Stereoselective rhodium-catalysed [2+2+2] Cycloaddition of linear allene-ene/yne-allene substrates: Reactivity and theoretical mechanistic studies

Article abstract of DOI:10.1002/chem.201304463

Allene-ene-allene (2 and 5) and allene-yne-allene (3 and 7) N-tosyl and O-linked substrates were satisfactorily synthesised. The [2+2+2] cycloaddition reaction catalysed by the Wilkinson catalyst [RhCl(PPh₃)₃] was evaluated. Substrates 2 and 5, which bear a double bond in the central position, gave a tricyclic structure in a reaction in which four contiguous stereogenic centres were formed as a single diastereomer. The reaction of substrates 3 and 7, which bear a triple bond in the central position, gave a tricyclic structure with a cyclohexenic ring core, again in a diastereoselective manner. All cycloadducts were formed by a regioselective reaction of the inner allene double bond and, therefore, feature an exocyclic diene motif. A Diels-Alder reaction on N-tosyl linked cycloadducts 8 and 10 allowed pentacyclic scaffolds to be diastereoselectively constructed. The reactivity of the allenes on [2+2+2] cycloaddition reactions was studied for the first time by density functional theory calculations. This mechanistic study rationalizes the order in which the unsaturations take part in the catalytic cycle, the reactivity of the two double bonds of the allene towards the [2+2+2] cycloaddition reaction, and the diastereoselectivity of the reaction.

Full text of DOI:10.1002/chem.201304463

DOI: 10.1002/chem.201304463
Full Paper
&
Reaction Mechanisms
Stereoselective Rhodium-Catalysed [2+2+2] Cycloaddition of
Linear Allene–Ene/Yne–Allene Substrates: Reactivity and
Theoretical Mechanistic Studies
Ewelina Haraburda,[a] ꢀscar Torres,[a] Teodor Parella,[b] Miquel Solꢁ,*[a] and Anna Pla-
Quintana*[a]
Abstract: Allene–ene–allene (2 and 5) and allene–yne–allene
(3 and 7) N-tosyl and O-linked substrates were satisfactorily
synthesised. The [2+2+2] cycloaddition reaction catalysed
by the Wilkinson catalyst [RhCl(PPh3)3] was evaluated. Sub-
strates 2 and 5, which bear a double bond in the central po-
sition, gave a tricyclic structure in a reaction in which four
contiguous stereogenic centres were formed as a single dia-
stereomer. The reaction of substrates 3 and 7, which bear
a triple bond in the central position, gave a tricyclic structure
with a cyclohexenic ring core, again in a diastereoselective
manner. All cycloadducts were formed by a regioselective re-
action of the inner allene double bond and, therefore, fea-
ture an exocyclic diene motif. A Diels–Alder reaction on N-
tosyl linked cycloadducts 8 and 10 allowed pentacyclic scaf-
folds to be diastereoselectively constructed. The reactivity of
the allenes on [2+2+2] cycloaddition reactions was studied
for the first time by density functional theory calculations.
This mechanistic study rationalizes the order in which the
unsaturations take part in the catalytic cycle, the reactivity
of the two double bonds of the allene towards the [2+2+2]
cycloaddition reaction, and the diastereoselectivity of the re-
action.
Introduction
fectively construct a variety of polysubstituted hetero- and
carbo-cycles. The list of possible p-components includes, for
example, nitriles as triple-bonded functional groups and al-
kenes, carbonyls, isocyanates or isothiocyanates as double-
bonded functional groups. Allenes, characterised by their two
perpendicular p bonds, have been recognised as versatile sub-
strates or intermediates in modern organic synthesis,[2] and,
due to the high density of unsaturation, have shown particular
promise in the development of cycloaddition and cycloisomeri-
sation reactions.[3] However, their use as substrates for
[2+2+2] cycloaddition reactions has not been explored in
depth. The nickel-catalysed trimerisation of allene was reported
in 1959,[4] but it was not until much more recently that Ma
et al., explored the cycloaddition of three allenes under trans-
[RhCl(CO)(PPh3)2] catalysis to construct steroid-like tricyclic skel-
etons.[5]
One of the main goals of modern organic synthesis is to devel-
op new reactions in which the molecular complexity is rapidly
increased. The transition-metal-catalysed [2+2+2] cycloaddi-
tion reaction of three unsaturated partners,[1] which allows for
the formation of three new bonds is a nice example of such
a reaction type. The reaction not only delivers products with
considerably increased structural and functional complexity
but also runs with total atom economy. Therefore, it offers sig-
nificant advantages in the development of economical, more
ecological syntheses of complex synthetic targets.
The [2+2+2] cycloaddition reaction was first described to
form benzene rings by the reaction of three alkynes, but other
functional groups can also participate providing a means to ef-
Allenes have been used as substrates in [2+2+2] cycloaddi-
tion reactions also in combination with other p-components,
especially alkynes. Aubert and Malacria et al. used stoichiomet-
ric amounts of [Co(CO)2(Cp)] (Cp=cyclopentadienyl) to obtain
steroid skeletons by a completely intramolecular [2+2+2] cy-
cloaddition reaction of allenediynes.[6] Cheng et al., explored
the completely and partially intramolecular [2+2+2] cycloaddi-
tion reaction of an allene with monoynes or diynes, respective-
ly.[7] Isomerisation of the cycloadduct formed furnished ben-
zene derivatives. However, non-isomerised products were
formed when in-situ generated arynes were reacted with al-
lenes by the same authors.[7d] Ma and Jiang also made an in-
cursion into the rhodium-catalysed cycloaddition of allenynes,
[a] E. Haraburda,+ ꢀ. Torres,+ Prof. Dr. M. Solꢁ, Dr. A. Pla-Quintana
Institut de Quꢂmica Computacional i Catꢁlisi (IQCC)
and Departament de Quꢂmica
Universitat de Girona (UdG), Facultat de Ciꢃncies
Campus de Montilivi, s/n, 17071 Girona (Spain)
Fax: (+34)972-41-81-50
E-mail: miquel.sola@udg.edu
anna.plaq@udg.edu
[b] Dr. T. Parella
Servei de Ressonꢁncia Magnꢃtica Nuclear
Universitat Autꢄnoma de Barcelona (UAB)
08193 Cerdanyola, Barcelona (Spain)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304463.
Chem. Eur. J. 2014, 20, 5034 – 5045
5034
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
describing a bimolecular cyclisation followed by a Diels–Alder
reaction, which furnished polycyclic bislactone structures.[8]
[2+2+2] Cycloaddition reactions involving allenes and p-
components other than alkynes have only been the focus of
research in the last few years. Efficient access to enantiomeri-
cally enriched dihydropyrimidine-2,4-diones by a nickel-cata-
lysed intermolecular [2+2+2] cycloaddition reaction of two
molecules of isocyanate and one molecule of allene has been
described by Murakami et al.[9] Alexanian et al. studied a stereo-
selective version of the [2+2+2] cycloaddition reaction in its
partially intermolecular version. Using [{RhCl(C2H4)2}2] treated
with silver salts in the presence of 2,2’-bis(diphenylphosphino)-
5,5’,6,6’,7,7’,8,8’-octahydro-1,1’-binaphthyl (H8-BINAP) to gener-
ate the catalytic species, ene–allenes and allenoates reacted to
give bicyclic dienes in moderate to good enantiomeric
ratios.[10] More recently, the same authors disclosed the diaste-
reoselective synthesis of carbocycles featuring five contiguous
stereocentres by a nickel-catalysed [2+2+2] cycloaddition of
eneallenes and alkenes.[11] Saito, Sato et al. described the [RuCl-
rocyclic substrates,[14] and to tricyclic skeletons starting from
linear triunsaturated substrates.[15] Our interest in the construc-
tion of topologically new polycyclic structures, prompted us to
explore the [2+2+2] cycloaddition reaction of linear allene–
ene–allene and allene–yne–allene derivatives. Here we describe
the synthesis of the substrates, its [2+2+2] cycloaddition reac-
tion to furnish tricyclic diene scaffolds, and the subsequent
Diels–Alder reaction efficiently affording pentacyclic com-
pounds. Mechanisms that account for the observed reactivity
have been proposed for [2+2+2] cycloaddition reactions in-
volving allenes but, to the best of our knowledge, a mechanis-
tic study for such a transformation has not been described.
