Degradation of the first-generation Grubbs metathesis catalyst with primary alcohols, water, and oxygen. Formation and catalytic activity of ruthenium(II) monocarbonyl species

Article abstract of DOI:10.1021/om0208218

The first-generation Grubbs metathesis catalyst (PCy₃)₂2Ru(=CHPh)(Cl)₂ (1) was reacted with primary alcohols at 70 °C to give the monohydride species (PCy₃)₂(CO)Ru(Cl)(H) (4) in ～20% yield. Addition of either an inorganic or organic base greatly facilitated the formation of the hydride, and 4 could be isolated in >70% yield in the case of 1-propanol and triethylamine. Analysis of the reaction products and labeling experiments revealed that 4 was formed via a noncatalytic alcohol dehydrogenation pathway. When benzyl alcohol was used, (PCy₃)₂(CO)Ru(Cl)(Ph) (5) was formed, this also being one of the products of the decomposition of 1 by oxygen in both solid and solution states. Complexes 4 and 5 were found to be active and selective alkene double-bond isomerization catalysts.

Full text of DOI:10.1021/om0208218

Organometallics 2003, 22, 1089-1095
1089
Degr a d a tion of th e F ir st-Gen er a tion Gr u bbs Meta th esis
Ca ta lyst w ith P r im a r y Alcoh ols, Wa ter , a n d Oxygen .
F or m a tion a n d Ca ta lytic Activity of Ru th en iu m (II)
Mon oca r bon yl Sp ecies
Maarten B. Dinger and J ohannes C. Mol*
Institute of Molecular Chemistry, Faculty of Science, Universiteit van Amsterdam,
Nieuwe Achtergracht 166, 1018 WV Amsterdam, The Netherlands
Received October 2, 2002
The first-generation Grubbs metathesis catalyst (PCy3)2Ru(dCHPh)(Cl)2 (1) was reacted
with primary alcohols at 70 °C to give the monohydride species (PCy3)2(CO)Ru(Cl)(H) (4) in
∼20% yield. Addition of either an inorganic or organic base greatly facilitated the formation
of the hydride, and 4 could be isolated in >70% yield in the case of 1-propanol and
triethylamine. Analysis of the reaction products and labeling experiments revealed that 4
was formed via a noncatalytic alcohol dehydrogenation pathway. When benzyl alcohol was
used, (PCy3)2(CO)Ru(Cl)(Ph) (5) was formed, this also being one of the products of the
decomposition of 1 by oxygen in both solid and solution states. Complexes 4 and 5 were
found to be active and selective alkene double-bond isomerization catalysts.
In tr od u ction
uct were not reported, its crystal structure was deter-
mined and assigned as the Ru(IV) dihydride (PCy3)2-
Ru(Cl)2(H)2 (“3”). The ruthenium appeared to occupy an
octahedral coordination environment, with the chlorine,
phosphine, and hydride pairs all mutually trans to their
counterparts. The structure was especially remarkable
in light of a previously characterized analogue of “3”,
(PPri)2Ru(Cl)2(H)2, which was found to have a distorted-
square-antiprismatic geometry.5
The olefin metathesis catalyst systems 1 and 2a have
attracted considerable attention over the past few years,
largely as a consequence of their excellent functional
group tolerance, ease of use, and relatively high catalytic
activity.1 Metathesis in the presence of hybrids of 1 and
We were intrigued by the formation of complex “3”
during the synthesis of 2b, since we have not observed
such a species in any of several preparations of 2b,
despite using very similar reaction protocols.3 Addition-
ally, we wondered if this material, as proposed by
Fu¨rstner,4 was indeed a candidate for the hitherto
unknown compound(s) responsible for the double-bond
isomerization seen in some of the reactions using
ruthenium metathesis catalysts.3,6 While various hy-
dride complexes have been prepared by reaction of 1
with dihydrogen,7 “3” had clearly formed without dihy-
drogen as a reagent. Grubbs has recently described the
formation of the square-pyramidal monohydride com-
plex (PCy3)2(CO)Ru(Cl)(H) (4) from the thermal decom-
position of (PCy3)2Ru(dCHOEt)(Cl)2 (eq 1),8 and we
wondered if there was more than a merely superficial
resemblance of this reaction and the byproduct “3”.
2a bearing solubilizing functional groups can even be
carried out in protic media, such as alcohols and water.2
Recently, we prepared and characterized the second-
generation ruthenium complex 2b, studied its olefin
metathesis activity, and compared this with that of 1
and 2a .3 Complex 2b had also been published by
Fu¨rstner, but its isolation was reported to be hindered,
owing to the formation of a byproduct.4 Although
experimental and spectroscopic details for this byprod-
(5) Gru¨nwald, C.; Gevert, O.; Wolf, J .; Gonza´lez-Herrero, P.; Werner,
H. Organometallics 1996, 15, 1960-1962.
(6) (a) Bourgeois, D.; Pancrazi, A.; Nolan, S. P.; Prunet, J . J .
Organomet. Chem. 2002, 643-644, 247-252. (b) Fu¨rstner, A.; Thiel,
O. R.; Ackermann, L.; Schanz, H.-J .; Nolan, S. P. J . Org. Chem. 2000,
65, 2204-2207.
* To whom correspondence should be addressed. Fax: +31 20
5256456. E-mail: jcmol@science.uva.nl.
(1) (a) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-
29. (b) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413-4450.
(2) (a) Connon, S. J .; Blechert, S. Bioorg. Med. Chem. Lett. 2002,
12, 1873-1876. (b) Ro¨lle, T.; Grubbs, R. H. Chem. Commun. 2002,
1070-1071. (c) Lynn, D. M.; Mohr, B.; Grubbs, R. H.; Henling, L. M.;
Day, M. W. J . Am. Chem. Soc. 2000, 122, 6601-6609.
