Deactivation of ruthenium metathesis catalysts via facile formation of face-bridged dimers

Article abstract of DOI:10.1021/om0110888

Reaction of RuCl(dcypb)(μ-Cl)₃Ru(dcypb)(N₂) (3) with an excess of tert-butylacetylene at ambient temperatures yields the dinuclear monovinylidene RuCl(dcypb)(μ-Cl)₃Ru(dcypb)(L) (4a; L = C=CHBu^t, dcypb = 1,4-bis(dicyclohexylphosphino)butane), rather than the expected mononuclear RuCl₂(dcypb)(L). Attempted synthesis of an allenylidene derivative via the corresponding reaction with 1,1-diphenyl-2-propyn-1-ol stops at the stage of hydroxyvinylidene 4b (L = C=CHC(OH)Ph₂). While formation of these dinuclear products may be an artifact of low solubility, the corresponding monoalkylidene species 4c (L = CHCH=CMe₂) is obtained on treating soluble RuH(dcypb)(μ-Cl)₂(μ-H)Ru(dcypb)(H₂) (5) with 3-methyl-3-chloro-1-butyne. Formation of the perchloro species 4c is consistent with facile homodimerization of the initially formed RuCl₂(dcypb)(CHCH-CMe₂) (2d), with expulsion of one alkylidene ligand as the free carbene. 2,7-Dimethylocta-2,4,6-triene, the formal product of carbene coupling, is observed by ¹H NMR. A minor product in this synthesis is proposed to be RuCl(dcypb)(μ-Cl)₂(μ₂,η¹-CHCH=CMe ₂)RuCl(dcypb) (8). While the low activity of 4a/4b in ring-opening metathesis polymerization of norbornene is attributable to their low solubility, that of 4c points toward the stability of the Ru₂(μ-Cl)₃ entity. The low activity and facile formation of 4c reveals an important deactivation pathway for catalysts of type 2d, with additional relevance to other such chlororuthenium complexes, including systems of the Grubbs type. Product identities were established by ¹H, ²H, ¹³C, and ³¹P NMR and IR spectroscopy and (for 4c and 5) by X-ray crystallography.

Full text of DOI:10.1021/om0110888

Organometallics 2002, 21, 3335-3343
3335
Dea ctiva tion of Ru th en iu m Meta th esis Ca ta lysts via
F a cile F or m a tion of F a ce-Br id ged Dim er s
Dino Amoroso, Glenn P. A. Yap, and Deryn E. Fogg*
Center for Catalysis Innovation and Research, Department of Chemistry, University of Ottawa,
Ottawa, Ontario, Canada K1N 6N5
Received December 27, 2001
Reaction of RuCl(dcypb)(µ-Cl)3Ru(dcypb)(N2) (3) with an excess of tert-butylacetylene at
ambient temperatures yields the dinuclear monovinylidene RuCl(dcypb)(µ-Cl)3Ru(dcypb)-
(L) (4a ; L ) CdCHBut, dcypb ) 1,4-bis(dicyclohexylphosphino)butane), rather than the
expected mononuclear RuCl2(dcypb)(L). Attempted synthesis of an allenylidene derivative
via the corresponding reaction with 1,1-diphenyl-2-propyn-1-ol stops at the stage of
hydroxyvinylidene 4b (L ) CdCHC(OH)Ph2). While formation of these dinuclear products
may be an artifact of low solubility, the corresponding monoalkylidene species 4c (L )
CHCHdCMe2) is obtained on treating soluble RuH(dcypb)(µ-Cl)2(µ-H)Ru(dcypb)(H2) (5) with
3-methyl-3-chloro-1-butyne. Formation of the perchloro species 4c is consistent with facile
homodimerization of the initially formed RuCl2(dcypb)(CHCHdCMe2) (2d ), with expulsion
of one alkylidene ligand as the free carbene. 2,7-Dimethylocta-2,4,6-triene, the formal product
1
of carbene coupling, is observed by H NMR. A minor product in this synthesis is proposed
to be RuCl(dcypb)(µ-Cl)2(µ2,η1-CHCHdCMe2)RuCl(dcypb) (8). While the low activity of 4a /
4b in ring-opening metathesis polymerization of norbornene is attributable to their low
solubility, that of 4c points toward the stability of the Ru2(µ-Cl)3 entity. The low activity
and facile formation of 4c reveals an important deactivation pathway for catalysts of type
2d , with additional relevance to other such chlororuthenium complexes, including systems
1
of the Grubbs type. Product identities were established by H, 2H, 13C, and 31P NMR and IR
spectroscopy and (for 4c and 5) by X-ray crystallography.
In tr od u ction
metathesis polymerization (ROMP),3 but will be mani-
fested only as a latency period in ring-closing metathesis
(RCM) or cross-metathesis. Indeed, a number of these
species have shown promising activity in ROMP or RCM
reactions.4
Catalytic olefin metathesis by ruthenium complexes
has received much attention, owing to the robustness
and functional-group tolerance of the metal, and extra-
ordinary success has accrued to benzylidene catalysts
of the type RuCl2LL′(CHPh) (1a , L ) L′ ) PCy3; 1b, L
) PCy3, L′ ) imidazol-2-ylidene).1 The high reactivity
of the phenyldiazomethane reagent used to install the
benzylidene ligand, and of the benzylidene functionality
itself, has prompted considerable recent interest in
alternative metal-carbon functionalities, including
readily accessible ruthenium vinylidene and allenyl-
idene derivatives.2 The slower “turn-on” of such cumulen-
ylidene species, vs benzylidene, may be advantageous
where controlled reactivity is critical. The expected
induction period in metathesis, arising from rates
of initiation being slower than those of propagation,
will increase polymer polydispersity in ring-opening
We5 and others6,7 have recently described the high
ROMP activity of a new class of metathesis catalysts
containing cis-chelating diphosphines. Among these,
benzylidene complexes of the type RuCl2(PP)(CHR) (R
) Ph; PP ) dcypb (1,4-bis(dicyclohexylphosphino)butane
(3) It may be noted that narrow polydispersities are not essential
in many polymer applications, particularly structural ones; indeed, the
precise specification of polymer properties that is associated with
narrow molecular weight distributions can hamper polymer processing.
See: Bo¨hm, L. L.; Enderle, H. F.; Fleissner, M. Stud. Surf. Sci. Catal.
1994, 89, 351.
(4) Representative examples in ROMP: (a) Schwab, P.; Grubbs, R.
H.; Ziller, J . W. J . Am. Chem. Soc. 1996, 118, 100. (b) Katayama, H.;
Ozawa, F. Chem. Lett. 1998, 67. (c) Katayama, H.; Yoshida, T.; Ozawa,
F. J . Organomet. Chem. 1998, 562, 203. (d) Del Rio, I.; van Koten, G.
Tetrahedron Lett. 1999, 40, 1401. (e) Saoud, M.; Romerosa, A.;
Peruzzini, M. Organometallics 2000, 19, 4005. (f) Louie, J .; Grubbs,
R. H. Angew. Chem., Int. Ed. 2001, 40, 1. In RCM: ref 4f. See also:
(g) Schanz, H.-J .; J afarpour, L.; Stevens, E. D.; Nolan, S. P. Organo-
metallics 1999, 18, 5187. (h) Fu¨rstner, A.; Liebl, M.; Lehmann, C. W.;
Picquet, M.; Kunz, R.; Bruneau, C.; Touchard, D.; Dixneuf, P. H. Chem.
Eur. J . 2000, 6, 1847. (i) Cetinkaya, B.; Demir, S.; Ozdemir, I.; Toupet,
L.; Semeril, D.; Bruneau, C.; Dixneuf, P. H. New J . Chem. 2001, 25,
519. Indenylidene complexes have also proved metathesis active in
RCM: ref 4g. See also: (j) Fu¨rstner, A.; Guth, O.; Duffels, A.; Seidel,
G.; Liebl, L.; Gabor, B.; Mynott, R. Chem. Eur. J . 2001, 7, 4811 and
references therein. (k) J afarpour, L.; Schanz, H.-J .; Stevens, E. D.;
Nolan, S. P. Organometallics 1999, 18, 5416.
