Rhodium(I)-catalysed intramolecular [2+2+2] cyclotrimerisations of 15-, 20- and 25-membered azamacrocycles: Experimental and theoretical mechanistic studies

Article abstract of DOI:10.1002/chem.200802548

A new series of 20- and 25- membered polyacetylenic azamacrocy- cles have been satisfactorily prepared and completely characterised by spec- troscopic methods. Various [2+2+2] cy- clotrimerisation processes catalysed by the Wilkinson's catalyst, [RhCl- (P

Full text of DOI:10.1002/chem.200802548

FULL PAPER
DOI: 10.1002/chem.200802548
Rhodium(I)-Catalysed Intramolecular [2+2+2] Cyclotrimerisations of 15-,
20- and 25-Membered Azamacrocycles: Experimental and Theoretical
Mechanistic Studies
Anna Dachs,[a] Anna Torrent,[b] Anna Roglans,*[b] Teodor Parella,[c] Sꢀlvia Osuna,[a] and
Miquel Solꢁ*[a]
Dedicated to Professor Josep Font on the occasion of his 70th birthday
Abstract: A new series of 20- and 25-
membered polyacetylenic azamacrocy-
cles have been satisfactorily prepared
and completely characterised by spec-
troscopic methods. Various [2+2+2] cy-
clotrimerisation processes catalysed by
the Wilkinsonꢀs catalyst, [RhCl-
bered azamacrocycle (like the previ-
ously synthesised 15-membered azama-
crocyle) led to the expected cyclotri-
merised compound in contrast to the
20-membered macrocycle, which is
characterised by its lack of reactivity.
The difference in reactivity of the 15-,
20- and 25-membered macrocycles has
been rationalised through density func-
tional theory calculations.
Keywords: cyclotrimerization
density functional calculations
·
·
macrocycles · reaction mechanisms ·
rhodium
ACHTUNGTRENNUNG(PPh3)3], were tested in the above-
mentioned macrocycles. The 25-mem-
Introduction
are formed in one step, is one of the most elegant methods
for the construction of polysubstituted aromatics, which
have important academic and industrial uses (for recent re-
views see references [1–4]) The Wilkinsonꢀs complex [RhCl-
Transition-metal-catalysed [2+2+2] cyclotrimerisation reac-
tions involving alkynes, in which three carbon–carbon bonds
AHCTUNTGRENNG(UN PPh3)3] has been widely employed in these reactions. Parti-
[a] A. Dachs, S. Osuna, Prof. M. Solꢁ
Institut de Quꢂmica Computacional and
Departament de Quꢂmica
ally intramolecular approaches or fully intramolecular cyclo-
addition processes represent efficient entries into various
multiple-fused ring compounds. Over the last few years, we
have developed an efficient rhodium(I)-catalysed [2+2+2]
cyclotrimerisation process of 15-, 16- and 17-membered
triynic and enediynic azamacrocycles of type 1 (to give com-
pounds 2) and have described the first examples in the liter-
ature of completed closed systems[5–8] (Scheme 1). To broad-
Universitat de Girona
Campus de Montilivi, 17071 Girona (Spain)
Fax : (+34)972-418-356
E-mail: miquel.sola@udg.edu
[b] A. Torrent, Dr. A. Roglans
Departament de Quꢂmica
Universitat de Girona
Campus de Montilivi, 17071 Girona (Spain)
Fax : (+34)972-418-150
E-mail: anna.roglans@udg.edu
[c] Dr. T. Parella
Servei de RMN
Universitat Autꢃnoma de Barcelona
Cerdanyola, 08193 Barcelona (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802548. It contains experi-
1
mental description of all new compounds shown in Scheme 2 and H
and 13C NMR spectra for compound 14 and details of the structure
determination, including atomic coordinates, bond lengths and angles,
thermal parameters, least-squares planes and interatomic contacts of
macrocycle 3a. Cartesian xyz coordinates and total energies of all sta-
tionary points located are also given.
Scheme 1. Cyclotrimerisation reactions of macrocyclic triynes and ene-
diynes.
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5289

en the scope of these cyclotri-
merisation reactions to other
macrocyclic systems and to
afford new kind of fused tetra-
cycles, we have prepared 20-
membered tetraacetylenic aza-
macrocycles 3a–c and 25-mem-
bered pentaacetylenic azama-
crocycle 4 (Scheme 2) and stud-
ied their cyclotrimerisation re-
actions catalysed by the Wilkin-
sonꢀs complex.
Later in this paper, we shall
discuss our finding that unlike
the 15- and 25-membered aza-
macrocycles, the 20-membered
tetraacetylenic azamacrocycles
3a–c do not lead to the expect-
ed [2+2+2] cyclotrimerisation
products. To understand the
origin of this lack of reactivity,
we decided to carry out theo-
retical calculations using densi-
ty functional theory (DFT) with
a hybrid functional.
After a pioneering semiem-
pirical study,[9] several DFT in-
vestigations into the reaction
mechanism of the transition-
metal-catalysed [2+2+2] cyclo-
trimerisation reactions involv-
ing alkynes have been reported
in the last decade. The reaction
mechanism for the alkyne cy-
clotrimerisation reaction cata-
Scheme 2. Synthesis of acetylenic azamacrocycles 3a, 3b, 3c and 4.
lysed by [CoCp(L)2] (Cp=cyclopentadiene; L=CO, PR3,
THF and olefin),[10–13] [RuCpCl][14–17] complexes and the
{RhCp}[18] and {RhInd} (Ind=indene)[18] fragments has been
analysed in these DFT studies. The generally accepted reac-
tion mechanism is drawn in Scheme 3. The reaction begins
with a pair of ligand–alkyne substitution reactions. Then, the
oxidative coupling of the two alkyne ligands generates a
metallacyclopentadiene IIIa ([CoCp(L)2], {RhCp} and
{RhInd}) or a metallacyclopentatriene IIIb ([RuCpCl]) with
a biscarbene structure. This step has been found to be the
rate-determining step with activation energies typically in
the range 11–14 kcalmolꢀ1. Subsequent coordination of a
third alkyne ligand to the metallacyclopentadiene or metal-
lacyclopentatriene intermediate is followed by either alkyne
insertion to form a planar aromatic metallacycloheptatriene
V (the so-called Schoreꢀs mechanism[19]) or metal-mediated
[4+2] cycloaddition to yield a 7-metallanorbornadiene com-
plex VI or [2+2] cycloaddition to give a metallabicyclo
[3.2.0]heptatriene VII. Finally, the arene is formed by the re-
ductive elimination of the metal. Although this is the gener-
al picture of the reaction mechanism, differences are found
between catalysts and the presence or absence of coordinat-
ing ligands, such as phosphine groups.[13]
Abstract in Catalan: S’ha sintetitzat i caracteritzat espectros-
cꢀpicament una nova sꢁrie de macrocicles nitrogenats polia-
cetilꢁnics de 20- i 25-membres. Amb aquests macrocicles
sꢂhan dut a terme les reaccions de ciclotrimeritzaciꢃ [2+2+2]
catalitzades pel catalitzador de Wilkinson, [RhClACHTNUTRGNE(UNG PPh3)3]. El
macrocicle nitrogenat de 25-membres (de la mateixa manera
que el macrocicle nitrogenat de 15-membres) permet l’ob-
tenciꢃ del compost ciclotrimeritzat. Per contra, el macrocicle
de 20-membres es caracteritza per la seva falta de reactivitat.
El diferent comportament de reactivitat dels macrocicles de
15-, 20-, i 25-membres ha estat estudiat mitjanÅant cꢄlculs teꢀ-
rics basats en la teoria del funcional de la densitat.
On the other hand, the cyclotrimerisation of nitriles and
acetylenes to afford pyridine rings has been also discussed
in a series of theoretical studies.[17,18,20,21] The rate-determin-
ing step of the overall catalytic cycle changes to become the
addition of the nitrile molecule to the metallacyclopenta-
diene or metallacyclopentatriene intermediate.[17,18] One of
5290
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300

