Nature of the active silane alcoholysis catalyst in the Ruw Clx(CO)y(PMe3)z (w, x, y, z = 1 or 2) system; Ru2 (μ-Cl)2 Cl2(CO)4(PMe3)2 as a new catalyst for silane alcoholysis in a polar solvent

Article abstract of DOI:10.1139/v00-193

The dimeric complex Ru₂(μ-Cl)₂Cl₂(CO)₄(PMe ₃)₂ (1) forms an active silane alcoholysis complex in the polar solvent N,N-dimethylamino-acetamide (DMA). The dynamic behaviour of 1 in DMF-d₇ solution has been investigated by variable temperature (VT) ¹H NMR. The solid state structures of 1 and cis,cis,trans-RuCl₂(CO)₂(PMe₃)₂ (3) have been determined by single crystal X-ray diffractometry. Using ethylene glycol - triethylsilane as a model system, the catalytic activity of 1 in DMA is compared to that of a series of known silane alcoholysis catalysts, including RhCl(PPh₃)₃, PdX₂ (X = Cl^-, OAc^-), and cis,cis,trans-[IrH₂S₂L₂]SbF₆ (L = PPh₃, S = THF). Complex 3 is not an active silane alcoholysis catalyst in this solvent, but is active in nonpolar solvents. Its structure and spectroscopic signature is, however, different from that of a previously reported catalyst of the same composition.

Full text of DOI:10.1139/v00-193

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
949
Nature of the active silane alcoholysis catalyst in
the RuwClx(CO)y(PMe3)z (w, x, y, z = 1 or 2) system;
Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 as a new catalyst for
silane alcoholysis in a polar solvent1
Mee-Kyung Chung, George Ferguson, Valerie Robertson, and Marcel Schlaf
Abstract: The dimeric complex Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 (1) forms an active silane alcoholysis complex in the polar
solvent N,N-dimethylamino-acetamide (DMA). The dynamic behaviour of 1 in DMF-d7 solution has been investigated
1
by variable temperature (VT) H NMR. The solid state structures of 1 and cis,cis,trans-RuCl2(CO)2(PMe3)2 (3) have
been determined by single crystal X-ray diffractometry. Using ethylene glycol – triethylsilane as a model system, the
catalytic activity of 1 in DMA is compared to that of a series of known silane alcoholysis catalysts, including
RhCl(PPh3)3, PdX2 (X = Cl–, OAc–), and cis,cis,trans-[IrH2S2L2]SbF6 (L = PPh3, S = THF). Complex 3 is not an ac-
tive silane alcoholysis catalyst in this solvent, but is active in nonpolar solvents. Its structure and spectroscopic signa-
ture is, however, different from that of a previously reported catalyst of the same composition.
Key words: silane alcoholysis, polar solvent, ruthenium catalyst.
Résumé : Dans un solvant polaire comme le N,N-diméthylacétamide (DMA), l’alcoolyse du complexe dimère
Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 (1) donne lieu à la formation d’un complexe de silane actif. Faisant appel à la RMN du H
1
à température variable, on a étudié le comportement dynamique de 1 en solution dans le DMF-d7. On a déterminé les
structures de 1 et du cis,cis,trans-RuCl2(CO)2(PMe3)2 (3) à l’état solide par diffraction des rayons X par un cristal
unique. Utilisant le système éthylèneglycol/triéthylsilane comme modèle, on a comparé l’activité catalytique de 1 dans
le DMA avec celles d’une série de silanes connus comme catalyseurs de l’alcoolyse, y compris RhCl(PPh3)3, PdX2 (X =
Cl–, OAc–) et cis,cis,trans-[IrH2S2L2]SbF6 (L = PPh3, S = THF). Dans ce solvant, ce composé 3 n’est pas un silane ac-
tif comme catalyseur de l’alcoolyse; il est toutefois actif dans les solvants non polaires. Sa structure et ses caractéristi-
ques spectroscopiques sont toutefois différentes de celles rapportées antérieurement pour un catalyseur de même
composition.
