Roles of a tetrahydroborate ligand in a facile route to ruthenium(II) ethyl hydride complexes, and a kinetic study of ethane reductive elimination

Article abstract of DOI:10.1039/b411600c

The tetrahydroborate ligand in [Ru(η²-BH₄)(CO) H(PMe₂Ph)₂], 1, allows conversion under very mild conditions to [Ru(CO)(Et)H(PMe₂Ph)₃], 7, by way of [Ru(η²-BH₄)(CO)Et(PMe₂Ph)₂], 4. Deprotection of the hydride ligand in 7 (by BH₃ abstraction) occurs only in the final step, thus preventing premature ethane elimination. A deviation from the route from 4 to 7 yields [Ru(η²-BH ₄)(COEt)(PMe₂Ph)₃], 6, but does not prevent ultimate conversion to 7. Modification of the treatment of 4 yields an isomer of 7, 10. Both isomers eliminate ethane at temperatures above 250 K: the immediate product of elimination, thought to be [Ru(CO)(PMe₂Ph)₃], 11, can be trapped as [Ru(CO)(PMe₂Ph)₄], 12, [Ru(CO)H ₂(PMe₂Ph)₃], 3a, or [Ru(CO)(C≡CCMe ₃)H(PMe₂Ph)₃], 13. The elimination is a simple first-order process with negative ΔS? and (for 7) a normal kinetic isotope effect (k_H/k_D = 2.5 at 287.9 K). These results, coupled with labelling studies, rule out a rapid equilibrium with a σ-ethane intermediate prior to ethane loss.

Full text of DOI:10.1039/b411600c

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Roles of a tetrahydroborate ligand in a facile route to ruthenium(II)
ethyl hydride complexes, and a kinetic study of ethane reductive
elimination
Simon B. Duckett,* J. Claire Lowe (née Stott), John P. Lowe and Roger J. Mawby
Department of Chemistry, University of York, Heslington, York, UK YO10 5DD.
E-mail: sbd3@york.ac.uk; Fax: +01904 432516; Tel: +01904 432564
Received 28th July 2004, Accepted 22nd September 2004
First published as an Advance Article on the web 12th October 2004
The tetrahydroborate ligand in [Ru(g2-BH4)(CO)H(PMe2Ph)2], 1, allows conversion under very mild conditions
to [Ru(CO)(Et)H(PMe2Ph)3], 7, by way of [Ru(g2-BH4)(CO)Et(PMe2Ph)2], 4. Deprotection of the hydride ligand
in 7 (by BH3 abstraction) occurs only in the final step, thus preventing premature ethane elimination. A deviation
from the route from 4 to 7 yields [Ru(g2-BH4)(COEt)(PMe2Ph)3], 6, but does not prevent ultimate conversion to
7. Modification of the treatment of 4 yields an isomer of 7, 10. Both isomers eliminate ethane at temperatures
above 250 K: the immediate product of elimination, thought to be [Ru(CO)(PMe2Ph)3], 11, can be trapped as
[Ru(CO)(PMe2Ph)4], 12, [Ru(CO)H2(PMe2Ph)3], 3a, or [Ru(CO)(CCCMe3)H(PMe2Ph)3], 13. The elimination is
a simple first-order process with negative DS‡ and (for 7) a normal kinetic isotope effect (kH/kD = 2.5 at 287.9 K).
These results, coupled with labelling studies, rule out a rapid equilibrium with a r-ethane intermediate prior to
ethane loss.
rather than a phenyl ligand by these methods were unsuccessful.
Introduction
Work which we have recently published,17 however, suggested an
alternative route. We found that the g2-tetrahydroborate complex
[Ru(g2-BH4)(CO)H(PMe2Ph)2], 1, reacted at low temperatures
with nucleophiles L {(a), L = PMe2Ph; (b), L = CO; (c), L =
4-methylpyridine (4-MePy)} by displacement of the bridging
hydrogen trans to the hydride ligand, giving products [Ru(g1-
BH4)(CO)H(L)(PMe2Ph)2], 2a–2c. Further treatment with L
yielded dihydrido-complexes [Ru(CO)H2(L)(PMe2Ph)2], 3a–3c,
by abstraction of BH3 as the adduct H3B·L. Complex 1 also
reacted with C2H4, but here initial (and reversible) formation
of [Ru(g1-BH4)(CO)(C2H4)H(PMe2Ph)2] was followed by
combination of ethene and hydride ligands, allowing a reversion
to g2-binding of the tetrahydroborate ligand in the product ethyl
complex [Ru(g2-BH4)(CO)Et(PMe2Ph)2], 4.
These results demonstrated (i) that the ease with which
the tetrahydroborate ligand changes its mode of bonding
to ruthenium allows conversion of a hydride ligand into an
alkyl ligand under extremely mild conditions, and (ii) that
an g2-tetrahydroborate ligand can be regarded as a protected
hydride ligand, from which the protection can be removed at
an appropriate point, again under very mild conditions, by
treatment with two molar equivalents of a nucleophile. In
combination, these features seemed to offer a relatively simple
route by which complexes containing both an alkyl and a
hydride ligand might be prepared without risking premature
decomposition by alkane elimination. We therefore studied the
reaction of 4 with the nucleophile PMe2Ph, hoping to generate
a complex containing carbonyl, ethyl and hydride ligands. We
did, however, envisage the possibility that complications might
arise from side-reactions involving combination of the ethyl and
carbonyl ligands in 4.
For many years chemists have been interested in transition metal
complexes which contain both alkyl and hydride ligands. Such
complexes are known (or believed) to be involved in catalytic
cycles, and the very reaction which usually severely limits their
stability (alkane elimination) is itself believed to be the final step
in the catalytic hydrogenation of alkenes.1
One of the first complexes of this type to be prepared (by
Chatt and Hayter2) was the ruthenium(II) complex [Ru(Me)H-
(Ph2PCH2CH2PPh2)2], in which the methyl and hydride ligands
were conclusively shown to be mutually cis. Subsequently
Wilkinson and his co-workers reported the preparation of
[Ru(Me)H(OEt2)2(PPh3)2]3 and cis-[Ru(R)H(PMe3)4] (R = Me4
and Et5), and Bergman obtained trans-[Ru(Me)H(Me2PCH2-
PMe2)2].6 With the discovery that simple alkanes may be acti-
vated by some transition metal complexes, however, the focus
has to some extent shifted away from ruthenium to adjacent
metals in the periodic table (for example iridium,7,8 rhodium,9,10
osmium,11 iron12 and rhenium13), for all of which complexes
containing both alkyl and hydride ligands have now been
obtained from direct reactions with simple alkanes.
We were interested in synthesising and studying the chemistry
of ruthenium(II) complexes containing a carbonyl ligand as
well as an alkyl and a hydride ligand, because its presence
increases the range of reactions which the complexes can
potentially undergo. We recognised, however, that the carbonyl
ligand was likely also to increase the ease with which the
complexes would decompose by alkane elimination, but we
hoped to avoid decomposition by generating the complexes
under mild conditions. Such an approach was used by Jones
and Feher9 in their studies of alkane activation by rhodium(I)
complexes, when they employed an indirect route to generate
the rhodium(III) species [Rh(g5-C5Me5)(Me)H(PMe3)], and
were then able to determine its stability with respect to
methane elimination. Their synthetic route involved halide
abstraction from [Rh(g5-C5Me5)(Me)Cl(PMe3)] with Ag+,
followed by low-temperature treatment of the resulting
cation with a source of hydride ion. We were able to obtain
[Ru(CO)2(Ph)H(PMe2Ph)2] by a similar two-stage route from
[Ru(CO)2(Ph)Cl(PMe2Ph)2],14,15 and a second isomer of the
same complex from [Ru(CO)2Cl(H)(PMe2Ph)2] and LiPh,15,16
but our attempts to synthesise similar complexes with an alkyl
Apart from the original interest in the mechanism of alkane
elimination in the context of catalytic cycles, the fact that the
reaction represents the exact reverse of alkane elimination
has attracted the attention of theoretical chemists18,19 and
encouraged kinetic studies of the elimination process from
well-characterised complexes containing both alkyl and
hydride ligands.20–32 The success of the synthetic approach
outlined above, through which we obtained two isomers of
[Ru(CO)(Et)H(PMe2Ph)3], prompted us not only to determine
the conditions under which alkane elimination occurred
from the complexes but also to carry out a kinetic study of
3 7 8 8
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

View Article Online
the elimination process. Both synthetic and kinetic work are
reported in this paper.
