and yields of the carbonylative cycloaddition reactions between different functionalized ace-
tylenes and allyl halides under standard conditions.[a]
Product 10 displays high re-
action versatility: Conjugate ad-
ditions to activated double
Acetylene
Allyl
Product (Yield [%])
bonds,
allylic
substitutions,
ketone condensations, reduc-
tion of the carbonyl and car-
boxy moieties, and, by using a
monosubstituted allyl halide, it
should be possible to be used in
dendrimer chemistry as a multi-
functional “core”.[33]
1
2
1a
2a
3a (95%)
1b
2a
3b (80%)
3
4
Conclusion
1c
1d
2a
2a
3c (80%)[b]
3d (54%)[c]
Efforts to make the previous
stoichiometric metal-mediated
carbonylative cycloaddition of
allyl halides and alkynes cata-
lytic in nickel, based on experi-
mental mechanistic observa-
tions, have been reported
herein. Although the active cat-
alytic species was an in situ
generated NiI complex, the un-
expected involvement of iron
and water generated from the
solvent (acetone) in the final
release of the organic cycload-
duct has been rationalized.
Thus, iron acts in all cases as a
dehalogenating agent whereas
water is required only with non-
functionalized allyl halides to
give an adduct that is stable
under the experimental condi-
tions.
5
1e
1 f
2a
2a
3e (81%)
3 f (87%)
6
7
8
1g
2a
3g (79%)
1h
1i
2a
2a
3h (55%)
9
–
10
1j
2a
–
The catalytic version of the
carbonylative cycloaddition of
alkynes and allyl halides using a
controlled amount of water in-
stead of a Lewis acid shows a
wide tolerance towards func-
tional and steric substitution on
the alkyne. Steric factors, how-
ever, are much more important
in the allylic component,
thwarting the termination step
and leading to adducts with
alkyne polyinsertion. The ea-
gerness of the allyl–nickel inter-
mediate for triple bonds is so
great that multicentered reac-
tions with high yields are possi-
ble with polyacetylenes.
11
1a
1a
2b
2c
3i (76%)[d]
3j (84%)
12
13
1a
1a
2d
2e
5b (21%)
5c (40%)
14
Chem. Eur. J. 2010, 16, 9193 – 9202
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9199

S. Ricart, J. M. Moretꢀ et al.
ganic phase with water until neutrali-
zation, the organic layer was treated
with a solution of Na2S2O3 (to remove
any I2 produced by oxidation), washed
again with water, and dried on
MgSO4. After removal of the solvent
and flash chromatography (CH2Cl2/
CH3OH) the products were obtained.
Table 2. (Continued)
Acetylene
Allyl
Product (Yield [%])
15
1a
2 f
3k (61%)
Compound characterization
3a: 1H NMR (300 MHz, CDCl3): d=
16
17
2.47 (ddd, J
CH2), 2.58 (dd, J
CH2), 2.85–2.95 (m, 1H; CH), 2.97
(dd, J(H,H)=17, 3 Hz, 1H; CH2), 3.02
(ddd, J(H,H)=19, 6, 3 Hz, 1H; CH2),
7.30–7.45 (m, 3H; Ph), 7.65–7.75 (m,
2H; Ph), 7.80 ppm (dd, (H,H)=3,
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
1a
2g
2a
3 l (73%)
9 (81%)[d]
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
–
JACHTUNGTRENNUNG
3 Hz, 1H; CH); 13C NMR (75 MHz,
CDCl3): d=33.2 (t), 34.9 (t), 42.4 (d),
127.0 (d), 128.5 (d), 128.6 (d), 131.4
(s), 142.2 (s), 157.2 (d), 177.5 (s),
[a] Standard conditions: 0.25 mmol NiBr2, 1.5 mmol NaI, 5 mmol Fe, 5 mmol acetylene, 5 mmol allyl halide,
and 5 mmol H2O in acetone as solvent at room temperature under a CO atmosphere (1.0 bar pressure).
[b] Two-fold alkyne was used. [c] Estimated from the reaction crude. [d] Only one stereoisomer was detected.
207.2ppm (s); IR (CH2Cl2): nn˜ =3600–
À1
À
3000 (O H), 1737 (C=O), 1703 cm
Experimental Section
(C=O); HRMS (EI): m/z: calcd for [M]+: 216.0783; found: 216.0785.