(3) Dinger, M. B.; Mol, J . C. Adv. Synth. Catal. 2002, 344, 671-
677.
(7) (a) Fogg, D. E.; Amoroso, D.; Drouin, S. D.; Snelgrove, J .; Conrad,
J .; Zamanian, F. J . Mol. Catal. A: Chem. 2002, 190, 177-184. (b)
Drouin, S. D.; Zamanian, F.; Fogg, D. E. Organometallics 2001, 20,
5495-5497. (c) Drouin, S. D.; Yap, G. P. A.; Fogg, D. E. Inorg. Chem.
2000, 39, 5412-5414. (d) Oliva´n, M.; Caulton, K. G. Inorg. Chem. 1999,
38, 566-570. (e) Sutton, A. E.; Seigal, B. A.; Finnegan, D. F.; Snapper,
M. L. J . Am. Chem. Soc. 2002, 124, 13390-13391.
(4) Fu¨rstner, A.; Ackermann, L.; Gabor, B.; Goddard, R.; Lehmann,
C. W.; Mynott, R.; Stelzer, F.; Thiel, O. R. Chem. Eur. J . 2001, 7, 3236-
3253.
(8) Louie, J .; Grubbs, R. H. Organometallics 2002, 21, 2153-2164.
10.1021/om0208218 CCC: $25.00 © 2003 American Chemical Society
Publication on Web 01/30/2003

1090 Organometallics, Vol. 22, No. 5, 2003
Dinger and Mol
even with the added bases; the very low yields of hydride
detected are possibly attributable to adventitious water
in the reagents. In all instances, the heating of the
reaction proved beneficial for both the high conversion
of 1 and moderate yields of the hydride, although when
bases were used the hydride was generated even at
room temperature.
The isolated hydride complex was a bright yellow
powder and was air stable in the solid state for a few
hours before noticeable decomposition occurred; solid
samples could be easily handled in air. Solutions of the
hydride under nitrogen were thermally stable, but
exposure to air, even at room temperature, resulted in
their rapid decomposition.
We decided to investigate these curious results fur-
ther and have discovered that 4 can also be synthesized
by reaction of 1 with primary alcohols in the absence of
dihydrogen. We now report the results of these studies.
While full spectroscopic data for the dihydride com-
plex “3” have not been published, the hydride peak in
1
the H NMR was reported to appear at -24.38 ppm.4
We had therefore initially assigned our ruthenium
hydride species as the dihydride complex “3”. However,
certain aspects of this designation were found to be
1
inconsistent. The integration of the H NMR spectrum
of our complex clearly showed a hydride to tricyclohexy-
lphosphine ratio of 1:2, as opposed to the 1:1 expected
for “3”. Furthermore, the elemental analysis (EA) data
for our complex did not fit those required for (PCy3)2-
Ru(Cl)2(H)2 (“3”), giving a carbon percentage that was
too high. Finally, IR data strongly indicated the pres-
ence of a carbonyl group. The NMR, EA, and IR data
together suggested the formula of our hydrido species
was in fact that of the monohydride monochloride
(PCy3)2(CO)Ru(Cl)(H) (4).
Resu lts a n d Discu ssion
When catalyst 1 was reacted with an excess of
methanol in toluene at 70 °C for 2 days, the initially
purple-red solution gradually became clear dark orange.
The progress of the reaction was monitored by 31P NMR,
which showed the slow disappearance of the signal at
37.3 ppm from 1 concomitant with the emergence of a
new peak at 47.5 ppm, among others.9 Similarly, in the
1H NMR, the benzylidene resonance from 1 at 20.6 ppm
was gradually replaced by a hydride signal at -24.3
ppm. Whereas ethanol and 1-propanol were also found
to readily generate the hydride, 2-propanol and water
(vide infra) were found to be ineffective and only other,
unknown, products were observed in the 31P and 1H
NMR spectra.
Fogg et al. have recently suggested that complex 4
might be formed as a byproduct from the stoichiometric
hydrogenolysis
of
1
with
methanol/dichloro-
methane in the presence of triethylamine, although it
was not isolated.7a,b Our results verify this observation,
but it is noteworthy that hydrogen is in fact not
necessary for the formation of 4. The species 4 is a
relatively well-known system that has been previously
reported and crystallographically characterized by both
Grubbs8 and Nolan10 and can be readily synthesized
from the reaction of [(cod)Ru(Cl)2] (cod ) 1,5-cycloocta-
diene), PCy3, and ethanol.11 The reported spectroscopic
properties of 48,11 were indistinguishable from those
obtained from our samples generated by the degradation
of 1.
Addition of a suitable base (organic or inorganic) was
found to greatly facilitate the hydride-forming reaction,
in terms of both rate of formation and yield. When 1
1
equiv of sodium methoxide or sodium hydroxide or
/
2
Mech a n ism for th e F or m a tion of 4. Notwithstand-
ing that a number of aspects regarding the decomposi-
tion of catalyst 1 have been studied in some detail,12
the identity of the approximately eight inorganic phos-
phorus-containing species observed in the 31P NMR
have not yet been elucidated. Therefore, the mechanism
for the reaction of 1 with an alcohol to form 4 is of
interest, since it may shed some light on a possible
decomposition pathway for 1, particularly when primary
alcohols and/or water are present.
equiv of potassium carbonate, dissolved in methanol,
was added to a toluene solution of 1 at 70 °C, the
reaction was complete in 15 min, giving an ∼50% yield
of the hydride by 31P NMR (35% isolated). The use of
an organic base also stimulated hydride formation, and
∼85% yield by 31P NMR (73% isolated) of the hydrido
species was obtained when using 1-propanol together
with triethylamine. In this case, however, the reaction
rate was significantly lower, with ∼24 h needed for the
complete consumption of 1. Interestingly, in the pres-
ence of triethylamine, water could be used in lieu of a
primary alcohol, although only a modest amount (∼30%)
of hydride was formed in this case. 2-Propanol remained
essentially unreactive with regard to hydride formation
We wondered what the actual source of the hydride
atom was: the alcohol group, the benzylidene moiety,
and alcohol dehydrogenation13 were all considered.