Mots clés : alcoolyse de silanes, solvant polaire, catalyseur de ruthénium.
[Traduit par la Rédaction] Chung et al. 957
Introduction
The high enthalpic driving force of the oxygen–silicon
bond formation, together with the entropically favourable
generation of 1 equivalent of hydrogen gas, generally results
in clean and quantitative conversions. The reaction is cata-
lyzed by a large variety of transition metal complexes (1–7).
Specific examples of highly active catalysts are cis,cis,trans-
The silane alcoholysis reaction (eq. [1]) constitutes an ex-
ceptionally mild and efficient method of protecting hydroxyl
functions by trialkyl or aryl-silyl groups.
[IrH2S2L2]SbF6 (L
= PPh3, S = THF) (8), all-cis-
[1]
RuCl2(CO)2(PMe3)2 (9), Co2(CO)8 (10), RhCl(PPh3)3 (11–13),
IrX(CO)L2 (L = PR3) (14), and [IrCl(C8H14)2]2 (14). Most
simply, palladium metal, generated in situ by reduction of
PdX2 (X = OAc–, Cl–) (15) also catalyzes the reaction. An-
other commercially available and metal-free catalyst system
is B(C6F5)3, recently described by Blackwell et al. (16).
Common to all these catalyst systems is that, to date, they
have only been tested in nonpolar, non-coordinating organic
solvents such as benzene, toluene, pentane, or methylene
chloride. We are interested in applying the silane alcoholysis
reaction to highly polar polyol and sugar substrates with
multiple hydroxyl functions, which are insoluble in these
solvents, but soluble in chemically largely inert amides such
as N,N-dimethylformamide (DMF), N-methyl-pyrrolidin-2-
one (NMP), and N,N-dimethylacetamide (DMA). In order to
establish which of a selection of the known catalyst systems
Recieved August 20, 2000. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on July, 11, 2001.
Dedicated to Professor Brian R. James on the occasion of his
65th birthday.
M.-K. Chung, G. Ferguson, V. Robertson, and M. Schlaf.2
Department of Chemistry & Biochemistry, University of
Guelph, Guelph, ON N1G 2W1, Canada.
1Contribution from the Guelph–Waterloo Centre for Graduate
Work in Chemistry (GWC)2.
2Corresponding author (telephone: (519) 824-4120 ext. 3002;
fax: (519) 766-1499; e-mail: mschlaf@uoguelph.ca).
Can. J. Chem. 79: 949–957 (2001)
DOI: 10.1139/cjc-79-5/6-949
© 2001 NRC Canada

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
950
Can. J. Chem. Vol. 79, 2001
Fig. 1. Photochemically and thermally induced stereochemical
rearrangements observed for RuCl2(CO)2(PMePh2)2.
would tolerate the anticipated coordinative inhibition by the
donor atoms (O, N) in these solvents, we conducted an ac-
tivity evaluation of a series of catalysts in DMA using ethyl-
ene glycol and triethylsilane as a model system. In the
course of this study we found that the most active
homogeneous catalyst system in DMA is derived from the
general composition RuwClx(CO)y(PMe3)z (w, x, y, z = 1
or 2), but that the structures of the active catalyst precur-
sors are in fact Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 (1) or
cis,cis,trans-RuCl2(CO)2(PMe3)2 (3) rather than all-cis-
RuCl2(CO)2(PMe3)2 (2) as previously reported (17). The dy-
namic solution behaviour of dimer 1 in amide solvents, as
well as the solid state structures of the complex 1 and 3 as
determined by single crystal X-ray structure analysis, are re-
ported.
alcoholysis reactions using the material isolated by us,
prompted us to reinvestigate the RuwClx(CO)y(PMe3)z sys-
tem in more detail.
Results
The conditions and reagents for the synthesis of the cata-
lyst system RuwClx(CO)y(PMe2)z as published are shown in
eq. [2] (17).