* To whom correspondence should be addressed. E-mail: dfogg@
science.uottawa.ca. Fax: (613) 562-5170.
(1) Recent reviews: (a) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res.
2001, 34, 18. (b) J ørgensen, M.; Hadwiger, P.; Madsen, R.; Stu¨tz, A.
E.; Wrodnigg, T. M. Curr. Org. Chem. 2000, 4, 565. (c) Roy, R.; Das, S.
K. J . Chem. Soc., Chem. Commun. 2000, 519. (d) Buchmeiser, M. R.
Chem. Rev. 2000, 100, 1565. (e) Fu¨rstner, A. Angew. Chem., Int. Ed.
2000, 39, 3012.
(2) For leading references to Ru allenylidene species, see: (a) Bruce,
M. I. Chem. Rev. 1998, 98, 2797. (b) Cadierno, V.; Gamasa, M. P.;
Gimeno, J . Eur. J . Inorg. Chem. 2001, 571. Vinylidenes: (c) Bruneau,
C.; Dixneuf, P. H. Acc. Chem. Res. 1999, 32, 311. (d) Bruce, M. I. Chem.
Rev. 1991, 91, 197.
(5) (a) Amoroso, D.; Fogg, D. E. Macromolecules 2000, 33, 2815. (b)
Amoroso, D.; Yap, G. P. A.; Fogg, D. E. Can. J . Chem. 2001, 79, 958.
10.1021/om0110888 CCC: $22.00 © 2002 American Chemical Society
Publication on Web 07/10/2002

3336 Organometallics, Vol. 21, No. 16, 2002
Amoroso et al.
Sch em e 1
(2a )), dppb 1,4-bis(diphenylphosphino)butane; binap,
(2,2′-bis(diphenylphosphino)-1,1′-binaphthyl)) represent
the first Ru-diphosphine catalysts to exhibit high
metathesis activity without ligand loss.5 A minimum
turnover number (TON) of 2400 h-1 was found for
ROMP of norbornene via 2a ,5b in which the diphosphine
is the electron-rich, bulky, but flexible ligand dcypb. In
contrast, comparably basic but rigid four- or five-
membered chelate complexes (R ) CHCHdCMe2, PP
) bis(di-tert-butylphosphino)methane (2b);6a,b R ) Ph,
PP ) 1,2-bis(di-tert-butylphosphino)ethane (2c);6c R )
Ph, PP ) 1,2-bis(dicyclohexylphosphino)ethane6d) dis-
played limited or zero activity prior to abstraction of
chloride. We attribute the difference in catalytic behav-
ior to the energetic accessibility in 2a of a square-
pyramidal geometry in which alkylidene lies in the basal
plane and is thus cis to the entering monomer. Such a
structure is supported by modeling and reactivity data.
In contrast, spectroscopic, crystallographic, and model-
ing studies of the four- and five-membered chelate
complexes indicate a strong structural preference for
apical alkylidene, blocking the cis sites.6 While the
activity of 2a could potentially arise from dechelation
of one end of the diphosphine ligand, the cumulative
evidence suggests otherwise. Thus, no dangling phos-
phine is spectroscopically observable, no induction
period is required in catalysis, and the polydispersity
of the polymers produced (PDI ≈ 1.05) is lower than that
found with 1a .5 The last point implies that catalysis
proceeds via a single species, with initiation rates
exceeding rates of propagation. Preliminary indications
of enhanced steric definition of the active site come from
an increase in cis content of the polymer on use of the
binap catalyst. Offsetting the high activity of these
benzylidene species, however, was an instability that
necessitated their generation in situ. Multiple decom-
position products were observed in stoichiometric ex-
periments, accompanied by extrusion of alkylidene as
stilbene, consistent with the high reactivity of this
functionality. We were thus interested in the possibility
of installing alternative, more stable alkylidene or
cumulenylidene ligands.
heightened reactivity can also be manifested as a
susceptibility to alkylphosphine dehydrogenation, par-
ticularly in coordinatively unsaturated Ru precursors
containing labile donors.5b,10 Thus, while dimeric [RuCl-
(dppb)]2(µ-Cl)2 is a tractable, readily isolable starting
material,11 its dcypb analogue undergoes exhaustive
dehydrogenation of the alkylphosphine ligand under
vacuum or argon.5b “Placeholder” ligands are essential
to restrain decomposition of the “RuCl2(dcypb)” entity
until a targeted ligand set (CHR, CdCndCHR, etc.) can
be installed. Mixed-phosphine complexes such as RuCl2-
(dcypb)(PPh3) are limited in their utility, however, by
the requirement for removal of liberated PPh3.5a Atom
efficiency in such transformations is especially impor-
tant in view of the tendency of alkylphosphine com-
plexes to extremes of solubility, which can impede
purification by recrystallization or trituration.4b,c,5 Dini-
trogen is unique as a placeholder ligand in providing a
labile, innocuous donor that does not offer alternative,
undesired reaction pathways and does not contaminate
the product. Dinuclear RuCl(dcypb)(µ-Cl)3Ru(dcypb)(N2)
(3) thus provides an ideal entry point into dcypb
chemistry.
Vin ylid en e Der iva tive. Vinylidene complexes, the
simplest class of metallocumulenylidenes Md(Cd)nCR2
(n ) 1), are readily accessible via reactions of transition-
metal precursors with 1-alkynes.2 Reaction of 3 with
excess 3,3-dimethyl-1-butyne cleanly forms the di-
nuclear monovinylidene RuCl(dcypb)(µ-Cl)3Ru(dcypb)-
(CdCHtBu) (4a ), with evolution of N2 gas (Scheme 1).
The identity of 4a is established by detailed spectro-
scopic analysis and microanalytical data; no evidence
of a mononuclear vinylidene species analogous to 2 is
observed. Installation of vinylidene proceeds smoothly
at room temperature (22 °C), despite the insolubility of
both starting material and product. Thus, while addition
of the alkyne reagent (0.5-4 equiv) to a stirred orange
suspension of 3 in benzene did not effect dissolution,
Resu lts a n d Discu ssion
Ruthenium complexes containing bulky, electron-rich
phosphines exhibit exceptionally high reactivity in
metathesis1a,5b,6b and hydrogenation8,9 catalysis. This
an aliquot removed after 12 h showed no remaining 31
P
NMR signals for 3. Concentration of the solution and
addition of pentane permitted isolation of orange 4a
in 86% yield. The proposed structure is supported
by observation of a sharp, medium-intensity infrared
band at 1633 cm-1, a location characteristic2d of the
(6) Several groups have now described Ru-alkylidene complexes of
type 2, containing four- or five-membered chelating diphosphines,
which are activated upon abstraction of halide. See: (a) Hansen, S.
M.; Rominger, F.; Metz, M.; Hofmann, P. Chem. Eur. J . 1999, 5, 557.
(b) Hansen, S. M.; Volland, M. A. O.; Rominger, F.; Eisentrager, F.;
Hofmann, P. Angew. Chem., Int. Ed. 1999, 38, 1273. (c) Volland, M.
A. O.; Straub, B. F.; Gruber, I.; Rominger, F.; Hofmann, P. J .
Organomet. Chem. 2001, 617, 288. (d) Werner, H.; J ung, S.; Gonzalez-
Herrero, P.; Ilg, K.; Wolf, J . Eur. J . Inorg. Chem. 2001, 1957. More
surprisingly, trigonal-bipyramidal analogues containing a 1,1′-bis-
(diphenylphosphino)ferrocene ligand6d were also inactive prior to
treatment with trimethylsilyl triflate. This is likely due, however, to
the lower reactivity of arylphosphine derivatives, in conjunction with
use of the low-ring-strain substrate cyclooctene.
1
vinylidene νCdC vibration. H NMR analysis reveals a
triplet for the vinylidene proton (δH 3.06, 4J HP ) 3.8 Hz),
accompanied by a peak for the tert-butyl protons at δH
1.24, visible as a sharp singlet above the broad, unre-
solved dcypb resonances (1.0-3.2 ppm). A diagnostic
downfield triplet for the vinylidene R-carbon appears in
the 13C{1H} NMR spectrum (δC 352.1, J CP ) 16 Hz),
2
(7) Six, C.; Beck, K.; Wegner, A.; Leitner, W. Organometallics 2000,
19, 4639.