3b: 1H NMR (300 MHz, CDCl3): d=0.98 (t, J
ACHTNUTRGNE(NUG H,H)=8 Hz, 3H; CH3),
1H NMR and 13C NMR spectra were recorded in CDCl3 on a Bruker
ARX 300 (300 and 75.5 MHz) spectrometer, on a Bruker AC 250, and on
a Bruker AC 500. Chemical shifts are given in ppm relative to TMS (1H
d=0.0 ppm) or CDCl3 (13C d=77.0 ppm). IR spectra were recorded on a
Perkin Elmer Spectrum One FTIR spectrometer, using CH2Cl2 as the sol-
vent. Mass spectra were obtained on an AutoSpec-Q mass spectrometer.
Elemental analyses were performed using a Carlo Erba 1106 apparatus.
Standard Schlenck techniques were used. Acetone was distilled under Ar
just before use. THF was freshly distilled from sodium benzophenone
under argon. Iron powder (10mm diameter (reduced by hydrogen)), was
provided by Merck-Darmstadt (product no.: 1.03819.0100). Unless other-
wise stated, all common reagents and solvents were used as obtained
from commercial suppliers without further purification.
1.14 (t, JACHTNUTRGEN(UNG H,H)=8 Hz, 3H; CH3), 2.10–2.55 (m, 6H; CH2), 2.70–2.80 (m,
1H; CH), 2.82–2.92 ppm (m, 2H; CH2); 13C NMR (75 MHz, CDCl3): d=
11.9 (q), 13.2 (q), 16.2 (t), 23.9 (t), 35.3 (t), 35.7 (t), 42.9 (d), 140.1 (s),
À
174.3 (s), 176.9 (s), 209.9 ppm (s); IR (CH2Cl2): nn˜ =3600–3000 (O H),
1727 (C=O), 1705 cmÀ1 (C=O); MS (EI): m/z (%): 195.9 (11) [M+1]+,
194.9 (100) [M]+, 151.0 (4).
3c: 1H NMR (300 MHz, CDCl3): d=1.30–2.00 (m, 10H; CH2), 2.19 (brs,
2H; CH2), 2.20–2.50 (m, 3H; CH, CH2), 2.60–2.70 (m, 1H; CH), 2.74
(dd, JACTHUNTRGENN(UG H,H)=16, 3 Hz, 1H; CH2), 2.88 (ddd, JHCATUNGTRENN(UGN H,H)=22, 6, 3 Hz, 1H;
CH2), 3.64 (brs, 2H; CH2), 4.06 (brs, 2H; CH2), 7.28 ppm (m, 1H; CH);
13C NMR (75 MHz, CDCl3): d=23.8 (t), 23.9 (t), 24.1 (t), 28.3 (t), 29.6
(t), 32.2 (t), 33.6 (t), 35.2 (t), 41.7 (d), 62.5 (t), 64.4 (t), 68.1 (s), 84.0 (d),
145.3 (s), 156.0 (d), 172.0 (s), 209.5 ppm (s); MS (EI): m/z (%): 293.2
(18) [M+1]+, 292.2 (100) [M]+.
3d: From the reaction crude: 1H NMR (300 MHz, CDCl3): d=1.36 (t, J-
AHCTUNGTRENG(NNU H,H)=9 Hz, 3H; CH3), 2.20–2.60 (m, 2H; CH2), 2.60–3.00 (m, 3H; CH,
General procedure: NiBr2 (55 mg, 0.25 mmol), NaI (225 mg, 1.5 mmol),
and iron powder (300 mg) were introduced into a 50 mL three-necked
flask, provided with an efficient magnetic stirring bar. The three necks
were connected to a gas burette with vaseline oil, a source of CO gas,
and a rubber septum (through a stopcock). The internal atmosphere was
also filled with CO, the internal pressure set equal to atmospheric pres-
sure. The chosen solvent (usually acetone, 2.0 mL) was introduced into
the reaction flask by means of a syringe and the stirrer was set to its max-
imum speed in such a way that the iron powder keeps turning around, re-
leased from the magnetic bar by centrifugal forces. The initial deep-red
solution turned pale red or grey.
CH2), 3.91 (q, JACTHNUGRTNENUG(H,H)=9 Hz, 2H; CH2), 6.32 ppm (dd, JACHTUGNTREN(NUNG H,H)=3, 3 Hz,
1H; CH); 13C NMR (75 MHz, CDCl3): d=14.2 (q), 34.7 (t), 36.4 (t), 41.5
(d), 65.4 (t), 126.3 (d), 155.6 (s), 174.3 (s), 202.8 ppm (s).