(10) Lee, H.-M.; Smith, D. C.; He, Z.; Stevens, E. D.; Yi, C. S.; Nolan,
S. P. Organometallics 2001, 20, 794-797.
(9) Other significant resonances in the 31P NMR spectrum were two
single peaks at 36.1 and 25.9 ppm, together with multiple peaks around
73 ppm. Some free PCy3 (10.5 ppm) was also evident. The NMR yield
of 4 was ∼30%.
(11) Yi, C. S.; Lee, D. W.; Chen, Y. Organometallics 1999, 18, 2043-
2045.
(12) Ulman, M.; Grubbs, R. H. J . Org. Chem. 1999, 64, 7202-7207.

Ruthenium(II) Monocarbonyl Species
Organometallics, Vol. 22, No. 5, 2003 1091
Sch em e 1
Initially we postulated that the alcohol group might
add to the benzylidene in 1, eliminating benzene, among
other products. The resulting intermediate species could
then eliminate RCl, in a fashion similar to that recently
described by Grubbs for the decomposition of ruthenium
Fischer-type carbene complexes of the type (PCy3)2Ru-
(dCHOR)(Cl)2.8,14 However, Fischer-type carbene com-
plexes show good stability;8 still, we had not observed
this intermediate species in the NMR spectra of the
decomposition reaction mixtures. Furthermore, GC
analysis of the reaction mixture when 1-nonanol was
used as an alcohol source revealed that no chlorononane
was formed. In fact, octane was detected, this strongly
suggesting that alcohol dehydrogenation, possibly by
partially decomposed 1, was the primary source of the
hydride moiety in 4. The dehydrogenation of alcohols
provides a source of aldehydes, which can react with
ruthenium species via decarboxylation to give carbonyl
and hydride moieties.13 The continuing presence of the
relatively small excess of alcohol (5-10 equiv), even
after prolonged reaction times, showed that the dehy-
drogenation was noncatalytic. To add further insight
into the mechanism for the reaction of 1 with alcohols,
a number of additional experiments were carried out.
To trace the origin of the hydride group in 4, this
being most likely from the CH2 group adjacent to the
alcohol moiety,13a,b we reacted 1 with CH3CH2OD. As
predicted, predominantly nondeuterated 4 formed, al-
though some deuterated material, (PCy3)2(CO)Ru(Cl)-
(D) (4-d ), was also present. The formation of some 4-d
could be explained by the observation that the hydride
moiety in isolated samples of 4 readily underwent
deuterium exchange (to give 4-d ) when treated with
excess CH3CH2OD under the reaction conditions used
for the generation of 4 from 1.
We also analyzed the organic byproducts of the
various degradation experiments by NMR and GC/MS.
The composition of the nonvolatile reaction residues was
found to be very complex, due to the presence (and
decomposition) of unknown ruthenium compounds, and
the only product that could be unambiguously identified
(by comparison of its spectral data with that of authentic
samples) was cis- and trans-stilbene (ratio ∼90:10).
While stilbene was also found to be a thermal decom-
position product of 1 by Grubbs,12 we were not wholly
convinced that this was the only fate of the benzylidene
group. Therefore, the deuterated analogue of the Grubbs
catalyst, (PCy3)Ru(dCDPh)(Cl)2 (1-R-d ), was synthe-
sized and reacted with methanol (10 equiv) in the
presence of triethylamine (10 equiv), and the reaction
products were subsequently analyzed with 2H NMR
spectroscopy.
spectrum, indicating that some exchange had taken
place. This was later verified in a H NMR experiment,
1
where we clearly observed gradual loss in the deuterium
enrichment of the HR.15 However, exchange with the
alcohol was not the main fate of the deuterium label.
2
The dominant feature of the H NMR spectrum was
a resonance at 7.2 ppm (∼25%). We assigned this signal
predominantly to C6H5D and possibly some trans-
styrene-d2 (7.1 ppm), although the cis form (6.6 ppm)
was not detected. Interestingly, the remainder of the
deuterium atoms (∼60%) were present in a number of
products showing chemical shifts ranging from 2.2 to
0.9 ppm, with a peak at 2.1 ppm being especially
prominent. GC of the crude reaction mixture showed
that toluene was a major reaction product; therefore,
we assigned the identity of this latter signal as PhCH2D.
GC also verified the formation of benzene as one of the
reaction products.
The beneficial role of added bases was clarified by
fractional crystallization of the reaction residue, which
resulted in colorless crystals of triethylamine hydro-
chloride (characterized by NMR). The formation of
triethylamine hydrochloride also identifies the fate of
the “lost” chlorine.
Additional proof that alcohol dehydrogenation was
responsible for the formation of 4 was provided by using
1-13C-labeled ethanol. Thus, when 1 was reacted with
CH313CH2OH in the usual way, 13C-carbonyl-labeled 4
((PCy3)(13CO)Ru(Cl)(H)) resulted. This was determined
by 13C NMR, which showed in a very short period of
time only a triplet attributable to the carbonyl group.
1
Moreover, the H NMR showed the hydride moiety as
a complicated multiplet (as opposed to a simple triplet),
confirming the presence of the 13C label adjacent to the
hydride.
Unfortunately, from these results, the mechanism by
which 4 forms from 1 could still not be unambiguously
elucidated. Particularly, the role of the benzylidene
group is not fully clear. Scheme 1 shows a proposed
mechanism that is consistent with the observations for
the degradation of 1 by methanol.
