X-ray crystal structure of an alkene-pentacarbonyldicobalt-alkyne complex: Isolation of a stable magnus-type Pauson-Khand reaction intermediate

Article abstract of DOI:10.1002/anie.200605171

Caught in the act: The [(dibenzosuberenol)Co₂(CO)₆] cluster 1 (see scheme) loses a carbonyl ligand to yield 2, the first structurally characterized alkyne-dicobaltpentacarbonyl-alkene complex. This finding provides experimental support for the first step of the proposed mechanism of the Pauson-Khand reaction and yields crucial structural data relevant to computational studies of the process. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200605171

Angewandte
Chemie
DOI: 10.1002/anie.200605171
Cobalt Alkyne Complexes
X-ray Crystal Structure of an Alkene–Pentacarbonyldicobalt–Alkyne
Complex: Isolation of a Stable Magnus-Type Pauson–Khand Reaction
Intermediate**
Emilie V. Banide, Helge Müller-Bunz, Anthony R. Manning, Paul Evans,* and
Michael J. McGlinchey*
The Pauson–Khand reaction (PKR), the formal
[2+2+1] cycloaddition of an alkyne 1, an alkene 2, and
carbon monoxide, is a highly convergent method for the
alkene to form the metallocycle 7, alkyl migration to
incorporate the carbonyl group, as in 8, and reductive
elimination of the cyclopentenone from the final intermediate
9 complete the process.
preparation of the important cyclopentenone ringmotif
3
(Scheme 1).[1] Initially reported in 1973,[2] early examples of
Since 1985, several reports concerningthis mechanism
have appeared, for example, 5 has
been detected by usingIR spectros-
copy[4] and is isolable when the
alkyne contains an S atom capable
of coordinating, forming a chelate to
the vacant site.[5] More recently, elec-
trospray ionization mass spectrome-
try has been employed to detect, in
one example, an alkene association
complex of the type 6.[6] However,
apart from the initially formed hex-
acarbonyldicobalt–alkyne complex 4
none of the proposed intermediates
alongthis pathway have previously
been isolated and characterized as
stable entities.
Our involvement in this area
arose somewhat serendipitously as a
continuation of earlier studies on
metal-cluster-stabilized nonclassical
or antiaromatic carbocations.[7] We
Scheme 1. Proposed mechanism of the Pauson–Khand reaction. See text for details.
this process focused on the use of octacarbonyldicobalt as a
reaction mediator and source of carbon monoxide, with
thermal promotion. Since then, many variants have enhanced
the scope and utility of the process, but the mechanism
continues to attract discussion. The chemically reasonable
reaction pathway proposed by Magnus and Príncipe[3] in 1985
invokes initial formation of an alkyne–hexacarbonyldicobalt
tetrahedral cluster 4 that loses a ligand to generate the
pentacarbonyl cluster 5, thus allowingincorporation of the
alkene moiety as in 6. Subsequent couplingof the alkyne and
chose
to
prepare
(5-phenylethynyl-5H-dibenzo-
[a,d]cyclohepten-5-ol)hexacarbonyldicobalt (10) with the
aim of investigating the extent of the interaction between
the potential cationic center and the cobalt vertices.[8]
However, when a solution of the alkynol ligand in THF was
allowed to react with [Co2(CO)8] at room temperature for
15 h, two products were isolated: the major product (73%)
was the anticipated hexacarbonyl cluster 10, and the minor
product (15%), separable by flash column chromatography,
proved to be the pentacarbonyldicobalt complex 11, whereby
the third ligand site on one of the cobalt centers was then
=
occupied by the C10 C11 bond of the seven-membered ring
[*] E. V. Banide, Dr. H. Müller-Bunz, Prof. Dr. A. R. Manning,
Dr. P. Evans, Prof. Dr. M. J. McGlinchey
School of Chemistry and Chemical Biology
University College Dublin
(Scheme 2). Moreover, further investigation of this unprece-
dented decarbonylation/alkene-coordination process demon-
strated it to be rather facile. Thus, a mixture of 10 and 11
(85:15) in [D]chloroform gradually metamorphosed into a
mixture containingequal amounts of 10 and 11 after 24 h at
room temperature.
