The [2+2] cycloaddition of alkynes at a Ru-P π-bond

Article abstract of DOI:10.1039/c002765k

The RuP double bond in the 5-coordinate terminal phosphido complex [Ru(η⁵-indenyl)(PR₂)(PPh₃)] undergoes regioselective [2+2] cycloaddition with simple and activated alkynes to give metallaphosphacyclobutene complexes. These unusual examples of alkyne insertion into a metal-heteroatom bond represent potentially important intermediates in stereoselective routes to new phosphine reagents and ligands.

Full text of DOI:10.1039/c002765k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
The [2+2] cycloaddition of alkynes at a Ru–P p-bondw
Eric J. Derrah,a Robert McDonaldb and Lisa Rosenberg*a
Received 9th February 2010, Accepted 21st April 2010
First published as an Advance Article on the web 10th May 2010
DOI: 10.1039/c002765k
The RuQP double bond in the 5-coordinate terminal phosphido
complex [Ru(g5-indenyl)(PR2)(PPh3)] undergoes regioselective
[2+2] cycloaddition with simple and activated alkynes to give
metallaphosphacyclobutene complexes. These unusual examples
of alkyne insertion into a metal–heteroatom bond represent
potentially important intermediates in stereoselective routes to
new phosphine reagents and ligands.
of a series of ruthenium complexes for hydrophosphination of
terminal alkynes by PPh2H relies on activation of the alkyne to
give a ruthenium vinylidene species that undergoes subsequent
nucleophilic attack by free phosphine to form the new P–C
bond (Scheme 1c).6 We recently described [2+2] cycloaddition
reactions of alkenes at the RuQPR2 bond of a terminal
phosphido ruthenium complex (1a-b8), and showed that these
1,2-insertion reactions occur via a concerted, metathesis-like
mechanism.9 We now report the analogous [2+2] cyclo-
addition reactions of alkynes at 1a-b to give new phospha-
metallacyclobutene complexes. These reactions illustrate a
new mode of P–C bond formation at late metal centres, and
point to new routes to the regio- and stereoselective prepara-
tion of alkenyl phosphine reagents and ligands.
Metal-catalyzed addition of the P–H bond in primary or
secondary phosphines to alkynes represents a potentially
regio- and stereoselective route to important alkenyl phosphine
reagents and ligands.1 Despite the considerable synthetic effort
directed toward the development of new chiral phosphine
scaffolds, examples of such metal mediated reactions remain
quite limited.2–7 Lanthanide- and calcium-catalyzed inter-
molecular hydrophosphination of alkynes by PPh2H,2,3 and intra-
molecular hydrophosphination of primary alkynyl phosphines,4
involve insertion of alkyne into the metal–phosphorus bond of
a phosphido intermediate (M–PR2), leading to carbon-bound
metal phosphine complexes (Scheme 1a). However, such 1,2-
insertion of alkynes into a M–P bond has not yet been observed
for late metal catalysts. The group 10 metal-catalyzed hydro-
phosphination of alkynes by PPh2H shows tunable regio-
selectivity that is rationalized by a mechanism involving
alkyne insertion into a M–H bond (Scheme 1b).5 In this case
reductive elimination of a phosphido ligand and the resulting
alkenyl ligand is responsible for P–C bond formation. The activity
Terminal phosphido complex 1a-b reacts rapidly and regio-
selectively with one equivalent of phenylacetylene to give
the unusual [2+2] cycloaddition product, [Ru(Z5-indenyl)-
(k2-PhCQCHPR2)(PPh3)] (2a-b, Scheme 2). The phospha-
metallacyclobutene moiety in 2a-b shows diagnostic upfield
31P{1H} NMR chemical shifts (ꢀ37 ppm (a), ꢀ27 ppm (b))
similar to those previously observed for the four-membered
metallacycle in [Ru(Z5-indenyl){k2-(o-C6H4)PPh2)}(PR2H)]
1
(Scheme 3, 7a-b).8b HSQC NMR for 2a-b shows J correla-
tions between signals for the distinct olefinic proton on the
metallacycle (7.20 ppm (a), 7.36 ppm (b)) and those due to the
b-carbon of the metallacycle (121.2 ppm (a), 122.9 ppm (b)),
1
which also show large JCP coupling (45 Hz (a), 44 Hz (b)) to
Scheme 2
Scheme
1
Examples of metal-catalyzed hydrophosphination of
alkynes.
a Dept. of Chemistry, University of Victoria, P.O. Box 3065,
Victoria, BC, Canada V8W 3V6. E-mail: lisarose@uvic.ca;
Tel: +1-250-721-7173
b X-Ray Crystallography Laboratory, Department of Chemistry,
University of Alberta, Edmonton, AB, Canada T6G 2G2
w Electronic supplementary information (ESI) available: PDF with
experimental and spectroscopic details. CCDC 765663 and 765664.