(8) (a) Burk, M. J .; Martinez, J . P.; Feaster, J . E.; Cosford, N.
Tetrahedron 1994, 50, 4399. (b) Burk, M. J . Acc. Chem. Res. 2000, 33,
363.
(9) (a) Drouin, S. D.; Zamanian, F.; Fogg, D. E. Organometallics
2001, 20, 5495. (b) Drouin, S. D.; Amoroso, D.; Yap, G. P. A.; Fogg, D.
E. Organometallics 2002, 21, 1042.
(10) (a) Borowski, A.; Sabo-Etienne, S.; Christ, M. L.; Donnadieu,
B.; Chaudret, B. Organometallics 1996, 15, 1427. (b) Six, C.; Gabor,
B.; Gorls, H.; Mynott, R.; Philipps, P.; Leitner, W. Organometallics
1999, 18, 3316.
(11) J oshi, A. M.; Thorburn, I. S.; Rettig, S. J .; J ames, B. R. Inorg.
Chim. Acta 1992, 198-200, 283.

Deactivation of Ru Metathesis Catalysts
Organometallics, Vol. 21, No. 16, 2002 3337
accompanied by a singlet for Câ at δC 119.7. 31P{1H}
NMR data provide unequivocal evidence for the di-
nuclear monovinylidene structure. Four distinct sets of
resonances of equal integrated intensity are observed,
signifying the presence of four inequivalent phosphorus
nuclei within a dinuclear structure devoid of symmetry
Sch em e 2
2
2
(δP 50.6, 49.6 (ABq, J PP ) 37 Hz); 45.9, 40.1 (AB, J PP
) 23 Hz)).
Microanalytical data are in good agreement with the
proposed structure, although low solubility precluded
reprecipitation of 4a from THF or aromatic (benzene,
toluene) solvents. Purification by reprecipitation from
methylene chloride or chloroform, in which the solubility
is higher, is thwarted by the susceptibility of dcypb
complexes to chlorination in such solvents. Mixed-
valence, paramagnetic Ru2Cl5(dcypb)2 has been crys-
tallographically identified among the products formed
on dissolving 3 in CDCl3.5b A new 31P{1H} NMR singlet
(δ 46.8; ∼2-5%) is evident immediately after dissolving
4a in CDCl3, though the timescale of decomposition is
sufficiently slow that 1H NMR data can be obtained
without much difficulty on saturated solutions, and
weak 31P{1H} NMR signals for 4a can be discerned even
after 7 days in solution.
3-H yd r oxyvin ylid en e Der iva t ive. Spontaneous
dehydration of propyn-1-ols by transition-metal com-
plexes provides an important route to allenylidene
derivatives,2a,b,12 and in some cases their indenylidene
isomers.4k,13 The reaction commonly proceeds via a 3-
hydroxyvinylidene species, which on loss of water yields
the allenylidene complex.2a,b Addition of 1,1-diphenyl-
2-propyn-1-ol (2 equiv) to a stirred suspension of 3 at
room temperature caused a color change from orange
to brown without dissolution; the 31P NMR signals for
starting 3 disappeared completely within 12 h. The light
brown product was isolated in 88% yield by diluting the
suspension with hexanes, filtering, and washing the
powder with hexanes. As with 4a , the insolubility of the
complex precluded reprecipitation, and purification was
effected by trituration with hexanes.
at 4.38 and 3.99 ppm, each giving an integration of
1H. 1H-13C HMQC experiments correlate the signal at
δH 4.38 with an sp2 carbon at 120 ppm. The signal at
δH 3.99 does not correlate with any carbon nucleus; its
identity as a hydroxyl proton is confirmed by exchange
with D2O and by observation of a medium-intensity
ν(OH) band at 3483 cm-1 in the IR spectrum of the
protio derivative. From these data, we infer that instal-
lation of the allenylidene ligand has halted at the stage
of hydroxyvinylidene 4b.2a,b,14,15 A particular resistance
of electron-rich Ru-hydroxyvinylidene species to de-
hydration has been suggested.2a,14 The 1H NMR
data, as well as the 13C NMR shift position for Câ
deduced from 1H-detected HMQC experiments, are in
excellent agreement with values reported for the related
species [Cp*Ru{CdCHC(OH)Ph2}(PiPr2CH2CH2PiPr2)]-
[BPh4].14 Attempts to confirm the assignment by direct
measurement of the 13C{1H} NMR spectrum were
thwarted by the poor solubility of 4b in benzene. The
spectrum in CDCl3, though complicated by competing
decomposition over the 24 h timescale required for good
signal-to-noise ratios (vide supra), reveals the diagnostic
downfield signals for the allenylidene moiety (CR, δ
308.7, unresolved t; Câ, δ 243.4, s; Cγ, δ 150.4, s).
Dissolution-triggered tautomerization of related hydroxy-
vinylidene species was recently described.14 This indi-
rect evidence, with the cumulative weight of the 1H
NMR, HMQC, and infrared data, provides strong sup-
port for hydroxyvinylidene structure 4b. Microanalytical
data are in good agreement with this formulation.
Or igin of Din u clea r P r od u cts. In principle, the
dinuclear products of type 4 may be generated by an
initially formed mononuclear species of the type RuCl2-
(PP)(CdCdCHR) via homodimerization (Scheme 3, path
i), or via cross-dimerization16 with unreacted 3 (path ii).
In view of the poor solubility of both 3 and 4a /4b,
however, we cannot rule out the alternative possibility
that reaction takes place within the dinuclear frame-
work, at a site vacated by N2 (path iii).
31P{1H} NMR analysis of the product reveals four sets
of resonances of equal intensity, in a pattern closely
analogous to that for 4a (51.5, 50.1 (ABq, 2J PP ) 38 Hz);
44.7 (d, unresolved), 40.9 (d, unresolved)), indicating
formation of a complex of the type RuCl(dcypb)(µ-Cl)3Ru-
(dcypb)(L). The infrared spectrum, however, does not
show the diagnostic, strong allenylidene νCdCdC band
between 1870 and 1970 cm-1 2a,b
Indeed, this region
.
of the spectrum is featureless: instead, a sharp, medium-
intensity band is found at much lower energy (1644
cm-1), suggesting the presence of a vinylidene (or
indenylidene) ligand. The aromatic region of the 1H
NMR spectrum contains only a first-order splitting
pattern for two equivalent, monosubstituted benzene
rings, ruling out the possibility of isomerization of the
allenylidene ligand to form a 3-phenyl-1-indenylidene4k
derivative (Scheme 2). The remainder of the spectrum
reveals, in addition to the expected envelope for the
dcypb protons from 0.8 to 3.5 ppm, two broad singlets
In this context, installation of an alkylidene ligand
via treatment of “RuHCl(dcypb)” 17 with the propargyl
chloride derivative 3-chloro-3-methyl-1-butyne18 (Scheme
(14) Bustelo, E.; J ime´nez-Tenorio, M.; Puerta, M. C.; Valerga, P. Eur.
J . Inorg. Chem. 2001, 2391.
(15) (a) Touchard, D.; Dixneuf, P. H. Coord. Chem. Rev. 1998, 178-
180, 409. (b) Cadierno, V.; Gamasa, M. P.; Gimeno, J .; Gonzalez-Cueva,
M.; Lastra, E.; Borge, J .; Garcia-Granda, S.; Perez-Carreno, E. Organo-
metallics 1996, 15, 2137. (c) Le Lagadec, R.; Roman, E.; Toupet, L.;
Mu¨ller, U.; Dixneuf, P. H. Organometallics 1994, 13, 5030.
(16) Ohm, M.; Schulz, A.; Severin, K. Eur. J . Inorg. Chem. 2000,
2623.
(12) Selegue, J . Organometallics 1982, 1, 217.
(13) (a) Touchard, D.; Pirio, N.; Toupet, L.; Fettouhi, M.; Ouahab,
L.; Dixneuf, P. H. Organometallics 1995, 14, 5263. (b) Pirio, N.;
Touchard, D.; Toupet, L.; Dixneuf, P. H. J . Chem. Soc., Chem.