Results and discussion
The NMR data for all new ruthenium complexes are collected
in Table 1. All 13C, 31P and 11B data refer to spectra recorded
with full proton-decoupling. The 1H spectra were recorded with
full 31P-coupling, but 1H{31P} spectra were also run, with both
broad-band and selective decoupling. The NMR data estab-
lished that all the new ruthenium(II) complexes described below
contain either two mutually trans PMe2Ph ligands or a mer
arrangement of three PMe2Ph ligands. For complexes contain-
1
ing either an ethyl or a propanoyl ligand, H{1H} decoupling
and/or 1H–1H COSY experiments were employed to confirm the
coupling between CH2 and CH3 protons.
(i) The reaction of [Ru(g2-BH4)(CO)Et(PMe2Ph)2], 4, with
PMe2Ph
At 193 K in CD3COCD3 solution, the reaction between
equimolar quantities of 417 and PMe2Ph led to the
almost quantitative formation of a single product, [Ru(g1-
BH4)(CO)Et(PMe2Ph)3], 5. Resonances for the ethyl group were
Scheme 1 The route to isomer
7
of [Ru(CO)(Et)H(PMe2Ph)3]
(L = PMe2Ph).
1
found in the H NMR spectrum at d 1.10 (q, CH3CH2) and d
BH4)(COEt)(PMe2Ph)3], 6, a process accelerated by raising the
temperature to ca. 240–250 K. NMR spectra for 6 were similar
in both solvents, but those obtained in C6D5CD3 (at 240 K) were
of rather better quality. The ethyl group was now represented by
resonances in the 1H spectrum at d 1.22 (t, CH3CH2) and d 2.92
(q, CH3CH2), and in the 13C spectrum at d 11.5 (s, CH3CH2) and
d 53.6 (s, CH3CH2). Comparison with 5 revealed major changes
in the methylene chemical shifts and the loss of all splittings
due to the 31P nuclei, both indicating the incorporation of the
ethyl group into an acyl ligand.34 This was confirmed by the
detection of the acyl resonance at d 261.8 in the 13C spectrum.34
The resonances for the protons in the now g2-bonded
tetrahydroborate ligand were very reminiscent of those for both
4 and 1,17 with a broad resonance at d 4.79 for the two equivalent
terminal protons and separate broad resonances at d −4.85 and
d −7.41 for the two inequivalent bridging protons. One distinct
difference, however, was that the resonance at d −7.41 showed a
large doublet splitting (|2JPH| = 39.6 Hz) by the 31P nucleus in the
unique PMe2Ph ligand. In 1 and 4, both PMe2Ph ligands are cis
to each of the bridging tetrahydroborate hydrogens, whereas in 6
(see Scheme 1) one bridging hydrogen must of necessity lie trans
to the unique PMe2Ph ligand. We assume that it is this hydrogen
whose resonance shows the doublet splitting. Letts et al.36 have
observed the same phenomenon for [Ru(g2-BH4)H{PhP(CH2-
CH2CH2PPh2)2}]. As in the case of 1 and 4, the resonances
for the hydrogens in the tetrahydroborate ligand broadened
with increasing temperature (a process which was reversed on
cooling), indicating a scrambling of bridging and terminal
hydrogens, presumably via an g1-tetrahydroborate intermediate.
The onset of broadening occurred at a significantly lower
temperature for the bridging hydrogen trans to the propanoyl
ligand than for that trans to PMe2Ph, implying that the Ru–HB
bond trans to the acyl ligand breaks more readily than that trans
to PMe2Ph. The two steps which connect 5 and 6 in Scheme 1
account both for the conversion of 5 to 6 and for the fluxionality
of the g2-tetrahydroborate ligand in 6.
0.98 (m, CH3CH2), the latter chemical shift indicating that this
group was directly attached to the metal,17,33 and not part of
an acyl ligand.34 Whereas the methylene protons were coupled
to the methyl protons and to all three 31P nuclei, the apparent
quartet splitting of the methyl resonance was shown to be due to
couplings of roughly equal magnitude to the unique 31P nucleus
and the two methylene protons. In 4, where the ethyl ligand is cis
to both PMe2Ph ligands, coupling of the methyl protons in this
ligand to the 31P nuclei is too weak to detect.17 This implied that
in 5 the ethyl ligand was trans to the unique PMe2Ph ligand.
The 13C NMR spectrum of 5 recorded in CD3COCD3 solu-
tion was of rather poor quality. On repeating the reaction in
1
C6D5CD3 solution, we concluded from H and 31P spectra that
5 was again formed, and obtained a better 13C spectrum. The
methylene carbon chemical shift confirmed that the ethyl group
was directly attached to the metal17,33 and not part of an acyl
ligand,34 and the fact that the doublet splitting of the resonance
was considerably greater than the triplet splitting provided
further evidence that the ethyl group was trans to the unique
PMe2Ph ligand.33 In contrast, the appearance of the resonance
for the carbonyl ligand in 5, at d 206.5, made it clear that the
coupling constants to the three 31P nuclei must all be of similar
magnitude, placing this ligand cis to all three PMe2Ph ligands.
1
A variable temperature H NMR study of 5 in CD3COCD3
solutionrevealedthatthetetrahydroborateligandwas g1-bonded
to the metal and was undergoing some kind of fluxional motion.
At 193 K, a broad resonance integrating for a single proton was
observed at d −12.3, a chemical shift similar to that of d −11.9
for the bridging hydrogen in [Ru(g1-BH4)(CO)H(PMe2Ph)3],
2a.17 We were unable to detect the (presumably extremely broad)
resonance for the three terminal protons. When the solution was
warmed, the resonance at d −12.3 rapidly broadened, becoming
undetectable at 223 K. Further increase in temperature resulted
in the appearance of a very broad resonance at d −2.3, which
sharpened significantly as the temperature was raised to 253 K.
Integration showed this resonance to be due to four protons,
indicating that the fluxional process was exchanging bridging
and terminal tetrahydroborate protons.35 The corresponding
resonance for 2a was at d −1.7.17 Apart from the effects of a
further reaction of 5 (see below), the changes in the appearance
of the 1H NMR spectrum with temperature were fully reversible,
as were those for 2a. We concluded that the ligand arrangement
in 5 was that shown in Scheme 1.
In both CD3COCD3 and C6D5CD3, the initial effect of
treating 5 with at least an equimolar amount of PMe2Ph and
then warming the solution to 250 K was to produce a mixture
of 6 and a further species 7. The ratio of 6 to 7 slowly decreased,
with eventual quantitative formation of 7. As 7 appeared, so
also did an equimolar quantity of H3B·PMe2Ph.17 7 could also
be obtained, together with H3B·PMe2Ph, by treating 6 with
PMe2Ph at 243 K. In addition to three PMe2Ph ligands, it