Synthesis and isolation
The reaction of eq. [2] gives an amorphous solid from
which two different products of different crystal habit are
obtained by fractional crystallization, first from CH2Cl2–Et2O
and subsequently from EtOH–H2O. The latter solvent sys-
tem is also the one used by Singer et al. (17). Single crys-
tal X-ray diffraction analysis revealed the structure of the
first product to be the dimeric complex Ru2(µ-
Cl)2Cl2(CO)4(PMe3)2 (1) and that of the second component
PMe3
Cl
Cl
PMe3
reflux
[2]
RuCl3 + 2 PMe3 + CO(g)
Me = 2-methoxy-ethanol
Ru
CO
CO
2
Following this procedure we obtained materials that gave
inconsistent results with respect to catalytic activity and
whose NMR and IR spectra in no case matched the previ-
ously reported data. Singer et al. (17) assigned the all-cis
to
be
the
mononuclear
complex
cis,cis,trans-
RuCl2(CO)2(PMe)2 (3). The yields for both products (1 and
3) are between 30 and 40%, with a total yield of about 70%
with respect to ruthenium. By H and 31P NMR, both com-
1
1
configuration 2 to their reaction product based on H NMR
pounds are present in about the same ratio in the crude prod-
uct, i.e., the product identity and distribution is not an
artifact of the fractional crystallization.
and IR data.3 They reported a single doublet at 1.78 ppm
with JH,P = 12 Hz in CDCl3 and two ν(CO) stretching fre-
quencies at 2085 and 2022 cm–1 in Nujol. While the IR data
and symmetry considerations support a cis position of the
two CO ligands, the assignment does not match with the po-
sition of the phosphine ligands. In an all-cis arrangement
one of the phosphines must be trans to the σ-donor – π-ac-
ceptor CO, while the other must be trans to the σ-donor –
π-donor Cl–. Two different chemical environments and
chemical shifts should, therefore, result for the phosphorus
and by extension carbon and proton nuclei on the two phos-
phine ligands, unless their shifts are fortuitously equivalent.
This, however, appears to be unlikely, as the structurally
very closely related complex all-cis-RuCl2(CO)2(PMePh2)2
shows two separate doublets for the methyl substituents on
the phosphine ligands (18). Also the all-cis configuration of
this complex is thermally unstable against isomerization to
cis,cis,trans-RuCl2(CO)2(PMePh2)2 and only accessible
through a photochemical reaction via the even more unstable
all-trans form (Fig. 1).
Structure determinations
ORTEP diagrams (19) of the structures of the dimer
Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 (1) and that of the mononuclear
complex cis,cis,trans-RuCl2(CO)2(PMe)2 (3) are shown in
Figs. 2 and 3, respectively.
Selected bond angles and distances of complexes 1 and 3
are summarized in Tables 1 and 2, respectively. The dimer 1
has a center of inversion in the plane defined by the two ru-
thenium centers and the bridging chloride ligands. The two
Ru—Cl bond distances in this plane are, however, not equiv-
alent, but slightly shorter (2.457 Å) for the bond trans to the
strong π-acceptor CO and slightly longer (2.496 Å) for the
bond trans to the phosphine, resulting in a distorted diamond
shape with Cl-Ru-Cl and Ru-Cl-Ru bond angles in this plane
of 83.95 and 96.04°, respectively. The Ru—Cl bond distance
of the terminal chloride ligand is very close to that of an av-
erage value of 2.409 Å as reported by Orpen et al. (20) The
other bond angles around ruthenium closely approach that of
an idealized octahedral environment with bond lengths in the
expected ranges.
1
The expected H NMR of the all-trans form, i.e., a dou-
blet of doublet or pseudo triplet for the methyl signals due to
coupling to the two phosphorus nuclei, equally would not
match the literature spectral data. These considerations to-
gether with wide variations of catalyst activity in the silane
All ruthenium ligand bond lengths in the mononuclear
complex 3 are pairwise equivalent within the 3σ level and
3 The original formulation in the paper by Singer et al. is: “The dérivé III (i.e., RuCl2(CO)2(PMe3)2) isolé possède tous ses ligands en position
cis.”