3e: 1H NMR (300 MHz, CDCl3): d=2.43 (dd, J
CH2), 2.47 (dd, J(H,H)=19, 10 Hz, 1H; CH2), 2.74 (dddd, J
5, 3 Hz, 1H; CH), 2.88 (dd, J(H,H)=19, 5 Hz, 1H; CH2), 3.01 (dd, J-
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNG(UN H,H)=22, 7 Hz, 1H; CH2), 3.31 (s, 3H; CH3), 3.39 (s, 3H; CH3), 4.11 (s,
2H; CH2), 4.43 ppm (s, 2H; CH2); 13C NMR (75 MHz, CDCl3): d=34.2
After 30 min the CO uptake had reached up to 10–12 mL and a solution
of the corresponding allyl halide (5.8 mmol) and the corresponding acety-
lene (5.0 mmol) was added dropwise by means of a syringe pump and a
polypropylene cannula through the septum and the stopcock hole over
3 h (if water or an alcohol was used as an additive, the required amount
of the additive was added to the reagents together with acetone (1.0 mL)
to homogenize the solution). After the addition of a few drops to the
flask, the suspension turned turbid red and some gas evolution was ob-
served. Soon after, the uptake of CO restarted. After 1 h, the reaction
mixture had become a deep red-brown solution and the uptake increased
dramatically, up to nearly 1 mLminÀ1. At around 200 mL of CO, the rate
decreased until a standstill at 220 mL when all the substrate had fully re-
acted.
(t), 34.4 (t), 40.8 (d), 58.2 (q), 58.7 (q), 63.0 (t), 69.7 (t), 135.4 (s), 173.3
À
(s), 175.8 (s), 208.1 ppm (s); IR (CH2Cl2): nn˜ =3600–3000 (O H), 1730
(C=O), 1705 cmÀ1 (C=O); HRMS (EI): m/z: calcd for [M]+: 228.0993;
found: 228.0988.
3 f: 1H NMR (300 MHz, CDCl3): d=1.50–1.80 (m, 4H; CH2), 2.10–2.15
(m, 4H; CH2), 2.35 (ddd, J
(H,H)=20, 9 Hz, 1H; CH2), 2.75–2.90 (m, 1H; CH), 2.93 (dd, J
20, 5 Hz, 1H; CH2), 2.99 (ddd, J(H,H)=22, 6, 4 Hz, 1H; CH2), 6.92 (t, J-
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGERTG(NNUN H,H)=4 Hz, 1H; CH), 7.35 ppm (dd, JACHNUTGTREG(NNUN H,H)=4, 4 Hz, 1H; CH);
13C NMR (75 MHz, CDCl3): d=21.7 (t), 22.3 (t), 25.3 (t), 26.2 (t), 32.5
(t), 34.9 (t), 42.6 (d), 128.0 (s), 128.8 (d), 141.9 (s), 153.3 (d), 176.8 (s),
À
208.1 ppm (s); IR (CH2Cl2): nn˜ =3600–3000 (O H), 1730 (C=O),
1703 cmÀ1 (C=O); HRMS (EI): m/z: calcd for [M]+: 220.1095; found:
After 4–5 h the reaction was over (no more CO uptake, approx. 220 mL
of CO). The solvent was removed and the contents of the flask were
transferred to a separation funnel, washing the flask (with the remaining
iron) with dichloromethane. The reaction mixture darkened deeply on
contact with air. The solution was treated with portions of 5n HCl solu-
tion until no further discoloration was observed. After washing the or-
220.1089.
3g: 1H NMR (300 MHz, CDCl3): d=1.30–1.60 (m, 4H; CH2), 2.00–2.30
(m, 4H; CH2), 2.35 (ddd, J
(H,H)=16, 8 Hz, 2H; CH2), 2.60–2.70 (m, 2H; CH), 2.81 (dd, J
16, 3 Hz, 2H; CH2), 2.89 (ddd, J(H,H)=19, 6, 3 Hz, 2H; CH2), 7.29 ppm
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
9200
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9193 – 9202

Cycloaddition of Allyl Halides and Acetylenes
FULL PAPER
(dd, J
G
d=33.7 (t), 34.5 (t), 42.8 (d), 78.8 (d), 127.0 (d), 128.5 (d), 128.6 (d),
128.7 (d), 129.5 (d), 131.4 (s), 132.2 (s), 142.5 (s), 147.5 (d), 157.7 (d),
171.4 (s), 208.0 ppm (s); IR (CH2Cl2): nn˜ =3059 (C=C), 1755 (C=O), 1699
(C=O).
