Phosphine-alkene ligands as mechanistic probes in the pauson-khand reaction

Article abstract of DOI:10.1002/chem.201000357

An alkyne tetracarbonyl dicobalt complex with a chelated phosphine-alkene ligand, in which the phosphorus atom and the alkene from the ligand are attached to the same cobalt atom has been prepared, isolated, and characterized by X-ray crystallography. The

Full text of DOI:10.1002/chem.201000357

DOI: 10.1002/chem.201000357
Phosphine–Alkene Ligands as Mechanistic Probes in the Pauson–Khand
Reaction
Catalina Ferrer,[a] Jordi Benet-Buchholz,[b] Antoni Riera,*[a] and Xavier Verdaguer*[a]
Dedicated to Prof. Santiago Olivella on the occasion of his 65th birthday
Abstract: An alkyne tetracarbonyl di-
cobalt complex with a chelated phos-
phine–alkene ligand, in which the phos-
phorus atom and the alkene from the
ligand are attached to the same cobalt
atom has been prepared, isolated, and
characterized by X-ray crystallography.
The complex serves as a mechanistic
model for an intermediate of the
Pauson–Khand (PK) reaction. Al-
though the alkene fragment is located
in an equatorial coordination site with
an appropriate orientation, and, there-
fore, should undergo insertion, it failed
to give the PK product upon either
thermal or N-methylmorpholine N-
phine–alkene complex that contains a
terminal alkene readily provided the
corresponding PK product. We attri-
bute this change in reactivity to the dif-
ferent ability of each olefin to undergo
1,2-insertion. These results provide fur-
ther insights into the factors that
govern a crucial step in the PK reac-
tion, the olefin insertion.
oxide activation. However,
a phos-
Keywords: alkene ligands · cobalt ·
insertion · Pauson–Khand reaction ·
phosphane ligands
Introduction
The Pauson–Khand (PK) reaction is a cobalt-catalyzed or
-mediated [2+2+1] cycloaddition of an alkyne, an alkene,
and carbon monoxide to form a cyclopentenone.[1] It has
been widely used to synthesize natural products containing
a cyclopentane ring.[2,3] The commonly accepted mechanism
of the PK reaction (Scheme 1) was originally proposed by
Magnus.[4] The first step is the reaction of an alkyne unit
and dicobalt octacarbonyl to form the hexacarbonyl com-
plex I. Thermal activation yields loss of a CO molecule (II)
and the resulting vacant position is occupied by an alkene
[a] Dr. C. Ferrer, Prof. A. Riera, Prof. X. Verdaguer
Unitat de Recerca en Sꢁntesi Asimꢂtrica (URSA)-PCB
Institute for Research in Biomedicine (IRB) Barcelona and
Departament de Quꢁmica Orgꢀnica
Scheme 1. The Pauson–Khand reaction mechanism proposed by Magnus.
Universitat de Barcelona
c/Baldiri Reixac 10, 08028 Barcelona (Spain)
Fax : (+34)403-70-95
E-mail: antoni.riera@irbbarcelona.org
xavier.verdaguer@irbbarcelona.org
Homepage: www.ursapcb.org
(III). The alkene undergoes 1,2-insertion at the less hin-
À
dered C Co bond of the tetrahedral cluster. This is followed
by coordination of a new CO molecule to form cobalt
A
AHCTUNGTREGcNNNU ycle IV, which undergoes CO insertion (V) and reductive
elimination to give VI. Finally, reductive elimination of
[Co2(CO)6] affords the cyclopentenone VII.
This mechanism is supported by theoretical studies per-
formed by Nakamura,[5] Gimbert,[6] and Pericꢀs.[7] Neverthe-
less, little experimental evidence for the proposed inter-
[b] Dr. J. Benet-Buchholz
Institute of Chemical Research of Catalonia (ICIQ),
Avda. Paꢃsos Catalans 16, 43007 Tarragona (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000357.
8340
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8340 – 8346

FULL PAPER
mediates has been found. In fact, of the intermediates that
have been isolated and characterized, nearly all are of
type I, although those of type II have been detected by IR
spectroscopy,[8] and our group has isolated type II intermedi-
ates in reactions featuring an alkyne that contains a sulfur
atom which can occupy the vacant coordination site.[9] Fur-
thermore, one example of an intermediate of type III has
been detected by electrospray ionization mass spectrosco-
py.[10] During failed attempts at inserting cyclopropene into
Co2–alkyne complexes, Fox et al. identified various side-
products.[11] More recently, Evans and McGlinchey isolated
and characterized (by X-ray diffraction) compounds 1a–e,
type III complexes in which an olefin is intramolecularly co-
ordinated to a cobalt atom (Scheme 2).[12] Interestingly, com-
plexes 1 do not undergo alkene insertion to give the corre-
sponding PK adducts; the authors have attributed this lack
of reactivity to two factors: the severe strain in the final PK
adduct and the pseudoequatorial position of the alkene in
complexes 1.