contained a hydride ligand, characterised by a quartet resonance
Although 5 was the kinetic product of the reaction between 4
and PMe2Ph in both CD3COCD3 and C6D5CD3, it subsequently
rearranged (in both solvents) to a new complex, [Ru(g2-
(|2JPH| = 25.0 Hz) at d −6.52 in the H NMR spectrum, and a
1
carbonyl ligand, responsible for a doublet of triplets (|2JPC| = 7.3
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
3 7 8 9

View Article Online
Table 1 NMR data for new complexesa
Complex
Nucleus, temperature
d/ppm (multiplicity, area)
Assignment
Coupling constants/Hz
Assignment
5
31P, 213 Kb
1H, 213 Kb
−10.8 (t)
PMe2Ph
PMe2Ph
RuHBH3
CH3CH2
CH3CH2
28.5
28.5
|2JPP
|2JPP
|
|
5.5 (d)
−12.3 (br, 1)c
0.98 (m, 2)d
1.10 (q, 3)
6.8
6.8
6.8
7.1
6.6
|3JHH
|3JHH
|
|
trans-|4JPH
|
1.14 (d, 6)
1.56 (t, 6)
1.63 (t, 6)
5.9 (dt)d
14.8 (t)
15.9 (t)
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2
CO
PMe2Ph
PMe2Ph
RuH2BH2
RuH2BH2
PMe2Ph
PMe2Ph
CH3CH2CO
PMe2Ph
CH3CH2CO
RuH2BH2
CH3CH2CO
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2CO
CH3CH2CO
PMe2Ph
|2JPH
|
|2JPH + 4JPH
|2JPH + 4JPH
|
|
6.0
13C, 210 K
49.2
28.8
29.5
21.5
trans-|2JPC
|
|1JPC + 3JPC
|1JPC + 3JPC
|
|
17.9 (d)
|1JPC
|
21.8 (s)
206.5 (m)d
6.9 (d)
6
31P, 240 K
1H, 240 K
35.5
35.5
39.6
|2JPP
|2JPP
|
|
20.8 (t)
e
−7.41 (br d, 1)
−4.85 (br, 1)
1.09 (t, 6)
1.16 (d, 6)
1.22 (t, 3)
1.70 (t, 6)
2.92 (q, 2)
4.79 (br, 2)
11.5 (s)
trans-|2JPH
|
f
5.6
8.7
7.2
5.9
7.2
|2JPH + 4JPH
|2JPH
|3JHH
|2JPH + 4JPH
|3JHH
|
|
|
|
|
13C, 240 K
13.9 (t)
29.5
28.8
|1JPC + 3JPC
|1JPC
|
18.7 (d)
|
ca. 20.4 g
53.6 (s)
261.8 (m)d
2.6 (t)
7
31P, 240 K
1H, 240 K
21.2
21.2
25.0
6.0
7.5
5.4
5.6
5.9
7.5
48.6
10.4
27.7
30.5
20.8
|2JPP
|2JPP
|
|
18.4 (d)
PMe2Ph
RuH
PMe2Ph
CH3CH2
PMe2Ph
PMe2Ph
CH3CH2
−6.52 (q, 1)
0.92 (d, 6)
1.21 (m, 2)d
1.48 (t, 6)
1.50 (t, 6)
1.89 (dt, 3)
|2JPH
|2JPH
|
|
|3JHH
|
|2JPH + 4JPH
|2JPH + 4JPH
|
|
trans-|4JPH
|
|3JHH
trans-|2JPC
|
|
13C, 240 K
−3.5 (dt)
CH3CH2
cis-|2JPC
|
19.4 (t)
19.7 (t)
21.7 (d)
25.2 (s)
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2
CO
|1JPC + 3JPC
|1JPC + 3JPC
|
|
|1JPC
|
206.0 (dt)
7.3
10.8
|2JPC
|2JPC
|
|
8
31P, 220 K
1H, 220 K
10.1 (s)
PMe2Ph
RuHBH3
CH3CH2
−9.87 (br, 1)h
0.74 (sext, 2)
7.3
7.3
7.3
6.2
|3JHH
|
|
|3JPH
|3JHH
|
1.52 (t, 3)
1.56 (t, 6)
1.60 (s, 3)
1.64 (t, 6)
4.70 (br, 3)h
7.34 (d, 2)
8.79 (d, 2)
−29.0 (br)
14.9 (s)
CH3CH2
PMe2Ph
4-MePy
PMe2Ph
RuHBH3
4-MePy, H3,5
4-MePy, H2,6
RuHBH3
PMe2Ph
RuH
|2JPH + 4JPH
|
|
5.7
|2JPH + 4JPH
5.1
5.1
|3JHH
|3JHH
|
|
11B, 200 K
31P, 250 K
1H, 250 K
9
−14.74 (t, 1)
0.50 (tq, 2)
24.7
10.8
7.5
5.9
5.7
|2JPH
|3JPH
|3JHH
|
|
CH3CH2
|
1.62 (t, 6)
1.63 (t, 6)
1.66 (s, 3)
1.82 (t, 3)
6.00 (d, 2)
7.85 (d, 2)
11.2 (t)
PMe2Ph
PMe2Ph
4-MePy
|2JPH + 4JPH
|2JPH + 4JPH
|
|
CH3CH2
7.5
5.9
5.9
11.7
27.6
27.6
|3JHH
|3JHH
|3JHH
|
|
|
4-MePy, H3,5
4-MePy, H2,6
CH3CH2
13C, 250 K
|2JPC
|
15.9 (t)
18.8 (t)
PMe2Ph
PMe2Ph
|1JPC + 3JPC
|1JPC + 3JPC
|
|
20.0 (s)
23.9 (s)
125.5 (s)
150.3 (s)
153.0 (s)
4-MePy
CH3CH2
4-MePy, C3,5
4-MePy, C4
4-MePy, C2,6
3 7 9 0
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7

View Article Online
Table 1 (Contd.)
Complex
Nucleus, temperature
d/ppm (multiplicity, area)
Assignment
Coupling constants/Hz
Assignment
200.9 (t)
−4.1 (t)
11.0 (d)
CO
10.8
22.3
22.3
94.4
28.9
8.8
7.5
7.4
6.2
5.6
5.2
7.5
10.8
19.4
27.7
31.2
|2JPC
|
10
31P, 245 K
1H, 245 K
PMe2Ph
PMe2Ph
RuH
|2JPP
|2JPP
|
|
−8.05 (dt, 1)
trans-|2JPH
|
cis-|2JPH
|3JPH
|
0.51 (tqd, 2)
CH3CH2
|
|
|
|
|3JHH
|3JPH
|2JPH
0.95 (d, 6)
1.44 (t, 6)
1.51 (t, 6)
1.73 (t, 3)
0.5 (q)
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2
CH3CH2
PMe2Ph
PMe2Ph
PMe2Ph
CH3CH2
CO
|2JPH + 4JPH
|2JPH + 4JPH
|
|
|3JHH
|
13C, 245 K
|2JPC
|1JPC
|
|
16.2 (d)
16.5 (td)d
22.6 (td)d
26.4 (s)
|1JPC + 3JPC
|1JPC + 3JPC
|
|
202.9 (td)
11.4
8.0
|2JPC
|2JPC
|
|
12
13
31P, 260 K
1H, 260 K
13C, 283 K
3.7 (s)
1.32 (t)i
23.9 (tt)
PMe2Ph
PMe2Ph
PMe2Ph
4.1
23.2
3.5
|2JPH + 4JPH
|1JPC + 3JPC
cis-|3JPC
|
|
|
223.9 (quin)
−6.0 (t)
6.0 (d)
CO
4.5
|2JPC
|
31P, 250 K
1H, 250 K
PMe2Ph
PMe2Ph
RuH
21.5
21.5
92.5
27.0
6.8
|2JPP
|2JPP
|
|
−7.69 (dt, 1)
trans-|2JPH
|
cis-|2JPH
|2JPH
|
1.08 (d, 6)
1.54 (s, 9)
1.71 (t, 6)
1.76 (t, 6)
16.8 (d)
PMe2Ph
CMe3
|
PMe2Ph
PMe2Ph
PMe2Ph
PMe2Ph
PMe2Ph
CMe3
6.0
6.7
20.8
|2JPH + 4JPH
|2JPH + 4JPH
|
|
13C, 250 K
|1JPC
|
20.5j
22.1 (t)
29.8 (s)
31.7
|1JPC + 3JPC
|
32.9 (s)
101.2 (td)
CMe3
RuCC
21.4
18.8
|2JPC
|2JPC
|
|
116.0 (s)
204.4 (q)
RuCC
CO
10.9
|2JPC
|
a Except where indicated otherwise, spectra were recorded in C6D5CD3 solution. Resonances for phenyl protons and carbon atoms omitted. b Recorded
in CD3COCD3 solution. c Recorded at 193 K: see text. d One or more splittings not fully resolved. e Trans to PMe2Ph. f Trans to EtCO. g Partly masked
by solvent resonances. h Recorded at 195 K: see text. i Central peak of triplet broadened: see text. j Obscured by solvent resonances: located by a 1H–13C
COSY experiment.
and 10.8 Hz respectively) in the 13C spectrum at d 206.0. The sizes
of the coupling constants indicated that both ligands were cis to
all three PMe2Ph ligands.37 The remaining coordination site in 7
was occupied by an ethyl ligand. The 1H and 13C resonances for
ligand and abstraction of BH3) is uncertain: both possibilities
are shown in the scheme.
Thus [Ru(g2-BH4)(CO)H(PMe2Ph)2], 1, is indeed a conve-
nient starting material from which to make a complex contain-
ing carbonyl, alkyl and hydride ligands. As we had anticipated,
the stability of 7 proved to be distinctly limited, with ethane
elimination occurring at a significant rate at temperatures above
250 K. In addition, none of the intermediates between 1 and 7
was stable at room temperature (although 4 can be stabilised by
storage under C2H417), but the fact that each step could be car-
ried out at low temperature meant that this was not a handicap.
Insurance against ethane elimination occurring earlier in the
reaction sequence was provided by the fact that the hydride
ligand required for the elimination was “protected” (by being
part of a tetrahydroborate ligand) until the BH3 was removed
at the end of the sequence.
1
this ligand, which were linked by a H–13C COSY experiment,
showed splitting patterns very similar to those for the ethyl
ligand in 5, confirming that here again the ethyl ligand was trans
to the unique PMe2Ph ligand. Thus 7 was our desired product,
[Ru(CO)(Et)H(PMe2Ph)3], with the ligand arrangement shown
in Scheme 1.