© 2001 NRC Canada

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Chung et al.
951
Fig. 2. Molecular structure of the dimer complex
Ru2(µ-Cl)2Cl2(CO)4(PMe3)2 (1) as determined by single crystal
X-ray diffraction (ORTEP plot at 30% probability level).
Fig. 3. Molecular structure of complex cis,cis,trans-
RuCl2(CO)2(PMe3)2 (3) as determined by single crystal X-ray
diffraction (ORTEP plot at 30% probability level).
fall within the expected ranges and bond angles approach
that of an idealized octahedral environment around the ru-
thenium center. In the solid state the two PMe3 ligands are
staggered about the P-Ru-P axis by 63.93° breaking the oth-
erwise high symmetry of the ligand arrangement.
18.0 ± 0.4 kcal mol–1 and ∆S‡ = –2.7 ± 0.9 cal mol–1 K–1 for
the reaction of the minor to the major component (kreverse
Spectroscopy
:
The spectroscopic data of 1, 1a, and 3 are summarized in
Table 3.
Dissolving an authentic sample of X-ray grade crystals of
1a → 1) and through Keq = kf/kr to ∆H‡ = 18.4 ± 0.9 kcal mol–1
and ∆S‡ = –3.0 ± 0.9 cal mol–1 K–1 for the reaction of the
major to the minor component (kforward: 1 → 1a). Within the
experimentally achievable error ranges the activation param-
eters are the same for the forward and the reverse reaction,
which reflects an equilibrium constant Keq close to unity. Keq
ranges from 0.45 to 0.49 in the temperature range used and
was determined from the integral ratios of the 31P decoupled
signals at low temperatures and extrapolated for the higher
temperatures, where beginning coalescence makes a reliable
integration impossible.
1
1 in either CDCl3 or DMF-d7 results in an H NMR spec-
trum that displays two doublets for the PMe3 ligands at the
chemical shifts listed in Table 3 in relative intensities of
0.42:1, with the lower shift component as the more abundant
one. Based on the high symmetry of the solid state structure
of 1, whose center of inversion would render the two
phosphines chemically and magnetically equivalent, this
spectral appearance immediately suggests a breakup of the
dimer in solution, most likely through solvent coordination
in the manner indicated in eq. [3], which retains the mutual
trans arrangement of the π-donor chloride and π-acceptor
carbonyl.
Further support for the reaction of eq. [3] comes from the
1
13C and H NMR spectra of the same sample at ambient
temperature. In addition to the 1:1:1 triplet for the carbonyl
carbon of free DMF-d7 at 162.7 ppm, a second small 1:1:1
triplet (JC,D = 30 Hz) is observed at 169.2 ppm, which we
assign to the coordinated DMF-d7 in complex 1a. This 13C
signal is matched by a very small singlet at 8.45 ppm in
1H NMR, which disappears with the beginning coalescence
of the methyl signals and which we assign to the aldehyde
proton of coordinated DMF-d6. The signal reversibly reap-
pears upon cooling to ambient temperature.
CO
Ru
Cl
Cl
CO
Ru
Cl
OC
Cl
Cl
PMe3
CO
OC
S
[3]
Ru
CO
+ 2 S
2
Me3P
Me3P
Cl
S = Solvent
1
1a
1
The variable temperature (VT) H NMR spectrum of the
methyl signals of a sample of 1 in DMF-d7 is shown in
Fig. 4 and reveals the existence of a dynamic process in so-
lution, which supports the above hypothesis.