Results and Discussion
Compound
2 is in the dibenzotropylidene phosphane
(TROPP) class of ligands, which were originally described
by Grꢅtzmacher et al.[13] We envisaged that once the phos-
phorus atom was coordinated to cobalt, the olefin moiety of
the ligand would also be accessible for coordination. This, in
turn, would enable olefin insertion, since the expected PK
product would be less strained than those derived from com-
plexes 1. We observed better results using borane-protected
2 (2-BH3) than if the free ligand was used. The free ligand
was prepared by a slight modification of a literature proce-
dure[13a] and was then treated with BH3·SMe2 to form 2-BH3.
In situ deprotection of 2-BH3 with 1,4-diazabicyclo-
AHCTUNGTREG[NNNU 2.2.2]octane (DABCO, followed by reaction with the trime-
thylsilylacetylene dicobalt complex 4a in toluene, at 708C,
afforded a red solution, which was not isolated. After a fur-
ther two hours, this complex yielded (as determined by
TLC) complex 5a (also red), which was isolated by chroma-
tography in 42% yield (Scheme 3). We obtained similar re-
sults for the reaction of 2-BH3 with phenylacetylene com-
plex 4b.
Scheme 2. Cobalt–alkene complexes characterized by Evans and McGlin-
chey; TMS=trimethylsilyl.
Scheme 3. The reaction of borane-protected 2 (2-BH3) with acetylene di-
cobalt complexes.
To determine if this lack of reactivity is, indeed, due to
the relative orientations of the cobalt-coordinated alkyne
and the alkene, and to gain mechanistic insights into the PK
reaction, we decided to study the coordination and reactivity
of phosphine–alkene ligands with alkyne–dicobalt com-
plexes. We envisaged that phosphine–alkene ligands would
be flexible enough to allow insertion of the olefin and
would provide valuable information on the PK reaction
mechanism. We chose two types of unsaturated phosphorous
We obtained suitable crystals of complex 5a for analysis
by X-ray diffraction (Figure 1).[14] The solid-state structure
obtained unambiguously showed that the phosphorus atom
and the alkene both coordinate to the same cobalt atom:
the former is positioned pseudoaxially, whereas the latter
occupies an equatorial coordination site anti to the TMS
group. As in 1, the seven-membered ring in 2 adopts a boat
conformation, providing a characteristic avian-type struc-
ture. However, in sharp contrast to the case of 1, described
by Evans and McGlinchey,[12] coordination of 2 to the
alkyne–dicobalt complex (5a) does not distort the ligand ar-
rangement around the Co2C2 cluster. Whereas in complex 1
the alkene and CO ligands adopt a staggered conformation
relative to the vicinal CO ligands, in complex 5a the bound
alkene and the remaining ligands show an eclipsed confor-
mation (Figure 2). Another important difference between 1
and 5a is the alkene–Co bond lengths: in complex 1 the
bonds are 2.14 and 2.18 ꢆ, whereas in complex 5a they are
2.13 and 2.16 ꢆ, indicating stronger bonding to the metal
center. Finally, and most interestingly, the alkene moiety is
ligand: dibenzo
ACHTUNGTRENNUNG
the 5H-dibenzoACHTUNGTRENNUNG
which resembles complexes 1; and the alkenyloxydiphenyl-
phosphines 3, featuring terminal alkenes. Herein, we report
our study into the coordination and PK reactivity of alkyne
dicobalt hexacarbonyl complexes with 2 and 3a–c.
À
parallel and close to the Co Calkyne bond, the position in
which olefin insertion should occur: the bond distances are
À
À
2.13 ꢆ (Co1 C11) and 2.92 ꢆ (C10 C2). Again, these
Chem. Eur. J. 2010, 16, 8340 – 8346
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8341

A. Riera, X. Verdaguer et al.
empty p* orbitals of the olefin.[15] X-ray data reveal that the
C=C bond length is 1.328 ꢆ in 2 and 1.406 ꢆ in 5a, corre-
sponding to a C=C bond lengthening of 0.077 ꢆ.[16] This in-
crease is greater than that calculated for norbornene (0.047
to 0.050 ꢆ) and suggests that effective backbonding occurs
between the metal and the alkene moiety in 5a.
Despite the seemingly optimal geometry between the
À
alkene and the Co C bond, we were unable to identify any
PK product. Both thermal activation (toluene, 1108C) and
N-oxide activation (N-methylmorpholine N-oxide (NMO),
CH2Cl2, RT) of complexes 5a and b only resulted in decom-
position products. However, the putative structures of the
PK adducts corresponding to 5a and b (6a and b, respec-
tively, Scheme 4) are not strained. Furthermore, despite the
Figure 1. The crystal structure of complex 5a (ORTEP diagram showing
50% probability ellipsoids).