The initial reaction between
4 and PMe2Ph could
have involved (a) abstraction of BH3 by the PMe2Ph, (b)
combination of ethyl and carbonyl ligands, or (c) a switch
to g1-coordination of the tetrahydroborate ligand. Each
would have left a vacant site on the metal to be occupied by
PMe2Ph, giving 7, 6 and 5 respectively. In the event, (c) proved
to be the kinetically favoured pathway, although the resulting
complex 5 was less stable than 6. On addition of more PMe2Ph
to 5, rearrangement to 6 still managed to compete with the
abstraction of BH3 from 5 to give 7. Nevertheless the ultimate
result was (as we had hoped) complete conversion to 7. Given
the evidence for the extreme lability of the Ru–H bonds to the
tetrahydroborate ligand in 6, it seems reasonable that the first
step on the route from 6 to 7 should be the formation of the five-
coordinate species [Ru(g1-BH4)(COEt)(PMe2Ph)3] (see Scheme
1). The order of the remaining two steps (breakdown of the acyl
(ii) The reaction of 4 with 4-MePy: a route to an isomer of 7
Like PMe2Ph, 4-MePy proved to be capable of displacing
one bridging hydrogen from the metal in 4, yielding [Ru(g1-
BH4)(CO)Et(4-MePy)(PMe2Ph)2], 8. The reaction was carried
out in C6D5CD3 at 220 K, and 8 was sufficiently stable at this
temperature to allow 1H and 31P NMR spectra to be obtained.
These established the presence of a single 4-MePy ligand and
an ethyl group in 8. On the basis of the chemical shift for the
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
3 7 9 1

View Article Online
methylene protons, d 0.74, the ethyl group was evidently directly
attached to the metal. No resonances for the tetrahydroborate
ligand could be detected at 220 K: on cooling to 210 K a very
broad resonance integrating for one proton was observed at
d −9.87, and this sharpened considerably on further cooling to
195 K. At this temperature a further, extremely broad, resonance
had become visible at d 4.70. Comparison with the spectra of
the g1-tetrahydroborate complexes 2a, 2c17 and 5 indicated that
the resonance at d −9.87 was due to the bridging hydrogen in
an g1-bonded tetrahydroborate ligand, and it seemed likely
that the resonance at d 4.70 was due to the three non-bridging
hydrogens. The collapse of the resonances on restoring the
temperature to 220 K was presumably due to increasingly rapid
scrambling of the bridging and terminal hydrogens, as observed
for 2a, 2c and 5, but the limited stability of 8 thwarted attempts
to obtain spectra at high enough temperatures to detect the
averaged resonance for all four hydrogens. An 11B spectrum of
8, recorded at 200 K, contained a broad resonance at d −29.0,
close to the value of d −28.9 for 2a.17 We inferred the presence
of the carbonyl ligand in 8 from the fact that 8 was formed from
4 and (see below) reacted with more 4-MePy to form 9, both of
which definitely contained a carbonyl ligand.
quite different from that (d 1.21) for its isomer 7, in which the
ethyl ligand is trans to PMe2Ph rather than to CO, was very
similar to the corresponding value of d 0.50 for 9, as expected
since both complexes have CO trans to the ethyl ligand.
The route for the conversion of 4 to 10 is shown in
Scheme 2, but it should be noted that the ligand arrangement
in 8 is uncertain. We suspect that the weakness of the bond
between the metal and the 4-MePy ligand, demonstrated by the
ease of conversion of 9 into 10, also facilitates a switch in the
positions of the CO and 4-MePy ligands, but cannot be sure
whether this occurs during the formation of 8 or (as implied in
Scheme 2) in the process of its conversion into 9.
(iii) Reactions of isomers 7 and 10 of [Ru(CO)(Et)H-
(PMe2Ph)3]
In C6D5CD3 solution both isomers of [Ru(CO)(Et)H(PMe2Ph)3]
decomposed at a significant rate at temperatures above 250 K.
1
When the decomposition was monitored by H and 31P NMR
spectroscopy, it was clear from the steady growth of a singlet
1
resonance at d 0.80 in the H NMR spectrum that the process
involved the reductive elimination of ethane, but the 31P spectrum
1
The reaction of 8 with more 4-MePy was carried out at
250 K, and it yielded H3B·4-MePy17 and a single ruthenium
complex [Ru(CO)(Et)H(4-MePy)(PMe2Ph)2], 9. Complex 9 was
sufficiently stable for good quality NMR spectra to be obtained
at 250 K, and these established the presence of all the ligands.
We had expected a ligand arrangement analogous to that in 7,
with mutually trans carbonyl and hydride ligands, and the ethyl
ligand trans to 4-MePy. The clearest indication that this was not
the case came from the chemical shift for the hydride ligand,
d −14.74, very different from the values (d −4.46 to d −7.16) for
a hydride trans to CO in 2b, 3a–3c17 and 7, but similar to those
(d −15.01, d −11.43 and d −11.3) for a hydride trans to 4-MePy
and the remainder of the H spectrum became increasingly
complex. Evidently the initial ruthenium(0) product, tentatively
assumed to be the 16-electron species [Ru(CO)(PMe2Ph)3], 11,
decomposed with the formation of a variety of other products.
In order to simplify matters, various reagents were added in the
hope that each would trap 11, converting it into a single, stable
18-electron species.
Since 7 was formed by treating 4 with two molar equivalents
of PMe2Ph, and 10 by addition of PMe2Ph to 9, the simplest
method of trapping 11 seemed to be to perform each of the
decompositions in the presence of excess PMe2Ph. When
solutions of 7 and 10 containing free PMe2Ph were warmed
to 270 K, NMR studies confirmed that ethane elimination
was now accompanied by the formation of a single ruthenium
complex, and that both isomers gave the same complex, 12. The
complex was labile, with simultaneous broadening of resonances
for coordinated and free PMe2Ph in the 31P and 1H NMR spectra
of the solution, and we did not attempt to isolate it. The likeliest
formula for 12, however, was [Ru(CO)(PMe2Ph)4], since a 13C
NMR spectrum recorded at 283 K included a 1:4:6:4:1 quintet
at d 223.9, implying the presence of a carbonyl ligand coupled
to four apparently equivalent 31P nuclei. At this temperature,
the 31P resonance for 12 was a reasonably sharp singlet, and
38
in 2c, 3c17 and [Ru(CO)H(4-MePy)2(PPh3)2]BF4 respectively.
On this basis we assigned the ligand arrangement shown in
Scheme 2 to complex 9.
1
from the H and 13C spectra it appeared that all the methyl
substituents in the PMe2Ph ligands were equivalent. A ligand
arrangement consistent with these findings would be a square-
based pyramid (see Scheme 3), with CO at the apex and with
the four PMe2Ph ligands forming the base of the pyramid. The
methyl proton resonance was not well resolved (although it
became a sharp singlet on decoupling at the frequency of the
31P resonance for 12), but the methyl carbon resonance was
a clear triplet of triplets, implying that trans-|2JPP| was large
enough to result in “virtual coupling” (|1JPC + 3JPC| = 23.2 Hz)
between a given methyl carbon and both its “own” 31P nucleus
and the one trans to it,39 and that the resulting triplet was
further split by a weaker coupling (cis-|3JPC| = 3.5 Hz) to the
two 31P nuclei cis to it. It should be noted that the related Ru(0)
complex [Ru(PMe3)(Me2PCH2CH2PMe2)2] is known to have a
square-pyramidal ligand arrangement.40 In contrast, however,
[Ru(CO)(Me2PCH2CH2PMe2)2] is trigonal bipyramidal, and
Jones40 has suggested that this difference in geometry could be
linked to the preference of a strongly p-accepting ligand for an
equatorial position in the trigonal bipyramidal structure. If 12,
which also contains a carbonyl ligand, is trigonal bipyramidal, it
must be fluxional in solution, with pairs of axial and equatorial
PMe2Ph ligands rapidly exchanging positions at 283 K by the
Berry41 mechanism. NMR spectra of 12 recorded at 213 K did
show some broadening of the resonances for the 31P nuclei and
Scheme 2 The route to isomer 10 of [Ru(CO)(Et)H(PMe2Ph)3]
(L = PMe2Ph). The ligand arrangement in 8 is uncertain.
Since we had already decided to study the kinetics of
ethane elimination from 7, the possibility that 9 might exchange
its 4-MePy for PMe2Ph without change in the overall ligand
arrangement raised the prospect of obtaining, and studying
ethane elimination from, an isomer of 7. Indeed, treatment of
the C6D5CD3 solution of 9 with PMe2Ph at 250 K resulted in
the release of 4-MePy and the formation of a single complex
10 whose NMR spectra, recorded at 245 K, confirmed that it
was a second isomer of [Ru(CO)(Et)H(PMe2Ph)3] and clearly
indicated the ligand arrangement shown in Scheme 2. The
pattern of coupling constants for the resonance for the hydride
ligand at d −8.05 (dt, |2JPH| = 94.4 and 28.9 Hz respectively)
placed it trans to the unique PMe2Ph ligand,37 while those for
the carbon atom in the carbonyl ligand and for the methylene
carbon in the ethyl ligand indicated that both of these ligands
must be cis to all three PMe2Ph ligands. The chemical shift, d
0.51, for the methylene protons in the ethyl ligand of 10, while
3 7 9 2
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7

View Article Online
pared [Ru(g2-BD4)(CO)D(PMe2Ph)2], d5-1, by the method used
to obtain 1,17 but using deuterated reagents. The chemical shift
of the resonance for the 31P nuclei in d5-1 was virtually identical
with that for 1, and the same applied to the resonances for the
1
PMe2Ph ligands in the H NMR spectrum of d5-1. Very small
resonances were observed at the chemical shifts for the hydride
ligand and the tetrahydroborate hydrogens in 1, indicating that
deuteration was almost, but not quite, complete. Conversion
of d5-1 to [Ru(g2-BD4)(CO)(CH2CH2D)(PMe2Ph)2], d5-4, was
then attempted by treatment with C2H4 in C6D5CD3 at 250 K.