An unambiguous assignment of the larger and smaller sets
of peaks to 1 and 1a, respectively follows from the
13C NMR signals of the carbonyl ligands. In the dimer 1 the
two carbonyl ligands are chemically and magnetically
inequivalent and should give rise to two doublets due to cou-
pling to the phosphorus nucleus. In complex 1a both car-
bonyl ligands are chemically and magnetically equivalent
and should, thus, result only in a single doublet. Two larger
doublets of approximately equal intensity are observed at
198.12 and 192.93 ppm, the latter of which overlaps with a
smaller doublet at 193.08 ppm. The larger peaks must, there-
fore, belong to complex 1, bearing the two nonequivalent
The two doublets coalesce at ~100°C and reemerge as a
time averaged doublet above ~120°C. In order to simplify a
more quantitative analysis of this process through simula-
tions, the VT experiment was repeated with phosphorus de-
coupling, which collapses the doublets into two simple
coalescing singlets. Using the MEX program by Bain and
Duns (21) a set of spectra were simulated that matched the
these decoupled spectra over the temperature range from 65
to 124°C. From an Eyring plot of the rate constants calcu-
lated by the program vs. 1/T (in Kelvin) the approximate ac-
tivation parameters of the process were determined to ∆H‡ =
1
carbonyl ligands. By extension, the larger peaks in the H
and 31P NMR spectra are assigned to the dimer 1, which is
© 2001 NRC Canada

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
952
Can. J. Chem. Vol. 79, 2001
Table 1. Selected bond lengths (Å) and angles (°) in complex 1.
Bond lengths (Å)
Bond angles (°)
Ru(1)—C(1)
Ru(1)—C(2)
Ru(1)—P(1)
Ru(1)—Cl(1)
Ru(1)—Cl(11)
Ru(1)—Cl(11a)
Cl(11)—Ru(1a)
Cl(11a)—Ru(1a)
O(1)—C(1)
1.885(3)
1.855(2)
2.3018(6)
2.4018(7)
2.4573(6)
2.4967(6)
2.4967(6)
2.4573(6)
1.126(3)
1.128(3)
C(2)-Ru(1)-C(1)
C(2)-Ru(1)-P(1)
C(1)-Ru(1)-P(1)
C(2)-Ru(1)-Cl(1)
C(1)-Ru(1)-Cl(1)
P(1)-Ru(1)-Cl(1)
C(2)-Ru(1)-Cl(11)
C(1)-Ru(1)-Cl(11)
P(1)-Ru(1)-Cl(11)
Cl(1)-Ru(1)-Cl(11)
91.19(11)
89.85(8)
93.21(7)
89.16(9)
178.97(8)
87.77(2)
178.45(8)
89.89(7)
91.20(2)
89.74(2)
C(2)-Ru(1)-Cl(11a)
94.92(8)
90.70(7)
173.77(2)
88.30(2)
83.955(19)
96.043(19)
C(1)-Ru(1)-Cl(11a)
P(1)-Ru(1)-Cl(11a)
Cl(1)-Ru(1)-Cl(11a)
Cl(11)-Ru(1)-Cl(11a)
Ru(1)-Cl(11)-Ru(1a)
O(2)—C(2)
Table 2. Selected bond lengths (Å) and angles (°) in complex 3.
Bond lengths (Å)
Bond angles (°)
Ru(1)—Cl(1)
Ru(1)—Cl(2)
Ru(1)—P(1)
Ru(1)—P(2)
Ru(1)—C(1)
Ru(1)—C(2)
2.4403(9)
2.4323(6)
2.3888(11)
2.3905(9)
1.851(2)
Cl(1)-Ru(1)-Cl(2)
Cl(1)-Ru(1)-P(1)
Cl(1)-Ru(1)-P(2)
Cl(1)-Ru(1)-C(1)
Cl(1)-Ru(1)-C(2)
Cl(2)-Ru(1)-P(1)
91.31(4)
92.23(2)
85.64(4)
86.97(10)
178.77(9)
90.63(4)
Cl(2)-Ru(1)-P(2)
Cl(2)-Ru(1)-C(1)
Cl(2)-Ru(1)-C(2)
P(1)-Ru(1)-P(2)
P(1)-Ru(1)-C(1)
P(1)-Ru(1)-C(2)
87.55(4)
178.29(11)
88.86(10)
177.17(4)
89.45(10)
88.99(9)
1.861(3)
therefore the major component in solution at ambient tem-
perature.