Scheme 4. Neither thermal nor N-oxide activation of complexes 5 yielded
the desired Pauson–Khand products 6.
equatorially positioned alkene, the geometries of the cobalt
complexes seemed optimal. Thus, we attributed this lack of
reactivity to two possible factors: the inherent inability of
the dibenzoACTHNUGRTNEUNG[a,d]cycloheptene system to undergo 1,2-inser-
tion and the presence of a phosphine ligand at the same
cobalt atom as the alkene (phosphine ligands are known to
decrease the reactivity of alkyne–dicobalt carbonyl com-
plexes[17]). To resolve this matter, we decided to study the
phosphine–alkene ligands 3a–c, which all feature a sterically
unhindered, linear olefin.
Figure 2. A comparison of the conformation of cobalt complexes 1 (stag-
gered; shown on the left) and 5a (eclipsed, shown on the right); red=
oxygen, orange=phosphorus, grey=carbon, green=cobalt.
We prepared (prop-3-enyloxy)diphenyl phosphine (3a),
(but-3-enyloxy)diphenyl phosphine (3b), and (pent-4-eny-
loxy)diphenyl phosphine (3c) by reacting the corresponding
alkenyl alcohols with chlorodiphenyl phosphine. Reaction of
phosphine–alkene ligands 3a–c with phenylacetylene di-
AHCTUNGTREGcNNNU obalt complex 4b at 408C for 4 h afforded complexes 7a–c,
in which only the phosphine moiety is coordinated to the
cobalt atom (Scheme 5).
values contrast with those of 1, in which the alkene is
À
almost perpendicular to the Co[1] C[16] bond (Figure 2).
The theoretical study of Milet, Gimbert et al.[6d] estab-
lishes that the insertion of ethene into the propyne dicobalt
pentacarbonyl complex is favored if the olefin is in an axial
coordination site, rather than an equatorial one; however,
5a and b should readily overcome the energy barrier associ-
ated with reacting in the equatorial position (4.2 kcalmolÀ1
higher than the axial one). Moreover, rotation of the ML3
vertex could bring the olefin into an axial coordination site.
These authors also studied the insertion of several different
olefins and concluded that the LUMO of the coordinated
olefin is crucial to its reactivity in the PK reaction, as it de-
termines the degree of backbonding that occurs, which is of
Despite our efforts, we could not detect any complexes in
which the alkene was coordinated to the cobalt atom. In
one attempt, the reaction of phosphine 3a with phenylacety-
À
paramount importance for C C bond formation. Greater
backbonding should facilitate the subsequent insertion. In
terms of structure, the change in the C=C bond length in
complexed versus uncomplexed olefins reflects the amount
of backdonation from the filled d orbitals of Co into the
Scheme 5. The reaction of phosphine–alkene ligands 3a–c with phenyl-
AHCTUNGTRENGNaUN cetylene–hexacarbonyldicobalt.
8342
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8340 – 8346

The Pauson–Khand Reaction
FULL PAPER
lene–hexacarbonyldicobalt at 808C generated a novel red
complex (8), which, despite being unstable in solution, was
isolated by chromatography. X-ray, H NMR, and IR data
sequent 1,2-insertion. This finding agrees with those of
Krafft et al. for sulfide-directed PK reactions.[20]
1
The formation of compound 9a confirms that, in contrast
to the case of complexes 5a and b, the alkene in 7a coordi-
nates to cobalt and readily undergoes 1,2-insertion to give
the corresponding PK adduct. This reaction also indicates
that the failure of the suberene system to undergo 1,2-inser-
tion in complexes 5a and b is not due to the coordination of
a phosphorous atom, but to the olefinic system itself.[21]
Despite this progress, we were still perplexed as to why
compounds 5 would not give the PK products despite being
coordinated to cobalt and well positioned for olefin inser-
tion. Although there is a general consensus that the best PK
substrates are strained alkenes, there are some exceptions:
for instance, ethene is a reasonably good substrate for inter-
molecular PK reactions,[22] yet is not strained. Thus, in trying
to understand the reactivity of different olefins in PK
chemistry, other features (e.g., substitution patterns) must
also be considered. In this respect, we reasoned that the
thermodynamics of the overall PK reaction could provide a
rough estimate of the efficiency with which the energy con-
tained in the original double bond is released during the for-
mation of the cobaltacycle and that it should mimic the ex-
perimental reactivity observed for different alkenes.[23] To
test this hypothesis, we calculated the energy release at the
DFT level in PK reactions of acetylene with different olefins
(Table 1).[24]
were consistent with 8 being a bisphosphine–hexacarbonyl-
dicobalt complex in which the phosphines contained the PK
adduct.[18] The IR spectrum of 8 shows a strong band at
1952 cmÀ1, which is in full agreement with literature reports
for similar complexes.[19] Complex 8 is likely to be formed,
after the PK reaction, by rearrangement of the phosphine
groups into a more stable cobalt complex (Scheme 6).
Scheme 6. The reactions of phosphine–alkene ligands with dicobalt–
alkyne complexes.
To locate the PK adducts, we treated complexes 7a–c
with NMO at room temperature. Gratifyingly, complex 7a,
which contains an allyloxyphosphine group, provided the
corresponding PK adduct 9a in 27% yield (Scheme 6). The
structure of 9a was unambiguously assigned by X-ray crys-
tallography (Figure 3). As expected, under N-oxide condi-
tions the phosphinite group was oxidized to the correspond-
ing phosphinate.