The reaction is studied by means of DFT calculations, leading
us to propose a mechanism that rationalises the reactivity of
allenes in the [2+2+2] cycloaddition reaction.
Results and Discussion
Synthesis of the substrates
(cod)(Cp*)]-catalysed
(Cp*=pentamethylcyclopentadienyl;
cod=1,5-cyclooctadiene) completely intramolecular [2+2+2]
cycloaddition reaction of allene–yne–enes, which accounts for
the construction of fused-tricyclic skeletons.[12] Finally, Oonishi,
Sato et al. report the [2+2+2] cycloaddition of a linear allene–
alkyne–aldehyde substrates affording tricyclic compounds with
a pyran ring, catalysed by [{RhCl(cod)}2] and AgBF4.[13]
To evaluate the reactivity of allene–ene–allene and allene–yne–
allene systems in the completely intramolecular [2+2+2] cyclo-
addition reaction, four model substrates bearing either an N-
tosyl (NTs) or O-tether were designed. Two different synthetic
pathways were envisaged depending on the tether used. For
the NTs-tethered substrates, N-tosylbuta-2,3-dien-1-amine[16] (1)
was reacted with either (E)-1,4-dibromobutene or 1,4-dibromo-
butyne to furnish the desired substrates (Scheme 1).
In the course of our ongoing programme based on the syn-
thesis of polycyclic structures, we have explored the intramo-
lecular [2+2+2] cycloaddition reaction as an efficient way to
rapidly increase complexity. We have evaluated both methodo-
logically and mechanistically the [2+2+2] cycloaddition reac-
tion as an efficient entry to tetracycles by the reaction of mac-
Abstract in Catalan: S’ha preparat satisfactꢄriament compos-
tos contenint dues unitats d’al·lꢃ i un doble (2 and 5) o un triple
enllaÅ (3 and 7) amb unitats enllaÅants N-tosil i O. Aquests com-
postos s’han emprat com a substrats en processos de cicloaddi-
ciꢅ [2+2+2] catalitzats pel complex de Wilkinson [RhCl(PPh3)3].
Els substrats 2 i 5, contenint un doble enllaÅ en posiciꢅ central,
han donat compostos tricꢂclics en una reacciꢅ on s’han format
de forma diastereoselectiva quatre centres estereogꢃnics. La reac-
ciꢅ dels substrats 3 i 7 contenint un triple enllaÅ en posiciꢅ cen-
tral, han permꢃs l’obtenciꢅ de tricicles contenint un anell central
ciclohexꢁnic en una reacciꢅ de nou diastereoselectiva. Tots els ci-
cloadductes s’han format mitjanÅant la reacciꢅ regioselectiva del
doble enllaÅ intern de l’al·lꢃ. Aixꢂ s’han obtingut compostos con-
tenint un diꢃ exocꢂclic. Una reacciꢅ de Diels–Alder en els substrats
amb uniꢅ N-tosil 8 i 10, ha permꢃs la construcciꢅ diastereoselec-
tiva d’estructures pentacꢂcliques. La reactivitat dels al·lens en la
cicloaddiciꢅ [2+2+2] s’ha estudiat per primer cop mitjanÅant
cꢁlculs teꢄrics basats en la teoria del funcional de la densitat.
Aquest estudi mecanꢂstic ha permꢃs racionalitzar l’ordre en quꢃ
les insaturacions participen en la reacciꢅ, la reactivitat dels dos
dobles enllaÅos de l’al·lꢃ en la cicloaddiciꢅ [2+2+2] i la diaster-
eoselectivitat del procꢆs.
Scheme 1. Synthesis of the NTs-tethered substrates (NTs=p-toluenesulfonyl).
For the substrates bearing an oxygen atom on the tether,
a different pathway was followed that consisted in construct-
ing the allene by a Crabbꢃ homologation[17] on the corre-
sponding enediyne[18] (4) or triyne[19] (6) substrate. The terminal
alkynes were converted to terminal allenes by reaction with
paraformaldehyde in the presence of dicyclohexylamine and
CuI (Scheme 2), following the improved methodology of Ma
et al.[20]
Reactivity studies
After preparation of the substrates, their reactivity was evaluat-
ed (Table 1). The allene–yne/ene–allene substrates were treated
with the Wilkinson catalyst in toluene. We were pleased to find
that the reaction was completely regioselective with cycloaddi-
tion taking place only with the internal double bonds of the
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5035
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. The three diastereomers with trans protons coming from the
alkene that can be formed upon [2+2+2] cycloaddition of allene–ene–
allene substrates.
Scheme 2. Synthesis of the O-tethered substrates.
acted.[10a,11,12,14a] Therefore, only cycloadducts with a trans con-
figuration of the two protons coming from the double bond
were considered as possible products for the [2+2+2] cycload-
dition reaction (Figure 1).
Table 1. [2+2+2] cycloaddition of allene–yne/ene–allene substrates.
1D and 2D NMR spectroscopic experiments were used to
1
determine the product formed. H NMR spectra did not show
a symmetrical product, so i and ii, which both featured a C2-
axis, were discarded. Thus, diastereisomer iii was considered to
be the more likely diastereomer formed as was later confirmed
by 2D NMR spectroscopy. After a complete chemical shift as-
signment of the signals, the dipolar couplings observed in the
NOESY spectrum (Figure 2) were analysed and compared with
the distances obtained in the modelled structure of diastereo-
isomer iii for cycloadduct 8.
Entry
1
Substrate
t [h]
Product
Yield [%]
41
6.5
2
2
44
3
4
2
1
32
27
allene furnishing tricyclic dienes. Furthermore, the reaction was
diastereoselective and just one of the possible diastereoiso-
mers for each product was obtained in moderate to good
yields. It became evident during the optimisation that the reac-
tion time and temperature were having a significant influence
on the yield. The reaction at temperatures lower than 1008C
did not reach completion and degradation increased when the
temperature was raised further or the reaction time was in-
creased. Other solvents, such as dichloroethane (DCE) or etha-
nol, invariably led to a lower yield of the product. It is note-
worthy that the reaction proceeded satisfactorily without the
need for the electron-withdrawing CO2R group, which had
proven crucial in other [2+2+2] cycloaddition reactions involv-
ing allenes.[7a,c,8,10] The yields for the allene–yne–allene sub-
strates were lower than those for the allene–ene–allene sub-
strates due to the increased decomposition observed during
the reaction.
Figure 2. Representation of cycloadduct 8 showing the major NOESY corre-
lations that were observed and help to confirm the proposed geometry.
Olefinic proton signals on C1 and C12 were used to differen-
tiate the diastereotopic protons in the proximal methylenic
units C4 and C9 (plain arrows). Intense dipolar couplings be-
tween the axial protons on C5 and C8 and the H6 proton were
observed (dashed arrows). The strong NOE contacts observed
between the H7 proton and the H3 and H10 protons (bold
arrows), and the absence of NOE cross-peaks between the H6
proton with any of the H3, H7 and H10 protons, fully confirms
the relative configuration between the fused-C3, C6, C7 and
C10 stereocenters. The scalar couplings could not be com-
pletely determined due to the overlapping of signals.