The mechanism presented in Scheme 1 cannot, how-
ever, be immediately extended to include the longer
chain alcohols that were found to be equally effective
at forming 4. Scheme 1 would predict that longer chain
alcohols should give either an alkyl benzene in steps 3
and 4 or, alternatively, an alkyl complex (step 5) as
Confirming that HR from the benzylidene is not a
source for the hydride in 4, we observed no detectable
amounts of 4-d when reacting 1-R-d with methanol.
Noteworthy is the fact that we did observe small
2
amounts (∼12% of total D) of CH3OD in the H NMR
(13) (a) Chen, Y.-Z.; Chan, W. C.; Lau, C. P.; Chu, H. S.; Lee, H. L.
Organometallics 1997, 16, 1241-1246. (b) Portnoy, M.; Frolow, F.;
Milstein, D. Organometallics 1991, 10, 3960-3962. (c) Shinoda, S.;
Itagaki, H.; Saito, Y. J . Chem. Soc., Chem. Commun. 1985, 860-861.
(d) J ung, C. W.; Garoou, P. E. Organometallics 1982, 1, 658-666. (e)
Esteruelas, M. A.; Werner, H. J . Organomet. Chem. 1986, 303, 221-
231.
(14) Wu, Z.; Nguyen, S. T.; Grubbs, R. H.; Ziller, J . W. J . Am. Chem.
Soc. 1995, 177, 5503-5511.
(15) Lynn, D. M.; Grubbs, R. H. J . Am. Chem. Soc. 2001, 123, 3187-
3193.

1092 Organometallics, Vol. 22, No. 5, 2003
Dinger and Mol
opposed to a hydride. However, regardless of the alkyl
alcohol used, the benzylidene group was still found to
be fully hydrogenated to toluene and neither substituted
benzenes (e.g. such as ethyl benzene when ethanol was
reacted) nor alkyl complexes were formed.
1 dissolved in toluene with both water and dry air at
80 °C. While the reaction of 1 with water was relatively
slow, after 3 days a single phosphorus-containing spe-
cies, showing a 31P NMR peak at 49.2 ppm, had formed.
The identity of this compound has not yet been deter-
mined, but it showed no hydride or carbene resonances
in the 1H NMR. Conversely, when a small portion of
dry air was introduced to a solution of 1, the solution
rapidly became dark brown. 31P NMR of the crude
mixture showed a broad peak at 46.8 ppm (possibly due
to OdPCy3) and a sharp peak at 25.9 ppm, assigned to
complex 5. Fortuitously, while 5 is insoluble in metha-
nol, all of the other reaction products dissolved, this
allowing the isolation of 5 from the reaction (6% yield)
and thereby confirming its formation. When excess
triethylamine was added, higher yields of 5 (∼15%) were
observed.
Furthermore, Scheme 1 also does not explain the
modest success of using water as a source of hydride
atoms. In this case, we speculated that the water reacts
with the benzylidene group to form a primary alcohol
in situ, which in turn reacts with the ruthenium
fragment to produce 4. It seemed to us that the most
likely alcohol to form from the reaction of 1 with water
is benzyl alcohol. To test this idea, 1 was reacted directly
with benzyl alcohol in the presence of triethylamine.
However, instead of forming the hydride 4, the new
complex (PCy3)2(CO)Ru(Cl)(Ph) (5) was produced in 54%
isolated yield. This result was slightly surprising, since
the longer chain alkyl alcohols clearly had not resulted
in alkyl complexes. Nevertheless, in the benzyl alcohol
case at least, the moiety on the R-carbon of the alcohol
is transferred to the metal center (step 5, Scheme 1) and
not to the benzylidene group.
Remarkably, the reaction of oxygen with 1 also
occurred for solid samples. In fact, the reaction is much
more efficient, due to the suppression of decomposition
from the oxidation of the phosphine ligands which are
labile in solutions of 1 but not in solid samples. Thus, a
sample that was stored at 4 °C in air for a prolonged
period of time (over 4 years) slowly developed a peak at
25.9 ppm in the 31P NMR spectrum (24% conversion).
Complete conversion of 1 to 5 can be achieved overnight
when 1 is pressurized (40 bar) with pure oxygen at 60
°C, in 75% yield.19 Whereas addition reactions of
electron-rich Schrock-type carbenes with the other
chalcogens (sulfur, selenium, and tellurium) are known,20
to the best of our knowledge, reaction with oxygen as
the electrophile is unprecedented. While the mechanistic
details of the reaction of 1 with oxygen are not yet clear,
we presume the carbene ligand in 1 is first “oxygenated”
concomitant with the loss of HCl to initially give the
benzoyl complex (PCy3)2(Cl)RuC(O)Ph. This intermedi-
ate species could then deinsert CO to give 5.21
Species closely related to 5 have been synthesized
previously; bis-PBut Me16 and various bis-PPh3 ana-
17
2
logues are known. Unsurprisingly, we found that com-
plex 5 can be readily synthesized in 75% yield following
the literature methodology used for the synthesis of the
related compounds,16 by reaction of complex 4 with
diphenylmercury. In contrast to 4, solid 5 was com-
pletely air stable, and even exposed solutions decom-
posed only very slowly.
Of particular interest is that complex 5 gives its 31P
NMR resonance at 25.9 ppm. We immediately recalled
that this is exactly the same chemical shift observed for
a species in decomposed samples of 1 and, indeed, is
present in many of the crude reaction mixtures when 1
is involved, including metathesis.18 We wondered if
complex 5 is in fact also a decomposition product of
complex 1. To verify this, we measured the IR spectra
of a number of different reaction residues showing a
peak at 25.9 ppm and, as expected, a strong band at
1894 cm-1 was observed, the same location as for 5.
Careful comparison of the aromatic region of the 1H
NMR lent additional evidence, by way of the presence
of two triplets at 6.65 and 6.59 ppm, the same positions
seen in 5. While we were unable to isolate the complex
from the other reaction products, the 31P NMR, 1H
NMR, and IR data taken together provide very strong
evidence that one of the major decomposition products
of 1 could be 5.