Table 1. Calculated energy release in PK reactions of acetylene with var-
ious olefins.
[a]
[b]
Alkene
DE [kcalmolÀ1
]
DErel [kcalmolÀ1
]
norbornene
AHCTUNGTRENGN[UN 2,2]paracyclophane-1,9-diene
ethene
cycloheptene
cyclopentene
cyclohexene
(Z)-stilbene
suberene
À73.45
À70.62
À69.86
À67.50
À65.37
À61.52
À58.34
À52.67
0.0
2.8
3.6
5.9
8.1
11.9
15.1
20.8
[a] Reaction energies were calculated at the DFT level (B3LYP/6-31G*)
and include zero-point energy corrections. [b] DErel =DE-
(alkene)ÀDE(norbornene).
AHCTUNGTRENNUNG
As expected, norbornene—a reactive, strained alkene—
Figure 3. The crystal structure of 9a (ORTEP diagram showing 50%
probability ellipsoids).
leads to the most exoergic cyclization. Thus, we classified
the olefins according to the endoergicity of their corre-
sponding reactions relative to that of norbornene; thus, the
values are expressed as DErel. We were pleased to observe
that the calculated energies accurately parallel the observed
experimental reactivities. Ethene is among the most reactive
olefins in the PK reaction, with DErel =3.8 kcalmolÀ1 relative
to norbornene. We anticipated unusual reactivity of the
cyclic olefins—the calculated DErel for cycloheptene
Interestingly, if complexes 7b and c were treated with
NMO, under the same conditions as complex 7a, or heated
at 808C, only the starting materials were recovered. This in-
dicates that the PK reaction occurs intramolecularly and
that the length of the tether between the phosphorus group
and the alkene is crucial to the olefin coordination and sub-
Chem. Eur. J. 2010, 16, 8340 – 8346
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8343

A. Riera, X. Verdaguer et al.
(5.9 kcalmolÀ1) indicates than it is more reactive than cyclo-
pentene (8.1 kcalmolÀ1) and cyclohexene (11.9 kcal
molÀ1).[25] As a control experiment, we also calculated the
energy of a single PK reaction of [2,2]paracyclophane-1,9-
diene. De Meijere et al. have shown that [2,2]paracy-
clophane-1,9-diene, which is structurally related to suberene
(they are both (Z)-1,2-diarylalkenes), undergoes rapid cycli-
zation.[26] By thermodynamic analysis, we accurately predict-
ed that the cyclophane alkene ranks among the most active
alkenes for PK chemistry (Table 1). Gratifyingly, suberene
was the least reactive of all the alkenes that we studied
(DErel =20.8 kcalmolÀ1). This is in complete accordance with
our experimental results and is consistent with the fact that
complexes 5a and b do not undergo the PK reaction. We be-
lieve that this lack of reactivity probably stems from the
energy penalty associated to the loss of conjugation between
the two benzene rings that would occur upon olefin inser-
tion.[27]
ene (5 mL), at À158C. The solution was allowed to warm to room tem-
perature and stirred overnight. The excess thionyl chloride and the sol-
vent were evaporated under vacuum. The residue was dissolved in tolu-
ene and concentrated again to remove further excess thionyl chloride.
This process was performed twice. The corresponding cycloheptenyl chlo-
ride was obtained as a purple solid and used in the next reaction without
further purification.
A solution of Ph2PH (0.46 mL, 2.64 mmol) in toluene (4 mL) was added
to
a
solution of dibenzoACTHNGUTERNNU[G a,d]cyclohepten-5-yl chloride (544 mg,
2.40 mmol) in toluene (4 mL) and the mixture was heated at reflux over-
night. The reaction was then cooled to room temperature and borane di-
methylsulfide (0.29 mL, 3.12 mmol) was added. The reaction was stirred
for 2 h at room temperature and quenched with water. After workup by
extraction (EtOAc/water) the mixture was purified by silica gel column
chromatography (1:1, hexane/EtOAc) and 2-BH3 was obtained as a white
solid (422 mg, 45%). M.p. 187–1888C; 1H NMR (400 MHz, CDCl3): d=
5.05 (d, J=14 Hz, 1H), 6.33 (s, 2H), 7.10–7.18 (m, 4H), 7.21–7.25 (m,
4H), 7.28 (dd, J=7, 2 Hz, 4H), 7.38–7.50 ppm (m, 6H); 13C NMR
(75 MHz, CDCl3): d=56.3 (d, JP =24 Hz, CH), 127.7 (d, JP =2 Hz, 2CH),
128.2 (d, JP =10 Hz, 4CH), 128.8 (d, JP =1 Hz, 2CH), 129.2 (d, JP =
50 Hz, 2C), 129.9 (d, JP =2 Hz, 2CH), 131.0 (d, JP =2 Hz, 2CH), 131.2
(d, JP =5 Hz, 2CH), 132.0 (brs, 2CH), 133.8 (d, JP =8 Hz, 4CH), 134.1
(d, JP =1 Hz, 2C), 136.1 ppm (d, JP =4 Hz, 2C); 31P NMR (121 MHz,
CDCl3): d=25.9 ppm (brs); IR (film): n˜max =1062, 1433, 2387 cmÀ1
;
HRMS (ESI): m/z calcd for C27H23BP [MÀH]À: 389.1630; found:
Conclusion
389.1637.