Commun. 1991, 980. (c) Bruce, M. I.; Skelton, B. W.; White, A. H.;
Zaitseva, N. N. Inorg. Chem. Commun. 1999, 2, 17.

3338 Organometallics, Vol. 21, No. 16, 2002
Amoroso et al.
Sch em e 3
Sch em e 5
Ta ble 1. P r op er ties of Hyd r id e Sign a ls for 5 a t
194 K a n d 500 MHz
δH
-9.8
-12.6
2H
-20.3
1H
integration
T1, ms
assignt
1H
200
60
450
Ru(µ2,η1-H)
Ru(η2-H2)
Ru(η1-H)
studies, supported by X-ray crystallography, confirm the
formation of 5. This species is also formed on hydrolysis
and hydrogenolysis of the Ru-silylene derivative RuCl-
Sch em e 4
(η3-dcypb)(SiLN2) (SiLN ) 1,3-di-tert-butyl-1,3,2-diaza-
2
silol-2-ylidene).19
31P NMR analysis of 5 (C6D6, 22 °C) reveals two broad
singlets, at 65 and 52 ppm (1:1 ratio; ω1/2 ) 57.8 and
173.7 Hz, respectively). These signals do not couple to
each other (31P-31P COSY) but are correlated by 31P
EXSY NMR. The upfield resonance broadens into the
baseline on cooling to 276 K, while the signal at δP 65
sharpens. Little further change occurs down to 180 K,
at which temperature two broad signals (δP 78, 21)
emerge, each giving an integration of approximately half
the intensity of the resonance still apparent at 63 ppm.
This behavior is characteristic of Ru(H)(PP)(µ-Cl)2(µ-H)-
Ru(PP)(η2-H2) species: directly analogous spectra have
been described for triphenylphosphine and tri-p-tolylphos-
phine complexes.20 The broad upfield peak in these
species is assigned to the “Ru-H2” end of the molecule,
at which the terminal dihydrogen ligand undergoes
rapid exchange with bridging hydride. While the breadth
of this signal precludes observation of a hydride cor-
relation in 1H-31P HMQC experiments,21 1H NMR
analysis of 5 (C7D8, 22 °C) is in excellent agreement with
the literature reports.20 A single hydride resonance
appears (δH -13.8, br s), which on cooling to 194 K
coalesces into three broad resonances (Table 1). No
improvement in resolution occurs down to 170 K.
Integration and T1 values are consistent with the
presence of two hydride ligands and a single Ru(η2-H2)
moiety.
4) is attractive for its mechanistic unambiguity. The
extent of replacement of hydride by chloride offers direct
insight into the probability of homodimerization (path
i), vs paths ii and iii. Furthermore, installation of one
alkylidene concomitant with each chloride ligand means
that formation of monoalkylidene 4c (RuCl(dcypb)-
(µ-Cl)3Ru(dcypb)(CHCHdCMe2)) would necessarily im-
plicate homodimerization. Conversely, if formation of
4a /4b occurs via path iii and dimerization of the
mononuclear complexes RuCl2(PP)(L) is disfavored (as
indeed suggested by the stability6b of 2b,c), we consid-
ered that the propargyl chloride route might afford
access to RuCl2(dcypb)(CHCHdCMe2) (2d ).
Dcyp b Hyd r id es. Exploration of the propargyl chlo-
ride route requires prior synthesis of the hydrido chloro
dimer 5, Ru(H)(dcypb)(µ-Cl)2(µ-H)Ru(dcypb)(η2-H2).17
The latter is cleanly obtained by reaction of 3 with
KHBsBu3 (Scheme 5); disproportionation to a polyhy-
dride does not occur, in contrast to the behavior of
dichlorocarbonyl species [RuCl2(dcypb)(CO)]2.9b Suspen-
sions of 3 react rapidly with 2 equiv of KHBsBu3 in
benzene under 1 atm of H2 (room temperature, <5 min),
yielding a homogeneous orange solution. Detailed NMR
Cr ysta l Str u ctu r e of 5. Several small crystals of 5
formed at the gas-solution interface of a benzene
solution of 5 on storage under H2, one of which was
found suitable for X-ray analysis. An ORTEP represen-
tation appears in Figure 1, with crystallographic and
selected structural parameters in Tables 2 and 3,
respectively. All hydride ligands were located and
refined with a riding model. Complex 5 adopts a
diruthenium structure, in which the two metal centers
(19) Amoroso, D.; Haaf, M.; Yap, G. P. A.; West, R.; Fogg, D. E.
Organometallics 2002, 21, 534.
(20) (a) Hampton, C.; Cullen, W. R.; J ames, B. R.; Charland, J .-P.
J . Am. Chem. Soc. 1988, 110, 6918. (b) Hampton, C. R. S. M.; Butler,
I. R.; Cullen, W. R.; J ames, B. R.; Charland, J .-P.; Simpson, J . Inorg.
Chem. 1992, 31, 5509. (c) Hampton, C.; Dekleva, T. W.; J ames, B. R.;
Cullen, W. R. Inorg. Chim. Acta 1988, 145, 165.
(17) Neither the 14-electron, mononuclear species RuHCl(dcypb) nor
the corresponding 16-electron dimer is attainable, but the H2-stabilized
dimer Ru(H)(dcypb)(µ-Cl)2(µ-H)Ru(dcypb)(η2-H2) (5; cf. 3) provided a
suitable alternative (vide infra).
(18) Wilhelm, T. E.; Belderrain, T. R.; Brown, S. N.; Grubbs, R. H.
Organometallics 1997, 16, 3867.
(21) For successful correlation by HMQC experiments, the peak
width in Hz must be less than the value of the heteronuclear coupling
constant. While exchange masks the 1H-31P coupling within 5, the
2
expected J HP value of 2-4 Hz is very small relative to the peak width
of 40 Hz.

Deactivation of Ru Metathesis Catalysts
Organometallics, Vol. 21, No. 16, 2002 3339
Ta ble 3. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for 5
Ru(1)-P(1)
Ru(1)-P(2)
Ru(1)-H(1)
Ru(1)-H(3)
Ru(1)-Cl(1)
Ru(1)-Cl(2)
Ru(1)-Ru(2)
2.2681(5)
2.3854(5)
1.573(13)
1.768(18)
2.4725(5)
2.4298(5)
2.8595(3)
Ru(2)-P(3)
Ru(2)-P(4)
Ru(2)-H(2)
Ru(2)-H(3)
Ru(2)-Cl(1)
Ru(2)-Cl(2)
2.2543(5)
2.2453(5)
1.559(12)
1.84(2)
2.5531(5)
2.5034(5)
P(1)-Ru(1)-P(2) 100.384(17) P(4)-Ru(2)-P(3)
98.762(19)
Ru(1)-Cl(1)-Ru(2) 69.337(12) Ru(1)-Cl(2)-Ru(2) 70.832(14)
P(1)-Ru(1)-Cl(1) 171.687(17) P(4)-Ru(2)-Cl(2) 166.667(19)
P(2)-Ru(1)-H(3) 173.2
H(1)-Ru(1)-Cl(2) 168.3
Ru(1)-H(3)-Ru(2) 104.6(9)
P(3)-Ru(2)-H(3) 167.1
H(2)-Ru(2)-Cl(1) 169.5
disorder exists. Likewise, the strong trans effect of
hydride, relative to η2-H2, will normally cause lengthen-
ing of the Ru-Cl bond trans to the hydride ligand.
However, the presence in 5 of a statistical distribution
of hydride and dihydrogen ligands trans to µ-Cl results
in a narrower range in bond lengths relative to
those in 6: cf. values of 2.4298(5)-2.5531(5) Å in 5
vs 2.435(2)-2.620(2) Å in 6.
F igu r e 1. ORTEP diagram of Ru(H)(dcypb)(µ-Cl)2(µ-H)-
Ru(dcypb)(η2-H2) (5) with thermal ellipsoids shown at the
30% probability level. Non-hydridic hydrogen atoms and
solvate molecules are omitted for clarity.