Substrates 3 and 7 bearing a central triple bond gave tricy-
clic structures with a central cyclohexenic ring again featuring
the exocyclic diene motif. Only one diastereoisomer was ob-
Substrates 2 and 5 bearing a central double bond gave a tri-
cyclic structure with a central cyclohexanic ring with an exo-
cyclic diene. Four contiguous stereogenic centres were formed
as a single diastereomer. [2+2+2] cycloaddition reactions
involving double bonds are known to be stereoespecific
forming trans cycloadducts when trans double bonds are re-
1
served, which was symmetric as shown by the H NMR spectra.
Two different diastereomers can be formed in the cycloaddi-
tion showing either a symmetry plane or a C2 axis (Figure 3).
1D and 2D NMR spectroscopic experiments were examined to
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5036
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 3. The two diastereomers that can be formed upon [2+2+2] cycload-
dition of allene–yne–allene substrates.
determine the product formed but since both are symmetric
isomers, it was not possible to obtain an unambiguous answer
at this point. Further reaction of these cycloadducts (see
below) led us to confirm that, analogously to the reaction of
the allene–ene–allene substrates, the 1,4-H were in cis disposi-
tion and therefore diastereomer v was formed.
Scheme 3. Diels–Alder reaction of the tricyclic diene cycloadducts.
The combination of bisphosphine ligands and a cationic rho-
dium source was then studied as catalytic system for the cyclo-
addition of the allene–yne/ene–allene substrates with a view
to studying the enantioselectivity of the process. Various phos-
phines ((R)-H8-BINAP and (R)-BINAP), cationic rhodium sources
([Rh(cod)]BF4 or [{RhCl(C2H4)2}2] with AgBF4), solvents (dichloro-
methane, dichloroethane, methanol, and toluene) and temper-
atures (room temperature, 508C, and reflux) were systematical-
1
ly tested. H NMR spectroscopic analysis of the reaction mix-
ture invariably showed a mixture of diastereoisomers. The
combination of a biphosphine ligand and a cationic rhodium
source was a more active catalytic system (shorter reaction
times were necessary to consume the starting material) but it
was not diastereoselective. Therefore, the enantioselectivity
study was aborted.
Figure 4. Representation of cycloadduct 13 showing the major NOESY corre-
lations that were observed and help to confirm the proposed geometry.
Diels–Alder reaction
The regioselective formation of tricyclic dienes paved the way
for a Diels–Alder reaction on the substrate enabling increasing
complexity.[5,8,10] Cycloadducts 8 and 10 were treated with mal-
eimide in dichloromethane at reflux. Good yields of the cyclo-
adducts were obtained in short reaction times. The two penta-
cyclic compounds differ in the central cyclohexanic ring in that
one has a cyclohexene (12), whereas the other has a 1,4-cyclo-
hexadiene (13). The process was again diastereoselective and
just one diastereoisomer was formed in each of the reactions
(Scheme 3). Analysis of the NMR spectroscopic data for cyclo-
adduct 13, showed a symmetrical product proving that com-
pound 10 also had a symmetrical plane allowing us to con-
clude that diastereoisomer v for compound 10 (Figure 3) with
the two hydrogens in syn had been formed. Stereoselective
chemical shift assignments and relative configurations for all
centres in derivatives 12 and 13 were fully confirmed by NOE
spectroscopic data (Figure 4 and in the Supporting Informa-
tion).
tions at the M06-2X/cc-pVTZ//B3LYP/cc-pVDZ level of theory
(see the Computational Methods section for a more detailed
description of the method used) on the substrate with an O
atom as a tether, that is, substrate 5, and considering [RhCl-
(PPh3)] as the catalytically active species derived from the Wil-
kinson catalyst. The use of this active species in the [2+2+2]
cycloaddition reaction is justified by a previous study in which
some of us found that the catalytically active species in the ox-
idative addition process catalysed by Wilkinson catalyst is the
[RhCl(PPh3)] complex.[21]
The mechanism postulated for the [2+2+2] cycloaddition
reaction involving alkynes is composed of two main key steps
(Scheme 4). The first is the oxidative addition that occurs be-
tween two of the unsaturations initially coordinated to the
metal. The second is the insertion of the third unsaturation, al-
ready coordinated to the metal, which is followed by a ring-
closing step. The catalytic cycle is completed by the decoordi-
nation of the catalyst from the product. The exact nature of
the intermediates formed depends on the metal used, the
type of unsaturations participating in the reactions, and even
on the experimental conditions.
Computational calculations
After exploring the reactivity of our model systems in the labo-
ratory, we turned our attention to unravel the reaction mecha-
nism of the intramolecular [2+2+2] cycloaddition reaction in-
volving two allenes. With this aim, we performed DFT calcula-
[2+2+2] Cycloaddition reactions involving unsaturations
other than alkynes are postulated as following the same gener-
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5037
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
nal double bond of one allene and the internal double bond
of the other allene. The formation of oxidative coupling prod-
ucts B and G was discarded due to their endergonic nature
and relatively high barriers of the oxidative addition processes.
The rest of the couplings that are exergonic and 5.2E that has
the second lowest barrier of formation were considered for fur-
ther analysis.
Scheme 4. Generally accepted mechanism for the [2+2+2] cycloaddition re-
First, the allene (internal double bond)–alkene couplings
were analysed. Rhodacyclopentane 5.2A is the second most
stabilised and has the lowest barrier of formation (DG° =
18.7 kcalmolꢀ1). The barrier transforming 5.1B to 5.2B, which
differs from 5.2A by the location of the coordinated PPh3 and
action involving three alkynes.
al mechanistic scheme, with some particularities. In the case of
reactions involving allenes, such as the ones presented here,
there are three aspects that deserve special attention.
Firstly, it is necessary to determine the order in which
the unsaturations take part in the catalytic cycle (i.e.,
if the oxidative addition takes place between the two
allene moieties or between an allene and the double
bond). Secondly, due to the fact that the allenes have
two contiguous double bonds, we must determine
which of the double bonds is reacting and why. Final-
ly, it is important to explain why the reaction under
study is diastereoselective.
To address these three concerns we undertook
a DFT study to evaluate the energies involved in the
oxidative addition step leading to the rhodacyclopen-
tane intermediate. Of the many possibilities for the
oxidative cycloaddition, we decided to study not
only the ones that could explain the formation of the
carbon skeleton obtained in the experimental results
(5.2A–D, 5.2F, Figure 5), but also three cycloadditions
involving the terminal double bond of the allene that
did not lead to the observed product (5.2E, 5.2G and
5.2H, Figure 5). The oxidative addition intermediates
involving the ene and the internal double bond of
the allene A–D differ either in the relative disposition
of the PPh3 ligand (A and B) or the faces of the
double bond that had reacted giving different ring
fusion of the newly formed bicyclic structure (cis for
A and B, trans for C and D).
Figure 5. Structure of the different oxidative coupling products modelled in the present
study. The rhodacyclopentane compounds 5.2A–H have one PPh3 and a Cl attached to
the rhodium; these ligands have been omitted in the scheme for clarity.
The barriers for the transformations of the sub-
strate coordinated to the [RhCl(PPh3)] catalytic
system (5.1A–H) to the cycloaddition intermediates
(5.2A–H) are shown in Scheme 5. In this scheme, rela-
tive energies are given with respect to 5.0 that corre-
spond to free [RhCl(PPh3)3] and initial substrate 5
(Scheme 2). To calculate these relative energies, the
energies of two free PPh3 molecules have been
added to the energies of the intermediates, TSs and
products found in the potential energy surface. All of
the oxidative additions modelled are exergonic
except for three cases that are endergonic by about
4–15 kcalmolꢀ1. These three cases correspond to the
rhodacyclopentane 5.2E obtained by reaction of the
two external double bonds of the allene, 5.2B that is
the result of the oxidative addition of an internal
allene double bond and the ene group, and 5.2G
Scheme 5. Gibbs energy profile for the different oxidative coupling products modelled in
the present study. Gibbs energies at 298 K are given in kcalmolꢀ1. The rhodacyclopen-
tane compounds 5.1A–H and 5.2A–H have one PPh3 and a Cl attached to the rhodium;
that is obtained after oxidative addition of the exter- these ligands have been omitted in the scheme for clarity.