For ESI and crystallographic data in CIF or other electronic format
see DOI: 10.1039/c002765k
Scheme 3
ꢁc
This journal is The Royal Society of Chemistry 2010
4592 | Chem. Commun., 2010, 46, 4592–4594

View Article Online
Fig. 2 Molecular structure of 3a.15 The hydrogen atom attached to
P1 is shown with arbitrarily small thermal parameters; all other
hydrogen atoms are not shown. Selected interatomic distances (A)
and bond angles (1): Ru–P(1) = 2.2539(4), Ru–P(2) = 2.2890(4),
Ru–C(10) = 2.0143(15), Ru–C* = 1.930, P(1)–H1P = 1.327(18),
C(10)–C(11) = 1.219(2), D = 0.09; P(1)–Ru–P(2) = 92.553(15),
P(1)–Ru–C(10) = 78.79(4), P(2)–Ru–C(10) = 90.46(4), P(1)–Ru–C* =
129.3, P(2)–Ru–C* = 125.6, C(10)–Ru–C* = 126.1, Ru–C(10)–C(11) =
172.82(13), C(10)–C(11)–C(12) = 174.59(17).
Fig. 1 Molecular structure of 2a.15 The hydrogen atom attached to
C(9) is shown with arbitrarily small thermal parameters; all other
hydrogen atoms are not shown. Selected interatomic distances (A) and
bond angles (1): Ru–P(1) = 2.3323(6), Ru–P(2) = 2.2978(6), Ru–C(8) =
2.065(2), Ru–C* = 1.931, P(1)–C(9) = 1.783(2), C(8)–C(9) =
1.347(3), D = 0.12; P(1)–Ru–P(2) = 91.89(2), P(1)–Ru–C(8) =
65.63(6), P(2)–Ru–C(8)
=
92.45(6), P(1)–Ru–C*
=
135.4,
P(2)–Ru–C* = 125.6, C(8)–Ru–C*
=
126.9, Ru–P(1)–C(9) =
85.17(7), Ru–C(8)–C(9) = 109.30(15), Ru–C(8)–C(11) = 128.42(15),
C(9)–C(8)–C(11) = 122.2(2), P(1)–C(9)–C(8) = 98.80(15).
Both simple and internal alkynes also undergo [2+2]-
cycloaddition reactions with 1a-b. One equivalent of 1-hexyne
adds rapidly and regioselectively to 1a-b to give [Ru(Z5-indenyl)-
(k2-BunCQCHPR2)(PPh3)] (4a-b, Scheme 3) as the major
product, along with a small impurity of the alkynyl isomer,
[Ru(CꢂCBun)(Z5-indenyl)(PR2H)(PPh3)] (5a-b), as determined
by 31P{1H} NMR. The reaction of one equivalent of diphenyl-
acetylene with 1a-b was much slower (Scheme 3), but even-
tually gave up to 64% conversion to the metallacyclic complex
[Ru(Z5-indenyl)(k2-PhCQCPhPR2)(PPh3)] (6a-b), as determined
by 31P{1H} NMR. As for the terminal alkyne cycloaddition
products 2 and 4, complex 6a-b showed diagnostic upfield
31P chemical shifts for the phosphorus in the metallacycle
(ꢀ19 ppm (a), ꢀ8 ppm (b)). However, the longer reaction times
resulted in some decomposition to the orthometallated complex
7a-b,8b along with some minor, unidentified by-products.