Oehmichen and Singer (9). All other tested catalysts are
active in the DMA solvent and quantitative results as deter-
mined by direct quantitative GC analysis of the reaction
mixtures are summarized in Table 4. The data gives a direct
comparison of the relative activity of the catalyst systems in
the amide solvent, total yields of silylated products, as well
as their selectivity towards mono- and disilylation of the diol
substrate in the presence of one or two equiv of silane, re-
spectively. The reaction times given in Table 4 refer to the
end point of the reaction as indicated by the end of hydrogen
gas evolution and by the complete consumption of
triethylsilane as indicated by the GC trace of the reaction
mixture.
Of all the catalyst systems tested, by far the most active is
the dimer 1, which even surpasses the activity of the hetero-
geneous palladium systems. The true catalyst of unknown
structure and composition formed from 1 and triethylsilane
in DMA, therefore, does not suffer from coordinative inhibi-
tion by the solvent. This contrasts with the behaviour of
complex 3, which does suffer inhibition by the solvent to the
point of being inactive in the amide solvent, and the iridium
system [IrH2(PPh3)(THF)2]SbF6, which shows as many as
130 000 turnovers per hour in methylene chloride solvent at
25°C (8), but requires heating to 45°C to achieve acceptable
reaction rates in DMA. Consistent with the latter result,
Crabtree et al. (22) had previously shown that the same irid-
ium complex suffers strong inhibition by even trace amounts
of coordinating solvents when used as an olefin hydrogena-
tion catalyst. The active catalyst resulting from RhCl(PPh3)3
also must suffer coordinative inhibition, as it has been re-
ported to give complete conversion to silane alcoholysis
products within minutes in benzene solution at 20°C and
catalyst concentrations as low as 0.05 mol% (13).
The IR spectrum of the same sample of 1 dissolved in
CDCl3 displays only two well defined bands at 2068 and
2010 cm–1, i.e., does not show the presence of two compo-
nents in solution as observed by NMR. This can be rational-
ized by the fact that in both complexes (1 and 1a) both
carbonyl ligands are trans to the same π-donor chloride and
the ν(CO) stretching frequencies are, therefore, presumably
very close to or even identical with each other in both com-
plexes.
The NMR spectra of complex 3 show no dynamic phe-
nomena and display the expected coupling patterns for the
cis,cis,trans-arrangement. The pseudo triplet observed for
the methyl signals of 3, also establishes that it cannot have
been the compound originally isolated by Singer et al. (17),
whose identity remains unknown.
Evaluation of the catalysts in DMA
The catalytic activity of the new complexes 1 and 3 was
evaluated using the ethylene glycol – triethylsilane system as
the model reaction (eq. [4]) and compared to a series of cat-
alyst systems consisting of cis,cis,trans-[IrH2S2L2]SbF6 (L =
PPh3, S = THF) (8), RhCl(PPh3)3 (11–13), Co2(CO)8 (10),
and PdX2, (X = Cl–, OAc–) (15).
HOCH2CH2OSiEt3
[cat.]
HOCH2CH2OH + n Et3SiH
[4]
+
– n H2(g)
Et3SiOCH2CH2OSiEt3
In all reactions ethylene glycol (2.5 mmol) in DMA (6 mL)
were reacted with one or two equivalents of triethylsilane at
a catalyst concentration of 3 mol% with respect to metal
content. Complexes 3 and Co2(CO)8 are inactive in DMA,
however, 3 is an active catalyst in THF, the solvent used by
The yields of monosilylated product achieved by the cata-
lysts in DMA solvent are moderate to good, but in no case
© 2001 NRC Canada