A
2ACHTUNGTERNNUG(m-TMSC2H)(CO)4ACTHUNGTRENNUNG
(C27H21P)] (5a): DABCO (39 mg, 0.35 mmol) and
In summary, we have prepared, isolated, and characterized
(by X-ray crystallography) a new alkyne–tetracarbonyldico-
balt complex (5a) containing a chelated phosphine–alkene
ligand (2), in which the phosphorus atom and the alkene
from the ligand are both attached to the same cobalt atom.
Complex 5a serves as a mechanistic model for an intermedi-
ate in the PK reaction. Although the alkene fragment is lo-
cated in an equatorial coordination site with an appropriate
orientation and, therefore, should undergo insertion, it
failed to give the PK product upon either thermal or NMO
activation. Conversely, phosphine–alkene 3a, which contains
a terminal alkene, readily provided the corresponding PK
product. We attribute this reactivity to the differing ability
of each olefin to undergo 1,2-insertion. Although terminal
alkenes insert rapidly, the suberene moiety is not amenable
to insertion, because this transformation involves the unfav-
orable loss of conjugation between the two benzene rings.
These results underscore the premise that alkene coordina-
tion alone is not sufficient to enable alkene insertion in PK
chemistry. Our experimental results, in good agreement with
the theoretical calculations[5] and kinetic data,[16a] support
the fact that the rate-determining step in the PK reaction is
the alkene insertion.
a
solution of 2-BH3 (90 mg,
0.23 mmol) in toluene (2 mL). The reaction was heated at 708C for 2 h.
Then the solvent was evaporated and the mixture purified by silica gel
column chromatography (30:1, hexane/EtOAc) to give 5a as a red solid
(70 mg, 42%). 1H NMR (400 MHz, C6D6): d=0.05 (s, 9H), 4.53 (d, J=
15 Hz, 1H), 5.34–5.41 (m, 2H), 6.26 (d, J=8 Hz, 1H), 6.66–6.80 (m, 6H),
6.86–6.98 (m, 8H), 7.25–7.28 (m,3H), 7.39 ppm (dd, J=8.1 Hz, 1H); 13C-
AHCTUNGTRENNUNG
{31P} NMR (75 MHz, C6D6): d=1.3 (3CH3), 53.9 (CH), 69.7 (CH), 77.6
(CH), 94.7 (CH), 126.3 (CH), 126.4 (CH), 127.2 (CH), 127.3 (CH), 127.6
(2CH), 127.9 (2CH), 128.9 (CH), 129.1 (CH), 129.2 (CH), 129.3 (CH),
129.9 (2CH), 131.5 (C), 131.9 (C), 133.0 (2CH), 133.6 (2CH), 133.9 (C),
134.2 (C), 138.3 (C), 138.4 ppm (C) (one carbon signal is missing);
31P NMR (121 MHz, C6D6): d=84.9 ppm (s); IR (film): n˜max =2037,
1985 cmÀ1; HRMS (ESI): m/z calcd for C34H32Co2O2PSi [M+HÀ2CO]+:
649.0573; found: 649.0585.
AHCNUTRTGEGNN[UN Co2AHCUTNGETRN(NGUN m-PhC2H)(CO)5AHCTUNGTRENNUG(C15H15OP)] (7a): A solution of 3a (300 mg,
1.24 mmol) and 4b (528 mg, 1.36 mmol) in toluene (8 mL) was heated at
408C for 4 h. Then the solvent was evaporated and the mixture purified
by silica gel column chromatography (30:1, hexane/EtOAc) to give 7a as
a red oil (420 mg, 56%). 1H NMR (400 MHz, C6D6): d=3.79–3.83 (m,
1H), 3.86–3.92 (m, 1H), 4.85 (d, J=10 Hz, 1H), 5.00 (d, J=17 Hz, 1H),
5.33 (d, J=3 Hz, 1H), 5.39–5.46 (m, 1H), 6.88–6.94 (m, 9H), 7.33 (d, J=
6 Hz, 2H), 7.36–7.43 ppm (m, 4H); 13C NMR (75 MHz, C6D6): d=66.6
(d, JP =1 Hz, CH2), 70.9 (s, CH), 86.7 (s, C), 116.6 (s, CH2), 126.9 (s, CH),
128.1 (s, CH), 128.2 (d, JP =2 Hz, 2CH), 128.3 (d, JP =2 Hz, 2CH), 128.4
(s, 2CH), 130.4 (d, JP =12 Hz, 2CH), 130.6 (s, CH), 130.8 (s, 2CH), 130.9
(d, JP =13 Hz, 2CH), 133.2 (d, JP =9 Hz, CH), 137.3 (d, JP =42 Hz, C),
138.0 (d, JP =45 Hz, C), 138.5 ppm (s, C); 31P NMR (121 MHz, C6D6): d=
154.6 ppm (s); IR (film): n˜max =1962, 2006, 2061 cmÀ1; HRMS (ESI): m/z
calcd for C25H22O3Co2P [M+HÀ3CO]+: 518.9971; found: 518.9979.