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org 5038
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the internal double bond of the
allene for two reasons. First, we
have seen in the previous step
that the external allene double
bond is less reactive than the in-
ternal one. Second, addition to
the external double bond does
not lead to the experimentally
observed products. The insertion
of the internal double bond of
the allene can take place to give
a
rhodacycloheptane
or
a rhodabicyclo[3.2.0]heptane by
insertion into one of the RhꢀC
bonds of the rhodacyclopentane
(I, J, L, M) or to give a rhodanor-
bornane (K, N) by a Diels–Alder-
like transition state (Scheme 4
Figure 6. Optimised structures (B3LYP/cc-pVDZ) for TS(5.1A,5.2A) (left) and TS(5.1B,5.2B) (right).
Cl has the highest barrier (DG° =38.2 kcalmolꢀ1) showing the
importance of the coordination sphere of the metal on the
outcome of the reaction (see Figure 6). Formation of trans-
fused bicyclic rhodacyclopentadienes 5.2C and 5.2D showed
significantly higher Gibbs free energy barriers than the forma-
tion of 5.2A, and were therefore ruled out as possible reaction
pathways.
Next, the allene–allene coupling was evaluated. Only rhoda-
cyclopentane 5.2F, which has the two hydrogen atoms in syn
formation, and its corresponding transition state were located
when the two internal bonds of the two allenes were consid-
ered as reacting unsaturations. All attempts to locate the anti
stereoisomer were unsuccessful. Oxidative addition involving
the internal double bonds of the allene (5.2F) shows a higher
Gibbs free energy barrier than when the external double
bonds are involved (5.2E) (DG° =19.6 kcalmolꢀ1 for 5.2E and
DG° =31.7 kcalmolꢀ1 for 5.2F). The difference in Gibbs energy
barriers for the formation of 5.2A (allene–ene) and 5.2E
(allene–allene) is favourable to the allene–ene oxidative addi-
tion by only about 1 kcalmolꢀ1. However, 5.2E is endergonic
by 15.5 kcalmolꢀ1 and, therefore, thermodynamically disfav-
oured. The reason for the low stability of 5.2E has to be as-
cribed to the fact that it is the only oxidative coupling product
containing two trisubstituted Z double bonds. Finally, although
the 5.2H species that results from the allene (external double
bond)–ene is the most exergonic addition, it has a barrier that
is about 8 kcalmolꢀ1 higher than the one leading to 5.2A.
Therefore, we conclude that the preferred oxidative coupling
pathway was the alkene–allene (internal double bond) cou-
pling furnishing a cis ring fusion yielding 5.2A. The larger reac-
tivity of the internal double bond of the allene (compare 5.2A
and 5.2H) concurs with the fact that the occupied p molecular
orbital (MO) corresponding to this internal double bond is
0.4 eV destabilised relative to the occupied MO of the external
double bond.
Figure 7. Structure of the different approximations for the insertion of the
internal double bond of the allene to the rhodacyclopentane modelled in
the present study. In all cases (I–N) the rhodium has one PPh3 and a Cl at-
tached on the upper face that have been omitted in the scheme for clarity.
and Figure 7). In the case of the insertion, the rhodacyclopen-
tane is not symmetric, so two possibilities emerge, that is, in-
sertion on the RhꢀC(sp3) (I, L) or RhꢀC(sp2) (J, M) bond of the
rhodacyclopentane. Moreover, the two double bond faces of
the internal allene need to be considered. The insertion can
take place with the hydrogen atom on the allene pointing
inside the rhodacyclopentane (I, J, K) or outside the rhodacy-
clopentane (L, M, N). All these different attacks are schemati-
cally represented in Figure 7. It is worth noting here that just
one of the faces of the rhodium catalyst is free to react, since
the other one is hindered by the presence of the PPh3 and the
chloride.
Neither the transition state nor the rhodanorbornane inter-
mediate for the Diels–Alder mechanism (K, N) could be locat-
ed. This is not surprising taking into account the absence of
double bonds in the attacked rhodacyclopentane. However,
the four other possibilities for the insertion were modelled and
the results are represented in Scheme 6. As can be seen in this
scheme, the barriers for the insertion of the internal allene
double bond with the approaches J (42.9 kcalmolꢀ1) and L
Following the study from the 16-electron complex 5.2A, sev-
eral possibilities emerge for the insertion of the internal
double bond of the second allene. We have considered only
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5039
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 6. Gibbs energy profile for the different alkyne insertion and ring-closing steps modelled in the present study. Gibbs energies at 298 K are given in
kcalmolꢀ1. All intermediates and transition states have one PPh3 and a Cl attached to the rhodium atom. These ligands have been omitted in the scheme for
clarity.
(29.7 kcalmolꢀ1) are too high to be competitive and, therefore,
these two attacks were discarded. The approaches I (24.0 kcal
molꢀ1) and M (19.7 kcalmolꢀ1) are more favourable and, for
this reason, we have continued analysing the reaction path for
these two attacks.
When one depicts the whole Gibbs energy profile of the
5.0!9iii through approach M one finds that the turnover fre-
quency (TOF) determining transition state (TDTS) is
TS(5.7M,5.8M) corresponding to the ring-closing step
(Figure 9) and the TOF determining intermediate (TDI), the
most abundant reaction intermediate, is 5.7M with an energet-
The next step corresponds to the formation of the CꢀC
bond that generates the final cyclohexane ring. This process
takes place through TS(5.7,5.8). The barrier corresponding to
the 5.7M!5.8M transformation is 25.7 kcalmolꢀ1, whereas
that of the 5.7I!5.8I process is only 12.5 kcalmolꢀ1. The main
difference comes from the different stability of the 5.7M and
5.7I intermediates. In particular, 5.7M is 11.6 kcalmolꢀ1 more
stable than 5.7I because in the former there is an interaction
of the Rh with the allene terminal double bond that stabilises
this 16-electron complex (see Figure 8).
ic span of 25.7 kcalmolꢀ1 [22]
On the other hand, for approach I,
.
the TDTS is TS(5.5I,5.6I) corresponding to the insertion of the
second allene (Figure 9), the TDI is 5.2A and the energetic
span is 24.0 kcalmolꢀ1. From these energetic span values and
using the Kozuch and Shaik approximation to compute the
TOFs,[22] we find that the TOF for the I approach leading to the
experimentally found product 9iii is about 18-times higher
than that of the M approach generating the non-observed
product 9ii (ii and iii refer to the corresponding diastereomers
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5040
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
5.0 to 9iii transformation is exer-
gonic by 80.4 kcalmolꢀ1 and the
energetic span between TDI and
TDTS is 24.0 kcalmolꢀ1.
A comparative analysis can be
performed at this point between
linear allene–ene–allene and
yne–ene–yne systems[15d] when
submitted to
a rhodium-cata-
lysed [2+2+2] cycloaddition re-
action. Comparing the corre-
sponding postulated mecha-
nisms, several qualitative differ-
ences arise. First of all, the oxida-
tive addition step takes place
between the two triple bonds in
the
yne–ene–yne
systems,
Figure 8. Optimised structures (B3LYP/cc-pVDZ) for 5.7I (left) and 5.7M (right).
whereas the allene–ene–allene
substrate preferentially reacts by
allene–ene oxidative addition.