Reaction Mechanisms
FULL PAPER
Results and Discussion
Macrocycles 3 and 4 were pre-
pared following the synthetic
pathway outlined in Scheme 2
and they were completely char-
acterised
by
spectroscopic
methods (see Supporting Infor-
mation for full experimental de-
Scheme 3. Schematic [2+2+2] reaction mechanism with M=transition metal.
tails). One of the variants these
the key aspects in the cyclotrimerisation of nitriles and ace-
tylenes is the competition between arene and pyridine for-
mation. Relevant to our discussion are also several papers
that address the co-cyclisation of two acetylene molecules
with alkenes or CS2.[14,22–25] Finally, we can briefly mention
here that the uncatalysed thermal cyclotrimerisation is dis-
favoured entropically and has a high activation barrier.[26,27]
To the best of our knowledge, a theoretical study of the
reaction mechanism of the transition-metal-catalysed
[2+2+2] cyclotrimerisation re-
macrocycles can present is the nature of the aryl units of the
periphery. Firstly, two macrocycles 3a and 4 containing the
same arylic unit, p-tolylsulfonamide, were prepared. The
whole synthesis started with the reaction of N-tert-butyloxy-
carbonyl (Boc)-protected p-tolylsulfonamide 5[28] and
0.5 equiv of 2-butyne-1,4-diol under Mitsunobu reaction
conditions to give compound 6. The elimination of the Boc
groups in compound 6 (to give 7) and the subsequent treat-
ment with two equivalents of bromo derivative 8[5] resulted
action in a macrocyclic triyne
has yet to be performed. This
[2+2+2] reaction should be fav-
oured over cyclotrimerisation
reactions involving free acetyl-
ACHTUNGTRENNUNGene molecules for entropic rea-
sons.[15] In this paper we report
the results of our theoretical ex-
amination of the mechanism for
the cyclotrimerisation catalysed
by the Wilkinsonꢀs complex of
the 15-, 20- and 25-membered
acetylenic azamacrocycles de-
picted in Scheme 4 with two
main goals: 1) to investigate the
origin for the different reactivi-
ty of the 20-membered tetra-
ACHTUNGTRENNUNGacetylenic azamacrocycles and
2) to discuss the chemoselectivi-
ty of the trimerisation in the
case of the pentaacetylenic aza-
macrocycles for which more
than a single product can be ob-
tained as cycloaddition can
occur between three adjacent
or non-adjacent triple bonds.
This is, we believe, the first the-
oretical study that addresses the
mechanism of a cyclotrimerisa-
tion reaction in a series of ace-
tylenic macrocycles. Further-
more, this is the first time that
the Wilkinsonꢀs catalyst has
been investigated in a theoreti-
cal study of an intramolecular
cyclotrimerisation reaction.
Scheme 4. Studies on cyclotrimerisation reactions of 15-,[6] 20- and 25-membered acetylenic azamacrocycles 1,
3 and 4.
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5291