Belfield, Dublin 4 (Ireland)
Fax: ( +353)1-716-1185
E-mail: paul.evans@ucd.ie
michael.mcglinchey@ucd.ie
The IR spectrum of the hexacarbonyl cluster 10 exhibits
n(CO) peaks at 2090, 2056, and 2029 cmꢀ1, whereas, in the
pentacarbonyl complex 11 the n(CO) absorptions are found at
[**] We thank Science Foundation Ireland and University College Dublin
for generous financial support, and Dr. Xavier Verdaguer (Uni-
versitat de Barcelona) for helpful comments.
Angew. Chem. Int. Ed. 2007, 46, 2907 –2910
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2907

Communications
shieldingthe attached protons,
which resonate at d = 6.27 ppm
(7.23 ppm in the free ligand). The
carbonyl ligands in the {Co2(CO)6}
fragment in 10 are close to the
normally observed eclipsed saw-
horse conformation.
In the pentacarbonyl complex 11
it is evident that rotation of the
dicobalt–alkyne cluster about the
Scheme 2. Sequential formation of hexacarbonyl 10 and pentacarbonyl 11.
ꢀ
C5 C12 bond has occurred, thus
2075, 2023, and 1978 cmꢀ1. These data compare well with
placingone metal center close to the C10 C11 linkage. The
ꢀ
ꢁ
those reported for the photolysis of [(PhC CH)Co2(CO)6] in
boat conformation is now more pronounced,[12] thus generat-
inga distinctly avian-type structure. The alkyne carbon–
carbon linkage (1.340 Š) and the C C-C “bend-back” angles
an argon matrix that revealed strong IR peaks at 2099, 2062,
2036, and 2031 cmꢀ1 for the hexacarbonyl cluster, and at 2081,
2035, 2016, and 1981 cmꢀ1 after loss of a CO ligand.[4]
ꢁ
(142.68 and 144.78) lie within the normal ranges; as is normal,
ꢀ
The molecular structures of 10 and 11, determined by X-
ray crystallography,[9,10] are shown in Figures 1 and 2, which
illustrate clearly the boat conformations of the seven-
membered rings.[11] The {(phenylethynyl)Co2(CO)6} moiety
in 10 adopts a pseudoaxial position and is oriented such that
complexation to cobalt increases the C10 C11 double-bond
length (from 1.336(3) Š in 10 to 1.403(7) Š in 11).
Although the carbonyl ligands of the {Co(CO)3} moiety in
11 deviate only slightly from the conventional orientation
whereby one axial ligand is aligned with the midpoint of the
alkyne bond, the ligands of the {Co(CO)2(alkene)} moiety are
staggered rather than eclipsed with respect to the {Co(CO)3}
tripod (see Figure 3), and the alkene is clearly in a pseudo-
equatorial site.
=
the phenyl ringlies directly below the C10 C11 bond, thus
Figure 1. X-ray crystal structure of hexacarbonyl cluster 10. Thermal
ellipsoids are set at the 50% probability level.
Figure 3. View along the cobalt–cobalt bond in 11, emphasizing the
staggered conformation of the {Co(CO)3} and the {Co(CO)2(alkene)}
moieties.
The thermal instability of the hexacarbonyl cluster 10 is
remarkable. Even at room temperature, decarbonylation and
formation of the pentacarbonyl complex 11 is facile. This ease
could be a consequence of the intramolecular character of the
decarbonylation process, whereby loss of CO is compensated
by chelate formation. This result is important, since 11
represents the first structurally characterized h2-alkene–
pentacarbonyldicobalt–alkyne complex, a crucial structural
type in the widely invoked Magnus mechanism[3] of the
Pauson–Khand cyclopentenone synthesis.
Figure 2. X-ray crystal structure of the pentacarbonyl complex 11.
Thermal ellipsoids are set at the 50% probability level.