This might be ascribed to the
decreased reactivity of the al-
lenes relative to the triple bonds.
On the other hand, the rate-
determining step also varies for
the two systems. Whereas the
oxidative addition is the rate-de-
termining step for yne–ene–yne
systems, when two allenes are
involved in the place of the al-
kynes, insertion of the third un-
saturation becomes the rate-de-
termining step.
Conclusion
The reactivity of linear allene–
ene–allene and allene–yne–
allene substrates bearing either
an N-tosyl or O-tether was evalu-
ated in the rhodium-catalysed
[2+2+2] cycloaddition reaction.
Figure 9. Optimised structures (B3LYP/cc-pVDZ) for TS(5.5I,5.6I) (left) and TS(5.7M,5.8M) (right).
depicted in Figure 1). Although this is not a big difference, it
seems enough to justify the exclusive formation of 9iii. We
have also explored the effect on the Gibbs energy profiles of
increasing the temperature from 25 to 1008C. We have found
that the main conclusions remain the same, 9iii being the
most favourable product of the reaction. However, not unex-
pectedly, there is a decrease in the selectivity of the process
when the temperature increases. The different reaction paths
in relatively close proximity might possibly explain the de-
creased selectivity observed when changing the catalytic
system from Wilkinson’s catalyst to combinations of biphos-
phines and cationic rhodium complexes.
Wilkinson’s catalyst was able to promote the cycloaddition re-
gioselectively involving the internal double bonds of the allene
allowing for the preparation of scaffolds featuring exocyclic
dienes. Furthermore, the process is diastereoselective furnish-
ing cycloadducts with four contiguous stereocentres as only
one diastereomer for the allene–ene–allene substrates and two
stereocentres in the case of allene–yne–allene substrates. The
use of combinations of rhodium sources and biphosphines re-
sults in a more active but less selective cycloaddition reaction.
Pentacyclic compounds were obtained by reacting the cyclo-
adducts with maleimide in a Diels–Alder reaction in a reaction
that also proved to be diastereoselective.
The complete reaction mechanism for the Rh-catalysed reac-
tion of allene–ene–allene 5 that derives the present study is
shown in Scheme 7. The overall reaction corresponding to the
The reaction path of the rhodium-catalysed intramolecular
[2+2+2] cyloaddition of the O-linked allene–ene–allene sub-
strate was studied at the M06-2X/cc-pVTZ//B3LYP/cc-pVDZ
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5041
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 7. Gibbs energy profile for the [2+2+2] cycloaddition of allene–ene–allene 5 in which the catalytic species is [RhCl(PPh3)].
(E)-N,N-Di(buta-2,3-dienyl)-N,N-ditosylbut-2-ene-1,4-diamine
(2)
level of theory. The results show that initial oxidative addition
occurs between the internal double bond of the allene and
the ene in a cis ring fusion. Subsequent insertion of the
second allene takes place also with the internal double bond
yielding a rhodacycloheptane intermediate that evolves to a cy-
clohexane adduct by reductive elimination. Finally, release of
the catalytically active species yields the final product. In this
process, the TDTS is TS(5.5I,5.6I), corresponding to the inser-
tion of the second allene and the TDI is 5.2A, which is the in-
termediate of the initial oxidative addition reaction. The ener-
A
stirred mixture of N-tosylbuta-2,3-dien-1-amine (1) (0.30 g,
1.34 mmols), potassium carbonate (0.44 g, 3.20 mmols) and aceto-
nitrile (20 mL) was heated at reflux for 10 min. Then a solution of
(E)-1,4-dibromobut-2-ene, (0.14 g, 0.67 mmol) in acetonitrile (3 mL)
was added slowly to the reaction mixture. The reaction was heated
and monitored by TLC analysis until completion (6 h). The salts
were filtered off and the filtrate was evaporated. The residue was
purified by column chromatography on silica gel with hexanes/
ethyl acetate (8:2) to afford 2 (0.22 g, 56% yield) as a colourless
getic span between TDI and TDTS is 24.0 kcalmolꢀ1
.
1
solid. MW: 498.66 gmolꢀ1; m.p. 88–908C; H NMR (300 MHz, CDCl3,
5
258C, TMS): d=2.46 (s, 6H), 3.77–3.81 (m, 8H), 4.69 (dt, J=2.5 Hz/
3
3
4J=6.9 Hz, 4H), 4.79 (dq, J=6.0 Hz/4J=6.9 Hz, 2H), 5.47 (ddt, J=
1.52/3.53/5.09 Hz, 2H), 7.31 (d 3J=8.1 Hz, 4H), 7.67 ppm (d, 3J=
8.1 Hz, 4H); 13C NMR (75 MHz, CDCl3, 258C, TMS): d=21.52, 45.68,
47.82, 76.58, 85.42, 127.14, 129.31, 129.75, 137.32, 143.42,
209.49 ppm; IR (ATR): n˜ =2922, 1952, 1334, 1153 cmꢀ1; ESI-HRMS:
m/z: elemental analysis calcd (%) for [C26H30N2O4S2 +Na]+:
521.1539; found: 521.1557.
Experimental Section
General
Unless otherwise noted, materials were obtained from commercial
suppliers and used without further purification. All reactions requir-
ing anhydrous conditions were conducted in oven-dried glassware
under a dry nitrogen atmosphere. The solvents were removed
under reduced pressure with a rotary evaporator. Residues were
purified by chromatography on a silica gel column (230–400 mesh)
by using mixtures of hexane/ethyl acetate as the eluent.
(E)-N,N-Di(buta-2,3-dienyl)-N,N-ditosylbut-2-yne-1,4-diamine
(3)
A
stirred mixture of N-tosylbuta-2,3-dien-1-amine (1) (0.23 g,
1H and 13C NMR spectra were measured on a 400 or a 300 MHz
1.04 mmols), potassium carbonate (0.33 g, 2.36 mmols) and aceto-
nitrile (20 mL) was heated at reflux for 10 min. Then a solution of
1,4-dibromobut-2-yne, (0.10 g, 0.47 mmols) in acetonitrile (3 mL)
was added slowly to the reaction mixture. The reaction was heated
and monitored by TLC analysis until completion (20 h). The salts
were filtered off and the filtrate was evaporated. The residue was
purified by column chromatography on silica gel with hexanes/
ethyl acetate (8:2) to afford 3 (0.81 g, 35% yield) as a colourless
1
NMR spectrometer. H and 13C chemical shifts (d) are referenced to
internal solvent resonances and reported relative to SiMe4. The
chemical shifts were assigned on the basis of 2D COSY, NOESY,
HSQC and HMBC experiments performed under routine conditions.
N-Tosyl-2,3-butadien-1-amine (1),[16] (E)-1,4-bis(prop-2-ynyloxy)but-
2-ene (4),[18] and 1,4-bis(prop-2-ynyloxy)but-2-yne (6)[19] were pre-
pared as previously reported.
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5042
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
solid. MW: 496.64 gmolꢀ1; m.p. 86–888C; H NMR (300 MHz, CDCl3,
J=1.5 Hz, 1H), 5.07 (dd, J=1.5 Hz, 1H), 5.11 (d, J=1.5 Hz, 1H),
7.34 (t, J=8.1 Hz, 4H), 7.70 ppm (d, J=8.1 Hz, 4H); 13C NMR
(75 MHz, CDCl3, 258C, TMS): d=21.88, 21.91, 42.3, 43.0, 46.5, 46.6,
49.8, 50.7, 52.2, 52.5, 109.0, 114.7, 127.5, 130.2, 133.8, 134.8, 143.9,
144.3, 144.4, 144.5 ppm; IR (ATR): n˜ =2925, 1336, 1157 cmꢀ1; ESI-
258C, TMS): d=2.43 (s, 6H), 3.68 (dt, J=2.5/3J=6.9 Hz, 4H), 3.94
5
(s, 4H), 4.72 (dt, 5J=2.5, 4J=6.0 Hz, 4H), 4.90 (dq, 3J=6.9/4J=
3
3
6.0 Hz, 2H), 7.29 (d, J=8.4 Hz, 4H), 7.65 ppm (d, J=8.4 Hz, 4H);
13C NMR (75 MHz, CDCl3, 258C, TMS): d=21.54, 35.89, 45.48, 76.45,
78.30, 85.30, 127.14, 129.58, 136.18, 143.74, 209.65 ppm; IR (ATR):
n˜ =2917, 1945, 1340, 1157 cmꢀ1; ESI-HRMS: m/z: elemental analysis
calcd (%) for [C26H28N2O4S2 +Na]+: 519.1383; found: 519.1401.