Despite formation of a few small crystals of 5 as
described above, repeated efforts to isolate this species
proved unsuccessful. Its high solubility inhibited pre-
cipitation from all solvents investigated, including cold
(-35 °C) pentane. As with other dcypb complexes,
concentration to dryness results in decomposition. We
have described the susceptibility of the parent complex
3 to intramolecular dehydrogenation under vacuum or
argon.5b Intramolecular decomposition also occurs on
stripping solutions of 5 to dryness, as indicated by the
serendipitous crystallization of [Ru(H)(dcypb)(µ-Cl)3-
Ru(dcypb)(N2)] (7), containing an additional chloride
ligand, from the large number of species formed. We
have reported the crystal structure of 7.19 Attempts
to circumvent dehydrogenation by drying 5 under a
stream of H2 were unsuccessful, decomposition to
many products again being indicated by NMR analysis.
Microanalysis carried out on the solid products, on the
possibility that decomposition ensues on redissolution,
gave results consistently low in carbon. As with complex
3, however, 5 can be synthesized and handled without
difficulty in solutions saturated with an appropriate,
stabilizing donor ligand. For all subsequent synthetic
efforts, 5 was therefore prepared and reacted in situ
under H2.
Ta ble 2. Cr ysta l Da ta a n d Str u ctu r e Refin em en t
Deta ils for 5
empirical formula
fw
C68H120Cl2P4Ru2
1334.56
temp
203(2) K
wavelength
cryst syst, space group
unit cell dimens
0.710 73 Å
triclinic, P1h
a ) 11.2119(7) Å
b ) 13.6190(9) Å
c ) 24.6084(16) Å
R ) 77.5280(10)°
â ) 79.2950(10)°
γ ) 68.6670(10)°
3393.7(4) Å3
V
Z, calcd density
abs coeff
2, 1.306 mg/m3
0.656 mm-1
F(000)
1420
cryst size
θ range for data collecn
limiting indices
0.30 × 0.12 × 0.07 mm
1.63-28.73°
-14 e h e 14, -17 e k e 18,
0 e l e 33
no. of rflns collected/unique
completeness to θ ) 28.73°
abs cor
max and min transmissn
refinement method
26 599/15 369 (R(int) ) 0.0187)
87.3%
semiempirical from equivalents
0.928074 and 0.861365
full-matrix least squares on F2
no. of data/restraints/params 15 369/0/694
goodness of fit on F2
1.012
0.0273
0.0385
Ra
b
Rw
Alk ylid en e Der iva tives. Reaction of a homogeneous
orange solution of 5 with 3-chloro-3-methyl-1-butyne
resulted in immediate formation of the dinuclear
monoalkylidene RuCl(dcypb)(µ-Cl)3Ru(dcypb)(CHCHd
CMe2) (4c; Schemes 4 and 6), the identity of which was
confirmed by NMR and crystallographic analysis. In
contrast to 4a /4b, the complex exhibits good solubility,
aiding its spectroscopic characterization. It is isolated
in ca. 80% yield on reprecipitation from benzene-
hexanes, though contaminated with ca. 20% of coproduct
8 (vide infra) which could not be separated by repre-
cipitation or washing. No other products are spectro-
scopically evident. Four 31P{1H} NMR resonances, of
equal integrated intensity, are found for 4c, signifying
the presence of four inequivalent phosphorus nuclei
within a dinuclear complex devoid of symmetry. At room
temperature in C6D6, these resonances are broad,
a
b
2 1/2
R ) ∑||Fo| - |Fc||/∑|Fo|. Rw ) [∑wδ2/∑wFo
] .
have a distorted-octahedral geometry, bridged by two
chloride ligands and one hydride ligand. The diffraction
data indicate 50% site occupancy of two terminal
sites by hydride and η2-H2. The “Ru-H” (terminal)
distances (ca. 1.57(2) Å) are thus intermediate between
values expected for Ru-H and Ru-H2: cf. values of
1.75(3) Å for the Ru-(η2-H2) distance in the chloride-
bridged dimer Ru(η2-H2)(dppb)(µ-Cl)3RuCl(dppb)22 and
1.50(4) Å for the terminal Ru-H distance in Ru(H)-
(PPh3)2(µ-Cl)2(µ-H)Ru(L2)(η2-H2) (6; L2 ) Fe(η5-C5H3-
(CHMeNMe2)PiPr)2-1,2)(η5-C5H5)20), in which no such
(22) J oshi, A. M.; J ames, B. R. J . Chem. Soc., Chem. Commun. 1989,
1785.

3340 Organometallics, Vol. 21, No. 16, 2002
Amoroso et al.
Sch em e 6
Sch em e 7
phosphorus nuclei indicates a higher degree of sym-
metry than in 4c; this does not appear to be due to
averaging of environments through a fluxional process,
as low-temperature 31P NMR studies show no change
(other than minimal peak broadening) from 333 to 203
K. The NMR data are consistent with the square-
pyramidal structure 2d , containing apical alkylidene
and equivalent, cis phosphine ligands. However, this
mononuclear formulation is difficult to reconcile with
the instability toward dimerization implied by formation
of 4c, especially given the persistence of the singlet at
δP 43.4 in NMR spectra recorded over 7 days in solution.
The latter evidence strongly suggests that the species
responsible is not on the reaction path leading to 4c,
but that both it and 4c arise from dimerization of
mononuclear 2d .
An alternative structural possibility is an isomer of
4c, containing a bridging alkylidene ligand (8, Scheme
7), potentially formed by nucleophilic attack of the
electron-rich metal at CR. Precedent for such a rear-
rangement is found in crystallographically characterized
Cp*Ru(µ-Cl)2(µ,η1-CHCHdCPh2)RuCp*, in which the
bridging alkylidene lies perpendicular to the Ru-Ru
vector.24 Observation of a quartet for HR in 8 implies a
similar geometry, in which the plane containing CR and
the two terminal chloride ligands bisects the P-Ru-P
vector, and a near 90° dihedral angle exists between HR
and one 31P nucleus of each dcypb ligand.25 Attempts
to confirm this structure by 13C{1H} NMR analysis were
hampered by the low solubility of 8, but observation of
a poorly resolved multiplet containing five principal
lines at δC 308.6 for CR (cf. a well-resolved triplet for
the corresponding carbon nucleus in 4c) is consistent
unresolved singlets (δ 54.0, 46.1, 45.1, 42.9). On cooling
to 273 K (C7D8), the signal at ca. 43 ppm sharpens to a
doublet (2J PP ) 32 Hz), but the others remain unre-
solved: on cooling further, all four signals broaden until
(245 K) they are nearly lost in the baseline. In CDCl3,
these signals resolve into pairs of doublets (δ 59.0, 43.7
2
2
(AB, J PP ) 39 Hz); δ 46.9, 40.5 (AB, J PP ) 26 Hz)), in
a pattern closely similar (including J values) to those
found for 4a /4b, particularly for the upfield pair of
doublets. The latter resonances are therefore assigned
to the “Cl-end” of the dimer. Reaction with solvent also
occurs, however, affording an unidentified product
characterized by a pair of doublets at δP 46.4, 42.6 (2J PP
) 35 Hz). One of the accessible reaction paths involves
chlorination, as indicated by crystallization of paramag-
netic Ru2Cl5(dcypb)2 (identified by comparison of the
unit cell to that previously5b established) within 48 h
of dissolution in CDCl3.
1H NMR analysis of 4c in C6D6 reveals a quartet for
3
HR of the alkylidene (δH 16.92, q, J ) 11.5 Hz). The
multiplicity of this signal indicates coupling to two
3
3
phosphorus nuclei and Hâ with coincident J HH and J HP
values (as earlier found6b for 2b). Complex 4c is
observed irrespective of reaction time, stoichiometry (2-
20 equiv), or temperature (-35 °C or room temperature),
providing strong evidence for facile homodimerization
of initially formed 2d (Scheme 4, path i). The triene
with an A2B2X spin system, in which J P(A)-C * J P(B)-C
.