27.2 (t), 33.5 (t), 34.9 (t), 41.4 (d), 145.1 (s), 156.8 (d), 176.7 (s),
210.0 ppm (s); IR (CH2Cl2): nn˜ =3600–3000 (O H), 1731 (C=O),
1707 cmÀ1 (C=O); HRMS (EI): m/z: calcd for [M]+: 334.1410; found:
5b (mixture of isomers): 1H NMR (300 MHz, CDCl3): d=0.97 (s, 3H;
CH3), 1.04 (s, 3H; CH3), 1.10 (s, 3H; CH3), 1.18 (s, 3H; CH3), 2.50–3.10
(m, 3H; CH2), 5.52 (s, 1H; CH), 6.03 (s, 1H; CH), 7.00–8.10 ppm (m,
12H; Ph, CH); 13C NMR (75 MHz, CDCl3): d=19.5 (q), 20.2 (q), 20.3
(q), 20.4 (q), 29.8 (t), 40.4 (s), 40.6 (s), 51.2 (d), 52.8 (d), 84.5 (d), 85.6
(d), 127.0 (d), 128.6 (d), 129.3 (d), 131.1 (s), 133.0 (s), 143.7 (s), 146.2 (d),
146.3 (d), 157.1 (d), 157.7 (d), 171.7 (s) 171.8 (s), 207.3 (s), 207.5 ppm (s);
IR (CH2Cl2): nn˜ =1756 (C=O), 1694 cmÀ1 (C=O); HRMS (EI): m/z:
calcd for [M]+: 358.1569; found: 358.1565.
334.1403.
1
3h: H NMR (300 MHz, CDCl3): d=2.49 (ddd, J
G
CH2), 2.57 (dd, J
2.97 (dd, J
ACHTUNGTRENNUNG
E
ACHTUNGTRENNUNG
2H; CH2), 7.74 (s, 4H; Ph), 7.80 ppm (dd, JACTHNUTRGNEUNG(H,H)=3, 3 Hz, 2H; CH);
13C NMR (75 MHz, CDCl3): d=33.3 (t), 34.9 (t), 42.6 (d), 127.0 (d), 132.2
(s), 142.0 (t), 157.3 (d), 176.3 (s), 207.3 ppm (s); IR (CH2Cl2): nn˜ =3600–
À1
À
3000 (O H), 1735 (C=O), 1705 cm (C=O); HRMS (EI): m/z: calcd for
[M]+: 354.1098; found: 354.1107.
3i: 1H NMR (300 MHz, CDCl3): d=0.87 (t, J
1.15–1.45 (m, 10H; CH2), 1.65–1.75 (m, 2H; CH2), 2.66 (ddd, J
19, 2, 2 Hz, 1H; CH2), 2.87 (ddd, J(H,H)=19, 6, 2 Hz, 1H; CH2), 2.95–
3.05 (m, 2H; CH), 7.30–7.45 (m, 3H; Ph), 7.65–7.75 (m, 2H; Ph),
7.84 ppm (dd, J
(H,H)=2, 2 Hz, 1H; CH); 13C NMR (75 MHz, CDCl3):
G
5c (mixture of isomers): Major isomer: 1H NMR (300 MHz, CDCl3): d=
R
1.25–2.40 (m, 7H; CH, CH2), 2.95 (t, J
ACHTUNGTNER(NUNG H,H)=6 Hz, 1H; CH), 3.09 (brs,
G
1H; CH), 5.75 (d, J(H,H)=9 Hz, 1H; CH2), 7.30–7.90 (m, 10H; Ph),
ACHTUNGTRENNUNG
7.97 ppm (s, 1H; CH); 13C NMR (75 MHz, CDCl3): d=20.2 (t), 25.5 (t),
28.3 (t), 38.9 (d), 43.0 (d), 48.7 (d), 83.0 (d), 126.9 (d), 127.0 (d), 128.5
(d), 128.7 (d), 131.3 (s), 131.5 (s), 141.7 (s), 148.0 (d), 161.9 (d), 171.7 (s),
207.8 ppm (s); IR (CH2Cl2): nn˜ =1755 (C=O), 1698 cmÀ1 (C=O); HRMS
(EI): m/z: calcd for [M]+: 370.1569; found: 370.1595. Minor isomer:
1H NMR (300 MHz, CDCl3): d=1.25–2.10 (m, 7H; CH, CH2), 3.17 (brs,
ACHTUNGTRENNUNG
d=13.8 (q), 22.4 (t), 27.5 (t), 27.6 (t), 28.8 (t), 29.2 (t), 29.6 (t), 31.5 (t),
45.1 (d), 47.8 (d), 126.7 (d), 128.2 (d), 128.3 (d), 131.1 (s), 142.8 (t), 157.6
À
(d), 180.8 (s), 206.6 ppm (s); IR (CH2Cl2): nn˜ =3600–3000 (O H), 1724
(C=O), 1707 cmÀ1 (C=O); HRMS (EI): m/z: calcd for [M]+: 314.1875;
found: 314.1874.