HRMS: m/z: elemental analysis calcd (%) for [C26H30N2O4S2 +Na]+
:
521.1545; found: 521.1538.
Cycloadduct 9
4-[(E)-4-(Buta-2,3-dienyloxy)but-2-enyloxy]buta-1,2-diene (5)
Cycloadduct 9 (0.022 g, 44% yield) was obtained as a yellow oil.
MW: 192.25 gmolꢀ1 1H NMR (300 MHz, CDCl3, 258C, TMS): d=
A
mixture of (E)-1,4-bis(prop-2-ynyloxy)but-2-ene, 4, (1.00 g,
;
6.09 mmols), formaldehyde (0.46 g, 15.22 mmols), copper(I) iodide
(0.58 g, 3.05 mmols) and dioxane (30 mL) was heated at reflux for
10 min. Then dicyclohexylamine (2.18 mL, 10.96 mmols) was added
slowly to the reaction mixture. The reaction was heated and moni-
tored by TLC analysis until completion (3 h). The salts were filtered
off and the filtrate was evaporated. The residue was purified by
column chromatography on silica gel with hexanes/ethyl acetate
0.94–1.07 (m, 2H), 1.92–2.02 (m, 2H), 2.44 (s, 3H), 2.48 (s, 3H), 2.69
(dd, 1H), 2.85–2.93 (m, 1H), 3.03 (t, 1H), 3.20–2.56 (m, 4H), 3.53
(dd, 1H), 4.51 (d, J=1.5 Hz, 1H), 4.81 (d, J=1.5 Hz, 1H), 5.07 (dd,
J=1.5 Hz, 1H), 5.11 (d, J=1.5 Hz, 1H), 7.34 (t, J=8.1 Hz, 4H),
7.70 ppm (d, J=8.1 Hz, 4H); 13C NMR (75 MHz, CDCl3, 258C, TMS):
d=21.88, 21.91, 42.3, 43.0, 46.5, 46.6, 49.8, 50.7, 52.2, 52.5, 109.0,
114.7, 127.5, 130.2, 133.8, 134.8, 143.9, 144.3, 144.4, 144.5 ppm; IR
(ATR): n˜ =2925, 1639, 1038 cmꢀ1; ESI-HRMS: m/z: elemental analysis
calcd (%) for [C12H16O2 +Na]+ : 215.1043; found: 215.1027.
(8:2) to afford
192.25 gmolꢀ1
5 (0.68 g, 58% yield) as a yellow oil. MW:
;
1H NMR (300 MHz, CDCl3, 258C, TMS): d=4.01 (m,
8H), 4.78 (dt, J=2.5/4J=6.6 Hz, 4H), 5.23 (dq, J=6.6/3J=6.0 2H),
5.80 ppm (ddt, J=1.3/2.3/4.3 Hz, 2H); 13C NMR (100 MHz, CDCl3,
258C, TMS): d=67.85, 69.72, 75.67, 87.65, 129.48, 209.28 ppm; IR
(ATR): n˜ =2954, 1956, 1723, 1033 cmꢀ1; ESI-HRMS: m/z: elemental
analysis calcd (%) for [C12H16O2 +Na]+: 215.1043; found: 215.1053.
5
4
Cycloadduct 10
Cycloadduct 10 (0.016 g, 32% yield) was obtained as a colourless
solid. MW: 496.15 gmolꢀ1
;
m.p.: 185–1878C (dec.); 1H NMR
(400 MHz, CDCl3, 258C, TMS): d=2.43 (s, 6H), 2.72 (t, 2H), 3.28 (m,
2H), 3.49 (dd, 3H), 3.93 (m, 4H), 4.70 (d, 2H), 5.38 (d, 2H), 7.32 (d,
J=8.1 Hz, 4H), 7.70 ppm (d, J=8.1 Hz, 4H); 13C NMR (100 MHz,
CDCl3, 258C, TMS): d=21.57, 43.04, 49.02, 52.18, 110.63, 127.72,
128.22, 129.87, 132.94, 142.34, 143.94 ppm; IR (ATR): n˜ =2921,
1340, 1154, 1090 cmꢀ1; ESI-HRMS: m/z: elemental analysis calcd (%)
for [C26H28N2O4S2 +Na]+: 519.1383; found:519.1400.
4-[4-(Buta-2,3-dienyloxy)but-2-ynyloxy]buta-1,2-diene (7)
A
mixture of 1,4-bis(prop-2-ynyloxy)but-2-yne, 6, (0.32 g,
1.97 mmols), formaldehyde (0.15 g, 4.93 mmols), copper(I) iodide
(0.19 g, 0.99 mmols) and dioxane (20 mL) was heated at reflux for
10 min. Then dicyclohexylamine (0.70 mL, 3.55 mmols) was added
slowly to the reaction mixture. The reaction was heated and moni-
tored by TLC analysis until completion (4 h). The salts were filtered
off and the filtrate was evaporated. The residue was purified by
column chromatography on silica gel with hexanes/ethyl acetate
Cycloadduct 11
Cycloadduct 11 (0.010 g, 27% yield) was obtained as a yellow oil.
;
MW: 190.24 gmolꢀ1 1H NMR (300 MHz, CDCl3, 258C, TMS): d=
(9:1) to afford
7 (0.30 g, 51% yield) as a yellow oil. MW:
190.24 gmolꢀ1; H NMR (400 MHz, CDCl3, 258C, TMS): d=4.10 (dt,
5J=2.5/3J=6.9 Hz, 4H), 4.22 (s, 4H), 4.81 (dt, 5J=2.5/4J=6.6 Hz,
4H), 5.23 ppm (dq, 5J=6.6/Jꢀ1 =6.9 Hz, 2H); 13C NMR (75 MHz,
CDCl3, 258C, TMS): d=57.11, 67.38, 75.86, 82.23, 87.09,
209.56 ppm; IR (ATR): n˜ =2921, 1954, 1074 cmꢀ1; ESI-HRMS: m/z: el-
emental analysis calcd (%) for [C12H14O2 +Na]+: 213.0886; found:
213.0899.
1
3.46–3.52 (m, 4H), 4.23 (dt, 2H), 4.33–4.40 (m, 4H), 4.74 (dd, 2H),
5.47 ppm (dd, 2H); 13C NMR (75 MHz, CDCl3, 258C, TMS): d=44.79,
68.26, 72.11, 109.69, 129.21, 143.68 ppm; IR (ATR): n˜ =2925, 2854,
1726, 1039 cmꢀ1; ESI-HRMS: m/z: elemental analysis calcd (%) for
[C12H14O2+Na]+: 213.0886; found: 213.0879.