Cr ysta l Str u ctu r e of 4c. Crystals of 4c were ob-
tained by slow evaporation of toluene solutions layered
with hexanes. An ORTEP drawing is shown in Figure
2. Structural parameters are collected in Table 4 and
key bond lengths and angles in Table 5. Complex 4c
adopts a triply chloride bridged diruthenium structure,
in which the coordination geometry at each metal center
is distorted octahedral. The structure is unsymmetrical,
with Ru(1) bearing an alkylidene functionality and
Ru(2) a chloride ligand. Complex 4c represents the first
crystallographically characterized example of a di-
nuclear Ru phosphine complex containing a single,
terminal alkylidene moiety; importantly, it also dem-
onstrates the accessibility of the face-bridged geometry
in these systems. One prior report of a diruthenium
monoalkylidene complex has appeared, for which a
doubly chloride bridged structure, RuCl(p-cymene)-
(µ-Cl)2RuCl(PCy3)(CHCHdCMe2), was proposed.26 The
apparent failure of 2b,c to form face-bridged bioctahedra
of type 4 may be due to geometric and steric constraints
1
Me2CdCHCHdCHCHdCMe2 is observed by H NMR
as the organic coproduct.23 The latter “carbene coupling”
product is unlikely to arise from direct reaction of two
free, highly reactive carbenes, which are expected to be
present in low concentration,23a but may rather occur
via attack of carbene on precursor 2d (Scheme 7).
The byproduct 8 also contains an Ru(dcypb)(CHCHd
CMe2) entity, as indicated by observation of a 31P{1H}
1
NMR singlet at δP 43.4, which is correlated in H-31P
HMQC experiments with an alkylidene quartet (HR; δH
3
3
15.83, J HH ) J HP ) 12.0 Hz). The equivalence of the
(23) (a) Given the high reactivity of the carbene (Smith, M. B.;
March, J . March’s Advanced Organic Chemistry, 5th ed.; Wiley: New
York, 1994; p 252), other organic byproducts may also be expected.
However, the major peaks in the olefinic region are due to the complex
AA′BB′ patterns for 2,7-dimethylocta-2,4,6-triene (4E and 4Z isomers).
Chemical shifts agree with values reported (see, for example: tom
Dieck, H.; Keinzel, A. Angew. Chem., Int. Ed. Engl. 1979, 18, 324; Cao,
X.-P.; Chan, T.-L.; Chow, H.-F.; Tu, J . J . Chem. Soc., Chem. Commun.
1995, 1297), while the patterns correspond precisely to those calculated
using the ACD/HNMR software program. For a representation of
calculated and experimental data for this triene obtained in the course
of related work, see ref 23b. (b) Following submission of this paper,
identical dimerization chemistry was confirmed for the useful precursor
complex RuCl2(PPh3)2(CHCHdCMe2): Amoroso, D.; Snelgrove, J . L.;
Conrad, J .; Yap, G. P. A.; Fogg, D. E. Adv. Synth. Catal., in press.
(24) Gagne´, M. R.; Grubbs, R. H.; Feldman, J .; Ziller, J . W.
Organometallics 1992, 11, 3933.
(25) The alkylidene resonance for RuCl2(PCy3)2(CHPh) likewise
appears as a singlet, rather than the triplet found for the PPh3
analogue.4a
(26) Dias, E. L.; Grubbs, R. H. Organometallics 1998, 17, 2758.

Deactivation of Ru Metathesis Catalysts
Organometallics, Vol. 21, No. 16, 2002 3341
Ta ble 6. Ru -Ca ta lyzed ROMP of Nor bor n en ea
entry
cat.
additiveb
conversn (%)
time (min)
1
2
3
4
5
6
7
4c/8
4c/8
4c/8
4a
4a
4b
81
87
100
27
90
99
60
5
5
1140
1200
478
1353
TMS-OTf
PhCHN2
TMS-OTf
TMS-OTf
4b
92
a
Reaction conditions: CDCl3 solvent, [norbornene]:[Ru2] ) 225:
b
1, [Ru2] ) 3.2 mM, room temperature (22 °C). Additives: 3.2 mM
TMS-OTf or PhCHN2.
The Ru-C distance in 4c is similar to that described6b
for mononuclear alkylidene species 2b (1.888 Å for 4c
vs 1.858 Å for 2b). The most significant difference
between the two structures appears in the P-Ru-P
angles, which are ca. 20° greater in 4c, reflecting the
flexibility associated with the four-carbon, vs one-
carbon, diphosphine backbone. Non-alkylidene bond
lengths and angles within the RuCl(PP)(µ-Cl)3Ru(PP)-
(L) skeleton compare well with those reported for RuCl-
(dppb)(µ-Cl)3Ru(dppb)(L)] (L ) dmso,11 H222). Ru-P
bond lengths are very slightly longer for 4c (average
2.29 Å vs 2.28 Å); Ru(µ-Cl)3Ru angles are ca. 3° larger
(average 86.20°), reflecting the higher steric demand of
the dcypb entity.
F igu r e 2. ORTEP diagram of RuCl(dcypb)(µ-Cl)3Ru-
(dcypb)[CHCHdCMe2] (4c) with thermal ellipsoids shown
at the 30% probability level. Hydrogen atoms and solvate
molecule are omitted for clarity.
Ta ble 4. Cr ysta l Da ta a n d Str u ctu r e Refin em en t
Deta ils for 4c
empirical formula
fw
C64H119Cl4P4Ru2
1356.41
temp
203(2) K
wavelength
cryst syst, space group
unit cell dimens
0.710 73 Å
triclinic, P1h
Ca ta lysis. In striking contrast to the parent system
2a , dinuclear 4c/8 exhibits very low metathesis activity,
even toward reactive substrates such as norbornene
(NBE). Thus, while ROMP of 200 equiv of NBE in CDCl3
using 2a is complete before the time of first measure-
ment (5 min, minimum TON 2400 h-1; vide supra), an
identical catalytic run carried out with 4c/8 is still
incomplete after 1 h (Table 6, entry 1; TON 180 h-1).
We attribute this low reactivity to the thermodynamic
stability of the face-bridged structure. We note that
facile bridge cleavage, leading to very high metathesis
activity, has been reported for the edge-bridged dtbpm
dimer.6b High activity can be restored in the present
systems by abstraction of chloride from 4c/8 with
trimethylsilyl triflate (TMS-OTf; 1 equiv per dimer) or
by bridge cleavage via addition of 1.0 equiv of PhCHN2
per dimer,27 installing a second alkylidene functionality.
a ) 11.7069(16) Å
b ) 14.1411(19) Å
c ) 22.140(3) Å
R ) 95.066(3)°
â ) 93.018(2)°
γ ) 112.186(2)°
V
3365.9(8) Å3
Z, calcd density
abs coeff
2, 1.338 mg/m3
0.739 mm-1
F(000)
1438
cryst size
θ range for data collecn
limiting indices
0.12 × 0.10 × 0.01 mm
1.57-20.82°
-11 e h e 11, -14 e k e -14,
0 e l e 22
no. of rflns collected/unique
completeness to θ ) 20.82°
abs cor
max and min transmission
refinement method
26256/7055 (R(int) ) 0.1233)
99.8%
semiempirical from equivalents
0.928 076 and 0.598 612
full-matrix least squares on F2
no. of data/restraints/params 7055/579/667
Complexes 4b and 4a display even lower metathesis
activity, but their very low solubility precludes estab-
lishment of any correlation between alkylidene/cumu-
lenylidene structure and activity. In all cases, the
solutions became highly viscous as the reaction pro-
gressed, and the polymers were insoluble once isolated,
precluding molecular weight determination. These ob-
servations are consistent with slow rates of initiation
relative to propagation, with formation of very high
molecular weight polymers. Such behavior is expected
for 4a /4b, the low solubility of which results in a low
concentration of catalytically active species. Observation
of similar results for 4c/8, however, suggests an intrinsic
initiation barrier, presumably associated with the sta-
bility of the Ru2(µ-Cl)3 moiety. The very low activity of
the dinuclear alkylidene species, coupled with their ease
goodness of fit on F2
1.038
0.0540
0.1076
Ra
b
Rw
a
b
2 1/2
R ) ∑||Fo| - |Fc||/∑|Fo|. Rw ) [∑wδ2/∑wFo
] .