M. Solꢁ, A. Roglans et al.
in the isolation of compound 9 with an 80% yield. The elim-
ination of the Boc groups again with the same reaction con-
ditions as before (TFA in CH2Cl2) gave intermediate 10.
Compound 10, which already contains three acetylenic
chains and four sulfonamide units, was the key intermediate
for the preparation of both macrocycles. Cyclisation of 10
with 1,4-dibromo-2-butyne in the presence of K2CO3 as a
base afforded a 60% yield of macrocycle 3a. When inter-
mediate 10 was condensed with the dichloro derivative 11a,
previously prepared by us,[8] 25-membered pentaacetylenic
macrocycle 4 was obtained with an almost quantitative
yield. The p-tolylsulfonamide units give a high level of insol-
ubility in the most common organic solvents in the 20-mem-
bered macrocycle. This permitted suitable crystals to be ob-
tained to make X-ray diffraction analysis. Figure 1 shows
the Ortep-plot diagram for the compound together with its
labelling scheme. Compound 3a crystallises free of solvent
molecules with C1 symmetry. Bond lengths and angles are
within expected values. The four triple bonds at the macro-
cyclic ring are located on different planes.
After preparing the macrocycles, the next step was to
study how to obtain fused tetracycles 13–15 by [2+2+2] cy-
clotrimerisation reactions. There is one possible way for
macrocyles 3 to cyclise to afford a single 5,5,10-fused ben-
zene derivative. However, in the case of the 25-membered
ring 4 there are two possible ways of cyclisation; that is, cy-
cloaddition between three consecutive triple bonds to afford
compound 14 or between non-consecutive triple bonds to
afford compound 15. The Wilkinsonꢀs catalyst [RhClACHTUNGTRENNUNG(PPh3)3]
was selected as it had formerly given the best results in our
previous studies.[5–8] When 20-membered macrocycles 3a–c,
heated under reflux in toluene, were treated with [RhCl-
AHCTUNTGRENNG(UN PPh3)3], the reaction did not take place. In all three cases,
starting material together with decomposition products were
obtained. A stoichiometric amount of [CpCo(CO)2], anoth-
er typical catalyst for this chemistry,[1–4] was also tested. The
macrocycle was heated under reflux in toluene and the solu-
tion was exposed to light. However, this reaction also failed
and the starting macrocycle was recovered. In contrast,
when 25-membered macrocycle 4 was treated with a catalyt-
ical amount of rhodium complex, the cyclotrimerised com-
pound 14, resulting from the reaction of three contiguous al-
kynes, was obtained as the only product of the process. Con-
firmation of the structure of 14 was made by HMBC and
2D NOESY data (see the Supporting Information). Experi-
mentally we observe that 5,5,15-fused core is more favoura-
ble than 5,10,10-fused system.
As stated in the Introduction section, we performed
B3LYP/cc-pVDZ-PP calculations (see section on Computa-
tional Methods for a more detailed description of the
method used) to unravel the reaction mechanism of the in-
tramolecular [2+2+2] cyclotrimerisation catalysed by the
Wilkinsonꢀs complex of the 15-, 20- and 25-membered acety-
lenic azamacrocycles (MAAs) depicted in Scheme 4 in order
to understand the lack of reactivity of the 20-MAA. As we
will see, these DFT computations turned out to be very
helpful to rationalise the experimental outcome. To reduce
the computational effort required, the SO2-Ar moieties pres-
ent in the experimental 15-, 20- and 25-MAA and the three
Ph groups of the Wilkinsonꢀs catalyst were substituted by H
atoms. Before starting our theoretical study, we checked
that the thermodynamics of the 1a!2a conversion is not
significantly altered when going from the real system with
the p-tolylsulfonamide units to the model system. Our re-
sults show that the 1a!2a process in the real system is exo-
thermic by ꢀ140.3 kcalmolꢀ1 at the B3LYP/cc-pVDZ-PP
level, while in our model system is found to be exothermic
by ꢀ128.4 kcalmolꢀ1. Thus, substitution of the SO2-Ar moi-
eties by H atoms reduces the exothermicity of the reaction
by approximately 10%. Although this quantity is not negli-
gible, we think that it affects similarly the different reactions
studied and, therefore, the conclusions reached by compari-
son with our model systems should be still valid for the real
systems.
Figure 1. Ortep plot (50%) of macrocycle 3a.
Given that macrocycle 3a had a high level of insolubility
we decided to prepare two macrocycles, 3b and 3c, contain-
ing units of 2,4,6-triisopropylphenyl in order to increase sol-
ubility. To introduce an arylic unit that was different to the
other three and to prepare 3b, an alternative synthetic route
was used (see Scheme 2). Condensation between dichloro
derivative 11b,[8] which introduces the differential aryl unit,
and trisulfonamide 12a[6] afforded a 96% yield of macrocy-
cle 3b. Reaction of 12b[6] with 11b[8] gave macrocycle 3c
containing four units of 2,4,6-triisopropylphenyl.
We proceed now by examining the reaction mechanism
for the 15-MAA. The energy profile of the [RhCl
ACHTUNGTRENNUNG(PH3)3]-
catalysed cyclotrimerisation of this 15-MAA is presented in
5292
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300

Reaction Mechanisms
FULL PAPER
Figure 2. The xyz cartesian coordinates of all minima and
transition states (TSs) can be found in the Supporting Infor-
mation. As can be seen, initially the [RhClACHTNUGTRNEG(UN PH3)3] complex
10.9 kcalmolꢀ1 and
a
Gibbs free-reaction energy of
ꢀ31.9 kcalmolꢀ1. These values are in line with those ob-
tained for the conversion of a bisacetylene [RuClCp] com-
plex into a ruthenacyclopenta-
triene in three different studies
of about 13 and ꢀ35 kcalmolꢀ1,
respectively.[14,15,17] On the other
hand, similar barriers but lower
exothermicities were reported
for RhCp and CoCp-catalysed
cyclotrimerisations.[10,17,18]
As
found previously in similar Co
and Rh complexes, the 15-A3
complex exhibits p localisation
ꢀ
ꢀ
with Ca Cb and Cb Cbꢀ bond
lengths of 1.351 and 1.453 ꢅ,
respectively, despite the pres-
ence of metal lone pairs that
could favour the formation of
an aromatic ring.[10,18] The Rh
C bond length of 2.04 ꢅ is not
far from the average crystallo-
ꢀ
graphic
bond
length
of
2.01 ꢅ.[18] The Rh C bond
lengths for the C atoms of the
triple bond still present in this
complex, suggest that the p-
electronic structure interacts
with the metal in an h2 fashion
in 15-A3. This complex resem-
ꢀ
Figure 2. Reaction Gibbs free-energy profile for the [2+2+2] cyclotrimerisation of our model of macrocycle
1a. Relative free-energy values in kcalmolꢀ1. Imaginary frequencies [cmꢀ1] for the different transition states
are given in brackets. For the sake of clarity nonrelevant H atoms have been omitted.
weakly interacts externally with a triple bond to form the
18-electron complex 15-A1. After that, the distorted trigonal
bypiramidal complex 15-A2 is generated by replacing two
phosphines by two internal h2 interactions with adjacent ace-
tylenic units of the 15-MAA. This process is endoergonic by
10.9 kcalmolꢀ1, although this endoergonicity is probably
overestimated as a result of the substitution of the PPh3 li-
gands in the Wilkinsonꢀs catalyst by a stronger donor such
as PH3.[29] The substitution of the phosphines by acetylene
units in the formation of 15-A2 can follow a dissociative or
associative mechanism. Although we have not analysed this
point in detail, we consider that because of the small cavity
size in the 15-MAA, the associative mechanism is very un-
likely. Next, the oxidative coupling of the two alkyne groups
in 15-A2 leads to the distorted trigonal-bypiramidal bicyclo-
rhodacyclopentadiene 15-A3. The TS corresponding to this
transformation is depicted in Figure 3.
Figure 3. Optimised structure (B3LYP/cc-pVDZ-PP) for 15-TS
with the most relevant bond lengths [ꢅ] and angles [8].
ACHTUNGTRENNUNG(A2,A3)
ꢀ
The Cb Cbꢀ distance of 1.948 ꢅ is close to that found for
the corresponding TS of the transformation of a 1,6-diyne
bles the only reported example of a metalcyclopentadiene-
AHCTUNGTRENNUNG
(alkyne) species determined by X-ray diffraction.[30] The or-
into a similar ruthenabicyclo complex (2.040 ꢅ).[15] On the
ꢀ
ꢀ
other hand, a significant shortening of the Rh Ca bond
ganometallic complex 15-A3 shows a Ca Cb bond length of
1.351 ꢅ (to be compared with 1.358 ꢅ in the reported
ꢀ
lengths is observed (from 2.247 to 2.078 ꢅ), but the Ca Cb
(alkyne) system[30]), and a Cb Cbꢀ bond
metalACTHNGURTENNUcG yclopentadieneACTHUNGTRENNUNG
length of 1.453 ꢅ, almost identical to that previously report-
ꢀ
bonds are only slightly elongated with respect to 15-A2. The
conversion of 15-A2 into 15-A3 is the rate-determining step
of the overall process with a Gibbs free-energy barrier of
ed.[30]
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5293