An a lysis of th e F or m a tion of “3”. While we could
not duplicate the reaction conditions required to produce
the hydride “3” as reported by Fu¨rstner,22 we did
discover that “3” and 4 likely have more in common than
first appears. While examining the crystal structure
data for Fu¨rstner’s reported dihydride species “3”4
together with that of other crystallographically charac-
terized hydride species, we noticed that the structure
of “3” bore a remarkable resemblance to those of the
monohydride complexes 48,10 and (PCy3)2Ru(H)(Cl) (“6”).23
(19) A second, unidentified product that shows a very broad peak
at 50.4 ppm accounts for ∼25% of the total phosphorus content in the
31P NMR spectrum of the crude sample.
We presumed that the source of the oxygen required
for formation of the carbonyl group in 5 arises from
either adventitious water or oxygen present in the
glassware, solvents, and/or reagents. To ascertain which
of these two oxygen sources was responsible, we reacted
(20) (a) Werner, H.; Schwab, P.; Bleuel, E.; Mahr, N.; Windmu¨ller,
B.; Wolf, J . Chem. Eur. J . 2000, 6, 4461-4470. (b) Wolf, J .; Zolk, R.;
Schubert, U.; Werner, H. J . Organomet. Chem. 1988, 340, 161-178.
(c) Hill, A. F.; Roper, W. R.; Waters, J . M.; Wright, A. H. J . Am. Chem.
Soc. 1983, 105, 5939-5940.
(21) (a) Roper, W. R.; Taylor, G. E.; Waters, J . M.; Wright, L. J . J .
Organomet. Chem. 1979, 182, C46-C48. (b) Roper, W. R.; Wright, L.
J . J . Organomet. Chem. 1977, 142, C1-C6.
(16) Huang, D.; Streib, W. E.; Bollinger, J . C.; Caulton, K. G.;
Winter, R. F.; Scheiring, T. J . Am. Chem. Soc. 1999, 121, 8087-8097.
(17) Rickard, C. E. F.; Roper, W. R.; Taylor, G. E.; Waters, J . M.;
Wright, L. J . J . Organomet. Chem. 1990, 389, 375-388.
(18) After metathesis of excess 1-hexene with 1 (i.e. sufficient
1-hexene was used to ensure complete decomposition of 1) at 60 °C
and subsequent vacuum distillation of the product 6-dodecene, 31P
NMR (C6D6) of the residue revealed only five phosphorus-containing
species: δ 50.8 (4%), 46.9 (19%, possibly OdPCy3), 35.0 (19%), 25.9
(12%), and 10.9 (46%, PCy3). 1H NMR showed no hydride- or metal-
carbene-containing compounds.
(22) We have reacted catalyst 1 with most of the reagents (and
combinations thereof) used in the synthesis of 2b, including some of
the possible impurities that can be present in these starting materials.
For example, acetone, tert-butyl alcohol, benzaldehyde, and 1,3-bis-
(2,6-diisopropylphenyl)-4,5-dihydroimidazolinium chloride all failed to
react with 1, under the conditions used for the synthesis of 2b, to
produce hydride complexes. Interestingly, triethyl orthoformate, par-
ticularly in the presence of a base, produced modest amounts of
complex 4.
(23) van der Schaaf, P. A.; Kolly, R.; Hafner, A. Chem. Commun.
2000, 1045-1046, including Supplementary Information.

Ruthenium(II) Monocarbonyl Species
Organometallics, Vol. 22, No. 5, 2003 1093
Therefore, the crystal structure data for complexes “3”,24
4, and “6” were more carefully examined.
The original structure of “6” showed a 14-electron
mononuclear complex in which the ruthenium occupies
a square-planar coordination environment, with the two
phosphines trans to one another.23 However, a later
reevaluation of the structure showed that the compound
had actually formed as the 16-electron dinitrogen adduct
(PCy3)2(N2)Ru(H)(Cl) (7).25 The structure of 4 is very
similar to that of 7, with the carbonyl group simply
occupying the position of the dinitrogen molecule in 7.
A cursory examination of the unit cell parameters of the
Van der Schaaf structure of 7, the Nolan and Grubbs
structures of 4, and the Fu¨rstner structure of “3” reveals
that there is a very close similarity in the values.
Indeed, closer scrutiny of the three hydride structures
showed that all of the corresponding atomic coordinates
are in each case virtually identical: i.e., “3”, 4, and 7
appear to be isostructural. The disparity of the struc-
tural data for “3” seems to lie solely in the modeling of
the chlorine atom(s). Like “3”, the structures of 7 and 4
are centrosymmetric with only one unique phosphine
and chlorine in the asymmetric unit. The chlorine atom
and dinitrogen (for 7) and carbonyl (for 4) molecules are
disordered, each with the required site occupancy of
50%.10,23,25 Examination of the structure of “3” revealed
that the temperature factors for the full-occupancy
unique chlorine (in the asymmetric unit) are dispropor-
tionately large relative to the rest of the atoms. The
incorrect modeling of the chlorine atom as full occupancy
in “3” and the absence of the dinitrogen/carbonyl
molecule was made more successful than would nor-
mally have been the case, due to the very similar
additional electron density resulting from their super-
position.
F igu r e 1. Conversion as a function of reaction tempera-
ture for the isomerization of neat 1-octene (88 000 mol
equiv) in the presence of catalyst 4.
attention, since it was shown to be a potent catalyst for
the hydrogenation of alkenes10,28 and polybutadiene
rubber,29 while the bis-PPri3 system has been exploited
in a wide range of mostly stoichiometric reactions.30
However, small-alkene isomerization reactions using 4
or its analogues have, to the best of our knowledge, not
previously been specifically examined.