2H; CH), 6.04 (d, JACTHUNGTRENNU(G H,H)=8 Hz, 1H; CH2), 7.30–7.95 (m, 10H; Ph),
7.72 ppm (s, 1H; CH); 13C NMR (75 MHz, CDCl3): d=18.6 (t), 22.5 (t),
27.0 (t), 38.8 (d), 40.9 (d), 47.5 (d), 81.3 (d), 126.9 (d), 127.0 (d), 128.5
(d), 128.7 (d), 131.3 (s), 131.5 (s), 142.4 (s), 148.1 (d), 160.1 (d), 171.6 (s),
207.9 ppm (s); IR (CH2Cl2): nn˜ =1755 (C=O), 1698 (C=O); HRMS (EI):
m/z: calcd for [M]+: 370.1569; found: 370.1595.
3j: 1H NMR (300 MHz, CDCl3): d=2.48 (ddd, J
CH2), 2.56 (dd, J(H,H)=18, 10 Hz, 1H; CH2), 2.85–3.00 (m, 1H; CH),
2.96 (dd, J(H,H)=18, 4 Hz, 1H; CH2), 3.05 (ddd, J(H,H)=19, 7, 3 Hz,
1H; CH2), 3.72 (s, 3H; CH3), 7.30–7.45 (m, 3H; Ph), 7.65–7.75 (m, 2H;
Ph), 7.82 ppm (dd,
(H,H)=3, 3 Hz, 1H; CH); 13C NMR (75 MHz,
ACHTUNGTREN(NUNG H,H)=19, 3, 3 Hz, 1H;
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
JACHTUNGTRENNUNG
CDCl3): d=33.1 (t), 34.8 (t), 42.6 (d), 51.7 (q), 126.8 (d), 128.0 (d), 128.2
(d), 131.2 (s), 142.2 (s), 157.1 (d), 172.3 (s), 207.2 ppm (s); IR (CH2Cl2):
nn˜ =1726 (C=O), 1704 cmÀ1 (C=O).
6: 1H NMR (300 MHz, CDCl3): d=2.53 (ddd, J
CH2), 2.55–2.75 (m, 1H; CH2), 2.90–3.00 (m, 1H; CH), 2.95–3.05 (m,
1H; CH2), 3.06 (ddd, J(H,H)=19, 6, 3 Hz, 1H; CH2), 7.30–7.45 (m, 3H;
ACHTUNGTREN(NUNG H,H)=19, 3, 3 Hz, 1H;
AHCTUNGTRENNUNG
3k: 1H NMR (300 MHz, CDCl3): d=1.1–2.3 (m, 8H; CH2), 2.40 (dd, J-
ACHTUNGTRENNUNG
Ph), 7.65–7.75 (m, 2H; Ph), 7.82 ppm (dd, JACTHNUTRGNEUNG(H,H)=3, 3 Hz, 1H; CH);
13C NMR (75 MHz, CDCl3): d=33.5 (t), 43.6 (d), 127.0 (d), 128.4 (d),
128.5 (d), 131.6 (s), 142.5 (s), 157.5 (d), 207.9 ppm (s); IR (CH2Cl2): nn˜ =
À1
À
3600–3000 (O H), 3056 (C=C), 1733 (C=O), 1699 cm (C=O); MS (EI):
ACHTUNGTRENNUNG
m/z (%): 410.08 (100) [M]+.
(t), 31.6 (t), 33.3 (t), 41.4 (d), 42.1 (d), 55.6 (d), 127.1 (d), 128.4 (d), 128.5
(d), 132.1 (s), 141.7 (s), 159.9 (d), 180.1 (s), 207.3 ppm (s); IR (CH2Cl2):
9: 1H NMR (300 MHz, CDCl3): d=1.30–1.60 (m, 2H; CH2), 1.80–1.90
(m, 1H; CH2), 1.90–2.00 (m, 1H; CH2), 2.00–2.35 (m, 5H; CH, CH2),
À1
À
nn˜ =3600–3000 (O H), 1739 (C=O), 1704 cm (C=O); HRMS (EI): m/z:
calcd for [M]+: 270.1251; found: 270.1255.