The formation of phosphametallacyclobutenes 2, 4, and 6 is
an unusual example of the insertion of alkynes into a well-
characterized late M–P bond. It is distinct from the mecha-
nism of P–C bond formation proposed for the lanthanide- and
calcium-catalyzed alkyne hydrophosphinations described
above, in that this insertion is occurring at a MQPR2 double
bond instead of a single bond. In this respect these reactions
more closely resemble the observed [2+2]-cycloaddition of
alkynes at early metal phosphinidene16 and imido17 complexes,
the latter of which has been shown to be responsible for the
N–C bond forming step in the catalytic hydroamination of
alkynes by primary amines.18 Our observation of rapid [2+2]-
cycloaddition of the electron-rich 1-hexyne, as well as that of
the more activated substrate phenylacetylene, suggests that
these P–C bond-forming reactions do not rely on a step-
wise mechanism involving nucleophilic attack of the basic
phosphido ligand at an electrophilic carbon, and this is
supported by the high rate of cycloaddition relative to the
competing deprotonation of these terminal alkynes. These
results, and the much slower rate of cycloaddition we observe
3
the phosphorus contained in the metallacycle and small JCP
coupling (3 Hz (a), 2 Hz (b)) to the adjacent triphenyl-
phosphine ligand. The 13C signals for the a-carbon (180.3 ppm
(a), 182.2 ppm (b)) show moderate JCP coupling (25, 15 Hz
2
(a); 26, 14 Hz (b)) to the two Ru-coordinated phosphorus
nuclei.10 The solid state structure of 2a (Fig. 1)z confirms
the regiochemistry of addition, as well as the unsaturated character
of the metallacycle: short C(8)–C(9) bond distance (1.347(3) A)
and planarity at the metallated carbon (S C(8) + = 359.91) are
observed. The loss of double bond character in the Ru–PCy2 bond
is indicated by an increase in the Ru–P(1) bond length in 2a to
2.3323(6) A, relative to 2.1589(14) A in the crystal structure of
1a.8b The small internal bond angle for the metallacycle
(P(1)–Ru–C(8) = 65.63(6)1) points to considerable ring strain
within this metallacycle.
Along with complex 2a-b, the reaction of phenylacetylene
with 1 consistently produces a small amount of its structural
isomer, the alkynyl complex [Ru(CꢂCPh)(Z5-indenyl)(PR2H)-
(PPh3)] (3a: 6%; 3b: 7%), as determined by 31P NMR. The
identity of 3a was confirmed by its independent preparation
from the addition of phenylacetylene to [Ru(Cl)(Z5-indenyl)-
(PCy2H)(PPh3)] in the presence of methanolic KOH.11 This
allowed its characterization by X-ray crystallography
(Fig. 2)z and solution NMR (ESIw).12 IR spectroscopy of 3a
showed a strong diagnostic stretch due to the Z1-alkynyl ligand11
at 2071 cmꢀ1 as well as the expected nPH at 2143 cmꢀ1. The
formation of 3a-b in these reaction mixtures seems reasonable,
based on the established P-basicity of 1a-b,13 and the relatively
high acidity of the terminal proton in phenylacetylene
(pKa(H2O) E 20).14 The small relative amount of 3a-b that
we observe is interesting: since we have seen no evidence for
reversibility of these reactions, we presume that this is a kinetic
product distribution, which points to a much higher rate for
[2+2]-cycloaddition than for deprotonation.
ꢁc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 4592–4594 | 4593

View Article Online
for the internal alkyne diphenylacetylene, are consistent with a
possible concerted mechanism, similar to that we observe for
the cycloaddition of alkenes at 1a-b.9 Further experiments will
be required to confirm this and/or to investigate the alternative,
conventional migratory insertion, which would require prior
coordination of the alkyne ligand at ruthenium.19 We continue
to probe the mechanism of these reactions, to better understand
the regioselectivity of the addition of terminal alkynes.
In particular, complexes of the generic formula [Ru(Z1-alkenyl)-
(Z5-indenyl)L2] (L = tertiary phosphine) show 13C shifts in range
B150–200 ppm, with 2JCP E 12–16 Hz. (a) M. Bassetti, P. Casellato,
M. P. Gamasa, J. Gimeno, C. Gonzalez-Bernardo and B. Martin-
Vaca, Organometallics, 1997, 16, 5470; (b) V. Cadierno,
M. P. Gamasa, J. Gimeno, C. Gonzalez-Bernardo, E. Perez-Carreno
and S. Garcia-Granda, Organometallics, 2001, 20, 5177.
11 L. A. Oro, M. A. Ciriano and M. Campo, J. Organomet. Chem.,
1985, 289, 117.