We therefore decided to determine the efficacy of 4
in the double-bond isomerization of 1-octene at various
temperatures. We found that the use of an additional
solvent was unnecessary and that surprisingly low
catalyst loadings could be used. Figure 1 graphically
depicts the results of this study.
It is clear that complex 4 is a capable double-bond
isomerization catalyst. No metathesis products were
detected in any of the reactions using 4. Furthermore,
the catalyst showed a high degree of selectivity toward
the formation of 2-octene, even when high conversions
had been attained. Reaction of 88 000 mol equiv of
1-octene with 4 at 100 °C gave 97% conversion with 92%
selectivity for 2-octene after 3 h. At 120 °C, selectivity
was compromised, decreasing to only 56% after 3 h. The
cis:trans ratio of the 2-octene formed was found to be
independent of the reaction temperature, consistently
being 30:70.
We wished to examine the selectivity as a function of
reaction time for 4, since we noticed that the selectivity
was significantly reduced after long reaction times,
particularly at higher temperatures. Figure 2 shows the
conversion and selectivity for 2-octene as a function of
time at 100 °C.
From this information it is our contention that the
crystal structure reported by Fu¨rstner for “3” may have
been incorrectly assigned, and the complex’s real iden-
tity is that of either 4 or 7. Of the two, we believe that
4 is the more probable, given the identical chemical
shifts of the hydride in the 1H NMR spectrum,26
together with its likely origins of formation revealed in
the present article.
Isom er iza tion Activity of 4. Fu¨rstner suggested
that hydrides such as 4 (or 7) might be responsible, at
least in part, for the carbon-carbon double-bond isomer-
ization sometimes seen as byproducts of metathesis
reactions.4 This proposition is not unfounded; a number
of ruthenium complexes, particularly hydrides, have
been found to exhibit catalytic double-bond isomeriza-
tion activity toward olefins.7e,27 It should be noted that
To further test the selectivity of 4, we investigated
the double-bond isomerization of trans-2-octene at high
temperatures. Using 20 000 mol equiv relative to cata-
lyst, we saw only 1% conversion to 3-octene at 80 °C
after 24 h. At higher temperatures more isomerization
was observed: viz. 6% and 33% for 100 and 120 °C,
respectively. The relatively low 3-octene formation
implies that 2-octene does not easily decompose the
4 and its bis-PPri analogue are extremely versatile
3
reagents; 4 has been recently receiving increasing
(24) Cambridge Crystallographic Data Centre, Supplementary Pub-
lication No. CCDC-154 932.
(25) van der Schaaf, P. A.; Kolly, R.; Hafner, A. Chem. Commun.
2001, 940. For the web edition, see the “Corrections/Additions”
amendment for ref 19.
(26) The chemical shift of the hydride in 7 is -17 ppm: van der
Schaaf, P. A. Personal communication.
(27) For some homogeneous examples see: (a) Cook, S. L.; Evans,
J . J . Chem. Soc., Chem. Commun. 1983, 713-715. (b) Ewing, D. F.;
Hudson, B.; Webster, D. E.; Wells, P. B. J . Chem. Soc., Dalton Trans.
1972, 1287-1293. (c) Vaglio, G. A.; Osella, D.; Valle, M. Transition
Met. Chem. 1977, 2, 94-97. (d) Castiglioni, M.; Giordano, R.; Sappa,
E. J . Organomet. Chem. 1987, 319, 167-181. (e) Krompiec, S.; Pigulla,
M.; Bieg, T.; Szczepankiewicz, W.; Kuz´nik, N.; Krompiec, M.; Kubicki,
M. J . Mol. Catal. A: Chem. 2002, 189, 169-185.
(28) (a) Yi, C. S.; Lee, D. W.; He, Z.; Rheingold, A. L.; Lam, K.-C.;
Concolino, T. E. Organometallics 2000, 19, 2909-2915. (b) Yi, C. S.;
Lee, D. W. Organometallics 1999, 18, 5152-5156.
(29) Martin, P.; McManus, N. T.; Rempel, G. L. J . Mol. Catal. A:
Chem. 1997, 126, 115-131.
(30) (a) Werner, H.; Stu¨er, W.; Weberndo¨rfer, B.; Wolf, J . Eur. J .
Inorg. Chem. 1999, 1707-1713. (b) Esteruelas, M. A.; Go´mez, A. V.;
Lahoz, F. J .; Lo´pez, E. O.; Oro, L. A. Organometallics 1996, 15, 3423-
3435. (c) Esteruelas, M. A.; Oro, L. A. Organometallics 1993, 12, 2377-
2379. (d) Esteruelas, M. A.; Sola, E.; Oro, L. A. J . Mol. Catal. 1989,
53, 43-52. (e) Esteruelas, M. A.; Sola, E.; Oro, L. A. J . Mol. Catal.
1988, 45, 1-5.

1094 Organometallics, Vol. 22, No. 5, 2003
Dinger and Mol
likely that this reaction led to the formation of the
hydride that was inadvertently isolated by Fu¨rstner in
the synthesis of 2b, quite possibly during the methanolic
workup.4
We have also shown that complex 5 is formed from 1
in the presence of oxygen (air), and this could also have
repercussions with regard to isomerization side reac-
tions. Because 5 was also found to be a capable isomer-
ization catalyst, secondary metathesis products, which
form as a result of isomerization, might be anticipated
in metathesis reactions that are not completely free of
oxygen, particularly when higher temperatures are
used.
We are currently investigating the reaction of the
“second-generation” Grubbs catalyst 2a with alcohols.
Preliminary results indicate that 2a also readily forms
hydride(s), with the expected mixed-ligand system.
F igu r e 2. Conversion and selectivity for 2-octene as
functions of reaction time for the isomerization of neat
1-octene (88 000 mol equiv) by 4 at 100 °C.
catalyst, probably because it reacts only poorly with this
substrate. In no instance was 1-octene observed; thus,
the isomerization reaction with 4 appears to be ef-
fectively irreversible.