M. Solꢁ, A. Roglans et al.
We have not investigated the possible formation of h4-cy-
clobutadiene-like complexes as a result of a thermal cyclodi-
merisation from 15-A2, because this process was found to
be kinetically quite unfavourable in similar species.[31] The
photochemical process instead is more favourable.[30] The
16-electron 15-A3 complex adds a phosphine molecule to
yield the 18-electron species 15-A4 with a Gibbs free-energy
stabilisation of 13.1 kcalmolꢀ1. The coordination of this PH3
promotes an elongation from 2.033 to 2.076 ꢅ of the bond
length between the Rh and the Ca on the opposite side of
this PH3. The process that converts 15-A4 into 15-A5 is an
intramolecular [4+2] cycloaddition of the coordinated
alkyne of the 15-MAA to the rhodacyclopentadiene. This
reaction has a small barrier of only 3.8 kcalmolꢀ1 and is
exoergonic by 55.4 kcalmolꢀ1 in line with previous stud-
ies.[10,18] The driving force for this reaction comes from the
which is an indication of the loss of ring aromaticity as com-
pared to the free benzene ring and of the significant ring–
metal interaction (both donation from the p system of the
benzene ring to the metal and back-donation). This ring ex-
hibits a hinge angle of 308, similar to the 378 found by
BLYP/TZ2P calculations in the [Rh(h4-benzene)Cp] spe-
cies[18] or the 428 measured in the X-ray structure of [Rh
ACHTUNGTRENNUNG
{h5-
C
N
N
CHTUNGTRENNUNG
from 15-A5 into distorted tetrahedral 15-A6 involves a ring
slippage with a change of the arene-ring hapticity from h4 to
h2. The barrier for this conversion is low (2.4 kcalmolꢀ1)
and, somewhat unexpected when going from the 18- to 16-
electron species, quite exoergonic by 18.6 kcalmolꢀ1. The
thermodynamic driving force of this process is likely the par-
tial recovery of aromaticity of the arene ring, which be-
comes planar and has a small BLA. Completion of the cata-
lytic cycle occurs upon exoergonic (by 20.3 kcalmolꢀ1) dis-
placement of the arene by a phosphine molecule to regener-
ate the [RhClACHTNUGTRENUNG(PH3)3] catalyst.
Next, we examined the two possible reaction mechanisms
for the [2+2+2] cyclotrimerisation catalysed by the [RhCl-
AHCTUNGTREN(NGUN PH3)3] complex in the 20-MAA. The Gibbs free-energy
profile of this cyclotrimerisation is drawn in Figure 4.
The overall transformation is also very exoergonic
(ꢀ122.0 kcalmolꢀ1), so the reason for the lack of reactivity
of the 20-MAA must be kinetic. Although, the reaction
ꢀ
formation of two new C C s-bonds and a partially aromatic
distorted six-carbon arene ring.[10] The possible formation of
rhodacycloheptatriene or 7-rhodanorbornadiene or rhodabi-
cyclo [3.2.0]heptatriene intermediates was also investigated.
None of these intermediates were located in our potential-
energy surface for the 15-membered macrocycle. In the 18-
electron 15-A5 complex the benzene formed ring is attached
to the RhI through an h4-interaction. The coordinated buta-
diene portion of the six-membered ring displays a significant
short–long bond length alternation (BLA of 1.357/1.469 ꢅ),
Figure 4. Reaction Gibbs free-energy profile for the [2+2+2] cyclotrimerisation of our model of macrocycle 3. Relative free-energy values in kcalmolꢀ1
Imaginary frequencies [cmꢀ1] for the different transition states are given in brackets. For the sake of clarity nonrelevant H atoms have been omitted.
.
5294
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300