Over a period of some 44 h, the 31P resonance for d5-1 was
completely replaced by a new resonance with almost exactly
the same chemical shift as that for 4,17 and the new resonances
for the protons in the PMe2Ph ligands also matched those for
4. In the early stages of the reaction, the growing resonances
for the a- and b-protons in the ethyl ligand, at d 1.30 and d 0.99
respectively (very close to the values for 4), were approximately
equal in area. The a-proton resonance was a quintet (due to
roughly equal splittings by the two 31P nuclei and by two b-
protons), and the b-proton resonance was a triplet (the only
detectable splittings being by the two a-protons), exactly as
expected for [Ru(g2-BD4)(CO)(CH2CH2D)(PMe2Ph)2]. As the
reaction progressed, however, these resonances became rather
poorly resolved and unsymmetrical in shape, and integration in-
dicated that the final distribution of deuterium between a- and
b-positions was roughly statistical. Clearly the reversibility of
the reaction between 1 and ethene17 was allowing the [Ru(g2-
BD4)(CO)(CH2CH2D)(PMe2Ph)2] to equilibrate with [Ru(g2-
BD4)(CO)(CHDCH3)(PMe2Ph)2], and probably also with some
[Ru(g2-BD4)(CO)(CH2CH3)(PMe2Ph)2] and small amounts of
species containing two or more deuterium atoms in the ethyl
ligand. We did not anticipate (see later) that the rate of elimina-
tion from 7 would be significantly affected by either the presence
or the position of deuterium atoms in the ethyl ligand, so this
mixture (which, for simplicity, we will call d5-4) was used in the
preparation of d2-7. This was undertaken in the same way as the
conversion of 4 to 7. NMR spectra of d2-7 were very similar to
those of 7, with the exception of the a- and b-proton resonances
in the ethyl ligand which (although at the expected chemical
shifts) were rather complex and poorly resolved, and similar in
their relative areas to those for the samples of d5-4.
Scheme 3 Trapping reactions following ethane elimination from
[Ru(CO)(Et)H(PMe2Ph)3] (L = PMe2Ph).
the methyl protons, perhaps as a result of the decreased rate of
such a fluxional process.
A C6D5CD3 solution of 7 was also treated with H2 at 250 K.
The temperature of the solution was raised to 270 K, and at
this temperature the weakening of the resonance for dissolved
H2 at d 4.50 was accompanied by the steady growth of the sin-
glet resonance for ethane at d 0.80. The other product of the
reaction was shown by NMR studies to be the known complex
[Ru(CO)H2(PMe2Ph)3], 3a.37 Theaffinityof [Ru(CO)(PMe2Ph)3],
11, for H2, and the lability of 12 were both illustrated by the fact
that 12 was found to react with H2 at 260 K to give complete
conversion to 3a and PMe2Ph.
The third trapping agent tried was HCCCMe3. Both 7 and
10 reacted with HCCCMe3 in C6D5CD3 to give slow but com-
plete conversion to ethane and the same ruthenium(II) species,
[Ru(CO)(CCCMe3)H(PMe2Ph)3], 13. Attempts to obtain a
solid sample of 13 were unsuccessful, but NMR characterisa-
tion confirmed the presence and arrangement of all the ligands.
The 1H NMR spectrum contained a resonance at d −7.69
(dt, |2JPH| = 92.5 and 27.0 Hz respectively), indicating that the
hydride ligand was trans to the unique PMe2Ph ligand,37 and
there was a singlet at d 1.54 for the CMe3 protons. A quartet at
d 204.4 in the 13C spectrum showed that the carbonyl ligand was
cis to all three PMe2Ph ligands, and the sizes of the coupling
constants for the metal-bound carbon atom in the alkynyl ligand
(d 101.2, td, |2JPC| = 21.4 and 18.8 Hz, respectively) confirmed
that the same was true for the alkynyl ligand. Thus the ligand
arrangement was that shown in Scheme 3. In many respects,
the NMR spectra of 13 closely resembled those of the known
complex [Ru(CO)2(CCCMe3)H(PMe2Ph)2].42
1
Integration of the small hydride resonance in the H NMR
spectrum of the sample of d2-7 used in the kinetic studies,
relative to those for the methyl protons in the mutually trans
pair of PMe2Ph ligands, indicated that the deuteration in this
position was about 94% complete. This integration was also used
to help determine the mechanism of ethane elimination from
7. There is substantial evidence for the existence of r-alkane
complexes of several transition metals, in which the alkane is
attached to the metal without cleavage of a C–H bond.46–48
Computational studies support the view that such species must
also act either as intermediates or as transition states both in
the activation of alkanes by transition metals and in the process
of alkane elimination from complexes containing alkyl and
hydride ligands.49–52 In the event that a r-alkane complex is
an intermediate in alkane reductive elimination, and is formed
reversibly and relatively rapidly from the original complex, prior
to a slower step involving detachment of the alkane from the
metal, this will allow hydrogen exchange to occur between alkyl
and hydride ligands, provided that the metal can switch its point
of attachment from one C–H bond to another. There is evidence
to show that this intramolecular switching process between C–H
bonds can occur with a very low activation energy, particularly
when both bonds involve the same carbon atom, and many
cases of intramolecular hydrogen exchange of this type have
been reported.21,23,25,27,29–31,48,53 In order to discover whether such
an exchange occurred between ethyl and hydride ligands in 7,
a sample of d2-7 (ca. 94% deuterated in the hydride position)
was stored in C6D5CD3 solution at 253 K, a temperature at
which the elimination of ethane was very slow. After three days,
Scheme3summarisesthereactionsof 7and10withthetrapping
agents. The mechanism of the initial step will be discussed later
in the light of the kinetic results. The stereochemistry shown
for 11 in the scheme is based on the structure of [Ru(CO)2-
{PMe(CMe3)2}2],43 which is a somewhat distorted trigonal
bipyramid in which one equatorial site is vacant. It should,
however, be noted that [Ru(Me2PCH2CH2PMe2)2]44 is believed
to be planar, and that the C–Ru–C angle in [Ru(CO)2(PMe3)2]45
must be close enough to 180° to account for the failure to
observe an IR band for the symmetric stretching mode of the
two carbonyl ligands.
(iv) The preparation and study of [Ru(CO)(C2H4D)D-
(PMe2Ph)3], d2-7
In order to study the effect on the rate of ethane elimination
from 7 of replacing the hydride ligand by deuteride, we pre-
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
3 7 9 3

View Article Online
Table 2 Kinetic data used to obtain activation parameters for ethane
elimination from isomers 7 and 10 of [Ru(CO)(Et)H(PMe2Ph)3]a
integration of the hydride resonance relative to those for the
methyl protons in the mutually trans pair of PMe2Ph ligands in
the d2-7 remaining in the solution indicated no significant drop
in the level of deuteration in the hydride position. We concluded
that, if an alkane complex [Ru(CO)(r-C2H6)(PMe2Ph)3] was
indeed an intermediate in the process of ethane elimination
from 7, its rate of reconversion to 7 must be very low relative to
the rate of ethane loss to form 11.
Complex
T/K
104 k/s−1
7
277.2
0.660(30)
0.747(22)
0.838(25)
1.37(4)
1.53(4)
1.47(5)
3.18(8)
3.35(4)
3.27(5)
5.23(17)
5.33(16)
4.94(20)
0.599(12)
0.784(12)
0.638(14)
1.46(6)
1.19(6)
1.07(3)
2.53(5)
2.39(13)
2.55(9)
282.6
287.9
293.3
266.3
271.9
277.2
282.6
(v) Kinetic studies of ethane elimination from isomers 7 and 10
of [Ru(CO)(Et)H(PMe2Ph)3]
Solutions of 7 required for the kinetic studies were obtained by
treating C6D5CD3 solutions of 4 with the required quantity of
PMe2Ph (i.e. a molar ratio of 1:2). At 250 K this resulted in
rapid and complete conversion to 7. Similar treatment of 4 with
4-MePy at 250 K yielded 9, which was then converted to 10 by
adding an equivalent amount of PMe2Ph, still at 250 K. Where
PMe2Ph was the chosen trapping agent, this was allowed for by
increasing the initial addition of PMe2Ph to the solution of 4 or
9. Otherwise, the appropriate trapping agent was added prior
to the transfer of the solution of 7 or 10 to the NMR probe.