Exp er im en ta l Section
From the data, we conclude that the decrease in
selectivity over time is largely attributable to the
decomposition of the active catalyst. NMR analysis of
the spent catalyst revealed a single broad peak in the
31P NMR spectrum at 47.2 ppm, with a shoulder at 47.5
ppm, while the 1H NMR showed the absence of any
hydrido species. Interestingly, the peak(s) in the 31P
NMR spectrum of decomposed 4 are also prominent in
the spectrum of exhausted 1 that we recorded after
activity in the metathesis of 1-octene had ceased.31
Whether the decomposition products are indeed the
same for 1 and 4 remains presently unknown. In any
case, they are most likely poorly defined ruthenium
complexes, possibly polymeric, that can also isomerize
1-octene but lack the activity and selectivity of 4.
We also tested the isomerization activity of complex
5 and found it to be a good precatalyst for this reaction,
although the activity was lower than for 4. We tenta-
tively suggest that the likely active catalytic species,
(PCy3)(CO)Ru(Cl)(H),28a is also formed in situ when 5
is reacted with a terminal alkene.
Unless otherwise stated, all manipulations were carried out
under a nitrogen atmosphere on a vacuum line using standard
Schlenk techniques. All solvents used were dried and distilled
under nitrogen. (PCy3)2Ru(dCHPh)(Cl)2 (1; Fluka), diphen-
ylmercury (Merck), 2-octene (Acros), and 1-octene (98%, Ald-
rich) were obtained from commercial sources. Styrene-R-d was
prepared by the literature procedure.32 NMR spectra of the
complexes were recorded on a Varian Mercury 300 spectrom-
eter, at 300.14, 75.48, and 121.50 MHz for the proton, carbon,
and phosphorus channels, respectively. Elemental analyses
were determined at H. Kolbe Mikroanalytiches Laboratorium.
(P Cy3)2Ru (dCDP h )(Cl)2 (1-r-d ). Styrene-R-d (0.75 mL, 6.5
mmol) was added to a solution of 1 (0.50 g, 0.608 mmol) in
CH2Cl2 (5 mL), and the resulting mixture was stirred for 1 h
at room temperature. Most of the CH2Cl2 was removed under
vacuum, and methanol (20 mL) was then added, producing a
purple solid. After the mixture was stirred for several minutes,
the solid was filtered in air and washed with methanol (10
mL), followed by pentane (5 mL). The product thus obtained
was subjected a second time to the above procedure to give
1-R-d (95% atom D (1H NMR); yield 0.31 g, 62%).
31P{1H} NMR (C6D6): δ 37.5. 1H NMR (C6D6): δ 8.69 (d,
3J HH ) 8.7 Hz, 2H, o-H C6H5), 7.23 (m, 3H, C6H5), 2.85 (s, br,
6H, P(C6H11)3), 1.93 (d, J ) 11.1 Hz, 12H, P(C6H11)3), 1.67-
1.52 (m, 30H, P(C6H11)3), 1.26-1.19 (m, 18H, P(C6H11)3). 13C-
{1H} NMR (C6D6): δ 152.9 (i-C C6H5), 131.2, 129.2 (C6H5), 32.4
(vt, J C,P ) 8.5 Hz, P(C6H11)3), 30.1 (s, P(C6H11)3), 28.1 (vt, J C,P
) 4.3 Hz, P(C6H11)3), 26.9 (s, P(C6H11)3), RudCD not resolved
after 3 days’ acquisition.
Con clu sion s
The reaction of the first-generation Grubbs catalyst
1 with primary alcohols gave, in part, the ruthenium
hydride 423 in modest to good yields. The reaction
pathway was determined to be alcohol dehydrogenation
followed by decarbonylation. Both inorganic and organic
bases were found to facilitate the reaction. When benzyl
alcohol was used in place of an aliphatic alcohol, the
analogous reaction took place to produce the phenyl-
substituted complex 5. Compounds 4 and 5 were found
to be good double-bond isomerization catalysts; in fact,
4 isomerized 1-octene more effectively than 1 is capable
of metathesizing it.3 It is, however, not conclusive if any
hydride species can also form during olefin metathesis
initiated by 1; certainly the carbonyl species 4 could
never form if substrates devoid of an oxygen-containing
functionality are used. The present results indicate that
the use of primary alcohols or water in the synthesis or
reactions of catalyst 1 could potentially lead to the
formation of 4, especially at elevated temperatures and
in the presence of a base (or basic groups). It is highly
(P Cy3)2(CO)Ru (Cl)(H) (4). 1-Propanol (0.1 mL, 1.34 mmol)
and triethylamine (0.17 mL, 0.121 mmol) were added to a
toluene solution (4 mL) of 1 (0.200 g, 0.243 mmol), and the
resulting mixture was heated at 75 °C overnight. During this
time the solution changed from dark purple-red to clear dark
orange-yellow. After the mixture was cooled to room temper-
ature, methanol (10 mL) was added, resulting in the rapid
formation of a yellow precipitate. After this mixture was
stirred for 1 h, the solid was collected by filtration, washed
with methanol (2 × 5 mL) followed by hexanes (5 mL), and
dried under vacuum to give 4 as a bright yellow powder (0.128
g, 73%). The same procedure, but using a methanolic solution
(0.122 M) of K2CO3, NaOH, or NaOMe (0.5 mL, 0.061 mmol)
instead of Et3N, also produced 4 (0.030 g, 35%). Spectra of the
isolated product are indistinguishable from those of an au-
thentic sample.11
(32) Choi, H.-S.; Kuczkowski, R. L. J . Org. Chem. 1985, 50, 901-
902.
(31) Dinger, M. B.; Mol, J . C. Unpublished results.