Reaction Mechanisms
FULL PAPER
mechanism is similar to that found for the 15-MMA, there
are three remarkable differences. First, 18-electron complex
20-A2 is generated by replacing a single (instead of two)
phosphine ligand by two internal h2 interactions with adja-
cent acetylenic units of the 20-MAA. This process is endoer-
gonic by 2.8 kcalmolꢀ1. Second, there are two possible ways
of coordinating two alkyne units in the macrocycle. Thus,
species 20-A2’, with two internal h2 interactions with non-
adjacent acetylenic units of the 20-MAA, can also be
formed but it is less stable than 20-A2 by 12.8 kcalmolꢀ1.
The conversion of 20-A2 and 20-A2’ into bicyclorhodacyclo-
pentadiene 20-A3 and 20-A3’, respectively, takes place by
oxidative coupling of the two coordinated alkyne moieties
through the TSs depicted in Figure 5 with Gibbs free-energy
barriers of 30.7 and 15.3 kcalmolꢀ1, respectively. These
values are clearly larger than those found for the same rate-
determining step in the 15-MAA. The reasons for these
higher energy barriers, which are in agreement with the ex-
perimental lack of formation of 13, will be discussed later.
Although the barrier for the transformation of 20-A2’ into
20-A3’ is lower than that for the equivalent 20-A2 into 20-
A3 conversion, the two TSs differ by only 2.6 kcalmolꢀ1 and
we decided to continue the analysis of the reaction mecha-
nism only from the 20-A3 species. The rearrangement con-
verting 20-A3 into the distorted octahedral 20-A4 complex
requires 3.6 kcalmolꢀ1 and attaches one uncoordinated
triple bond of the 20-MAA to the metal. Since no transition
state connecting 20-A3 to 20-A4 was found, the barrier for
this step is expected to be very low or absent. At this stage
the third important difference arises. The process that trans-
forms 20-A4 into 20-A5 is not a [4+2] cycloaddition, but
rather a [2+2] asymmetric addition that results in the forma-
tion
a non-planar rhodacycloheptatriene complex (see
Figure 6) through the TS depicted in Figure 7. This process
has a low barrier (4.7 kcalmolꢀ1) and is exoergonic by
20.3 kcalmolꢀ1. All attempts to find the TS corresponding to
the [4+2] cycloaddition were unsuccessful. Thus, even small
changes in the macrocycle can lead to important differences
in the reaction mechanism of the [2+2+2] intramolecular
cyclotrimerisation. A seven-membered ruthenacycle was
also found in the reaction mechanism of the [2+2+2] cyclo-
trimerisation of alkynes.[15] Experimentally, the crystal struc-
ture of an iridacycloheptatriene with a tub-shaped confor-
mation has been reported.[33] Our 20-A5 rhodacyclohepta-
Figure 6. Optimised structure (B3LYP/cc-pVDZ-PP) for 20-A5 with the
most relevant bond lengths [ꢅ] and angles [8].
Figure 5. Optimised structures (B3LYP/cc-pVDZ-PP) for 20-TS
ACHTUNGTRNE(NUNG A2,A3)
(top) and 20-TS(A2’,A3’) (bottom) with the most relevant bond lengths
N
Figure 7. Optimised structure (B3LYP/cc-pVDZ-PP) for 20-TSACHTUNGTRENNUNG(A4,A5)
[ꢅ] and angles [8].
with the most relevant bond lengths [ꢅ] and angles [8].
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5295

M. Solꢁ, A. Roglans et al.
triene complex has a tub-shaped conformation with a short–
long BLA of 1.347/1.475 ꢅ similar to that found in the irida-
cycloheptatriene of 1.344/1.484 ꢅ.[33] On the other hand, the
bond length between the two carbons atoms not linked is
2.648 ꢅ to be compared with the 2.885 ꢅ found in the irida-
cycloheptatriene.[33] Final ring closure in 20-A5 provides 20-
A6 in which the arene ring is coordinated in an h2 fashion.
The barrier for this reaction is only 1.3 kcalmolꢀ1 and the
exoergonicity is as large as 50.1 kcalmolꢀ1. Finally, the cata-
lytic cycle is closed upon exoergonic displacement (by
20.7 kcalmolꢀ1) of the arene by a phosphine molecule to re-
by 22.8 kcalmolꢀ1. It is likely that the greater stability of 14
relative to 15 is a result of the triple bonds in ten-membered
rings being particularly strained. Indeed the ]CCC angle in-
cluding the two carbon atoms linked by a triple bond is far
from linear (around of 1608 as compared to 1798 in the 15-
membered ring of 14). This is in line with the fact that the
exoergonicity of the trimerisation substantially decreases
when going from the product of the 15-MAA (no ten-mem-
bered rings with triple bonds) to that of the 20-MAA (one
ten-membered ring with triple bonds) to 15 (two ten-mem-
bered rings with triple bonds). Despite this, the thermody-
namics alone cannot explain the formation of only 14. As
can be seen in Figure 8, 18-electron complexes 25-A2 and
25-A2’ are formed by replacing a single phosphine ligand by
two internal h2 interactions either with adjacent or non-adja-
cent acetylenic units of the 25-MAA. For the non-adjacent
coordinated triple bonds species, the process is endoergonic
by 13.5 kcalmolꢀ1. The subsequent oxidative coupling takes
place through a very high Gibbs free-energy barrier of
52.9 kcalmolꢀ1. Therefore, this reaction pathway is inaccessi-
ble at the temperature of the reaction. Instead, coordination
of two adjacent triple bonds requires only 2.3 kcalmolꢀ1 and
the barrier for the oxidative coupling to yield the rhodacy-
clopentadiene 25-A3 is low (of only 6.4 kcalmolꢀ1). The TSs
for the transformation of 25-A2 into 25-A3 and 25-A2’ into
25-A3’ are depicted in Figure 9. We attribute the large barri-
generate the [RhClACHTUNGTRENNUNG(PH3)3] catalyst.
We shall now examine the reaction profile of the 25-
MAA depicted in Figure 8. Due to the size of the 25-MAA,
these calculations are extremely computationally demand-
ing. For this reason, we only located the TSs corresponding
to the first steps of the reaction mechanisms. As we have
seen for the 15- and 20-MAA, the barriers involved in the
last steps of the reaction mechanism are low and are expect-
ed to have little influence on the overall kinetics of the reac-
tion. As shown in Scheme 4, there are two possible products
(14 and 15) of the cyclotrimerisation of the 25-MAA. From
a thermodynamic point of view, the reaction is favoured for
the two products with exoergonicities of 121.6 and 98.8 kcal
molꢀ1 for 14 and 15, respectively. It is interesting, however,
that the product experimentally observed is the most stable
Figure 8. Reaction Gibbs free-energy profile for the [2+2+2] cyclotrimerisation of our model of macrocycle 4. Relative free-energy values in kcalmolꢀ1
Imaginary frequencies [cmꢀ1] for the different transition states are given in brackets. For the sake of clarity nonrelevant H atoms have been omitted.
.
5296
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300