2.35–2.50 (m, 1H; CH), 2.49 (dd, JACHNUTGTRENNUNG
3l: 1H NMR (300 MHz, CDCl3): d=0.9–2.3 (m, 10H; CH2), 2.42 (dd, J-
ACHTUNGTRENNUNG
J
J
U
ACHTUNGTRENNUNG
34.0 (t), 45.6 (d), 48.1 (d), 115.2 (t), 137.9 (d), 178.1 (s), 220.0 ppm (s); IR
ACHTUNGTRENNUNG
À1
À
(CH2Cl2): nn˜ =3600–3000 (O H), 1732 (C=O), 1714 cm (C=O); HRMS
(t), 26.3 (t), 26.4 (t), 34.4 (t), 41.6 (d), 43.8 (d), 53.3 (d), 126.9 (d), 128.1
(d), 128.6 (d), 131.9 (s), 140.6 (d), 161.8 (s), 179.5 (s), 207.8 ppm (s); IR
(EI): m/z: calcd for [M]+: 196.1095; found: 196.1098.
À1
1
À
(CH2Cl2): nn˜ =3600–3000 (O H), 1739 (C=O), 1705 cm (C=O); MS
10: H NMR (300 MHz, CDCl3): d=2.45 (ddd, J
U
(EI): m/z (%): 285.1 (18) [M+1]+, 284.1 (100) [M]+.
CH2), 2.56 (dd, J
2.94 (dd, J
ACHTUNGTRENNUNG
1
N
ACHTUNGTRENNUNG
4a: H NMR (300 MHz, CDCl3): d=2.41 (ddd, J
CH2), 2.84 (dd, J
3.12 (ddd, J
N
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
8.04 ppm (s, 3H; Ph); 13C NMR (75 MHz, CDCl3): d=33.0 (t), 34.7 (t),
42.5 (d), 51.6 (q), 125.3 (d), 131.6 (s), 141.6 (s), 157.9 (d), 172.1 (s),
207.0 ppm (s); IR (CH2Cl2): nn˜ =1725 (C=O), 1705 cmÀ1 (C=O); HRMS
(EI): m/z: calcd for [M]+: 534.1881; found: 534.1878.[32]
N
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
JACHTUNGTRENNUNG
(t), 42.4 (d), 125.7 (d), 126.9 (d), 128.2 (d), 128.3 (d), 128.9 (d), 130.5 (d),
131.7 (s), 134.2 (s), 142.2 (s), 143.1 (d), 157.5 (d), 197.9 (s), 208.2 ppm (s).
5a (mixture of isomers): Major isomer: 1H NMR (300 MHz, CDCl3): d=
1.67 (ddd, J
3 Hz, 1H; CH2), 2.57 (ddd, J
1H; CH), 3.08 (ddd, J(H,H)=16, 7, 3 Hz, 1H; CH2), 5.41 (ddd, JCAHTUNGTRENNNUG
A
ACHTUNGTRENUN(NG H,H)=14, 3,
AHCTUNGTRENNUNG
R
7, 3, 2 Hz, 1H; CH), 7.30–8.00 ppm (m, 12H; Ph, CH); 13C NMR
(75 MHz, CDCl3): d=34.3 (t), 35.6 (t), 44.0 (d), 80.0 (d), 127.0 (d), 128.5
(d), 128.7 (d), 129.4 (d), 131.4 (s), 131.6 (s), 142.3 (s), 147.8 (d), 157.8 (d),
171.5 (s), 208.1 ppm (s). IR (CH2Cl2): nn˜ =3059 (C=C), 1755 (C=O), 1699
(C=O); minor isomer: 1H NMR (300 MHz, CDCl3): d=2.10 (ddd, J-
Acknowledgements
The authors acknowledge Dr. Anna Roig for the Mçssbauer spectra and
Mrs. Maria Huerta-Fontela for the MS spectra. Also we acknowledge fi-
nancial support from the Spanish Direcciꢀn General de Investigaciꢀn
(DGI) (EMOCIONA project CTQ2006-06333/BQV) and the Consejo In-
terdepartamental de Investigaciꢀn e Innovaciꢀn Tecnolꢀgica (CIRIT) (D.
del Moral, grant no.: 2007FIC00554).
A
ACHTUNGTREN(NUNG H,H)=17, 6, 3 Hz, 1H;
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 9193 – 9202
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9201

S. Ricart, J. M. Moretꢀ et al.
[21] Even in entry 7 when an excess of iron was added to the reaction
mixture (5 mol of iron per mol of nickel, excluding any [Fe(CO)5]
formation).
[22] a) F. Guerrieri, G. Salerno, J. Organomet. Chem. 1976, 114, 339–
344; b) M. Durandetti, C. Gosmini, J. Pericꢀn, Tetrahedron 2007, 63,
1146–1153.