Ta ble 5. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for 4c
Ru(1)-C(1)
Ru(1)-P(1)
Ru(1)-P(2)
Ru(2)-P(3)
Ru(2)-P(4)
Ru(1)-Cl(2)
1.888(8)
2.304(3)
2.320(2)
2.257(3)
2.277(2)
2.471(2)
Ru(1)-Cl(3)
Ru(1)-Cl(4)
Ru(2)-Cl(1)
Ru(2)-Cl(2)
Ru(2)-Cl(3)
Ru(2)-Cl(4)
2.451(2)
2.543(2)
2.398(2)
2.400(2)
2.525(2)
2.532(2)
C(1)-Ru(1)-Cl(4) 170.0(3)
P(4)-Ru(2)-Cl(3)
170.33(9)
96.95(9)
88.43(8)
86.12(7)
84.04(7)
P(2)-Ru(1)-Cl(3)
P(1)-Ru(1)-Cl(2)
P(1)-Ru(1)-P(2)
173.33(8) P(3)-Ru(2)-P(4)
169.43(8) Ru(2)-Cl(2)-Ru(1)
94.31(9) Ru(1)-Cl(3)-Ru(2)
Cl(1)-Ru(2)-Cl(2) 166.89(8) Ru(2)-Cl(4)-Ru(1)
P(3)-Ru(2)-Cl(4) 167.92(7)
(27) The high activity that ensues for 4c/8 in CDCl3 implies that
the chlorocarbon-induced decomposition noted above does not occur
at a rate competitive with metathesis. The higher stability of the dcypb
complexes in C6D6 is offset by the slow rate of metathesis in this
solvent, as previously noted.5
associated with their rigid, bulky chelate rings. On
abstraction of halide from 2b, the edge-bridged bipyra-
midal dimer is formed.6b

3342 Organometallics, Vol. 21, No. 16, 2002
Amoroso et al.
Ru Cl(d cyp b)(µ-Cl)3Ru (d cyp b)(CdCHBu t) (4a ). An or-
ange suspension of 3 (180 mg, 0.14 mmol) in 5 mL of benzene
was treated with 3,3-dimethyl-1-butyne (52.2 µL, 0.32 mmol).
The suspension was stirred at 22 °C for 12 h, over which time
it darkened slightly but did not dissolve. Concentration and
addition of pentane (2 mL) afforded more of the orange solid,
which was filtered off, washed with pentane (3 × 1 mL), and
of formation, points toward a major catalyst deactivation
pathway in this chemistry.
Con clu sion s
The foregoing describes an efficient route to the
hydrido chloro dimer RuCl(dcypb)(µ-Cl)2(µ-H)Ru(dcypb)-
(H2) and utilization of this species and the closely
related chlororuthenium dimer RuCl(dcypb)(µ-Cl)3Ru-
(dcypb)(N2) as atom-efficient entry points into dcypb
chemistry. Dinuclear monoalkylidene, monovinylidene,
and mono(hydroxy)vinylidene derivatives were isolated
on reaction of the parent dimers with an excess of the
appropriate alkyne. In no case were mononuclear or
disubstituted dinuclear derivatives obtained. While
formation of vinylidene and hydroxyvinylidene species
of the type RuCl(dcypb)(µ-Cl)3Ru(dcypb)(L) may be an
artifact of the poor solubility of both the starting dimer
and these products, formation of the analogous alkyl-
idene complex from the soluble hydrido chloro precursor
indicates a strong driving force for homodimerization
of initially formed RuCl2(dcypb)(CHCHdCMe2) (2d ).
The low reactivity of the face-bridged alkylidene prod-
ucts, in conjunction with the facility with which such
dimers are formed, affords insight into an important
deactivation pathway accessible to catalysts of type 2.
We now have evidence that this process is likewise
operative for the Grubbs catalyst RuCl2(PPh3)2(CHCHd
CMe2).23b A related process, possibly involving loss of
one bulky L donor per Ru, may apply to systems of type
1. The accessibility of such deactivation pathways limits
the advantages of enhanced catalyst lifetime anticipated
from use of a robust late-transition-metal catalyst. Our
current efforts focus on development of pseudohalide
analogues of 2, which have potential for enhanced
selectivity as well as improved lifetime. The relative
stability of model edge- and face-bridged complexes is
also under investigation.
1
dried under vacuum. Yield: 161 mg (86%). H NMR (CDCl3):
4
δ 3.06 (t, RudCdCHBut, J HP ) 3.8 Hz, 1H), 3.1-3.2 (br,
aliphatic, 2H), 1.24 (s, C(CH3)3, 9H), 1.0-3.0 (br, aliphatic,
2
102H). 31P{1H} NMR (CDCl3): δ 50.6, 49.6 (ABq, J PP ) 37
2
2
Hz), 45.9 (d, J PP ) 23 Hz), 40.1 (d, J PP ) 23 Hz). 13C{1H}
2
NMR (C6D6): δ 352.1 (t, RuC, J CP ) 16 Hz), 119.7 (s, RuCd
C), 33.8 (s, C(CH3)3), 15-45 (aliphatic). IR (Nujol): ν(CdC)
1633 cm-1. Anal. Calcd for C62H114Cl4P4Ru2: C, 56.10; H, 8.66.
Found: C, 56.35; H, 9.04.
R u Cl(d cyp b )(µ-Cl)3R u (d cyp b )(CdCH C(OH )P h 2) (4b ).
An orange suspension of 3 (98 mg, 0.075 mmol) and 1,1-
diphenyl-2-propyn-1-ol (34 mg, 0.16 mmol) in 8 mL of benzene
was stirred for 12 h at 22 °C, over which time it darkened to
brown. Hexanes (4 mL) was added, and the brown solid was
filtered off, washed with hexanes (3 × 1 mL), and dried under
1
3
vacuum. Yield: 97 mg (88%). H NMR (C6D6): δ 7.51 (d, J HH
3
) 6.5 Hz, 4H, o-C6H5), 7.28 (t, J HH ) 7.5 Hz, 4H, m-C6H5),
7.18 (t, 3J HH ) 7.1 Hz, 2H, p-C6H5), 4.38 (br, RudCdCH, 1H),
3.99 (br, 1H, xch D2O, OH), 0.8-3.5 (br, aliphatic, 104H).
31P{1H} NMR (C6D6): δ 51.5 (d, J PP ) 38 Hz), 50.1 (d, J PP
)
2
2
38 Hz), 44.7 (d, unresolved), 40.9 (d, unresolved). 13C{1H} NMR
(C6D6): δ 308.7 (t, RuC, unresolved), 243.4 (s, RuCdC), 150.4
(s, RuCdCdC), 149.4-120.2 (aromatic), 53.1-14.0 (aliphatic).
IR (Nujol): ν(O-H) 3483; ν(CdC) 1644 cm-1. Anal. Calcd for
C
71H114Cl4P4Ru2: C, 58.67; H, 8.04. Found: C, 59.07; H, 8.34.