Reaction Mechanisms
FULL PAPER
Figure 9. Optimised structure (B3LYP/cc-pVDZ-PP) for 25-TSACHTNUTGRNEUNG(A2,A3)
Figure 10. Optimised structure (B3LYP/cc-pVDZ-PP) for 25-A5 (top)
and 25-A5’ (bottom) with the most relevant bond lengths [ꢅ] and angles
[8].
(top) and 25-TS
ACHTUNGTRENNUNG(A2’,A3’) (bottom) with the most relevant bond lengths
[ꢅ] and angles [8].
er found in the conversion of 25-A2’ into 25-A3’ in part to
the deformation of the 20-MAA required in order to form
two strained ten-membered rings. The deformation energy is
28.6 kcalmolꢀ1 higher in the TS for the 25-A2’ into 25-A3’
than for the 25-A2 into 25-A3 transformation. Again in 25-
A3 there are two possibilities for the coordination of a
triple bond. The triple bond can be adjacent to those in-
volved in the oxidative coupling that lead to the rhodacyclo-
pentadiene 25-A3 or non-adjacent. In the former process,
the complex 25-A4 is formed with a stabilisation of 18.2 kcal
molꢀ1, while in the latter the formation of the product 25-
A4’ releases only 3.5 kcalmolꢀ1. In addition, the subsequent
[2+2] cycloaddition to form the rhodacycloheptatriene com-
plexes 25-A5 and 25-A5’ depicted in Figure 10 through the
corresponding TSs (see Figure 11) is favoured in the case of
the adjacent triple bond (Gibbs free-energy barrier of
3.5 kcalmolꢀ1 with respect to 10.1 kcalmolꢀ1 for the non-ad-
jacent triple bond). Therefore, the reaction pathway involv-
ing addition of the three contiguous triple bonds is favoured,
which explains the experimental formation of 14 and not of
15. Final ring closure in 25-A5 provides 25-A6 in which the
arene ring is coordinated in an h4 fashion. Conversion of 25-
A6 into 25-A7 occurs after ring slippage and final displace-
ment of the arene by a phosphine molecule regenerates the
[RhClACHTNUGTRNEUGN(PH3)3] catalyst.
As reported in previous studies,[10–18] we have also found
that the rate-determining step in the [2+2+2] cyclotrimeri-
sation of ethynes is the initial oxidative coupling to yield the
metallacyclopentadiene or metallacyclopentatriene inter-
mediate. The Gibbs free-energy barriers obtained for this
process in the 15-, 20- and 25-MAA are calculated to be
21.8, 30.9 and 8.7 kcalmolꢀ1 with respect to separated reac-
tants. The values obtained are in line with the experimental
observations. To investigate the origin of the barriers, we di-
vided the energy difference between the TS and the separat-
ed reactants into deformation energy and interaction energy
(DEdef+DEint). The deformation energy (DEdef) is the energy
needed to modify the geometry of the free reactants to
attain the geometry they have in the TS. The interaction
energy (DEint) is the energy released when the two free de-
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5297

M. Solꢁ, A. Roglans et al.
Figure 12. 3D-representation of the HOMOs for the 20- (left) and 25-
MAA (right) at the geometry the macrocycles have in the 20- and 25-TS-
(A2,A3). Isosurface values are ꢁ0.192 and ꢁ0.172 au, respectively.
1) a more stable and delocalised HOMO orbital and 2) the
formation of a strained ten-membered ring.
Conclusions
In summary, we report an efficient stepwise preparation of
20- and 25-membered macrocycles featuring four and five
triple bonds, respectively, with different arylsulfonyl moiet-
ies in their structure. All new compounds are completely
characterised by spectroscopic methods and additional evi-
dence for structure was secured by X-ray diffraction for 3a.
An efficient rhodium(I)-catalysed [2+2+2] cyclotrimerisa-
tion of 25-membered pentaacetylenic azamacrocycle 4 che-
moselectively afforded the cyclotrimerised compound 14 re-
sulting from the reaction of three adjacent alkynes instead
of the cyclotrimerisation between non-adjacent triple bonds.
In contrast, the 20-membered tetraacetylenic azamacrocy-
cles 3 did not lead to the expected cyclotrimerised com-
pounds. DFT calculations to unravel the reaction mecha-
nism of the 15-, 20- and 25-MAA revealed that there are
two main factors that contribute to the lack of reactivity of
the 20-MAA: 1) the 20-MAA has a more stable and delo-
calised HOMO orbital and 2) the formation of a strained
ten-membered ring during the cyclotrimerisation of the 20-
MAA. These two factors increase the free-energy barriers
of the rate-determining step and difficult the intramolecular
cyclotrimerisation of the 20-MAA.
Figure 11. Optimised structure (B3LYP/cc-pVDZ-pp) for 25-TS
(top) and 25-TS(A4’,A5’) (bottom) with the most relevant bond lengths
[ꢅ] and angles [8].
ACHTUNGTRNE(NUNG A4,A5)
ACHTUNGTRENNUNG
formed reactants are brought to the position that they have
in the TS. The calculated deformation energies for the 20-
and 25-TS
ly. We have not included the 15-TSACHTNUGTRNE(NUG A2,A3) in the analysis
ACHTUNGTRENNUNG
(A2,A3) are 95.9 and 104.4 kcalmolꢀ1, respective-
as the different number of phosphine ligands hamper the
comparison. Clearly, the deformation energy in the TS is
not the origin of the larger barrier for the 20-TS
The interaction energies are in turn ꢀ62.4 and ꢀ95.7 kcal
molꢀ1 for the 20- and 25-TS
(A2,A3), respectively. The main
interaction in TS(A2,A3) occurs between the LUMO of the
ACHTUNGTRENN(UNG A2,A3).
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
catalyst and the HOMO of the macrocycle. By comparing
the HOMO and LUMO energies involved in the 20- and 25-
Experimental Section
TSACHTUNGTRENNUNG(A2,A3), it can be seen that the main difference is the
AHCTUNGTRENNUNG
energy of the HOMO of the 20-MAA (ꢀ0.192 au) as com-
pared to that of the 25-MAA (ꢀ0.172 au.). The HOMO–
LUMO overlaps are quite similar in the two cases. A picture
of the two HOMOs is given in Figure 12, showing a higher
delocalisation of the HOMO in the deformed 20-MAA,
which apparently is responsible for its greater stabilisation
and, as a consequence, its lower reactivity. So, we concluded
that the two factors that make the 20-MAA less reactive are
AHCTUNGTRENNUNG
heated under reflux for 30 h (TLC monitoring). The solvent was then
evaporated and the residue was purified by column chromatography on
silica gel with dichloromethane/ethyl acetate (20:1) to afford 14 (0.025 g,
50%) as a colourless solid. M.p. 194–1968C; 1H NMR (500 MHz, CDCl3,
258C, TMS): d=2.40 (s, 12H), 2.45 (s, 6H), 3.59 (s, 4H), 3.72 (s, 4H),
4.14 (s, 4H), 4.33 (s, 4H), 4.48 (s, 4H), 7.15 (AA’BB’ system, J=8.0 Hz,
4H), 7.35–7.38 (m, 10H), 7.63 (AA’BB’ system, J=8.2 Hz, 2H),
5298
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300