The progress of ethane elimination was normally monitored by
integrating the hydride resonance in the 1H NMR spectrum of
7 or 10. This resonance was chosen because it was well enough
separated from other reactant or product resonances to make
accurate integration straightforward.
10
4.30(7)
4.81(16)
4.59(7)
a For all these kinetic runs, the solvent was C6D5CD3 and the trapping
agent PMe2Ph.
In the initial kinetic run with 7, carried out at 282.6 K, no
trapping agent was added. Despite the range of ruthenium
decomposition products formed, the plot for the first-order
disappearance of 7 was essentially linear for over 2.5 half-lives,
giving a rate constant of 1.53(4) × 10−4 s−1. This simple kinetic
behaviour was, in itself, a strong indication that ethane reduc-
tive elimination occurred directly from 7, and did not require
the prior loss of some other ligand, or attack by an external
species. A second run, also at 282.6 K, was carried out using a
similar concentration of 7 (0.06 mol dm−3) but with a substantial
concentration of free PMe2Ph (0.23 mol dm−3), so that complete
conversion to 12 occurred during the run. Again the first-order
plot was essentially linear for over 2.5 half-lives, despite the
fact that the concentration of free PMe2Ph in the solution fell
by around 25% during the run. This fact, and the good agree-
ment between the rate constant obtained, 1.47(5) × 10−4 s−1, and
that obtained in the absence of free PMe2Ph indicated that the
PMe2Ph was not involved in the rate-determining step. We con-
cluded that this step was indeed the elimination of ethane, and
that it was not preceded or accompanied by PMe2Ph dissocia-
tion from (or addition to) 7.
A further check was made by determining the rate con-
stant at 282.6 K for the disappearance of 7 in the presence
of HCCCMe3. For this run, initial concentrations of 7 and
HCCCMe3 were 0.06 and 0.08 mol dm−3, respectively. Com-
plete conversion to 13 was observed, and values for the rate
constant were derived both by monitoring the disappearance
of the hydride resonance for 7 and by following the appearance
of that for 13. Despite the large variation in the concentration
of free HCCCMe3 during the run, both plots were essentially
linear for nearly 2.5 half-lives, ruling out any involvement of
HCCCMe3 in the rate-determining step. The rate constants
obtained, 1.42(3) × 10−4 s−1 and 1.55(8) × 10−4 s−1 respectively,
were in reasonable agreement with one another and with the val-
ues obtained with no trapping agent and with PMe2Ph as trap.
For isomer 10 of [Ru(CO)(Et)H(PMe2Ph)3], kinetic stud-
ies were carried out only for the reaction with PMe2Ph to
give ethane and 12. Three runs were carried out at 277.2 K,
the initial concentrations of 10 being 0.07, 0.04 and 0.04 mol
dm−3, and those of free PMe2Ph 0.14, 0.17 and 0.05 mol dm−3,
respectively. Each first-order plot was essentially linear for at
least 2.5 half-lives, despite the substantial variation in PMe2Ph
concentration during the runs (particularly the last of the three).
The rate constants obtained were 2.53(5) × 10−4, 2.55(9) × 10−4
and 2.39(13) × 10−4 s−1, respectively. As with 7, it appeared
that ethane elimination from 10 was a simple first-order pro-
cess, and that the only role of the PMe2Ph was to capture
[Ru(CO)(PMe2Ph)3], 11, the immediate ruthenium product of
the elimination.
Further runs were then carried out, all with PMe2Ph as the
trapping agent, to obtain values for the activation parameters
for ethane elimination from 7 and 10. The highest tempera-
tures at which we managed to obtain a satisfactory amount
of data within two or three half-lives were 293.3 K for 7 and
282.6 K for 10. At lower temperatures we were increasingly
hampered by the amount of instrument time required and by
poorer reproducibility of the values obtained for the rate con-
stants. Despite this, the data listed in Table 2 gave reasonably
satisfactory Eyring plots. From these, values of DH‡ and DS‡
of 80(6) kJ mol−1 and −37(21) J K−1 mol−1, respectively, were
obtained for 7, whereas for 10 the values were 71(7) kJ mol−1
and −59(25) J K−1 mol−1.
In order to determine the effect on the rate of ethane elimina-
tion from 7 of replacing the hydride ligand by a deuteride ligand,
westudiedeliminationfromd2-7,[Ru(CO)(C2H4D)D(PMe2Ph)3].
Our assumption (see earlier) that the rate of elimination from 7
would not be significantly affected by the presence of deuterium
in the ethyl ligand was supported by the work of Parkin and
Bercaw,30 who found that, whilst the rate of CH4 elimination
from [W(g5-C5Me5)2(Me)H] differed markedly from the rate
of CH3D elimination from [W(g5-C5Me5)2(Me)D], the rates of
elimination of CH3D and CD4 from [W(g5-C5Me5)2(Me)D] and
[W(g5-C5Me5)2(CD3)D], respectively, were virtually identical.
Since the kinetic studies of ethane elimination from 7 and
10 had been performed by monitoring the disappearance from
1
the H NMR spectrum of the resonance due to the hydride
ligand, a change in procedure was necessary for d2-7. We chose
to follow the disappearance of the resonances due to the methyl
protons in the mutually trans PMe2Ph ligands. In order to check
that this would be a satisfactory procedure, a further kinetic
run was carried out on 7 itself, at a temperature of 287.9 K,
and using initial concentrations of 7 and the trapping agent
PMe2Ph of 0.07 and 0.21 mol dm−3, respectively. Rate constants
were obtained using both the hydride resonance for 7 and the
methyl proton resonances mentioned above. The two plots, both
3 7 9 4
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7

View Article Online
essentially linear for at least 2.5 half-lives, gave rate constants of
3.27(5) × 10−4 s−1 and 3.10(7) × 10−4 s−1, respectively. Since the
two figures were in reasonable agreement, a run was then carried
out with d2-7 and PMe2Ph, using the same concentrations and
the same temperature as those for 7. Again the plot was linear for
over 2.5 half-lives, and the rate constant, 1.22(2) × 10−4 s−1, was
significantlylowerthanthosefor 7. Comparisonof thisvaluewith
the one obtained for 7 by the same technique, 3.10(7) × 10−4 s−1,
gave a figure of 2.5 for the ratio kH/kD at 287.9 K.
in such cases Parkin31 argues that a normal kinetic isotope
effect should be observed.
Our study of d2-7 {see Section (iv)} had clearly shown that, if
a r-ethane complex [Ru(CO)(r-C2H6)(PMe2Ph)3] is an interme-
diate in the process of reductive elimination from 7, its rate of
conversion to 7 must be very low relative to the rate of ethane
loss to give 11 {i.e. k2  k3 in eqn. (1)}. On this basis the rate
constant for the reductive elimination would simply be k1, and
our observation of a normal kinetic isotope effect (kH/kD = 2.5)
would be in accordance with Parkin’s proposals. The negative
entropies of activation we observed for ethane elimination from
both 7 and 10 could presumably be attributed to the develop-
ment of some degree of interaction between methyl and hydride
ligands en route to the transition state in the k1 step. It should
be noted, though, that our results are equally compatible with a
simpler mechanism in which there is no intermediate, and only
one transition state.
Normal kinetic isotope effects have been observed for alkane
reductive elimination from a number of other complexes, in-
cluding several of platinum(II).20,26,28 For these complexes, there
was no suggestion that an equilibrium between the platinum(II)
starting material and an intermediate platinum(0) r-alkane
complex might precede the loss of the alkane. An interesting
case is that of [Ir(CO)(Et)H2(Ph2PCH2CH2PPh2)], studied by
Deutsch and Eisenberg.32 As in the case of the two isomers, 7
and 10, of [Ru(CO)(Et)H(PMe2Ph)3], but in sharp contrast to
[Ir(g5-C5Me5)(C6H11)H(PMe3)],27 alkane elimination shows a
normal kinetic isotope effect (and also a negative entropy of
activation). Given that [Ir(CO)(Et)H2(Ph2PCH2CH2PPh2)] con-
tains a very similar ligand set to those in 7 and 10, it may well be
that the nature of the ligands in a complex can be as important
as the choice of metal in affecting the mechanism of alkane
elimination.55
(vi) The mechanism of ethane elimination from 7 and 10
The results of the reactions described in Section (iii) had sug-
gested that both 7 and 10 decomposed by ethane elimination to
give [Ru(CO)(PMe2Ph)3], 11, which could then be trapped as a
stable ruthenium(0) or ruthenium(II) species. In the case of 7,
the fact that ethane elimination occurred from 7 itself, and did
not require prior loss of some other ligand or attack by some
external species, was clearly shown by the simple first-order be-
haviour of the reaction, whose rate was not significantly affected
by the presence or absence of a reagent to trap 11, by variation
in the initial concentration of such a reagent, variation in its
concentration during a given run, or even a change in the reagent
used. The more limited study of 10 led to similar conclusions.