Ruthenium(II) Monocarbonyl Species
Organometallics, Vol. 22, No. 5, 2003 1095
IR (KBr, cm-1): 1906 (ν(CO)), 1445 (ν(RuH)). IR (toluene,
18 h. After the oxygen was vented off, the resulting brownish
solid was collected and thoroughly washed with methanol and
pentane to give pure 5 (0.036 g, 74%).
1
cm-1): 1902 (ν(CO)). 31P{1H} NMR (C6D6): δ 47.55. H NMR
(C6D6): δ 2.52 (br, 6H, P(C6H11)3), 2.22 (d, J ) 10.5 Hz, 6H,
P(C6H11)3), 2.02 (d, J ) 10.5 Hz, 6H, P(C6H11)3), 1.68 (m, 28H,
P(C6H11)3), 1.23 (18H, m, P(C6H11)3), -24.25 (t, J H,P ) 18.3 Hz,
1H, RuH). 13C{1H} NMR (C6D6): δ 34.9 (vt, J C,P ) 8.6 Hz,
P(C6H11)3), 31.2 (s, P(C6H11)3), 30.4 (s, P(C6H11)3), 28.0 (br s,
P(C6H11)3), 26.9 (s, P(C6H11)3), Ru-CO not resolved.
(P Cy3)2(CO)Ru (Cl)(P h ) (5). Meth od A. Benzyl alcohol (0.6
mL, 5.80 mmol) and triethylamine (0.2 mL, 1.42 mmol) were
added to a toluene (3 mL) solution of 1, and the subsequent
mixture was heated to 80 °C for ∼24 h. The course of the
reaction was followed by unlocked 31P NMR. When 1 was
completely consumed, the solvent was removed under reduced
pressure to leave a dark orange oil. Without further precaution
to preclude air, nondegassed ethanol (30 mL) was added to
the residue, and the mixture was vigorously stirred for 2 h.
The resulting pale orange solid was filtered and washed with
ethanol (2 × 10 mL) and pentane (2 mL) to give pure 5 (0.145
g, 60%).
Meth od B. Diphenylmercury (0.20 g, 0.564 mmol) was
added to a toluene (5 mL) solution of 4 (0.20 g, 0.275 mmol).
The reaction mixture was refluxed for 12 h, during which the
time the solution became dark orange. Without further efforts
to preclude air, the solution was filtered through a plug of
Celite to remove elemental mercury. The reaction was worked
up in the same way as for Method A, to give 5 as a pale orange
powder (0.221 g, 75%).
IR (KBr, cm-1): 1894 (ν(C)O). 31P{1H} NMR (CD2Cl2): δ
1
25.4. H NMR (CD2Cl2): δ 7.54 (v br, 2H, o-H C6H5), 6.65 (t,
2H, J ) 7.20 Hz, m-H C6H5), 6.59 (t, 1H, J ) 6.60 Hz, p-H
C6H5), 2.42 (br, 5H, P(C6H11)3), 2.00 (d, J ) 12.0 Hz, 5H,
P(C6H11)3), 1.68 (m, 23H, P(C6H11)3), 1.51-1.18 (m, 33H,
P(C6H11)3). The peak at 7.54 ppm resolves to a doublet at ∼80
°C. 13C{1H} NMR (CD2Cl2): δ 205.8 (t, J C,P ) 13.9 Hz, Ru-
CO), 156.1 (t, J C,P ) 10.0 Hz, i-C C6H5), 139.7 (v br, o-C C6H5),
128.9, 127.3 (s, o-C C6H5), 125.9 (br s, m-C C6H5), 120.8 (s,
p-C C6H5), 35.3 (vt, J C,P ) 9.2 Hz, P(C6H11)3), 30.5, 30.1 (s,
P(C6H11)3), 28.3 (m, P(C6H11)3), 27.1 (s, P(C6H11)3). Anal. Calcd
for C43H71OP2ClRu: C, 64.36; H, 8.92. Found: C, 64.39; H,
8.84.
Isom er iza tion Rea ction s. Immediately prior to use, the
octene was passed through a column (20 cm × 1.5 cm) of
neutral alumina (Acros, 50-200 µm), using 15 g of alumina
per 100 mL of octene, into a Schlenk flask. The octene was
then deoxygenated by a series of degassing (by evacuation of
the flask), followed by refilling with nitrogen. For each
reaction, 20 mL (127 mmol) of octene was used. The reaction
vessel was immersed in an oil bath and allowed to equilibrate
to the desired temperature. To the octene, 1.05 mg (six-figure
analytical balance) of catalyst was then added. No additional
solvents were used. All reactions were thoroughly stirred by
way of a magnetic stirrer bar. All reactions were allowed to
proceed to completion; i.e., the reported results are those
obtained when substrate consumption had ceased. The product
distribution of the isomerization reactions was measured by
GC/FID (Carlo Erba 8000 Top) using a ZB-5 (5% phenyl
polysiloxane) column (Zebron).
Meth od C. A toluene (5 mL) solution of 1 (0.20 g, 0.564
mmol) was stirred at 60 °C in the presence of dry air for 4 h.
During this time, the solution became dark brown-green. 31P
NMR of the reaction mixture showed two signals: a very broad
peak at 46.8 ppm and a sharp singlet at 25.9 ppm. The solvent
was completely removed under vacuum, and methanol (10 mL)
was introduced. The mixture was vigorously stirred for 6 h.
Filtration, followed by washing with methanol (4 × 5 mL), gave
5 (0.012 g, 6%).
Ack n ow led gm en t. We wish to thank SASOL for
funding this research. We also acknowledge Dr. Cathy
Dwyer for helpful discussions and Dr. Guido Fries for
providing a very old air-exposed sample of catalyst 1.
Meth od D. Finely powdered 1 (0.050 g, 0.061 mmol) was
deposited in an autoclave which was subsequently pressurized
with oxygen (40 bar). The autoclave was heated at 60 °C for
OM0208218