Reaction Mechanisms
FULL PAPER
7.78 ppm (AA’BB’ system, J=8.2 Hz, 4H); 13C NMR (50 MHz, CDCl3,
258C, TMS): d=22.2, 22.3, 37.6, 37.7, 45.8, 52.5, 53.8, 80.4, 80.6, 128.2,
128.4, 129.6, 130.1, 130.4, 130.6, 132.2, 134.0, 135.1, 136.2, 138.6, 144.7,
145.0, 145.3 ppm; IR n˜ =2923, 1335, 1157 cmꢀ1; MS (MALDI-TOF: m/z:
1106 [M+H]+, 1128 [M+Na]+, 1144 [M+K]+; elemental analysis calcd
(%) for C55H55N5O10S5.EtOAc (1194.495): C 59.33, H 5.32, N, 5.86;
found: C 59.14 and 59.10, H 5.21 and 5.26, N 6.18 and 6.21.
[6] A. Torrent, I. Gonzꢇlez, A. Pla-Quintana, A. Roglans, M. Moreno-
MaÇas, T. Parella, J. Benet-Buchholz, J. Org. Chem. 2005, 70, 2033–
2041.
[7] I. Gonzꢇlez, S. Bouquillon, A. Roglans, J. Muzart, Tetrahedron Lett.
2007, 48, 6425–6428.
[8] S. Brun, L. Garcia, I. Gonzꢇlez, A. Torrent, A. Dachs, A. Pla-Quin-
tana, T. Parella, A. Roglans, Chem. Commun. 2008, 4339–4341.
[9] C. Bianchini, K. G. Caulton, C. Chardon, M.-L. Doublet, O. Eisen-
stein, S. A. Jackson, T. J. Johnson, A. Meli, M. Peruzzini, W. E.
Streib, A. Vacca, F. Vizza, Organometallics 1994, 13, 2010–2023.
[10] J. H. Hardesty, J. B. Koerner, T. A. Albright, G.-Y. Lee, J. Am.
Chem. Soc. 1999, 121, 6055–6067.
X-ray data for 3a: CCDC-707902 (3a) contains the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Computational methods: All geometry optimisations were performed by
using the hybrid DFT B3LYP[34–36] method with the Gaussian 03[37] pro-
gram package. The geometry optimisations were performed without sym-
metry constraints. Analytical Hessians were computed to determine the
nature of stationary points (one or zero imaginary frequencies for transi-
tion 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.[38]
These last two terms were computed at 298.15 K and 1 atm to provide
the reported relative Gibbs free energies (DG298). Furthermore, the con-
nectivity between stationary points was established by intrinsic reaction
path[39] calculations. The all-electron cc-pVDZ basis set was used for P,
O, N, C and H atoms,[40,41] while for Rh we employed the cc-pVDZ-PP
basis set,[42] containing an effective core relativistic pseudopotential. Rel-
ative energies were computed taking into account the total number of
molecules present. The SO2-Ar moieties present in the experimental 15-,
20- and 25-membered acetylenic azamacrocycles and the Ph group in the
catalyst were substituted by H atoms to reduce the computational com-
plexity of the calculations involving these ligands. Substitution of PPh3 by
[11] A. A. Dahy, N. Koga, Bull. Chem. Soc. Jpn. 2005, 78, 781–791.
[12] A. A. Dahy, C. H. Suresh, N. Koga, Bull. Chem. Soc. Jpn. 2005, 78,
792–803.
[13] N. Agenet, V. Gandon, K. P. C. Vollhardt, M. Malacria, C. Aubert, J.
Am. Chem. Soc. 2007, 129, 8860–8871.
[14] K. Kirchner, M. J. Calhorda, R. Schmid, L. F. Veiros, J. Am. Chem.
Soc. 2003, 125, 11721–11729.
[15] Y. Yamamoto, T. Arakawa, R. Ogawa, K. Itoh, J. Am. Chem. Soc.
2003, 125, 12143–12160.
[16] R. Schmid, K. Kirchner, Eur. J. Inorg. Chem. 2004, 2609–2626.
[17] G. Dazinger, M. Torres-Rodrigues, K. Kirchner, M. J. Calhorda, P. J.
Costa, J. Organomet. Chem. 2006, 691, 4434–4445.
[18] L. Orian, J. N. P. van Stralen, F. M. Bickelhaupt, Organometallics
2007, 26, 3816–3830.
[19] N. E. Schore, Chem. Rev. 1988, 88, 1081–1119.
[20] Y. Yamamoto, K. Kinpara, T. Saigoku, H. Takagishi, S. Okuda, H.
Nishiyama, K. Itoh, J. Am. Chem. Soc. 2005, 127, 605–613.
[21] Y. Yamamoto, R. Ogawa, K. Itoh, J. Am. Chem. Soc. 2001, 123,
6189–6190.
PH3 is
a common procedure in theoretical organometallic chemis-
try.[10,31,43–47] In addition, we have checked that, despite the electronic and
steric differences, substitution of PPh3 by PH3 does not introduce signifi-
cant changes in the thermodynamics and kinetics of the cyclotrimerisa-
tion of three acetylene molecules.[48] 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 intermediates
and transition states in the reaction mechanism.[18] Because the reactions
studied are carried out in toluene, solvent effects have not been included
in the present calculations. Finally, since there is no experimental data
suggesting the presence of paramagnetic intermediates, our studies were
limited to the singlet-spin potential-energy surfaces. However, we did cal-
culations for the triplet form of several intermediates and they were
found always higher in energy than the singlet counterparts.
[22] J. A. Varela, S. G. Rubꢂn, L. Castedo, C. Saꢇ, J. Org. Chem. 2008, 73,
1320–1332.
[23] V. Gandon, N. Agenet, K. P. C. Vollhardt, M. Malacria, C. Aubert, J.
Am. Chem. Soc. 2006, 128, 8509–8520.
[24] M. M. Montero-Campillo, J. Rodrꢂguez-Otero, E. Cabaleiro-Lago, J.
Phys. Chem. A 2008, 112, 2423–2427.
[25] J. Rodrꢂguez-Otero, M. M. Montero-Campillo, E. Cabaleiro-Lago, J.
Phys. Chem. A 2008, 112, 8116–8120.
[26] A. Ioffe, S. Shaik, J. Chem. Soc. Perkin Trans. 2 1992, 2101–2108.
[27] R. D. Bach, G. J. Wolber, H. B. Schlegel, J. Am. Chem. Soc. 1985,
107, 2837–2841.
[28] B. R. Neustadt, Tetrahedron Lett. 1994, 35, 379–380.
[29] Indeed, at the same level of theory, substitution of two phosphine li-
gands by acetylene molecules in [RhClACTHUNGTRENNNUG
about 12 kcalmolꢀ1, while in [RhCl
ACHTUNGTRENNUNG
only 0.3 kcalmolꢀ1 (A. Dachs, S. Osuna, A. Roglans, M. Solꢁ, un-
published results).
Acknowledgements
[30] P. I. Dosa, G. D. Whitener, K. P. C. Vollhardt, A. D. Bond, S. J. Teat,
Org. Lett. 2002, 4, 2075–2078.
[31] L. F. Veiros, G. Dazinger, K. Kirchner, M. J. Calhorda, R. Schmid,
Chem. Eur. J. 2004, 10, 5860–5870.
We would like to thank the Spanish MEC (projects CTQ2005-04968-C02-
02/BQU;
CTQ2005-08797-C02-01/BQU;
CTQ2008-03077/BQU;
CTQ2008-05409-C02-02/BQU; CTQ2006-01080) and the Catalan DURSI
(projects 2005SGR-00238; 2005SGR-00305) for their financial support.
A.D. and S.O. thank the Spanish MEC for a doctoral fellowship. We also
acknowledge the Centre de Supercomputaciꢆ de Catalunya (CESCA) for
partial funding of computer time. We thank the Research Technical Serv-
ices of the UdG for spectral data, especially Xavier Fontrodona for X-
ray diffraction analysis.
[32] W. J. Bowyer, J. W. Merkert, W. E. Geiger, A. L. Rheingold, Orga-
nometallics 1989, 8, 191–198.
[33] E. ꢈlvarez, M. Gꢆmez, M. Paneque, C. M. Posadas, M. L. Poveda,
N. Rendꢆn, L. L. Santos, S. Rojas-Lima, V. Salazar, K. Mereiter, C.
Ruiz, J. Am. Chem. Soc. 2003, 125, 1478–1479.
[34] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[35] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[36] P. J. Stephens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys.
Chem. 1994, 98, 11623–11627.
[37] Gaussian 03, Revision C.01, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
[1] P. R. Chopade, J. Louie, Adv. Synth. Catal. 2006, 348, 2307–2327.
[2] V. Gandon, C. Aubert, M. Malacria, Chem. Commun. 2006, 2209–
2217.
AHCTUNGTREGeNNUN ry, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
[3] Y. Yamamoto, Curr. Org. Chem. 2005, 9, 503–519.
[4] S. Kotha, E. Brahmachary, K. Lahiri, Eur. J. Org. Chem. 2005, 4741–
4767.
[5] A. Pla-Quintana, A. Roglans, A. Torrent, M. Moreno-MaÇas, J.
Benet-Buchholz, Organometallics 2004, 23, 2762–2767.
Chem. Eur. J. 2009, 15, 5289 – 5300
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5299