2908 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2907 –2910

Angewandte
Chemie
High-level density functional theory studies, in which the
relative energies of the proposed Magnus intermediates have
been calculated, suggest that the most energetically demand-
ingstep is the loss of a carbon monoxide ligand from 4 to form
the pentacarbonyl complex 5.[14] These studies also indicate
that the two pseudoaxial carbonyl groups are more strongly
bound than their pseudoequatorial counterparts, thereby
implyingthat the alkenyl complex 6a (Scheme 1) will be
formed preferentially.[14,15] Irreversible bond migration within
6a subsequently generates metallocycle 7a. Milet and co-
workers have considered this key carbon–carbon bond-
formingstep and proposed an alternative scenario whereby
systems capable of undergoing rotation of the ML3 vertex
generate complex 6b, from which position olefin insertion to
form 7b is comparatively facile energetically.[16] The calcu-
molecular strain) was evident. This result may support the
proposal that alkene insertion occurs preferentially from an
axial position (6b!7b), and the existence of 11 as an arrested
Pauson–Khand intermediate is a consequence of the rigidity
of the structure. This lack of rotational freedom manifests
1
itself in the H and 13C NMR spectra of 10 and 11. In the
hexacarbonyl cluster 10, the time-averaged Cs symmetry of
the system renders the two benzo rings equivalent, and the
protons at C10 and C11 appear as a singlet at d = 6.27 ppm. In
contrast, in the pentacarbonyl complex 11, the mirror
symmetry is broken, the benzo rings are now inequivalent,
and the protons at C10 and C11 give rise to two doublets (J =
9.5 Hz) at d = 5.40 and 5.37 ppm. The gradual conversion of
10 into 11 is easily monitored by NMR spectroscopy, and the
results of a kinetic study of this system and a series of related
systems will be described in a future report.
lations of Yamanaka and Nakamura[14a] indicated that the
2
ꢁ
carbonyl ligands in [(HC CH)Co2(CO)5(h -C2H4)] should be
only slightly displaced from the eclipsed orientation generally
observed in [(alkyne)Co2(CO)6] systems; this findingcon-
trasts with the experimentally observed staggered orientation
in the pentacarbonyl complex 11. However, one must recall
that these experimental data are derived from an intra-
molecular rather than an intermolecular process. In all the
calculated geometries, it was found that the ethylene ligand
favors a pseudoequatorial site such that the distance between
an alkene carbon atom and the nearer alkyne carbon atom is
approximately 2.95 Š. Figure 4 depicts the core of the
Experimental Section
In a nitrogen atmosphere, a solution of (5-phenylethynyl-5H-dibenzo-
[a,d]cyclohepten-5-ol) (720 mg, 2.34 mmol) and [Co2(CO)8] (1.60 g,
4.67 mmol) in THF (25 mL) was stirred at room temperature for 15 h.
After removal of the solvent at low temperature, the crude material
was chromatographed on silica gel using pentane/dichloromethane
(90:10) as eluent to yield 10 (1.02 g, 1.72 mmol, 73%) and 11 (200 mg,
0.35 mmol, 15%) as brown solids that were subsequently recrystal-
lized from pentane/dichloromethane (80:20).
1
3
4
10: H NMR (400 MHz, CDCl3): d = 8.19 (dd, 2H, J = 8.0, J =
1.2 Hz; H4, H6), 7.42 (ddd, 2H, 3J = 8.0, 3J = 7.2, 4J = 1.6 Hz; H3, H7),
7.29 (td, 2H, 3J = 7.6, 4J = 1.2 Hz; H2, H8), 7.27–7.19 (m, 3H; phenyl
H), 7.17 (dd, 2H, 3J = 7.6, 4J = 1.6 Hz; H1, H9), 6.80–6.76 (m, 2H;
phenyl H), 6.27 (s, 2H; H10, H11), 2.84 ppm (s, 1H; OH); 13C NMR
(125 MHz, CDCl3): d = 198.7 (CO), 141.2 (C4A, C5A), 138.0 (Cipso),
132.2 (C9A, C11A), 130.2 (C10, C11), 128.8 (C1, C9), 128.7 (Cphenyl),
127.6 (C3, C7), 126.9 (Cphenyl), 126.5 (C2, C8), 126.1 (Cphenyl), 123.7 (C4,
C6), 109.1, 93.0 (C12, C13), 75.2 ppm (COH); IR (CDCl3): n˜ = 2090,
2056, 2029 cmꢀ1
.