We believe that phosphine–alkene ligands, in light of their
utility in this study, should continue to prove invaluable in
gaining further insight into the mechanism of the PK reac-
tion.
(2-Oxo-3-phenylcyclopent-3-enyl)methyldiphenyl phosphinate (9a):
A
solution of NMO (237 mg, 1.03 mmol) in CH2Cl2 (4 mL) was added to a
solution of 7a (244 mg, 0.41 mmol) in CH2Cl2 (4 mL). The reaction was
stirred at room temperature for 3.5 h. Then the solvent was evaporated
and the mixture purified by silica gel column chromatography (2:1,
hexane/EtOAc) to give 9a as a white solid with some traces of cobalt
(43 mg, 27%). M.p. 136–1378C; 1H NMR (400 MHz, C6D6): d=2.05
(ddd, J=10, 6, 2 Hz, 1H), 2.15–2.19 (m, 1H), 2.52 (dt, J=19, 2 Hz, 1H),
4.15–4.25 (m, 2H), 6.87–6.94 (m, 4H), 6.99 (t, J=3 Hz, 1H), 7.00–7.05
(m, 4H), 7.77–7.83 (m, 5H), 7.90–7.96 ppm (m, 2H); 13C NMR (75 MHz,
CDCl3): d=30.6 (s, CH2), 47.2 (s, CH), 64.2 (s, CH2), 127.2 (s, 2CH),
Experimental Section
DibenzoACHTUNGTRENNUNG[a,d]cyclohepten-5-yldiphenyl phosphine borane complex (2-
BH3): Phosphine 2-BH3 was prepared following a slightly modified litera-
ture procedure.[13a] Thionyl chloride (0.52 mL, 7.20 mmol) was added
dropwise to a solution of dibenzosuberenol (500 mg, 2.40 mmol) in tolu-
8344
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8340 – 8346

The Pauson–Khand Reaction
FULL PAPER
128.7 (s, 2CH), 128.8 (s, CH), 128.8–129.0 (m, 5CAr), 131.6 (s, C), 131.7–
132.5 (m, 7CAr), 143.4 (s, C), 159.1 (s, CH), 206.0 ppm (s, C); 31P NMR
(121 MHz, C6D6): d=30.9 ppm (s); IR (film): n˜max =1130, 1228, 1438,
1700 cmÀ1; HRMS (ESI): m/z calcd for C24H22O3P [M+H]+: 389.1307;
found: 389.1313.
K. Vekey, A. E. Greene, Y. Gimbert, A. Milet, J. Organomet. Chem.
2006, 691, 4281–4288.
[7] M. A. Pericꢀs, J. Balsells, J. Castro, I. Marchueta, A. Moyano, A.
Riera, J. Vꢈzquez, X. Verdaguer, Pure Appl. Chem. 2002, 74, 167–
174.
[8] C. M. Gordon, M. Kiszka, I. R. Dunkin, W. J. Kerr, J. S. Scott, J. Ge-
bicki, J. Organomet. Chem. 1998, 554, 147–154.
Dimer 8:
A solution of 3a (400 mg, 1.65 mmol) and 4b (640 mg,
1.65 mmol) in toluene (10 mL) was heated at 808C for 4 h. Then the sol-
vent was evaporated and the mixture purified by silica gel column chro-
matography (2:1, hexane/EtOAc) to give 8 as a red solid (150 mg, 8%).
Dimer 8 is rather unstable in solution and both the free PK adduct and
[9] 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; c) I. Marchueta, E. Montenegro, D. Panov, M. Poch, X. Verda-
guer, A. Moyano, M. A. Pericas, A. Riera, J. Org. Chem. 2001, 66,
6400–6409.
[10] Y. Gimbert, D. Lesage, A. Milet, F. Fournier, A. E. Greene, J. Tabet,
Org. Lett. 2003, 5, 4073–4075.
[11] M. K. Pallerla, G. P. A. Yap, J. M. Fox, J. Org. Chem. 2008, 73,
6137–6141.
[12] a) E. V. Banide, H. Mueller-Bunz, A. R. Manning, P. Evans, M. J.
McGlinchey, Angew. Chem. 2007, 119, 2965–2968; Angew. Chem.