Ru (H)(d cyp b)(µ-Cl)2(µ-H)Ru (d cyp b)(H2) (5). An orange
suspension of 3 (20 mg, 31.4 µmol Ru) in C6D6 (1 mL) was
stirred under 1 atm of H2 for 5 min at 22 °C, following which
a C6D6 solution (1 mL; H2-saturated) of KHBsBu3 (32 µL of a
1.0 M solution in Et2O) was added by cannula. A clear, slightly
darker orange solution formed within minutes. NMR analysis
was carried out by transferring the solution by cannula into a
Teflon-lined screw-cap NMR tube filled with H2. Attempts to
isolate the products, or indeed to handle them in the absence
of H2, resulted in extensive decomposition (NMR). We have
reported the crystal structure of one of the decomposition
products, [Ru(H)(dcypb)(µ-Cl)3Ru(dcypb)(N2)] (7).19 A few small
crystals of 5 formed at the solvent-gas interface over several
days under H2, one of which was found suitable for X-ray
Exp er im en ta l Section
Gen er a l P r oced u r es. All reactions were carried out at
room temperature (22 °C) under N2 using standard Schlenk
or drybox techniques, unless stated otherwise. All reactions
with H2 were carried out under 1 atm pressure. Dry, oxygen-
free solvents were obtained using an Anhydrous Engineering
solvent purification system and stored over Linde 4 Å molec-
ular sieves. CDCl3, C6D6, and toluene-d8 were dried over
activated sieves (Linde 4 Å) and degassed by consecutive
freeze/pump/thaw cycles. RuCl(dcypb)(µ-Cl)3Ru(dcypb)(N2)
(3)5b and phenyldiazomethane28 were prepared as previously
described. Norbornene was purchased from Aldrich and dis-
tilled from sodium under N2. Potassium tri(sec-butyl)boro-
hydride, 3,3-dimethyl-1-butyne, 3-chloro-3-methyl-1-butyne,
1,1-diphenyl-2-propyn-1-ol, and trimethylsilyl trifluoromethane-
sulfonate (TMS-OTf) were purchased from Aldrich and used
1
analysis. H NMR (C6D6): δ 0.6-3.0 (m, aliphatic), -13.8 (s,
H, H2). 31P NMR (C6D6): δ 65.1 (br s), 52.2 (br s). Hydride T1
min (C7D8, H2, 500 MHz, 284 K): 38 ms. IR (Nujol): ν(Ru-H)
2102 (m), 2066 (m) cm-1
.
Ru 2Cl4(d cyp b)2[CHCHdC(CH3)2] (4c + 8). A suspension
of 3 (186 mg, 0.29 mmol Ru) in 10 mL of toluene under H2
was treated with KHBsBu3 (292 µL of a 1.0 M solution in Et2O).
Over 3 h at 22 °C the suspension gave way to a dark orange
solution consisting solely of 5 (NMR evidence). Addition of a
solution of 3-chloro-3-methyl-1-butyne (66 µL, 0.58 mmol) in
2 mL of toluene caused immediate darkening of the solution
to green-brown. After 1 h of reaction, 31P{1H} NMR showed
complete conversion to two products, in a ratio of 4:1. The
major product can be unambiguously identified as RuCl-
(dcypb)(µ-Cl)3Ru(dcypb)(CHCHdCMe2] (4c). The second prod-
uct (8) is proposed to be an isomer containing a bridging
alkylidene (see text). Identical results were obtained on use
of a larger excess of the alkyne (20 equiv) or on carrying out
the reaction at -35 °C. The solution was filtered through Celite
and concentrated. Addition of hexanes resulted in a brown
precipitate, which was filtered off, washed with pentane (5 ×
1 mL), and then reprecipitated from benzene-hexanes.
Yield: 158 mg (82%). Repeated efforts to separate 4c and 8
by reprecipitation from other solvent mixtures (using various
combinations of hexanes, pentane, or 2-propanol with benzene,
1
as received. H NMR (200, 300, or 500 MHz), 31P NMR (121
MHz) and 13C NMR (75 MHz) spectra were recorded on a
Varian Gemini 200, Bruker Avance-300, or Bruker AMX-500
spectrometer. All 2D experiments were carried out on the
Avance-300 instrument. IR spectra were measured on
a
Bomem MB100 IR spectrometer. Microanalyses were carried
out inhouse, using a Perkin-Elmer Series II CHNS/O instru-
ment, and by Guelph Chemical Laboratories Ltd., Guelph,
Ontario, Canada.
(28) Creary, X. Organic Syntheses; Wiley: Toronto, 1990; Vol. VII,
p 438.

Deactivation of Ru Metathesis Catalysts
Organometallics, Vol. 21, No. 16, 2002 3343
toluene, or THF), or extraction were unsuccessful. 1H NMR
computationally stable results of refinement. The structures
were solved by direct methods, completed with difference
Fourier syntheses, and refined with full-matrix least-squares
procedures based on F2. Cocrystallized solvent molecules were
located in the asymmetric units of 4c (half a molecule of
n-hexane) and 5 (two molecules of benzene). The cocrystallized
hexane molecule in 4c is located at an inversion center.
In 5, hydrogen atomic positions H(1), H(2), and H(3) were
located from the difference map using low-angle data (<15°).
On the basis of trans bond distances, H(1) and H(2) were
assigned disordered H/H2 identities with a hydrogen occupancy
of 1.5 each. The bridging hydride ligand H(3) and the 50/50
disordered H/H2 sites H(1) and H(2) were refined with a riding
model. All other non-hydrogen atoms were refined with
anisotropic displacement parameters. All hydrogen atoms on
organic moieties were treated as idealized contributions. All
scattering factors and anomalous dispersion factors are con-
tained in the SHELXTL 5.10 program library.30
3
3
(C6D6): δ 16.92 (q, RuCH, J HH ) 11.5 Hz, J HP ) 11.5 Hz,
3
3
4c), 15.83 (q, RuCH, J HH ) 12.0 Hz, J HP ) 12.0 Hz, 8), 9.1
(m, RuCHCH, 4c and 8), 0.5-3.5 (br, aliphatic, 4c and 8).
13C{1H} NMR (C6D6): δ 308.6 (m, unresolved, RuC, 8), 292.7
2
(t, RuC, J PC ) 15 Hz, 4c). 31P{1H} NMR (C6D6): 4c, δ 54.0
(br s), 46.1 (br s), 45.1 (br s), 42.9 (br s); 8, δ 43.4 (s). 31P{1H}
2
2
NMR (CDCl3): 4c, δ 59.0 (d, J PP ) 39 Hz), 46.9 (d, J PP ) 26
2
2
Hz), 43.7 (d, J PP ) 39 Hz), 40.5 (d, J PP ) 26 Hz); 8, δ 44.4
2
(s). A new, unidentified species (δ 46.4 (d, J PP ) 35 Hz), 42.6
2
(d, J PP ) 35 Hz)) is also present in CDCl3, probably owing to
5b
reaction with the solvent, as crystals of Ru2Cl5(dcypb)2 form
on longer standing. IR (Nujol): ν(CdC) 1578 cm-1. Anal. Calcd
for C61H112Cl4P4Ru2 (4c + 8): C, 55.78; H, 8.60. Found: C,
55.08; H, 8.33. Crystals of 4c were obtained by layering a
toluene solution with hexanes.
Gen er al P r ocedu r e for P olym er ization of Nor bor n en e.
A solution of norbornene (34 mg, 0.36 mmol) in 299 µL of
CDCl3 was added to an NMR tube containing a CDCl3 solution
of the catalyst (1.6 µmol of RudC; 201 µL of an 8.1 mM stock
solution). The reaction was monitored by 1H NMR. Trimethyl-
silyl triflate or PhCHN2, if required, were added to the
monomer solution prior to addition to catalyst (TMS-OTf, 1.6
µmol, 10 µL of a 0.16 M TMS-OTf solution in CDCl3; PhCHN2,
0.2 µL, 1.6 µmol).
Str u ctu r a l Deter m in a tion of 4c a n d 5. Suitable crystals
were selected, mounted on thin glass fibers using paraffin oil,
and cooled to the data collection temperature. Data were
collected on a Bruker AX SMART 1k CCD diffractometer using
0.3° ω-scans at 0, 90, and 180° in φ. Unit cell parameters were
determined from 60 data frames collected at different sections
of the Ewald sphere. Semiempirical absorption corrections
based on equivalent reflections were applied.29 No symmetry
higher than triclinic was observed for 4c or 5, and solution in
the centrosymmetric option yielded chemically reasonable and
Ack n ow led gm en t. The kind assistance of Prof. Tze-
Lock Chan (Chinese University of Hong Kong) in
providing spectra of an authentic sample of the triene
is gratefully acknowledged. This work was supported
by the Natural Sciences and Engineering Research
Council of Canada, the Canada Foundation for Innova-
tion, and the Ontario Innovation Trust.
Su p p or tin g In for m a tion Ava ila ble: Tables of crystal
data and data collection and refinement parameters, atomic
coordinates, bond lengths and angles, anisotropic displacement
parameters, and hydrogen coordinates for 4c and 5. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OM0110888
(29) Blessing, R. Acta Crystallogr. 1995, A51, 33.
(30) Sheldrick, G. M. Bruker AXS, Madison, WI, 1997.