11: 1H NMR (500 MHz, CDCl3): d = 7.95 (d, 1H, 3J = 8.0 Hz;
H6), 7.89 (d, 1H, 3J = 8.0 Hz; H4), 7.58 (d, 1H, 3J = 7.5 Hz; H1), 7.44
(td, 1H, 3J = 7.5, 4J = 1.0 Hz; H3), 7.37 (d, 1H, 3J = 7.5 Hz; H9), 7.33
(t, 1H, 3J = 7.5 Hz; H7), 7.32 (t, 1H, 3J = 7.5 Hz; H2), 7.20 (t, 1H, 3J =
3
3
7.5 Hz; phenyl p-H), 7.18 (t, 1H, J = 7.5 Hz; H8), 7.14 (t, 2H, J =
7.5 Hz; phenyl m-H), 6.90 (d, 2H, J = 7.5 Hz; phenyl o-H), 5.40 (d,
3
1H, 3J = 9.5 Hz; H11), 5.37 (d, 1H, 3J = 9.5 Hz; H10), 2.54 ppm (s,
1H; OH); 13C NMR (125 MHz, CDCl3): d = 197.6 (CO), 142.6 (C5A),
142.3 (C4A), 136.4 (Cipso), 135.3 (C9A), 134.4 (C11A), 129.8 (Cortho),
129.7 (C1), 128.8 (C9), 128.5 (Cmeta), 128.1, 128.0 (C3, Cpara), 128.0
(C7), 127.4 (C2), 127.1 (C8), 123.6 (C12), 122.0 (C4), 91.3 (C13), 76.4
(COH), 74.2 (C11), 70.9 ppm (C10); IR (CDCl3): n˜ = 2075, 2023,
Figure 4. Cobalt–carbon bond lengths (a) and nonbonded carbon–
carbon separations (b) within the core of the pentacarbonyl
complex 11; values given in ꢀ.
1978 cmꢀ1
.
crystallographically determined structure of the h2-alkene–
pentacarbonyldicobalt–alkyne complex 11 and reveals
cobalt–carbon bond lengths of 2.139 and 2.184 Š for the
alkene ligand; more importantly, the alkene carbon atoms are
positioned only 2.817 and 2.858 Š from the nearer alkyne
carbon atom, somewhat closer than with the calculated
values. The dihedral angle between the alkene plane C9A-
C10-C11-C11A and the alkyne plane C5-C12-C13-C14 is 428.
Thermolysis of 11 in refluxingtoluene in a nitroegn
Received: December 21, 2006
Published online: March 13, 2007
Keywords: cobalt · density functional calculations · Pauson–
.
Khand reaction · reaction intermediates · structure elucidation
[1] For a very recent review of the Pauson–Khand reaction, see: N.
Jeongin Comprehensive Organometallic Chemistry III, Vol. 11
(Eds.: R. H. Crabtree, D. M. P. Mingos), Elsevier, Oxford, UK,
2006, pp. 335 – 366, and references therein.
[2] I. U. Khand, G. R. Knox, P. L. Pauson, W. E. Watts, M. I.
Foreman, J. Chem. Soc. Perkin Trans. 1 1973, 977 – 981.
atmosphere yielded dibenzosuberenone as the major organic
product, and no Pauson–Khand-derived cyclopentenone
product 12 (the formation of which may be precluded by
Angew. Chem. Int. Ed. 2007, 46, 2907 –2910
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 2909

Communications
[3] P. Magnus, L.-M. Príncipe, Tetrahedron Lett. 1985, 26, 4851 –
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.
4854.
[4] C. M. Gordon, M. Kiszka, I. R. Dunkin, W. J. Kerr, J. S. Scott, J.
Gebicki, J. Organomet. Chem. 1998, 554, 147 – 154.
[10] Crystal data for 11: C28H16O6Co2 (Mr = 566.27), monoclinic,
space group Cc (No. 9), a = 15.4109(17), b = 13.1859(14), c =
[5] a) M. E. Krafft, I. L. Scott, R. H. Romero, S. Feibelmann, C. E.
van Pelt, J. Am. Chem. Soc. 1993, 115, 7199 – 7207; b) X.
Verdaguer, A. Moyano, M. A. Pericas, A. Riera, V. Bernardes,
A. E. Greene, A. Alvarez-Larena, J. F. Piniella, J. Am. Chem.
Soc. 1994, 116, 2153 – 2154.