Int. Ed. 2007, 46, 2907–2910; b) S. A. Brusey, E. V. Banide, S. Dor-
rich, P. OꢉDonohue, Y. Ortin, H. Muller-Bunz, C. Long, P. Evans,
M. J. McGlinchey, Organometallics 2009, 28, 6308–6319.
[13] a) H. Schçnberg, S. Boulmaaz, M. Wçrle, L. Liesum, A. Schweiger,
H. Grꢅtzmacher, Angew. Chem. 1998, 110, 1492–1494; Angew.
Chem. Int. Ed. 1998, 37, 1423–1426; b) C. Bçhler, T. Bꢅttner, S.
Deblon, G. Frasca, G. Frison, J. Geier, C. Laporte, F. Lꢊng, S. Loss,
P. Maire, N. Pꢂ, H. Schçnberg, M. Scherer, C. Widauer, H. Grꢅtz-
macher, Chimia 2001, 55, 814–817.
1
dimer 8, are observed by NMR spectroscopy. H NMR (400 MHz, C6D6):
d=2.06–2.20 (m, 2H), 2.49–2.57 (m, 1H), 4.17–4.37 (m, 2H), 6.82–6.95
(m, 3H), 7.00–7.22 (m, 7H), 7.68–7.96 ppm (m, 5H); 13C NMR (75 MHz,
C6D6): d=30.6 (s, CH2, dimer+PK adduct), 47.3 (t, JP =4 Hz, CH,
dimer), 47.5 (d, JP =7 Hz, CH, PK adduct), 64.5 (d, JP =6 Hz, CH2,
dimer), 66.5 (s, CH2, PK adduct), 126.0–144.4 (m, dimer+PK adduct),
158.1 (s, CH, dimer), 158.6 (s, CH, PK adduct), 202.9 (t, JP =10 Hz,
[Co(CO)3], dimer), 205.3 (s, CO, dimer), 205.5 ppm (s, CO, PK adduct);
31P NMR (121 MHz, C6D6): d=30 (s, PK adduct), 167 ppm (s, dimer); IR
(film): n˜max =1436, 1703, 1952 cmÀ1
; HRMS (ESI): m/z calcd for
C24H22O3P [PK adduct+H]+: 389.1307; found: 389.1313.
Acknowledgements
We thank MICINN (CTQ2008-00763/BQU), IRB Barcelona, and the
Generalitat de Catalunya (2009SGR 00901) for financial support.
[14] CCDC-758950 (5a), 758951 (9a), and 761854 (8) contain the supple-
mentary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[15] C. Defieber, H. Grꢅtzmacher, E. M. Carreira, Angew. Chem. 2008,
120, 4558–4579; Angew. Chem. Int. Ed. 2008, 47, 4482–4502.
[16] J. Thomaier, S. Boulmaaz, H. Schonberg, H. Ruegger, A. Currao, H.
Grutzmacher, H. Hillebrecht, H. Pritzkow, New J. Chem. 1998, 22,
947–958.
[17] a) R. Cabot, A. Lledꢋ, M. Reves, A. Riera, X. Verdaguer, Organo-
metallics 2007, 26, 1134; b) See, for instance, the long reaction times
if diphosphinoamine ligands are used: Y. Gimbert, F. Robert, A.
Durif, M.-T. Averbuch, N. Kann, A. E. Greene, J. Org. Chem. 1999,
64, 3492–3497.
[18] Compound 8 crystallized in the form of extremely small crystals
(with axes <10 mm). After several attempts a crystal plate with the
dimensions 0.048ꢌ0.019ꢌ0.010 mm3 could be used. The obtained
data are of low resolution (2qmax: 278) and are not complete
(80%), but it was considered that the structure was of high enough
quality to confirm the identity of compound 8. Compound 8 crystal-
lizes with half a molecule in the unit cell (Ci symmetry). The five-
membered ring contained in the structure is disordered over two in-
verted positions with a ratio of 75:25.
[19] a) S. A. Llewellyn, M. L. H. Green, A. R. Cowley, Inorg. Chim. Acta
2006, 359, 3785–3789; b) F. Patcas, C. Maniut, C. Ionescu, S. Pitter,
E. Dinjus, Appl. Catal. B 2007, 70, 630–636.
[20] M. E. Krafft, C. A. Juliano, I. L. Scott, C. Wright, M. D. McEachin,
J. Am. Chem. Soc. 1991, 113, 1693–1703.
[21] The reactivity of alkyne cobalt complexes coordinated either with
phosphines, or bulky phosphites is fairly similar; see reference [17a].
[22] Ethene has not been widely used in PK reactions due to the difficul-
ties of using a gas and the low concentration of ethene in toluene at
low pressure; a) B. J. Rausch, H. Becker, R. Gleiter, F. Rominger,
Synlett 2002, 0723–0726; b) W. J. Kerr, M. McLaughlin, P. L.
Pauson, J. Organomet. Chem. 2001, 630, 118–124; c) W. J. Kerr, M.
McLaughlin, P. L. Pauson, S. M. Robertson, J. Organomet. Chem.
2001, 630, 104–117; d) N. Jeong, S. H. Hwang, Angew. Chem. 2000,
112, 650–652; Angew. Chem. Int. Ed. 2000, 39, 636–638; e) W. J.
