Regiochemistry in the Pauson-Khand reaction: Has a trans effect been overlooked?

Article abstract of DOI:10.1021/ja004170n

Alkyne-dicobalt hexacarbonyl complexes have been studied by DFT to examine whether electronic differences in the acetylenic substituents could play a role in determining the regiochemical outcome in the Pauson-Khand reaction. It appears that in some insta

Full text of DOI:10.1021/ja004170n

5396
J. Am. Chem. Soc. 2001, 123, 5396-5400
Regiochemistry in the Pauson-Khand Reaction: Has a Trans Effect
Been Overlooked?
Fr e´ d e´ ric Robert, Anne Milet, Yves Gimbert,* Den e´ s Konya, and Andrew E. Greene*
Contribution from the UniVersit e´ Joseph Fourier de Grenoble, Chimie Recherche (LEDSS),
3
8041 Grenoble, France
ReceiVed December 4, 2000
Abstract: Alkyne-dicobalt hexacarbonyl complexes have been studied by DFT to examine whether electronic
differences in the acetylenic substituents could play a role in determining the regiochemical outcome in the
Pauson-Khand reaction. It appears that in some instances the regiochemistry is, in fact, governed exclusively
by the electronic nature of the acetylenic substituents through the discriminant loss of a carbon monoxide
ligand.
Introduction
Scheme 1. Supposed Mechanism of the PK Reaction
The Pauson-Khand (PK) reaction, thanks in part to gains in
reactivity recently achieved in both the stoichiometric and
catalytic versions, is now a useful transformation that is widely
1
employed for the synthesis of complex organic compounds.
Although this reaction has been known for some 30 years, a
full, supported mechanistic elucidation has, surprisingly, yet to
2
be provided. Nonetheless, it is now largely accepted that the
alkyne-Co2(CO)6 complex, formed from an alkyne and
3
Co2(CO)8, initially suffers the loss of a CO ligand, and that
this is followed by olefin coordination at one of the two
enantiotopic Co atoms (Scheme 1). It has long been supposed,
but with only empirical foundation, that in the intermolecular
PK reaction the coordinating olefin for steric reasons positions
itself adjacent to the less bulky acetylenic substituent and that
the subsequent C-C bond formation occurs between an olefinic
carbon and this closer acetylenic carbon. Indeed, but with few
exceptions, the larger acetylenic substituent is found, following
insertion of CO and reductive elimination, at the R position in
the resulting cyclopentenone (Scheme 2).
Scheme 2. The Regiochemical Outcome of the PK Reaction
of Diversely Substituted Acetylenes and Norbornene
There are some results, though, that are difficult to explain
solely on the basis of steric effects: for example, and perhaps
4
most saliently, ethyl propiolate under PK conditions yields
mainly an R-carbethoxy-substituted cyclopentenone, whereas
5
ethyl butynoate under similar conditions reacts to form only a
â-carbethoxy-substituted derivative (Scheme 2). To rationalize
this, it has been suggested polarization effects that result from
electron density differences at the acetylenic carbons can
combine with steric effects to determine the regiochemistry of
the reaction. However, the manner in which these postulated
electronic effects might influence the resulting regiochemistry
of the PK reaction has nowhere been clearly indicated. Further-
more and more fundamentally, the occurrence of a specific
vacant site in the alkyne-Co2(CO)6 complex on loss of carbon
monoxide with attendant effect on the regiochemistry of the
PK reaction is a possibility that has not to date been addressed.
That electron effects engendered by different acetylenic sub-
stituents could in some way influence the position from which
carbon monoxide is lost and ultimately the regiochemistry of
the PK reaction seemed to be a possible unifying concept worthy
of study. A theoretical study has been carried out to probe this
idea.
*
Address correspondence to this author. Phone: (33) 4-76-51-46-91.
Fax: (33) 4-76-51-44-94. E-mail: Yves.Gimbert (or Andrew.Greene)@
ujf-grenoble.fr.
(1) For recent reviews, see: Geis, O.; Schmalz, H. G. Angew. Chem.,
Int. Ed. 1998, 37, 911-914. Jeong, N. Trans. Met. Org. Synth. 1998, 1,
5
60-577. Chung, Y. K. Coord. Chem. ReV. 1999, 188, 297-341.
2) Magnus, P.; Exon, C.; Albough-Robertson, P. Tetrahedron 1985, 41,
861-5869.
3) Gordon, G. M.; Kiszka, M.; Dunkin, I. R.; Kerr, W. J.; Scott, J. S.;
(
5
(
Gebicki, J. J. Organomet. Chem. 1988, 554, 147-154.
(
4) Krafft, M. E.; Romero, R. H.; Scott, I. L. Synlett 1995, 677-678.
(5) Krafft, M. E.; Romero, R. H.; Scott, I. L. J. Org. Chem. 1992, 57,
5
277-5278. For other reactions with 2-alkynoates, see: Fonquerna, S.;
Moyano, A.; Peric a` s, M. A.; Riera, A. Tetrahedron 1995, 51, 4239-4254.
1
0.1021/ja004170n CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/16/2001

Regiochemistry in the Pauson-Khand Reaction
J. Am. Chem. Soc., Vol. 123, No. 23, 2001 5397
Results and Discussion
A number of published spectroscopic studies have indicated
that the nature of the acetylenic substituents can markedly affect
electronic properties within the alkyne-Co2(CO)6 complex.
Carbon-13 NMR analysis of Co2(CO)6(RCtCR′) complexes
shows that the nature of the R and R′ groups, as before
coordination, but generally more so in the complex, impacts
6
on the chemical shifts of the acetylenic carbons. Assuming that
Figure 1. Lability of the carbon monoxide ligands (the cis and trans
positions are defined with respect to the position of substituent B).
the observed acetylenic carbon-13 chemical shifts reflect
7
electron density, then it is apparent that dissymmetry in the
8
acetylene induces an electronic distortion or polarization.
we belieVe to be responsable for a discriminant loss of CO. A
relative accumulation of density on one of these carbon atoms
should lead to discharge, via the metal atoms, to the strongly
π-accepting CO ligands, this back-donation serving to strengthen
the Co-CO bonds. Orbital symmetry considerations, though,
lead to the conclusion that the 6 carbon monoxide ligands will
not participate equally in this redistribution of electron density.
Carbon-13 NMR analysis of the CO carbons in the Co2(CO)6-
(RCtCR′) complexes indicates that these too are affected by
the nature of the R and R′ groups; however, the chemical shift
variations are smaller than those above. An increase in the
electron donating abilitity of the substituents results in downfield
shifts of the CO resonances. NMR, while not as useful as
9
-
infrared spectroscopy for gaging the effect of the R and R′
The pseudoequatorial, trans-positioned (with respect to C(δ ))
groups on the electronic nature of the Co-CO association,
carbon monoxide ligands should be the most receptive to back-
nonetheless clearly demonstrates a qualitative dependence of
donation, which would then make the pseudoequatorial, cis-
1
0
14
the δCO on the electronic properties of these groups.
The infrared spectra of Co2(CO)6(RCtCR′) complexes show
positioned ligands labile in comparison (Figure 1). The olefin
should then occupy the position of a labile CO and, conse-
quently, the A group should be found at the R position in the
resulting cyclopentenone. A Density Functional Theory (DFT)
-
1
5
or 6 absorptions in the carbonyl region (2000-2100 cm ),
1
1
which shift as a function of the R and R′ substituents, an
increase in the vibrational frequence υCO being observed with
an increase in the electronegativity of R and R′. Electronegative
substituents on the acetylene could be expected to strengthen
the acceptor properties of the bridging ligand, which would
result in reduced electron back-donation from the metal into
the π* orbitals of the carbonyls and produce the observed
shift.1
15
study has now been carried out to examine this novel proposal
that differences in the electron density at the acetylenic carbons
can translate into regiochemical preferences in the PK reaction.
In this study, all geometries were fully optimized through
analytic gradient calculations using the hybrid Hartree-Fock
16
DFT approach B3LYP. The basis sets used in these optimiza-
tions were as follows: the 10 innermost core electrons of the
cobalt atom were described by the non-relativistic pseudopo-
2,13
The spectroscopic findings are supportive of the possibility
that the acetylenic substituents may cause through transmitted
electronic effects the departure of a unique carbon monoxide
from the complex, and that this departure in turn governs the
regiochemical outcome of the PK reaction. The carbon-carbon
triple bond in a dissymmetric alkyne is most certainly polarized
due to the dissimilar electronic effects exerted on the acetylenic
carbons by the different substituents, and on coordination the
polarization appears to become more substantial. This difference
in electron density on the two acetylenic carbons in the complex
1
7
tential of Hay and Wadt and the remaining outer core and
valence electrons by a (8,5,5) 〈3,3,2〉 basis set. For the first row
atoms and hydrogen atoms the split valence basis sets of
1
8
Dunning were used. A d polarization function was added to
the first row atoms (RdO ) 0.85, RdC ) 0.75). Population
19
analysis was computed through the NBO scheme. All calcula-
(14) Cis-labilization has precedent in hexacoordinated, trigonal bipyramid
metal complexes. See: Atwood, J. D.; Brown, T. L. J. Am. Chem. Soc.
1976, 98, 3160-3166.
(15) To the best of our knowledge, no such theoretical studies have been
(
6) (a) Aime, S.; Milone, L.; Rossetti, R.; Stanghellini, L. Inorg. Chim.
performed to date on PK complexes; those that have been reported were
carried out at the semiempirical (PM3) level, followed at times by single
point calculations at the DFT level to improve the accuracy of the energy
differences among the various structures. See: Castro, J.; Moyano, A.;
Peric a` s, M. A.; Riera, A.; Alvarez-Larena, A.; Piniella, J. F. J. Am. Chem.
Soc. 2000, 122, 7944-7952. Fonquerna, S.; Rios, R.; Moyano, A.; Peric a` s,
M. A.; Riera, A. Eur. J. Org. Chem. 1999, 64, 3459-3478. Breczinski, P.
M.; Stumpf, A.; Hope, H.; Krafft, M. E.; Casalnuovo, J. A.; Schore, N. E.
Tetrahedron 1999, 55, 6797-6813. Verdager, X.; V a´ zquez, J.; Fuster, G.;
Bernardes-G e´ nisson, V.; Greene, A. E.; Moyano, A.; Peric a` s, M. A.; Riera,
A. J. Org. Chem. 1998, 63, 7037-7052. Recently, a DFT optimization of
complex 1a has appeared: Balsells, J.; V a´ zquez, J.; Moyano, A.; Peric a` s,
M. A.; Riera, A. J. Org. Chem. 2000, 65, 7291-7302. (Note added in
proof: A DFT study of the PK reaction pathway has just been published:
Yamanaka, M.; Nakamura, E. J. Am. Chem. Soc. 2001, 123, 1703-1708.)
(16) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789. This B3LYP functional
has previously been found to be of a suitable theoretical level for the study
of the binding energy of ligands, especially between carbon monoxide and
transition metals. Becke, A. D. J. Chem. Phys. 1992, 97, 9173-9177.
Bomberg, M. R. A.; Siegbahn, P. E. M.; Svensson, M. J. J. Chem. Phys.
1996, 104, 9546-9554. Ricca, A.; Bauschliber, C. W., Jr. J. Phys. Chem.
A 1997, 101, 8949-8955. Siegbahn, P. E. M.; Bomberg, M. R. A. Chem.
ReV. 2000, 100, 421-437.
Acta 1977, 22, 135-139. (b) Happ, B.; Bartik, T.; Zucchi, C.; Rossi, M.-
C.; Ghelfi, F.; Palyi, G.; Varadi, G.; Szalontai, G.; Horvath, I. T.; Chiesi-
Villa, A.; Guastini, C. Organometallics 1995, 14, 809-819. (c) Baxter, R.
J.; Knox, G. R.; McLaughlin, M.; Pauson, P. L.; Spicer, M. D. J. Organomet.
Chem. 1999, 579, 83-89.
(7) Although the C-13 chemical shifts are influenced by several factors,
comparison within a family is thought to allow a qualitative evaluation of
6
b
this property.
8) The interactions between R and R′ and C2Co2(CO)6 should be
governed by not only the donor/acceptor character of these groups but also
(
6
b
their orbital symmetry.
9) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
Wiley: New York, 1988.
(
(
10) Bodner, G. M.; Todd, L. J. Inorg. Chem. 1974, 13, 1335-1354.
(
11) Dickson, R. S.; Yawney, D. B. W. Aust. J. Chem. 1968, 21, 1077-
1
1
082. Bor, G.; Kettle, S. F. A.; Stanghellini, L. Inorg. Chim. Acta 1976,
8, L18-L20.
(12) For a similar analysis with bisphosphinoamine ligands, see: Gimbert,
Y.; Robert, F.; Durif, A.; Averbuch, M.-T.; Kann, N.; Greene, A. E. J.
Org. Chem. 1999, 64, 3492-3497.
(13) Co2(CO)6(RCtCR′) complexes can undergo alkyne exchange, which
is also influenced by the substituents and supports this interpretation. The
alkynes bearing the more electronegative R and R′ groups are preferentially
incorporated, which can be explained by an increase in the alkyne-cobalt
bond strength through enhanced electron back-donation from the d orbitals
of the cobalt into the π* orbitals of the preferred alkyne. See: Cetini, G.;
Gambino, O.; Rosseti, R.; Sappa, E. J. Organomet. Chem. 1967, 8, 149-
(17) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270-283.
(18) Dunning, T. H., Jr.; Hay, P. J. Method of Electronic Structure
Theory; Plenum Press: New York, 1977.
(19) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88,
899.
1
54.

5398 J. Am. Chem. Soc., Vol. 123, No. 23, 2001
Robert et al.
Table 1. Selected Bond Lengths (in Å) and Angles (in deg) in the
Optimized Complexes 1a-c and X-ray-Derived Values for the
2
2b
Di-tert-butylacetylene-Dicobalt Hexacarbonyl Complex
entry
X-Y
1a
1b
1c
X-ray22b
Figure 2. Complexes 1a-c.
1
2
3
4
5
6
7
8
9
Co1-Co2
Co1-C1ax
Co1-C2eq
Co1-C3eq
Co2-C6ax
Co2-C4eq
Co2-C5eq
2.468
1.780
1.818
1.827
1.780
1.818
1.827
1.342
1.965
1.979
1.965
1.979
2.470
1.796
1.828
1.823
1.790
1.827
1.824
1.355
1.944
1.956
1.955
1.949
2.465
1.792
1.832
1.815
1.783
1.833
1.821
1.352
1.962
1.966
1.976
1.943
2.463(1)
1.786(5)
1.816(4)
1.815(5)
1.786(5)
1.803(5)
1.823(5)
1.335(6)
1.994(4)
2.003(4)
1.992(4)
1.995(4)
R- R′
C C
Co1-C
Co1-CR′
Co2-C
Co2-CR′
R-CR-
R′-CR′-
R
Figure 3. Comparison of calculated bond lengths in 1a with those
10
1
1
1
1
1
2
3
4
R
(
given in parentheses) derived from X-ray of the di-tert-butyl acetylene-
dicobalt hexacarbonyl complex.
C
R′
142.6
143.8
140.0
138.8
141.6
138.6
144.8(4)
144.5(4)
C
R′
tions were carried out using the Gaussian 98 program package.20
The study was effected on the complex 1a derived from
propyne, taken as a typical alkyne, and on the complexes 1b
and 1c derived from methyl propiolate and methyl butynoate,
the results with which (ethyl esters) are arguably the most
puzzling in PK chemistry (see above) (Figure 2).21
Chart 1. MO Correlation Diagram for (CO)3CoCo(CO)3
with C2H2
The calculated geometry of 1a is in excellent agreement with
that found through X-ray diffraction analysis of similar com-
plexes.22 This concordance can be seen in Figure 3, in which
the calculated bond lengths in 1a are compared with the X-ray-
based bond lengths in the closely related di-tert-butylacetylene-
2
2b
dicobalt hexacarbonyl complex (indicated in parentheses).
With this assurance as to the validity of the calculations, the
Co2(CO)6(RCtCR′) complexes 1b and 1c were next similarly
optimized. Like complex 1a, complexes 1b and 1c exhibit a
bent, perpendicular acetylene, as found and discussed by
Hoffmann and co-workers.23 The different geometrical features
such as acetylene association, interatomic Co-Co distances, and
Co-CO bond lengths are in very good agreement with the X-ray
data of similar complexes. The most striking feature of the Co-
CO bond lengths is the calculated difference between the
pseudoaxial (apical) and pseudoequatorial (basal) values. This
revealing difference is observed for all three of the Co2(CO)6-
(
RCtCR′) complexes (see Table 1, entry 2-7), and is also
22
found in the crystallography-based structures that are available.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E., Jr.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J.
V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.6; Gaussian, Inc.:
Pittsburgh, PA, 1998.
The bonding between the bent acetylene fragment and
Co2(CO)6 is derived primarily from the stabilizing interaction
between the empty π* orbitals of the acetylene and the bonding
2
2
and antibonding combination of the dx -y orbitals of cobalt,
i.e., through back-donation to the acetylene (Chart 1). These
same d orbitals, however, are also involved in the σ-antibonding
to the equatorial CO ligands. Thus, back-donation to the
acetylene effectively increases the Co-COpseq distances.
The atomic charges on the acetylenic carbon atoms have been
calculated using natural population analysis and these “natural
charges” are presented in Table 2. It can be seen that the
different substituents on the acetylene give rise to different
(
21) The methyl esters were studied in place of the ethyl esters to simplify
the calculations.
22) (a) Dickson, R. S.; Fraser, P. J. AdV. Organomet. Chem. 1974, 12,
(
3
23-377. (b) Cotton, F. A.; Jamerson, J. D.; Stulb, B. R. J. Am. Chem.
Soc. 1976, 98, 1774-1779. See also: Baert, F.; Guelzim, A.; Poblet, J.
M.; Wiest, R.; Demuynck, J.; B e´ nard, M. Inorg. Chem. 1986, 25, 1830-
1
841. (c) D’Agostino, M. F.; Frampton, C. S.; McGlinchey, M. J.
Organometallics 1990, 9, 2972-2984.
23) Hoffman, D. M.; Hoffmann, R.; Fisel, C. R. J. Am. Chem. Soc.
982, 104, 3858-3875.
(
1

Regiochemistry in the Pauson-Khand Reaction
J. Am. Chem. Soc., Vol. 123, No. 23, 2001 5399
-
Table 2. Atomic Charges (e ) Based on Natural Population
Table 3. First Carbonyl Bond Dissociation Energies (kcal/mol) of
Complexes 1a-c
Analysis of the “Acetylenic” Carbons in Compounds 1a-c
entry
complex
q(C
R
)
q(CR′
)
∆q
entry
initial complex
2
3
4
1
2
3
1a
1b
1c
-0.31
-0.36
-0.14
-0.18
-0.33
-0.34
-0.23
-0.03
+0.20
1
2
3
1a
1b
1c
a
37.1
37.9
b
26.4
24.5
24.2
26.7
c
a
Optimized structure (38.5 kcal/mol) exhibits a small imaginary
b
frequency and reoptimization yields 4a. Optimization of this structure
yields 4a. c Optimization of this structure yields 3a.
complexes 3 and 4 depends on the particular acetylenic
substituents. For each of the initial complexes 1a and 1c, only
one complex lacking a COpseq could be optimized.
Figure 4. Different complexes missing a CO (the cis and trans positions
are defined with respect to the position of substituent R′).
The results of this theoretical study are consistent with those
25
of the above-mentioned spectroscopic studies and fully support
the proposal that electronic effects caused by different acetylenic
substituents can influence the position from which carbon
monoxide is lost and ultimately the regiochemistry of the PK
reaction. It bears repeating that the loss of carbon monoxide
from the Co2(CO)6(RCtCR′) complex provides for the incom-
ing alkene a coordination site, the position of which is assumed
to fix the regiochemistry of the resulting cyclopentenone adduct.
Let us now consider separately the complexes 1a-c.
electronic densities on its carbon atoms, CR and CR′, which
results in polarization of the CR-CR′ bond. Neither the intensity
nor the sign of this polarization is constant. Significantly, the
calculated values are in qualitative agreement with what could
be predicted based on the free alkynes. It could be reasoned,
for example, that in the alkyne of complex 1a, due to the
donating effect of the methyl group, the carbon atom (CR)
bonded to the hydrogen should have a larger electron density
than the carbon atom (CR′) that is methyl substituted, and indeed
this reflects what is calculated for the complex. It might,
however, initially appear doubtful that the propiolate of complex
In complex 1a, the CR-CR′ bond must be strongly polarized
(Table 2, entry 1). The larger charge density is found on CR
and the longer Co-COpseq bond length is calculated for the
carbon monoxides that are trans-positioned with respect to R′
1
b, for which calculations indicate an absence of polarization,
(Table 1, entries 3 and 4). Therefore, it is reasonable to expect
could be devoid of polarization. Although the ester group is
electronegative and thus has a tendency to polarize a carbon-
carbon triple bond, the acetylene does not carry a second
substituent capable of participating in polarization of the bond,
hydrogen’s 1s orbital having no electronic interaction with such
a π system, and therefore very little or no polarization should,
in fact, exist. The situation is different, however, with the
butynoate, where the corresponding CR-CR′ bond in the
complex is clearly polarized. In this case, the methyl group
overlaps to a degree with the π system and enriches it
electronically; in conjunction with the electronegative ester
group, a significant polarization could be expected. The
relevance of the orbital symmetry of acetylenic substitutents
on their capacity to interact with the acetylenic system has
previously been discussed.6b
these carbon monoxide ligands to dissociate more easily, an
expectation that is supported by the calculated bond dissociation
energies (Table 3, entry 1). The olefin (norbonene) then occupies
the vacant coordination site, which leads to the observed
R-methylcyclopentenone. Importantly, the calculations show that
even if one of the strongly bound COpsax ligands were to
dissociate from the complex and thus create a vacant site that
could, potentially, also produce the observed cycloadduct, a
migration of a COpseq,trans to the axial position would result to
give once again a vacant trans pseudoequatorial coordination
2
6
site.
In complex 1b, where the CR-CR′ bond appears to lack
significant polarization (Table 2, entry 2), the Co-COpseq bond
lengths and bond dissociation energies should be nearly identi-
cal, and this is exactly what the calculations indicate (Table 1,
entries 3 and 4; Table 3, entry 2). The COpseq ligands should
thus be able to dissociate with equal facility, and therefore steric
effects, as opposed to electronic ones, should now play the
decisive role. In the absence of significant electronic effects,
the larger substituent on an acetylene should ultimately be found
at the R position in the cyclopentenone, and the bigger ester
group, in the present case, is indeed found there.
In the Co2(CO)6(RCtCR′) complexes, three different carbon
monoxide ligands can dissociate: loss of COpsax will lead to 2,
whereas departure of the two equatorial ligands, COpseq,cis and
COpseq,trans, will produce 3 and 4, respectively (Figure 4).
Type 2 complexes exhibit relatively nondeformed structures
in comparison with the parent complexes, Co6(CO)6(RCtCR′).
In contrast, for the complexes 3 and 4, which have suffered the
loss of an equatorial CO, large structural changes have been
calculated: most significantly, a shortening of the Co-Co bond,
the acetylene-cobalt bonds, and the two remaining Co-CO
bonds of the involved cobalt. Energetically, the calculated values
are in reasonable agreement with available experimental and
calculated values for CO dissociation in mononuclear complexes
Finally, in complex 1c, in which the R and R′ substituents
are sterically similar, the CR-CR′ bond should again be strongly
polarized, but with CR′ in this case having the higher electron
density (Table 2, entry 3). Thus it is now a COpseq,cis that would
(25) The 13C NMR chemical shifts (in ppm) of the coordinated sp carbon
Table 3).24 It is noteworthy that type 2 complexes are found
atoms are: in 1a,6b 73.0 (C-1) and 90.8 (C-2); in 1b,6b 76.9 (C-2) and 73.5
(
(
C-3); in 1c, 80.0 (C-2) and 95.3 (C-3).
to be about 10 kcal/mol higher in energy than the two others,
which is consistent with the shorter Co-COax bond lengths.
Also clear from the results in this table is that the fate of
(26) However, axial complexation has been suggested: Derdau, V.;
Laschat, S.; Jones, P. G. Eur. J. Org. Chem. 2000, 681-689. Recent DFT
results indicate that in certain cases axial complexation can, in fact, be
produced by pseudorotation and that cobaltacycle formation can then occur
through a low-energy pathway (see: de Bruin, T. J. M.; Milet, A.; Robert,
F.; Gimbert, Y.; Greene, A. E. J. Am. Chem. Soc., submitted for publication).
In addition, preliminary DFT results indicate that the loss of a pseudoequa-
torial ligand can also lead to rearrangement of the coordination sphere to
produce a CO-bridged structure. The formation of both intermediates,
however, should also be subject to these electronic effects.
(24) For example, the experimental value of the first dissociation energy
for Mo(CO)6 is 34 kcal/mol. (Ehlers, A. W.; Frenking, G. J. Am. Chem.
Soc. 1994, 116, 1514-1520). Good agreement is found between the
calculated and experimental values for the dissociation energy of CO in
complexes in which acetylene and CO are on the same metal atom. (Decker,
S. A.; Klobukowski, M. J. Am. Chem. Soc. 1998, 120, 9342-9355.)

5400 J. Am. Chem. Soc., Vol. 123, No. 23, 2001
Robert et al.
Scheme 3. Structure of the Cycloadduct Resulting from the
PK Reaction of Complex 1d and Norbornene and an ORTEP
Representation (40% probability displacement ellipsoids)
exclusiVely in the R position in the resultant cyclopentenone 5
2
8
(Scheme 3).
In summary, a DFT study has been carried out on alkyne-
dicobalt hexacarbonyl complexes to probe for the first time
whether electronic differences in the acetylenic substituents
could impact on the regiochemistry of the PK reaction. It has
been found that the regiochemistry in certain cases appears, in
fact, to be entirely governed by electronic differences in the
acetylenic substituents, which translate, through discriminant
carbon monoxide loss, into regioselective cyclopentenone
formation. This new insight provides an alternative to the “steric
rationale” for mechanistic interpretation of this important aspect
of the PK reaction.
Acknowledgment. We thank Dr. M.-T. Averbuch and Dr.
C. Philouze for the X-ray structure determination and Dr. T. de
Bruin for help in preparing the manuscript. We are in addition
thankful to the Universit e´ Joseph Fourier and the Centre IDRIS
of the CNRS in Orsay (project 001219) for providing the
computer facilities.Financial support from Universit e´ Joseph
Fourier and the CNRS (UMR 5616) and a fellowship award
from M.E.S.R. to F.R. are gratefully acknowledged.
Supporting Information Available: Tables of Cartesian
coordinates for the theoretically optimized structures 1-4 and
cristallographic data for cycloadduct 5 (CIF). This material is
available free of charge via the Internet at http://pubs.acs.org.
be expected to dissociate, an expectation that is clearly supported
by the calculations (Table 1, entries 3 and 4; Table 3, entry 3).
Norbornene should coordinate to cobalt at a cis pseudoequatorial
coordination site, which would then lead to the product
cyclopentenone with the ester function at the â position: this
is precisely what is observed. Thus, the dichotomous PK results
with ethyl propiolate and ethyl butynoate can now be under-
stood.
JA004170N
(27) Ethyl 4-(4-methylphenylethynyl)benzoate was prepared from ethyl
-iodobenzoate and 1-ethynyl-4-methylbenzene by using the Sonogashira
4
procedure. See: Sonogashira, K. In Metal-Catalyzed Cross-Coupling
Reactions; Diederich, F., Stang, P. J., Eds.; Wiley-VCM: New York, 1997;
Chapter 5. Complexation (80% yield) and the PK reaction (65% yield) were
realized according to ref 12.
(
28) Crystal data for C26H26O3: colorless, triclinic, P 1h , a ) 5.827(5),
Å, b ) 9.154(6) Å, c ) 19.59(1) Å, R ) 90.76(3)°, â ) 92.54(3)°, γ )
What is the predictive power of this analysis? With an
acetylene bearing substituents that sterically are virtually
identical, but electronically different, can the regiochemical
81.82(3)°, V ) 1033(1) Å3, Z ) 2, d
) 1.242 g/cm3, F(000) ) 412,
calc
-
1
µ(Mo KR) ) 0.80 cm , θ range 2-26° (Mo KR), 4182 measured
reflections on a Enraf-Nonius CAD4 diffractometer, 4055 independent
reflections [R(int) ) 0.021], the structure determination using direct methods
2
7
outcome of the PK reaction now be foreseen? Complex 1d
is a good test case; gratifyingly, the tolyl group is found
(
SIR92), R(F) ) 0.062, Rw ) 0.057 for 2012 reflections with I > 2σ(I),
GOF ) 1.987.