12 Examples of other ruthenium indenyl complexes containing alkynyl
ligands: (a) M. P. Gamasa, J. Gimeno, B. M. Martinvaca, J. Borge,
S. Garciagranda and E. Perezcarreno, Organometallics, 1994, 13,
4045; (b) T. Tanase, H. Mochizuki, R. Sato and Y. Yamamoto,
J. Organomet. Chem., 1994, 466, 233; (c) V. Cadierno,
M. P. Gamasa, J. Gimeno, M. C. Lopez-Gonzalez, J. Borge and
S. GarciaGranda, Organometallics, 1997, 16, 4453.
13 Complex 1a-b rapidly and quantitatively deprotonates acetonitrile
(pKa = 25, R. Stewart, The Proton: Applications to Organic
Chemistry, Academic Press, Orlando, 1985) to give the secondary
phosphine complex [Ru(CH2CN)(Z5-indenyl)(PR2H)(PPh3)] (ref. 8a).
14 A. C. Lin, Y. Chiang, D. B. Dahlberg and A. J. Kresge, J. Am.
Chem. Soc., 1983, 105, 5380.
Currently we are investigating conditions allowing cleavage
of the Ru–C bond in these metallaphosphacyclobutenes,
including putative catalytic hydrophosphination conditions
involving excess secondary phosphine and alkyne,20 and the
addition of electrophilic reagents to promote the stoichiometric,
stereoselective formation of new alkenyl phosphines. We note
also the broader synthetic possibilities that are presented by the
a,b-unsaturation in these strained 2-phosphinovinyl metalla-
cycles: addition reactions are envisaged that may allow the
systematic elaboration of new tertiary phosphine structures.
We thank the NSERC of Canada for financial support.
15 Non-hydrogen atoms are represented by Gaussian ellipsoids at the
20% probability level. C* denotes the centroid of the plane defined
by C(7A)–C(1)–C(2)–C(3)–C(3A), and D (indenyl slip distortion) is
d[Ru–C(7A),C(3A)] ꢀ d[Ru–C(1),C(3)].
Notes and references
16 T. L. Breen and D. W. Stephan, Organometallics, 1996, 15, 5729.
17 Examples include: (a) A. M. Baranger, P. J. Walsh and
R. G. Bergman, J. Am. Chem. Soc., 1993, 115, 2753; (b) E. Blake,
D. M. Antonelli, L. M. Henling, W. P. Schaefer, K. I. Hardcastle
and J. E. Bercaw, Organometallics, 1998, 17, 718; (c) J. L. Polse,
R. A. Andersen and R. G. Bergman, J. Am. Chem. Soc., 1998, 120,
13405; (d) B. D. Ward, A. Maisse-Francois, P. Mountford and
L. H. Gade, Chem. Commun., 2004, 704.
z Crystallographic data: yellow crystals of [Ru(Z5-indenyl)-
(k2-PhCQCHPR2)(PPh3)]ꢃ0.5C6H6 (2a) were obtained via slow diffu-
sion of toluene and acetonitrile into a benzene solution of the
compound. Crystal data for 2a (CCDC 765663): C50H53P2Ru, M =
816.93, monoclinic, space group P21/c (No. 14), a = 10.1439(9) A,
b = 20.4714(18) A, c = 19.6785(18) A, b = 104.5062(14)1, V =
3956.2(6) A3, Z = 4, T = 193(1) K, 30 708 reflections measured, 8114
unique (Rint = 0.0373) which were used in all calculations, R1(F) =
0.0302 (6898 reflections with I Z 2s(I)), wR2(F2) = 0.0832 (all data).
Orange crystals of [Ru(Z5-indenyl)Ru(CꢂCPh)(PHCy2)(PPh3)] (3a)
were obtained via slow diffusion of hexanes into a dichloromethane
solution of the compound. Crystal data for 3a (CCDC 765664):
18 T. E. Muller, K. C. Hultzsch, M. Yus, F. Foubelo and M. Tada,
Chem. Rev., 2008, 108, 3795.
19 The empty coordination site required for alkyne coordination at
Ru could arise from loss of planarity at the phosphido ligand, Z5 to
Z3 ring-slippage of the indenyl ligand, or dissociation of the
triphenylphosphine ligand in 1a-b. Our previous studies (ref. 8)
suggest that adduct formation involving change of the phosphido
ligand from a planar, 3eꢀ donor to a pyramidal, 1eꢀ donor is too
sensitive to the bulk of the incoming donor to allow alkyne
coordination. Although hapticity change has been observed for
many indenyl complexes, and is commonly invoked to explain
associative substitution mechanisms at these complexes, previous
studies of the parent complex, [Ru(Cl)(Z5-indenyl)(PPh3)2],
indicate that its phosphine substitution reactions proceed dissocia-
tively, with no evidence for formation of Z3-indenyl intermediates,
at faster rates than for the corresponding Cp or Cp* derivatives
(M. P. Gamasa, J. Gimeno, C. Gonzalez-Bernardo and B. M.