Kerr, M. McLaughlin, P. L. Pauson, S. M. Robertson, Chem.Com-
[1] For general reviews, see: a) S. Laschat, A. Becheanu, T. Bell, A.
Baro, Synlett 2005, 2547–2570; b) S. E. Gibson, N. Mainolfi, Angew.
Chem. 2005, 117, 3082–3097; Angew. Chem. Int. Ed. 2005, 44, 3022–
3037; c) S. E. Gibson (nꢇe Thomas), A. Stevenazzi, Angew.Chem.
2003, 115, 1844–1854; Angew.Chem. Int. Ed. 2003, 42, 1800–1810;
d) L. V. R. Bonaga, M. E. Krafft, Tetrahedron 2004, 60, 9795–9833;
e) T. Shibata, Adv. Synth. Catal. 2006, 348, 2328–2336; f) M. R.
Rivero, J. Adrio, J. C. Carretero, Synlett 2005, 26–41; g) D. Strꢅbing,
M. Beller, Top. Organomet. Chem. 2006, 18, 165–178.
[2] Selected examples of intermolecular PK reactions, see: a) A. Vaz-
quez-Romero, L. Cardenas, E. Blasi, X. Verdaguer, A. Riera, Org.
Lett. 2009, 11, 3104–3107; b) A. Vazquez-Romero, J. Rodriguez, A.
Lledo, X. Verdaguer, A. Riera, Org. Lett. 2008, 10, 4509–4512;
c) M. Iqbal, P. Evans, A. Lledo, X. Verdaguer, M. A. Pericas, A.
Riera, C. Loeffler, A. K. Sinha, M. J. Mueller, ChemBioChem 2005,
6, 276–280.
[3] Selected examples of intramolecular PK reactions, see: a) P. Gao, P.
Xu, H. Zhai, J. Org. Chem. 2009, 74, 2592–2593; b) A. Nakayama,
N. Kogure, M. Kitajima, H. Takayama Org. Lett. 2009, 11, 5554–
5557; c) Y. Hayashi, N. Miyakoshi, S. Kitagaki, C. Mukai, Org. Lett.
2008, 10, 2385–2388; d) K. Kaneda, T. Honda, Tetrahedron 2008, 64,
11589–11593; e) T. Kozaka, N. Miyakoshi, C. Mukai, J. Org. Chem.
2007, 72, 10147–10154; f) S. Min, S. J. Danishefsky, Angew. Chem.
2007, 119, 2249–2252; Angew. Chem. Int. Ed. 2007, 46, 2199–2202;
g) J. Tormo, A. Moyano, M. A. Pericas, A. Riera, J. Org. Chem.
1997, 62, 4851–4856; h) J. Castro, A. Moyano, M. A. Pericas, A.
Riera, A. E. Greene, A. Alvarez-Larena, J. F. Piniella, J. Org. Chem.
1996, 61, 9016–9020; i) V. Bernardes, N. Kann, A. Riera, A.
Moyano, M. A. Pericas, A. E. Greene, J. Org. Chem. 1995, 60, 6670–
6671.
[4] P. Magnus, L. M. Principe, Tetrahedron Lett. 1985, 26, 4851–4854.
[5] M. Yamanaka, E. Nakamura, J. Am. Chem. Soc. 2001, 123, 1703–
1708.
[6] a) T. J. M. de Bruin, A. Milet, F. Robert, Y. Gimbert, A. E. Greene,
J. Am. Chem. Soc. 2001, 123, 7184–7185; 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, A. E. Greene, Y. Gimbert,
J. Org. Chem. 2004, 69, 1075–1080; d) T. J. M. de Bruin, C. Michel,
Chem. Eur. J. 2010, 16, 8340 – 8346
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8345

A. Riera, X. Verdaguer et al.
mun. 1999, 2171–2172; f) J. G. Donkervoort, A. R. Gordon, C. John-
stone, W. J. Kerr, U. Lange, Tetrahedron 1996, 52, 7391–7420.
[23] We based our hypothesis on Hammondꢉs postulate, which relates
the relative energy of the transition states to the relative energy of
reactants and products.
[26] H. Buchholz, O. Reiser, A. de Meijere, Synlett 1991, 20–22.
[27] This is in good agreement with the fact that TROPP ligands do not
insert hydrogen even if they are being used in catalytic hydrogena-
tions. P. Maire, S. Deblon, F. Breher, J. Geier, C. Bçhler, H. Rꢅeg-
ger, H. Schçnberg, H. Grꢅtzmacher, Chem. Eur. J. 2004, 10, 4198–
4205.
[24] Spartan ’06, Wavefunction, Inc., Irvine, CA.
[25] J. A. Brown, S. Irvine, W. J. Kerr, C. M. Pearson, Org. Biomol.
Chem. 2005, 3, 2396–2398.
Received: February 10, 2010
Published online: June 11, 2010
8346
www.chemeurj.org
ꢄ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8340 – 8346