M. Solꢁ, A. Roglans et al.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[43] Q. Cui, D. G. Musaev, K. Morokuma, Organometallics 1997, 16,
1355–1364.
[44] Q. Cui, D. G. Musaev, K. Morokuma, Organometallics 1998, 17,
742–751.
[45] Q. Cui, D. G. Musaev, K. Morokuma, Organometallics 1998, 17,
1383–1392.
[46] W. Zheng, A. Ariafard, Z. Lin, Organometallics 2008, 27, 246–253.
[47] Y. Abe, K. Kuramoto, M. Ehara, H. Nakatsuji, M. Suginome, M.
Murakami, Y. Ito, Organometallics 2008, 27, 1736–1742.
[48] For the cyclotrimerisation of three acetylene molecules catalysed by
[38] P. Atkins, J. De Paula in Physical Chemistry, Oxford University
Press, Oxford, 2006.
[RhClACHNUTGTRENNUNG
H and Ph3, while the energy barrier for the rate-determining step is
11.9 and 12.9 kcalmolꢀ1 for R=H and Ph3, respectively (A. Dachs,
S. Osuna, A. Roglans, M. Solꢁ, unpublished results).
[39] C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154–2161.
[40] T. H. Dunning, Jr., J. Chem. Phys. 1989, 90, 1007–1023.
[41] D. E. Woon, T. H. Dunning, Jr., J. Chem. Phys. 1993, 98, 1358–1371.
[42] K. A. Peterson, D. Figgen, M. Dolg, H. Stoll, J. Chem. Phys. 2007,
126, 124101.
Received: December 5, 2008
Published online: April 22, 2009
5300
www.chemeurj.org
ꢄ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 5289 – 5300