This simple behaviour mirrors that exhibited by a variety
of other complexes, containing metals such as rhodium,23,25
iridium,32 platinum,20,28 rhenium21 and tungsten.29–31 There
are, however, exceptions: for example, Flood11 has presented
evidence to show that one pathway for CMe4 elimination
from [Os(CH2CMe3)H(PMe3)4] involves prior loss of PMe3,
and Bercaw24 has proposed that ligand loss (either chloride or
a solvent molecule) precedes alkane elimination from some
platinum(IV) complexes.
It should be appreciated that the difference between the rate
constants for ethane elimination from 7 and 10 at each of the
two temperatures common to both studies was only a factor of
about 3, corresponding to a very small difference in DG‡ values.
Even the rather larger differences between the pairs of DH‡ and
DS‡ values were not statistically significant, so the conclusion
must be that the mechanism of the elimination is probably simi-
lar for the two isomers.
In summary, our results show that reductive elimination of
ethane from 7 and 10 is a simple first-order process, requiring
neither prior or simultaneous loss of another ligand nor attack
by some external species. They do not eliminate the possibility
that a r-ethane complex acts as an intermediate (rather than
simply as a transition state) in the process, but unequivocally
rule our a rapid pre-equilibrium with such an intermediate prior
to ethane loss.
There has been much interest in the fact that the reductive
elimination of alkanes from some transition metal complexes is
associated with an inverse kinetic isotope effect (i.e. elimination
from a particular hydride complex is slower than that from the
corresponding deuteride complex). In a recent paper on methane
elimination from [W{(g5-C5Me4)2SiMe2}(Me)H], Parkin and co-
workers31 have emphasised that the key to this effect lies in the
existence of an intermediate alkane r-complex {see eqn. (1)}:
Experimental
All experimental work (except the preparations of 4 and d5-4)
was carried out under an atmosphere of N2. The NMR spectra
(including those used to obtain kinetic data) were recorded on
a Bruker AMX 500 spectrometer. The preparation of complex
1 and its conversion to 4 have been described in the literature:17
the conversion was carried out at 273 K. The same methods were
used to obtain d5-1 and d5-4: for the former, NaBD4 and EtOD
were used instead of NaBH4 and EtOH, and for the latter (see
previous section) the conversion was carried out at 250 K. When
4 or d5-4 was stored in C6D5CD3 solution, this was performed at
253 K under an atmosphere of ethene. Before the solution was
used, the ethene was removed by purging the solution with N2.
k1
k3


k2
[M](R)(H)
[M](σ −HR) → [M] + RH
(1)
where [M] represents the metal together with ligands not
directly involved in the reductive elimination. In cases where
k2  k3, the rate constant for the overall reductive elimination
is given by k1k3/k2, and Parkin argues that it is the shift of
the equilibrium between starting material and intermediate
r-complex on replacing hydrogen by deuterium which causes
the inverse kinetic isotope effect (i.e. k1D/k2D > k1H/k2H). This
type of kinetic behaviour should therefore be associated with
deuterium scrambling between alkyl and hydride ligands, and
Parkin has shown that such scrambling does indeed occur
in [W{(g5-C5Me4)2SiMe2}(Me)D].31 Bergman and his co-
workers have established a similar link between an inverse
kinetic isotope effect and deuterium scrambling for [Rh(g5-
C5Me5)(Et)H(PMe3)]25 and [Ir(g5-C5Me5)(C6H11)H(PMe3)],27
as have Flood et al.23 for [Rh(Cn)(Me)H(PMe3)]+ {Cn =
(MeNCH2CH2)3} and Jones et al.22,54 for a range of complexes
[Rh(Tp)(CNCH2CMe3)(R)H] {Tp = tris-(3,5-dimethylpyra
zolyl)borate, R = alkyl}. In contrast, where k2  k3, the rate
constant for the overall reductive elimination is simply k1, and
The routes from 4 to 7 and from 4 to 10
The reactions were carried out in NMR tubes, typically using
17 mg (0.04 mmol) of 4 in 1 cm3 of C6D5CD3 or CD3COCD3.
Details of the other reactants, molar ratios of reactants and tem-
peratures employed have been given earlier in the text. Because
of the sensitivity of 7 and 10 to ethane elimination, no attempt
was made to isolate them from solution in a pure state.
Trapping reactions of 7 and 10
Solutions of
7 required for these reactions could be
straightforwardly prepared in an NMR tube by treating 4
{typically between 16 mg (0.04 mmol) and 36 mg (0.09 mmol)}
in C6D5CD3 solution (1 cm3) with two molar equivalents of
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
3 7 9 5

View Article Online
PMe2Ph at 250 K, using NMR spectroscopy to check that
conversion to 7 was complete. Similar treatment of 4 with two
molar equivalents of 4-MePy at 250 K yielded 9: after checking
that conversion to 9 was complete, one molar equivalent of
PMe2Ph was added, also at 250 K, to convert 9 into 10.
each spectrum the area of the resonance monitored was divided
by the area of the resonance for the small amount of CD2HC6D5
present in the solution. Attempts were also made to obtain
rate constants by monitoring the resonance for free ethane.
Unfortunately the growth of this resonance tailed off as kinetic
runs progressed, and the resonance actually decreased somewhat
in area towards the end of the reaction. Presumably this was due
to some loss of ethane into the gas phase above the solution.
It should be noted that the solutions used for the kinetic runs
also contained the adducts H3B·PMe2Ph or H3B·4-MePy. Two
runs carried out at 277.2 K with similar PMe2Ph concentra-
tions, but with markedly different concentrations (0.07 and
0.04 mol dm−3) of 10 (and therefore also of H3B·4-MePy)
(i) Trapping with PMe2Ph. For these reactions only, the
preparations of 7 from 4 and of 10 from 9 were carried out
using more PMe2Ph than the amounts given above, so that ca.
two molar equivalents of PMe2Ph remained in the solutions
of 7 and 10. When the temperature of the NMR probe was
raised to 270 K, both 7 and 10 were slowly converted to 12 with
elimination of ethane. Because of its lability (see earlier), 12 was
characterised only by NMR spectroscopy.
gave rate constants of 2.53(5) × 10−4 s−1 and 2.55(9) × 10−4 s−1
,
respectively, suggesting that the adduct had no significant effect
on reaction rate (and also that reaction rate did not depend on
the initial concentration of the ruthenium complex used).
(ii) Trapping with H2. A solution of 7, in an NMR tube fit-
ted with a Young’s tap, was connected to the vacuum manifold
of a Schlenk line and subjected to three freeze–pump–thaw
cycles in order to achieve complete degassing. The solution was
refrozen, and H2 was introduced into the NMR tube by open-
ing the tap to the gas manifold of the Schlenk line, which had
been filled with H2 at 1 atm. pressure. After closing the tap, the
solution was allowed to warm up to 250 K and shaken to ensure
thorough mixing. The tube was then placed in the NMR probe,
pre-cooled to 250 K. When the probe temperature was raised to
270 K, ethane elimination occurred over a period of hours, with
formation of 3a, identified by comparison of its NMR spectra
with those of an authentic sample of the complex.37 Complex
3a was also formed even when the solution of 7 used contained
free PMe2Ph, and when a solution of 12 (obtained as described
above) was allowed to react with H2 at 260 K.
Acknowledgements
We thank Johnson Matthey PLC (“JM”) for a generous loan
of ruthenium trichloride, and Professors Odile Eisenstein and
Robin Perutz for most helpful discussions.
References
1 See, for example, G. W. Parshall and S. D. Ittel, Homogeneous
Catalysis, 2nd edn., Wiley-Interscience, New York, 1992.
2 J. Chatt and R. G. Hayter, J. Chem. Soc., 1963, 6017.
3 D. J. Cole-Hamilton and G. Wilkinson, J. Chem. Soc., Dalton Trans.,
1977, 797.
4 J. A. Statler, G. Wilkinson, M. Thornton-PettandM. B. Hursthouse,
J. Chem. Soc., Dalton Trans., 1984, 1731.
5 W.-K. Wong, K. W. Chiu, J. A. Statler, G. Wilkinson, M. Motevalli
and M. B. Hursthouse, Polyhedron, 1984, 3, 1255.
6 J. F. Hartwig, R. A. Anderson and R. G. Bergman, Organometallics,
1991, 10, 1710.