Martin-Vaca, Organometallics, 1996, 15, 302). Further control
experiments (e.g. alkyne addition in the presence of excess PPh3)
are required to probe the susceptibility of the triphenylphosphine
ligand to substitution by incoming alkyne substrates, although we
note that free PPh3 (trace) is observed only in the slow diphenyl-
acetylene reactions.
20 Encouragingly, preliminary experiments show that 2a is a major
ruthenium-containing product formed when catalytic amounts of
[Ru(Cl)(Z5-indenyl)(PCy2H)(PPh3)] (the precursor to 1a) and
KOBut are added to a 1 : 1 mixture of HPCy2 and PhCCH.
However, no free alkenylphosphine hydrophosphination product
is observed in this reaction mixture, and we saw no reaction of
either excess HPCy2 or excess phenylacetylene with isolated 2a in
separate experiments at RT (see ESIw). More forcing conditions
may be required to ‘‘turn over’’ this reaction. Alternatively, the
use of less donating and/or bulky substituents at the secondary
phosphine may be required. Previous studies have shown that more
reactive diaryl (ref. 8b) and alkylaryl (G. L. Gibson, B.Sc. Honour
Thesis, University of Victoria, 2009) analogues of 1a-b can be
generated in situ.
ꢀ
C47H50P2Ru, triclinic, space group P1 (No. 2), a = 9.8234(11) A,
b = 10.0881(11) A, c = 22.744(2) A, a = 70.4887(12)1, b = 84.1365
(12)1, g = 62.9065 (11)1, V = 1888.0 (4) A3, Z = 2, T = 173(1) K,
14796 reflections used in calculations (twinned dataset), R1(F) = 0.0220
(14 570 reflections with I Z 2s(I)), wR2(F2) = 0.0628 (all data).
1 F. Alonso, I. P. Beletskaya and M. Yus, Chem. Rev., 2004, 104, 3079.
2 K. Takaki, K. Komeyama, D. Kobayashi, T. Kawabata and
K. Takehira, J. Alloys Compd., 2006, 408, 432.
3 M. R. Crimmin, A. G. M. Barrett, M. S. Hill, P. B. Hitchcock and
P. A. Procopiou, Organometallics, 2007, 26, 2953.
4 A. M. Kawaoka, M. R. Douglass and T. J. Marks, Organo-
metallics, 2003, 22, 4630.
5 M. A. Kazankova, I. V. Efimova, A. N. Kochetkov, V. V. Afanas’ev,
I. P. Beletskaya and P. H. Dixneuf, Synlett, 2001, 497.
6 F. Jerome, F. Monnier, H. Lawicka, S. Derien and P. H. Dixneuf,
Chem. Commun., 2003, 696.
7 Hydrophosphination strategies to prepare unsaturated phosphine
reagents are complemented by efforts to develop metal-catalyzed
phosphination of activated vinylic and allylic halides. (a) P. Butti,
R. Rochat, A. D. Sadow and A. Togni, Angew. Chem., Int. Ed.,
2008, 47, 4878; (b) S. R. Gilbertson, Z. Fu and G. W. Starkey,
Tetrahedron Lett., 1999, 40, 8509; (c) M. A. Kazankova,
E. A. Chirkov, A. N. Kochetkov, I. V. Efimova and
I. P. Beletskaya, Tetrahedron Lett., 1999, 40, 573.
8 (a) E. J. Derrah, K. E. Giesbrecht, R. McDonald and L. Rosenberg,
Organometallics, 2008, 27, 5025; (b) E. J. Derrah, D. A. Pantazis,
R. McDonald and L. Rosenberg, Organometallics, 2007, 26, 1473.
9 E. J. Derrah, D. A. Pantazis, R. McDonald and L. Rosenberg,
Angew. Chem., Int. Ed., 2010, 49, 3367.
10 Comparable, downfield 13C chemical shifts for Ca have been
observed for other examples of Ru-bound, sp2 hybridized carbon.
ꢁc
This journal is The Royal Society of Chemistry 2010
4594 | Chem. Commun., 2010, 46, 4592–4594