[23] The attempted esterification of the crude product with methanol at
reflux and a catalytic amount of p-toluenesulfonic acid led again to
3a. All attempts to silylate, methylate, or acylate the hydroxy func-
tionality also failed.
[24] Iodine is known to catalyze the condensation of acetone to diace-
tone alcohol: J. B. Conant, N. Tuttle, Org. Synth. 1921, 1, 53.
[25] a) H. des Abbayes, J. Y. Salaꢇn, Dalton Trans. 2003, 1041–1052;
b) P. Cabon, R. Runmin, J. Y. Salaꢇn, S. Triki, H. des Abbayes, Or-
ganometallics 2005, 24, 1709–1717; c) M. Periasamy, C. Rameshku-
mar, U. Radhakrishnan, J. J. Brunet, J. Org. Chem. 1998, 63, 4930–
4935; d) M. Periasamy, A. Devasagayaraj, U. Radhakrishnan, Orga-
nometallics 1993, 12, 1424–1428.
[1] a) B. M. Trost, Acc. Chem. Res. 2002, 35, 695–705; b) A. Dçmling, I.
Ugi, Angew. Chem. 2000, 112, 3300–3344; Angew. Chem. Int. Ed.
2000, 39, 3168–3210.
[2] a) S. E. Gibson, N. Mainolfi, Angew. Chem. 2005, 117, 3082–3097;
Angew. Chem. Int. Ed. 2005, 44, 3022–3037; b) S. E. Gibson (nꢄe
Thomas), A. Stevanazzi, Angew. Chem. 2003, 115, 1844–1854;
Angew. Chem. Int. Ed. 2003, 42, 1800–1810; c) C. Aubert, L. Fen-
sterbank, V. Gandon, M. Malacria, Top. Organomet. Chem. 2006, 19,
259–294; d) N. Jeong in Modern Rhodium-Catalyzed Reactions
(Ed.: P. A. Evans), Wiley-VCH, Weinheim, 2005, pp. 215–240; e) J.
Pꢄrez-Castells, Top. Organomet. Chem. 2006, 19, 207–257.
[3] a) F. Camps, J. Coll, J. M. Moretꢀ, J. Torras, J. Org. Chem. 1989, 54,
1969–1978; b) F. Camps, J. M. Moretꢀ, Ll. Pagꢁs, Tetrahedron 1992,
48, 3147–3162; c) L. Pagꢁs, A. Llebarꢅa, F. Camps, E. Molins, C.
Miravitlles, J. M. Moretꢀ, J. Am. Chem. Soc. 1992, 114, 10449–
10461; d) G. Garcia-Gꢀmez, J. M. Moretꢀ, Chem. Eur. J. 2001, 7,
1503–1509.
[4] G. P. Chiusoli, L. Cassar, Angew. Chem. 1967, 79, 177–186.
[5] a) W. Oppolzer, T. H. Keller, D. L. Kuo, W. Pachinger, Tetrahedron
Lett. 1990, 31, 1265–1268; b) W. Oppolzer, Angew. Chem. 1989, 101,
39–53; Angew. Chem. Int. Ed. Engl. 1989, 28, 38–52.
[26] Although almost all the spectral data and chemical behavior is indi-
cative of this structure, one of the referees disagrees with us.
[27] A low-temperature 13C NMR experiment showed no significant dif-
ferences in the spectrum.
[6] N. Tsukada, S. Sugawara, K. Nakaoka, Y. Inoue, J. Org. Chem. 2003,
68, 5961–5966.
[7] F. Camps, J. Coll, A. Llebaria, J. M. Moretꢀ, Tetrahedron Lett. 1988,
29, 5811–5814.
[8] M. L. Nadal, J. Bosch, J. M. Vila, G. Klein, S. Ricart, J. M. Moretꢀ,
J. Am. Chem. Soc. 2005, 127, 10476–10477.
[9] D. del Moral, J. M. Moretꢀ, E. Molins, S. Ricart, Tetrahedron Lett.
2008, 49, 6947–6950.
[28] a) D. H. Hall, E. Hurley, Jr., Can. J. Chem. 1969, 47, 1238; b) an ex-
periment performed under carbonylative conditions in the absence
of the nickel salt gave only 1,5-hexadiene; c) see also the Fe/FeCl3
system for the generation of organic radicals: L. Forti, E. Ghelfi, E.
Libertini, U. M. Pagnoni, E. Soragni, Tetrahedron 1997, 53, 17761.
[29] Comprehensive Organic Synthesis, Vol. 2 (Ed.: B. M. Trost, I. Flem-
ing), Pergamon Press, Oxford, 1991, pp. 723–726.
[30] Polynuclear iron–nickel and iron–iron species in different oxidation
states are thought to be the active centers in biologically significant
processes of free-radical nature. Interestingly, the different metal
centers in these characterized complexes are stabilized by carbonyl
ligands (or the corresponding isoelectronic isonitrile ligands) and
bridged by halide, thiolate, enolate, hydroxy, or even water ligands
No other metals seem to associate so efficiently with nickel as iron,
with the association stabilizing the NiI oxidation state. a) Encyclo-
pedia of Inorganic Chemistry, Vol. 5 (Ed.: R. B. King), 1995, Wiley,
New York, pp. 2386–2388 and Encyclopedia of Inorganic Chemistry,
Vol. 5 (Ed.: R. B. King), 1995, Wiley, New York, pp. 2416–2419;
b) M. Y. Darensbourg, E. J. Lyon, J. J. Smee, Coord. Chem. Rev.
2000, 206–207, 533–561; c) A. C. Marr, D. J. E. Spencer, M. Schrçd-
er, Coord. Chem. Rev. 2001, 219–221, 1055–1074; d) E. Bowman,
R. K. Henderson, A. L. Spek, J. Reedijk, Eur. J. Inorg. Chem. 1999,
217–219.
[31] Unlike externally added alcohols, an alkynol can coordinate the
active metal center, its distal hydroxy functionality approaching the
appropriate intermediate. Hence a high yield of the ester may arise.
[32] Plain unconjugated olefins like 1,5-hexadiene do not undergo the
carbonylative cycloaddition reaction with allyl halides. Electrophilic
attack of an allylic iron halide by allyl halide would result in the re-
lease of uncoordinated 1,5-hexadiene.
[10] R. A. Sheldon, J. Chem. Technol. Biotechnol. 1997, 68, 381–388.
[11] P. T. Anastas, M. M. Kirchhoff, Acc. Chem. Res. 2002, 35, 686–694.
[12] D. del Moral, A. M. Banet Osuna, A. Cꢀrdoba, J. M. Moretꢀ, J.
Veciana, S. Ricart, N. Ventosa, Chem. Commun. 2009, 4723–4725 .
[13] S. Das, S. Chandrasekar, J. S. Yaday, R. Gree, Chem. Rev. 2007, 107,
3286–3337.
[14] H. Kiyota, T. Koike, E. Higashi, T. Oritani, Flavour Frag J. 2001, 16,
175–179.
[15] G. P. Chiusoli, G. Bottaccio, A, Cameroni, Chim. Ind. (Milan)
1962, 44, 131–135.
[16] We tried other bases: Calcium carbonate was not as efficient as
sodium acetate and proton scavengers (tertiary amines, proton
sponge, or 2,3-dimethylbut-2-ene) either inhibited the reaction com-
pletely, giving a strong blue color (amines), and/or gave very small
yields of the adduct.
[17] a) J. K. Kochi, Pure Appl. Chem. 1980, 52, 571–605; b) F. M. Dayrit,
J. Schwartz, J. Am. Chem. Soc. 1981, 103, 4466–4473; c) L. S. Hege-
dus, D. H. P. Thompson, J. Am. Chem. Soc. 1985, 107, 5663–5669;
d) T. T. Tsou, J. K. Kochi, J. Am. Chem. Soc. 1979, 101, 7547–7560.
[18] a) V. B. Phapale, D. J. Cꢆrdenas, Chem. Soc. Rev. 2009, 38, 1598–
1607; b) E. DuÇach, D. Franco, S. Olivero, Eur. J. Org. Chem. 2003,
1605–1622.
[19] M. Kuroboshi, M. Tanaka, S. Kishimoto K. Goto M. Mochizuki, H.
Tanaka, Tetrahedron Lett. 2000, 41, 81–84.
[33] a) F. Vçgtle, G. Richardt, N. Werner, Dendrimer Chemistry, Wiley-
VCH, Weinheim, 2009; b) A. W. Bosman, H. M. Janssen, E. W.
Meijer, Chem. Rev. 1999, 99, 1665–1688.
[20] The extreme purity of the cycloadduct obtained with this reductant
is remarkable. At the end of the four cycles an insoluble red precipi-
tate was formed, whereas the supernatant was colorless and CO
uptake was definitely stopped. Plausibly, the resulting tetraaminodi-
carbenium dihalide trapped the degraded NiII in an insoluble poly-
meric cationic complex, although this has not been further investi-
gated.
Received: November 30, 2009
Revised: April 23, 2010
Published online: June 23, 2010
9202
www.chemeurj.org
ꢃ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 9193 – 9202