7 A. H. Janowicz and R. G. Bergman, J. Am. Chem. Soc., 1983, 105,
3929.
8 J. K. Hoyano, A. D. McMaster and W. A. G. Graham, J. Am. Chem.
Soc., 1983, 105, 7190.
9 W. D. Jones and F. J. Feher, Organometallics, 1983, 2, 562.
10 R. A. Periana and R. G. Bergman, Organometallics, 1984, 3, 508.
11 T. G. P. Harper, R. S. Shinomoto, M. A. Deming and T. C. Flood,
J. Am. Chem. Soc., 1988, 110, 7915.
(iii) Trapping with HCCCMe3. Solutions of 7 and 10 were
treated with an equimolar quantity of HCCCMe3 at 250 K,
and then placed in the probe of the spectrometer at 250 K. As in
the case of the reactions with PMe2Ph, the effect of raising the
probe temperature to 270 K was to cause ethane elimination.
Both 7 and 10 yielded the same product, 13. Attempts to obtain
a solid sample of 13 by column chromatography and crystallisa-
tion were unsuccessful, but the complex was characterised by
NMR spectroscopy.
12 M. V. Baker and L. D. Field, J. Am. Chem. Soc., 1987, 109, 2825.
13 R. G. Bergman, P. F. Seidler and T. T. Wenzel, J. Am. Chem. Soc.,
1985, 107, 4358.
14 C. F. J. Barnard, J. A. Daniels and R. J. Mawby, J. Chem. Soc.,
Dalton Trans., 1976, 961.
15 J. P. Lowe and C. P. Orme, unpublished work.
16 J. M. Bray and R. J. Mawby, J. Chem. Soc., Dalton Trans., 1987,
2989.
17 B. Chamberlain, S. B. Duckett, J. P. Lowe, R. J. Mawby and J. C.
Stott, J. Chem. Soc., Dalton Trans., 2003, 2603.
18 J. C. Green and C. N. Jardine, J. Chem. Soc., Dalton Trans., 1998,
1053.
Kinetic studies
Solutions of 7 or 10, typically containing between 0.04 and 0.07
mmol of the complex, were prepared from 4 as described under
“Trapping reactions of 7 and 10”. The procedure to obtain a
solution of d2-7 from d5-4 was the same as that for obtaining
7 from 4. For kinetic runs with PMe2Ph as the trapping agent,
the amount of PMe2Ph added was in excess of that required to
produce 7, d2-7 or 10, and the concentration of PMe2Ph actually
present in the solution at the start of the kinetic run was calcu-
lated by allowing for the amount consumed in the production
of 7, d2-7 or 10. For the run in which HCCCMe3 was used as
the trapping reagent, a solution of 4 was treated with exactly the
amount of PMe2Ph required to convert it to 7. The appropriate
amount of HCCCMe3 was then added to the solution of 7.
The NMR tube was then transferred to the probe of the spec-
trometer, which had been pre-cooled to the desired temperature.
For each setting of the variable temperature unit of the NMR
spectrometer used in kinetic runs, the true probe temperature
was determined by calibration using a methanol capillary held
in an NMR tube containing C6D5CD3.56
19 R. L. Martin, J. Am. Chem. Soc., 1999, 121, 9459.
20 L. Abis, A. Sen and J. Halpern, J. Am. Chem. Soc., 1978, 100,
2915.
21 G. L. Gould and D. M. Heinekey, J. Am. Chem. Soc., 1989, 111,
5502.
22 T. O. Northcutt, D. D. Wick, A. J. Vetter and W. D. Jones, J. Am.
Chem. Soc., 2001, 123, 7257.
23 C. Wang, J. W. Ziller and T. C. Flood, J. Am. Chem. Soc., 1995, 117,
1647.
24 S. S. Stahl, J. A. Labinger and J. E. Bercaw, J. Am. Chem. Soc., 1996,
118, 5961.
25 R. A. Periana and R. G. Bergman, J. Am. Chem. Soc., 1986, 108,
7332.
After allowing a few minutes for the sample to reach probe
temperature, spectra were recorded at appropriate intervals. For
each point on a given kinetic plot, eight scans were accumulated:
the total time required represented only a very small fraction
of the overall reaction time, even at the highest temperature
used. The resonances used to monitor the reactions have been
identified in the Results section. As a safeguard against any
variations in instrument performance during a kinetic run, in
26 M. Hackett, J. A. Ibers and G. M. Whitesides, J. Am. Chem. Soc.,
1988, 110, 1436.
27 J. M. Buchanan, J. M. Stryker and R. G. Bergman, J. Am. Chem.
Soc., 1986, 108, 1537.
28 R. A. Michelin, S. Faglia and P. Uguagliati, Inorg. Chem., 1983, 22,
1831.
29 R. M. Bullock, C. E. L. Headford, K. M. Hennessy, S. E. Kegley
and J. R. Norton, J. Am. Chem. Soc., 1989, 111, 3897.
30 G. Parkin and J. E. Bercaw, Organometallics, 1989, 8, 1172.
3 7 9 6
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7

View Article Online
31 D. G. Churchill, K. E. Janak, J. S. Wittenberg and G. Parkin, J. Am.
Chem. Soc., 2003, 125, 1403.
32 P. P. Deutsch and R. Eisenberg, J. Am. Chem. Soc., 1990, 112, 714.
33 M. Stephenson and R. J. Mawby, J. Chem. Soc., Dalton Trans., 1981,
2112.
45 R. J. Mawby, R. N. Perutz and M. K. Whittlesey, Organometallics,
1995, 14, 3268.
46 C. Hall and R. N. Perutz, Chem. Rev., 1996, 96, 3125.
47 X.-Z. Sun, D. C. Grills, S. M. Nikiforov, M. Poliakoff and M. W.
George, J. Am. Chem. Soc., 1997, 119, 7521.
34 C. F. J. Barnard, J. A. Daniels, P. R. Holland and R. J. Mawby,
J. Chem. Soc., Dalton Trans., 1980, 2418.
35 C. Bianchini, P. J. Perez, M. Perruzzini, F. Zanobini and A. Vacca,
Inorg. Chem., 1991, 30, 279.
36 J. B. Letts, T. J. Mazanec and D. W. Meek, J. Am. Chem. Soc., 1982,
104, 3898.
37 J. W. Bruno, J. C. Huffman and K. G. Caulton, Inorg. Chim. Acta,
1984, 89, 167.
38 D. F. Mullica, J. M. Farmer, J. A. Kautz, S. L. Gipson, Y. F. Belay
and M. S. Windmiller, Inorg. Chim. Acta, 1999, 285, 318.
39 R. K. Harris, Can. J. Chem., 1964, 42, 2275.
40 W. D. Jones and E. Libertini, Inorg. Chem., 1986, 25, 1794.
41 R. S. Berry, J. Chem. Phys., 1960, 32, 933.
48 S. Geftakis and G. E. Ball, J. Am. Chem. Soc., 1998, 120, 9953.
49 J. C. Green, J. N. Harvey and R. Poli, J. Chem. Soc., Dalton Trans.,
2002, 1861.
50 T. R. Cundari, T. R. Klinckman and P. T. Wolczanski, J. Am. Chem.
Soc., 2002, 124, 1481.
51 K. E. Janak, D. G. Churchill and G. Parkin, J. Chem. Soc., Chem.
Commun., 2003, 22.
52 H. Gerard, O. Eisenstein, D.-H. Lee, J. Chen and R. H. Crabtree,
New J. Chem., 2001, 25, 1121.
53 C. L. Cross and G. S. Girolami, J. Am. Chem. Soc., 1998, 120,
6605.
54 D. D. Wick, K. A. Reynolds and W. D. Jones, J. Am. Chem. Soc.,
1999, 121, 3874.
42 J. M. Bray and R. J. Mawby, J. Chem. Soc., Dalton Trans., 1989, 589.
43 M. Ogasawara, S. A. MacGregor, W. E. Streib, K. Folting,
O. Eisenstein and K. G. Caulton, J. Am. Chem. Soc., 1996, 118,
10189.
44 C. Hall, W. D. Jones, R. J. Mawby, R. Osman, R. N. Perutz and
M. K. Whittlesey, J. Am. Chem. Soc., 1992, 114, 7425.
55 A referee notes that the difference in behaviour no doubt relates to
a difference in geometry between [Ir(g5-C5Me5)(PMe3)] and the spe-
cies [Ir(CO)(H)(Ph2PCH2CH2PPh2)] and [Ru(CO)(PMe2Ph)3], both
expected to be approximately square planar.
56 C. Ammann, P. Meier and A. E. Merbach, J. Magn. Reson., 1982,
46, 319.
D a l t o n T r a n s . , 2 0 0 4 , 3 7 8 8 – 3 7 9 7
3 7 9 7