Diels–Alder cycloadduct 12
General procedure for the [2+2+2] cycloaddition reactions
A mixture of cycloadduct 8 (0.020 g, 0.040 mmols) and maleimide
(0.004 g, 0.040 mmols) was purged with nitrogen and dissolved in
dichloromethane (3 mL). The mixture was stirred at reflux for 4 h
(TLC monitoring). The solvent was evaporated and the crude was
purified by column chromatography on silica gel using a mixture
of hexane/ethyl acetate (9:1) as the eluent to afford 12 (0.022 g,
In a 10 mL two-necked round-bottomed flask, a mixture of the
substrate (0.200 mmols) and tris(triphenylphosphine)rhodium(I)
chloride (0.018 g, 0.020 mmols) was purged with nitrogen and dis-
solved in anhydrous toluene (3 mL). The mixture was stirred at
reflux for the time required (TLC monitoring). The solvent was re-
moved and the crude was purified by column chromatography on
silica gel using a mixture of hexane/ethyl acetate (9:1) as the
eluent to give cycloisomerised product.
92% yield) as a colourless solid. MW: 595.73 gmolꢀ1 1H NMR
;
(400 MHz, CDCl3, 258C, TMS): d=1.38–1.45 (m, 1H), 2.06–2.23 (m,
4H), 2.30–2.50 (m, 9H), 2.55–2.64 (m, 1H), 2.74–2.87 (m, 3H), 3.08–
3.16 (m, 3H), 3.21 (dd, J=6.2 Hz, 1H), 3.44–3.50 (m, 2H), 3.67 (dd,
J=7.4 Hz, 1H), 7.34 (dd, J=8.1 Hz, 4H), 7.68 (dd, J=8.1 Hz, 4H);
8.14 ppm (s, 1H); 13C NMR (100 MHz, CDCl3, 258C, TMS): d=21.56,
21.59, 25.73, 27.21, 39.80, 40.31, 40.70, 42.95, 44.36, 44.72, 50.01,
50.93, 51.04, 51.52, 127.14, 127.51, 129.91, 129.9, 130.44, 131.06,
132.74, 134.77, 143.66, 144.18, 179.22, 179.31 ppm; IR (ATR): n˜ =
2921, 1716, 1338, 1158 cmꢀ1; ESI-HRMS: m/z: elemental analysis
calcd (%) for [C30H33N3O6S2 +Na]+: 618.1703; found: 618.1684.
Cycloadduct 8
Cycloadduct 8 (0.041 g, 41% yield) was obtained as a colourless
solid. MW: 498.66 gmolꢀ1; m.p. 181–1838C; 1H NMR (300 MHz,
CDCl3, 258C, TMS): d=0.94–1.07 (m, 2H), 1.92–2.02 (m, 2H), 2.44 (s,
3H), 2.48 (s, 3H), 2.69 (dd, 1H), 2.85–2.93 (m, 1H), 3.03 (t, 1H),
3.20–2.56 (m, 4H), 3.53 (dd, 1H), 4.51 (d, J=1.5 Hz, 1H), 4.81 (d,
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5043
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Diels–Alder cycloadduct 13
Keywords: allenes
·
cycloaddition
·
density functional
A mixture of cycloadduct 10 (0.012 g, 0.024 mmols) and maleimide
(0.002 g, 0.024 mmols) was purged with nitrogen and dissolved in
dichloromethane (3 mL). The mixture was stirred at reflux for 1 h
(TLC monitoring). The solvent was removed and the crude was pu-
rified by column chromatography on silica gel using a mixture of
hexane/ethyl acetate (9:1) as the eluent to afford 13 (0.010 g, 75%
calculations · reaction mechanisms · rhodium
[1] For selected recent reviews, see: a) T. Shibata, K. Tsuchikama, Org.
Biomol. Chem. 2008, 6, 1317–1323; b) B. R. Galan, T. Rovis, Angew.
Chem. 2009, 121, 2870–2874; Angew. Chem. Int. Ed. 2009, 48, 2830–
2834; c) K. Tanaka, Chem. Asian J. 2009, 4, 508–518; d) P. A. Inglesby,
P. A. Evans, Chem. Soc. Rev. 2010, 39, 2791–2805; e) G. Dominguez, J.
Pꢃrez-Castells, Chem. Soc. Rev. 2011, 40, 3430–3444; f) N. Weding, M.
Hapke, Chem. Soc. Rev. 2011, 40, 4525–4538.
[2] a) S. Ma, Chem. Rev. 2005, 105, 2829–2871; b) C. Aubert, L. Fensterbak,
P. Garcia, M. Malacria, A. Simonneau, Chem. Rev. 2011, 111, 1954–1993.
[3] F. Lꢆpez, J. L. MascareÇas, Chem. Eur. J. 2011, 17, 418–428.
[4] For the trimerisation of allenes, see: a) R. E. Benson, R. V. Lindsey Jr., J.
Am. Chem. Soc. 1959, 81, 4247–4250; for the cycloaddition of allenes
and alkynes, see b) R. E. Benson, R. V. Lindsey Jr., J. Am. Chem. Soc.
1959, 81, 4250–4253.
1
yield) as a colourless solid. MW: 593.71 gmolꢀ1; H NMR (400 MHz,
CDCl3, 258C, TMS): d=2.11–2.17 (m, 2H), 2.44 (s, 6H), 2.48 (d, 2H),
2.58 (dd, 2H), 3.04 (brs, 2H), 3.15 (t, 2H), 3.64 (d, 2H), 3.81–3.81
(m, 4H), 7.33 (t, J=8.1 Hz, 4H), 7.70 ppm (d, J=8.1 Hz, 4H);
13C NMR (100 MHz, CDCl3, 258C, TMS): d=21.60, 26.14, 40.67,
43.52, 48.63, 50.98, 127.54, 127.72, 127.72, 129.13, 129.98, 133.67,
144.00, 178.83 ppm; IR (ATR): n˜ =2918, 1707, 1337, 1156,
1091 cmꢀ1
; ESI-HRMS: m/z: elemental analysis calcd (%) for
[C30H31N3O6S2 +Na]+: 616.1546; found: 616.1531.
[5] a) S. Ma, P. Lu, L. Lu, H. Hou, J. Wei, Q, He, Z. Gu, X. Jiang, X. Jin,
Angew. Chem. 2005, 117, 5409–5412; Angew. Chem. Int. Ed. 2005, 44,
5275–5278; b) S. Ma, L. Lu, Chem. Asian J. 2007, 2, 199–204; c) P. Lu, S.
Ma, Org. Lett. 2007, 9, 5319–5321; d) G. Chen, X. Jiang, C. Fu, S. Ma,
Chem. Lett. 2010, 39, 78–81.
[6] a) D. Llerena, O. Buisine, C. Aubert, M. Malacria, Tetrahedron 1998, 54,
9373–9392; b) M. Petit, C. Aubert, M. Malacria, Org. Lett. 2004, 6, 3937–
3940; c) M. Petit, C. Aubert, M. Malacria, Tetrahedron 2006, 62, 10582–
10593.
[7] a) M. Shanmugasundaram, M.-S. Wu, M. Jeganmohan, C.-W. Huang, C.-
H. Cheng, J. Org. Chem. 2002, 67, 7724–7729; b) M. S. Wu, M. Shanmu-
gasundaram, C.-H. Cheng, Chem. Commun. 2003, 718–719; c) M. Shan-
mugasundaram, M.-S. Wu, C.-H. Cheng, Org. Lett. 2001, 3, 4233–4236;
d) J.-C. Hsieh, D. Kumar Rayabarapu, C.-H. Cheng, Chem. Commun. 2004,
532–533.
[8] X. Jiang, S. Ma, J. Am. Chem. Soc. 2007, 129, 11600–11607.
[9] T. Miura, M. Morimoto, M. Murakami, J. Am. Chem. Soc. 2010, 132,
15836–15838.