12.1179(13) Š, b = 109.525(2)8, V= 2320.8(4) Š3, Z = 4, 1calcd
=
1.621 gcmꢀ3, m = 1.473 mmꢀ1, T= 100 K, crystal size 1.00 ” 1.00 ”
0.80 mm3, R1 = 0.0547, wR2 = 0.1496, GOF = 1.163 with I >
2s(I).[9b]
[6] Y. Gimbert, D. Lesange, A. Milet, F. Fournier, A. E. Greene, J.-
C. Tabet, Org. Lett. 2003, 5, 4073 – 4075.
[11] The seven-membered boat is defined by C4A-C5A-C9A-C11A
[plane 1], C4A-C5-C5A [plane 2] and C9A-C10-C11-C11A
[plane 3]. In 10, the interplanar angles [plane 1]–[plane 2] and
[plane 1]–[plane 3] are 50.48 (49.88) and 23.58 (22.28), respec-
tively, in the two independent molecules in the unit cell; the
interplanar angle between the two benzo rings is 47.58 (50.38).
[12] In 11 the interplanar angles [plane 1]–[plane 2] and [plane 1]–
[plane 3] are 57.58 and 29.38, respectively; moreover, the
interplanar angle between the two benzo moieties in 11 has
been increased to 59.58, as opposed to 51.18 in the free ligand.[13]
[13] a) T. Steiner, E. B. Starikov, M. Tamm, J. Chem. Soc. Perkin
Trans. 2 1996, 67 – 71; b) T. Steiner, S. A. Mason, M. Tamm, Acta
Crystallogr. Sect. B 1997, 53, 843 – 848.
[14] a) M. Yamanaka, E. Nakamura, J. Am. Chem. Soc. 2001, 123,
1703 – 1708; b) F. Robert, A. Milet, Y. Gimbert, D. Konya, A. E.
Greene, J. Am. Chem. Soc. 2001, 123, 5396 – 5400; c) T. J. M.
de Bruin, A. Milet, F. Robert, Y. Gimbert, A. E. Greene, J. Am.
Chem. Soc. 2001, 123, 7184 – 7185; d) T. J. M. de Bruin, A. Milet,
A. E. Greene, Y. Gimbert, J. Org. Chem. 2004, 69, 1075 – 1080.
[15] V. Derdau, S. Laschat, P. G. Jones, Eur. J. Org. Chem. 2000, 681 –
689.
[7] a) L. E. Harrington, N. Reginato, I. Vargas-Baca, M. J. McGlin-
chey, Organometallics 2003, 22, 663 – 669; b) J. A. Dunn, R.
Ruffolo, S. S. Rigby, M. A. Brook, M. J. McGlinchey, Organo-
metallics 1999, 18, 3372 – 3382; c) M. Gruselle, H. El Hafa, M.
Nikolski, G. Jaouen, J. Vaissermann, L. Li, M. J. McGlinchey,
Organometallics 1993, 12, 4917 – 4925; d) M. Kondratenko, M.
Gruselle, H. El Hafa, J. Vaissermann, G. Jaouen, M. J. McGlin-
chey, J. Am. Chem. Soc. 1995, 117, 6907 – 6913.
[8] a) M. J. McGlinchey, L. Girard, R. Ruffolo, Coord. Chem. Rev.
1995, 143, 331 – 381; b) A. J. M. Caffyn, K. M. Nicholas in
Comprehensive Organometallic Chemistry II, Vol. 12 (Eds.:
E. W. Abel, F. G. A. Stone, G. Wilkinson), Pergamon, Oxford,
UK, 1995, pp. 685 – 702; c) H. El Amouri, M. Gruselle, Chem.
Rev. 1996, 96, 1077 – 1104.
[9] a) Crystal data for 10: C29H16O7Co2 (Mr = 594.28), triclinic,
¯
space group P1 (no. 2), a = 10.5878(11), b = 13.3942(13), c =
18.1754(18) Š, a = 74.179(2), b = 81.926(2), g = 89.863(2)8, V=
2453.6(4) Š3, Z = 4, 1calcd = 1.609 gcmꢀ3, m = 1.401 mmꢀ1, T=
100 K, crystal size 0.82 ” 0.40 ” 0.20 mm3, R1 = 0.0396, wR2 =
0.1083, GOF = 1.035 with I > 2s(I). b) CCDC-629903 (10) and
CCDC-629902 (11) contain the supplementary crystallographic
[16] T. J. M. de Bruin, C. Michel, K. Vekey, A. E. Greene, Y.
Gimbert, A. Milet, J. Organomet. Chem. 2006, 691, 4281 – 4288.
2910 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 2907 –2910