Computational details
All geometry optimisations were performed without symmetry
constraints using the hybrid DFT B3LYP[23] method with the all-elec-
tron cc-pVDZ basis set for P, Cl, O, C, and H[24] and the cc-pVDZ-PP
basis set containing an effective core relativistic pseudopotential
for Rh.[25] Single-point energy calculations were carried out with
the M06–2X functional[26] by using the cc-pVTZ basis set for P, Cl,
O, C, and H and cc-pVTZ-PP basis set for Rh (i.e. the M06–2X/cc-
pVTZ//B3LYP/cc-pVDZ method). The M06–2X functional was found
to provide excellent results in thermochemical kinetics tests[26] and
relatively accurate energies for cycloaddition reactions.[27] Analytical
Hessians were computed at the B3LYP/cc-pVDZ level of theory to
determine the nature of stationary points (one or zero imaginary
frequencies for transition states and minima, respectively) and to
calculate unscaled zero-point energies (ZPEs) as well as thermal
corrections and entropy effects by using the standard statistical-
mechanics relationships for an ideal gas.[28] These last two terms
were computed at 298.15 K and 1 atm to provide the reported rel-
ative Gibbs energies. Furthermore, the connectivity between sta-
tionary points was established by intrinsic reaction path calcula-
tions.[29] All calculations were performed with the Gaussian 09[30]
program package. A previous study found that solvent effects due
to toluene and acetonitrile in [2+2+2] cyclotrimerisations are
minor, likely due to the absence of charged or polarised intermedi-
ates and transition states in the reaction mechanism.[31] Because
the reactions studied are carried out in toluene solution, solvent ef-
fects have not been included in the present calculations.
[10] a) A. T. Brusoe, E. J. Alexanian, Angew. Chem. 2011, 123, 6726–6730;
Angew. Chem. Int. Ed. 2011, 50, 6596–6600; b) A. T. Brusoe, R. V. Edwan-
kar, E. J. Alexanian, Org. Lett. 2012, 14, 6096–6099.
[11] N. N. Noucti, E. J. Alexanian, Angew. Chem. 2013, 125, 8582–8585;
Angew. Chem. Int. Ed. 2013, 52, 8424–8427.
[12] N. Saito, T. Ichimaru, Y. Sato, Chem. Asian J. 2012, 7, 1521–1523.
[13] Y. Oonishi, Y. Kitano, Y. Sato, Tetrahedron 2013, 69, 7713–7718.
[14] a) A. Torrent, I. Gonzꢇlez, A. Pla-Quintana, A. Roglans, M. Moreno-MaÇas,
T. Parella, J. Benet-Buchholz, J. Org. Chem. 2005, 70, 2033–2041; b) S.
Brun, L. Garcia, I. Gonzꢇlez, A. Torrent, A. Dachs, A. Pla-Quintana, T. Par-
ella, A. Roglans, Chem. Commun. 2008, 4339–4341; c) A. Dachs, A. Tor-
rent, A. Roglans, T. Parella, S. Osuna, M. Solꢁ, Chem. Eur. J. 2009, 15,
5289–5300; d) A. Pla-Quintana, A. Roglans, Molecules 2010, 15, 9230–
9251; e) S. Brun, A. Torrent, A. Pla-Quintana, A. Roglans, X. Fontrodona,
J. Benet-Buchholz, T. Parella, Organometallics 2012, 31, 318–326.
[15] For the cycloaddition of triynes, see: a) I. Gonzꢇlez, A. Pla-Quintana, A.
Roglans, Synlett 2009, 2844–2848; For the cycloaddition of enediynes
see: b) S. Brun, M. Parera, A. Pla-Quintana, A. Roglans, T. Leon, T.
Achard, J. Solꢁ, X. Verdaguer, A. Riera, Tetrahedron 2010, 66, 9032–
9040; c) A. Dachs, A. Roglans, M. Solꢁ, Organometallics 2011, 30, 3151–
3159; d) A. Dachs, A. Pla-Quintana, T. Parella, M. Solꢁ, A. Roglans, Chem.
Eur. J. 2011, 17, 14493–14507; e) T. Leꢆn, M. Parera, A. Roglans, A. Riera,
X. Verdaguer, Angew. Chem. 2012, 124, 7057–7061; Angew. Chem. Int.
Ed. 2012, 51, 6951–6955; for the cycloaddition of cianodiynes, see: f) L.
Garcia, A. Pla-Quintana, A. Roglans, T. Parella, Eur. J. Org. Chem. 2010,
3407–3415.
Acknowledgements
Financial support from the Spanish Ministry of Education and
Science (MINECO) (projects nos.: CTQ2011–23121/BQU,
CTQ2011–23156/BQU and CTQ2012–32436), the DIUE of the
Generalitat de Catalunya (project no.: 2009SGR637), and the
FEDER fund (European Fund for Regional Development) for the
grant UNGI08–4E-003 is gratefully acknowledged. E.H. thanks
Universitat de Girona for a predoctoral grant. Excellent service
by the Centre de Serveis Cientꢄfics i Acadꢅmics de Catalunya
(CESCA) is gratefully acknowledged. Support for the research
of M.S. was received through the ICREA Academia 2009 prize
for excellence in research funded by the DIUE of the Generali-
tat de Catalunya.
[16] H. Ohno, T. Mizutani, Y. Kadoh, A. Aso, K. Miyamura, N. Fujii, T. Tanaka, J.
Org. Chem. 2007, 72, 4378–4389.
[17] P. Rona, P. Crabbꢃ, J. Am. Chem. Soc. 1969, 91, 3289–3292.
[18] R. Grigg, R. Scott, P. Stevenson, J. Chem. Soc. Perkin Trans. 1 1988,
1365–1369.
[19] R. Grigg, R. Scott, P. Stevenson, Tetrahedron Lett. 1982, 23, 2691–2692.
[20] J. Kuang, S. Ma, J. Org. Chem. 2009, 74, 1763–1765.
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5044
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[21] A. Dachs, S. Osuna, A. Roglans, M. Solꢁ, Organometallics 2010, 29, 562–
569.
[30] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Na-
katsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G.
Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Ha-
segawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Ra-
ghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakr-
zewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Dan-
iels, ꢈ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian
09, revision A.02; Gaussian, Inc., Pittsburgh, PA, 2009.
[22] a) S. Kozuch, S. Shaik, J. Am. Chem. Soc. 2006, 128, 3355–3365; b) S.
Kozuch, S. Shaik, J. Phys. Chem. A 2008, 112, 6032–6041; c) S. Kozuch, S.
Shaik, Acc. Chem. Res. 2011, 44, 101–110.
[23] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785–789; c) P. J. Stephens, F. J. Devlin,
C. F. Chabalowski, M. J. Frisch, J. Phys. Chem. 1994, 98, 11623–11627.
[24] a) T. H. Dunning, Jr., J. Chem. Phys. 1989, 90, 1007–1023; b) D. E. Woon,
T. H. Dunning, Jr., J. Chem. Phys. 1993, 98, 1358–1371.
[25] K. A. Peterson, D. Figgen, M. Dolg, H. Stoll, J. Chem. Phys. 2007, 126,
124101.
[26] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[27] a) S. N. Steinmann, C. Corminboeuf, J. Chem. Theory Comput. 2011, 7,
3567–3577; b) R. S. Paton, J. L. Mackey, W. H. Kim, J. H. Lee, S. J. Dani-
shefsky, K. N. Houk, J. Am. Chem. Soc. 2010, 132, 9335–9340.
[28] P. Atkins, J. De Paula in Physical Chemistry, Oxford University Press,
Oxford, 2006.
[31] L. Orian, J. N. P. van Stralen, F. M. Bickelhaupt, Organometallics 2007, 26,
3816–3830.
Received: November 14, 2013
[29] C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154–2161.
Published online on March 11, 2014
Chem. Eur. J. 2014, 20, 5034 – 5045
www.chemeurj.org
5045
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim