Pairwise ligand exchange reactions in tetrahedral and pseudo-tetrahedral transition metal complexes

Article abstract of DOI:10.1039/b308062e

Ligand exchange reactions between various transition-metal complexes of the form [Mo(Q)₂X₂], [M(Q)CpX₂] and [Mo(Q)₂CpX], where Q and X are di-anionic and mono-anionic ligands, respectively, and M is Nb or Ta, ha

Full text of DOI:10.1039/b308062e

View Article Online / Journal Homepage / Table of Contents for this issue
Pairwise ligand exchange reactions in tetrahedral and pseudo-
tetrahedral transition metal complexes†
Vernon C. Gibson,* Andrew J. Graham, Matthew Jolly and Jonathan P. Mitchell
Department of Chemistry, Imperial College, South Kensington, London, UK SW7 2AY
Received 15th July 2003, Accepted 10th September 2003
First published as an Advance Article on the web 30th September 2003
Ligand exchange reactions between various transition-metal complexes of the form [Mo(Q)2X2], [M(Q)CpX2] and
[Mo(Q)2CpX], where Q and X are di-anionic and mono-anionic ligands, respectively, and M is Nb or Ta, have been
studied. The formal electron count at the metal, and also the steric and electronic properties of the supporting
ligands, are found to influence exchange rates, while ligand steric and electronic properties affect equilibrium
constants.
Introduction
The exchange of anionic ligands between electropositive metal
centres forms a cornerstone of inorganic synthetic method-
ology. One of the best known and most widely exploited
examples is the metathetical exchange of a halide ligand for a
group such as alkyl or aryl using main group compounds such
as Grignard reagents or their alkali metal relatives.1,2 Analo-
gous processes are also widely documented for transition-metal
species, usually involving mono-anionic groups such as halide,
alkoxide or amide, etc.3–7 More recently examples involving
dianionic groups such as oxo, imido or alkylidene have been
described.8–15
Some time ago we initiated a programme of study to learn
more about these fundamentally important and synthetically
useful transformations.16 A potential challenge in studying this
type of process for transition metals is the enormous variety,
not only of metals, but also of coordination numbers and
ligand combinations available; the task is then one of selecting
a suitable coordination geometry, ligand types and metals
which would allow the factors contributing to the exchange
process to be delineated.
Following initial observations of relatively clean and facile
exchange processes in some molybdenum and niobium systems,
we decided to focus on pseudo-tetrahedral complexes of the
type [M(Q)2X2] and [MЈ(Q)X3] where X is a mono-anionic
ligand such as halide, alkoxide, amide or even cyclopentadienyl
and where Q is a dianionic group such as oxo, imido or alkyl-
idene. This coordination geometry offers several advantages: (i)
a wide variety of closely related derivatives can be readily syn-
thesised, (ii) a tetrahedral or pseudo-tetrahedral geometry is
maintained in the reactants and products,‡ and (iii) the steric
and electronic environment around the metal can be varied in a
controlled manner through appropriate choice of Q and X
ligands. An additional facet of this system is that it has the
potential to allow competitive effects between the ligands to
be assessed since there exists a direct competition between the
ligands for the available dπ-symmetry metal orbitals in a tetra-
hedral geometry.17
Fig. 1 Generalised representations of the three classes of ligand
exchange systems.
the cyclopentadienyl moiety tend to be more sterically
hindered; this is partly due to the fact that five ligand (carbon)
atoms are interacting with the metal centre compared with only
one donor atom for an oxo or imido group. Secondly, the elec-
tron count of the complexes effectively increases in the order I
< II < III. For example, if the Cp ligand is considered to contri-
bute five electrons, and the dianionic Q and monianionic X
ligands their minimum quota of two and one electrons (as
neutral ligands) respectively, the electron count proceeds from
12 to 14 to 16 for Classes I through III. It has to be recognised,
however, that for ligand donor atoms possessing lone pairs of
electrons (which is the case for both the Q and X ligands dis-
cussed here), additional electron density will usually be donated
to the metal centre via pπ to dπ donation depending on the
electronegativity of the donor atom and the availability of
empty metal d-orbitals of suitable symmetry.§ Such ligands
inevitably lead to increased electron density at the metal centre
over and above the minimum electron count provided by the
ligand set. Nevertheless, the trend of increasing electron count
from Classes I through III is clear.
Earlier investigations
In a previous study we examined the exchange of the oxo and
imido ligands in four-coordinate molybdenum complexes of
Class I. For the reactants Mo(N-2,6-iPr2C6H3)2(OtBu)2 and
Mo(O)2(OtBu)2 (eqn. (1)), the exchange reaction was found to
proceed cleanly and at a convenient rate for a kinetic study by
NMR spectroscopy. However, the rate data for this reaction
were later found to be difficult to reproduce, and appeared to be
sensitive to the drying procedure for the glassware and different
batches of solvent. This led us to suspect that a proton source,
possibly derived from moisture, may be influencing the reac-
tion rate which, in turn, led us to examine more generally the
effect of proton acid catalysis on the pairwise ligand exchange
process. In a subsequent investigation, a clear role for acid in
The three systems forming the basis of this study are shown
in Fig. 1. Their steric and electronic characteristics are worth
pointing out at this stage. Firstly, the two systems containing
† Electronic supplementary information (ESI) available: Experimental
procedures for synthesis of new compounds and for NMR-scale ligand
exchange reactions as well as 1H NMR spectroscopic data characteris-
ing products of these reactions. See http://www.rsc.org/suppdata/dt/b3/
b308062e/
‡ X-Ray crystal structure determinations on representative examples of
[Mo(Q)2X2] and [Nb(Q)CpX2] compounds, including [Mo(O)2(OtBu)2]
and [Nb(NtBu)Cp(OCMe2CF3)2], reveal mononuclear structures: M.
Jolly, PhD thesis, University of Durham, 1994.
§ For tetrahedral and pseudo-tetrahedral metal complexes, it is possible
to form only three pπ–dπ bonds and so the ligands are effectively in
competition for the available metal orbitals.
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 3
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4457

View Article Online
catalysing the exchange reaction was established.18 It also
appeared that the oxo–imido exchange reaction shown in
eqn. (1) may be particularly sensitive due to the propensity
Table 1 Species of the type [Mo(Q)2X2]
(Q)2
X2
(O)2
(NtBu)2
(NAr)2
(NAr)(CHCMe2Ph)
(OtBu)2
(OtBuF3
(OtBuF6
(OiPr)2
Cl2
117
5
8
2
6
317
421
721
(1)
)
)
2
2
9
1021
11
1222
1323
of [Mo(O)2(OtBu)2] to undergo autocatalytic decomposition
upon exposure to trace amounts of moisture. The product
of bulk decomposition of [Mo(O)2(OtBu)2] is a blue solid,19
the so-called ‘molybdenum blues’, which comprise Mo()
oxo–hydroxide cluster species.20 Although a blue solid is not
observed during the exchange reaction, it is possible that trace
decomposition could occur with the generation of a proton
source which then affects the rate of the exchange reaction. We
report here the results of a further kinetic study on this system
using more rigorous drying procedures, along with a study of
exchange reactions involving a wide variety of tetrahedral and
pseudo-tetrahedral species of types I to III.
(NHtBu)2
Ar = 2,6-iPr2C6H3; tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2.
Syntheses of uncited compounds are described in the present work.
Table 2 Species of the type [M(Q)(η5-C5R5)X2]
(Q)
X2
(NAr)
(NtBu)
(NMe)
Cl2
Cp 1424 Cp* 15;24 Ta 15a24
1624
Cp 1724
Cp* 1824
(OtBu)2
(OtBuF3
(OtBuF3
(OiPr)2
Br2
19
Results
)
)
20
2
2
21
Introduction
22
Eight main classes of exchange reaction have been addressed;
these are outlined in Scheme 1. Type A entails exchange between
two different [Mo(Q)2X2] (Class I) species; either of their di-
anionic (Type A(i)) ligands or their monoanionic (Type A(ii))
ligands. Similar exchange processes (Types B(i) and B(ii)) may
be envisaged between two species of the type [M(Q)(η5-C5R5)-
X2] (Class II) {where R = H or Me and M is a metal from
Group 5, namely Nb or Ta}. We have also studied dianionic
ligand exchange reactions between Class I and Class II com-
pounds (C), between Class I and compounds of the type
23
(NHtBu)2
(Pyrrolyl)2
2425
25
Ar = 2,6-iPr2C6H3; tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2;
pyrrolyl = (NC4H4); M = Nb except where stated otherwise. Syntheses
of uncited compounds are described in the present work.
Table 3 Other reactants
[Mo(NtBu)2CpCl]
2626
[Mo(O)2Cl2(DME)]
2727,28
2827,28
[Mo(NtBu)2Cl2(DME)]
DME = 1,2-Dimethoxyethane.
[Mo(Q)2(η5-C5R5)X] (Class III, D) and between compounds of
Classes II and III (E).
For convenience, the starting species involved in the ligand
exchange reactions studied are labelled with numerals and
grouped together in Tables 1–3. Similarly, mixed ligand species
produced by the pairwise exchange process are labelled with
capital letters and subscripts to indicate the two starting species
from which the product is derived (Tables 4 and 5).
1
The ligand exchange reactions were monitored by H NMR
spectroscopy. This is an advantageous technique since the
species involved in, and synthesised by, these exchange reac-
1
tions tend to afford clean H NMR spectra with sharp and
readily assignable resonances (see for examples Figs. 2 and 5).
Thus accurate concentrations of all the solution species can be
determined at all times by comparison of the integrals of
selected peaks, chosen for maximum clarity and comparability.
A(i): Dianionic ligand exchange between complexes of the form
[Mo(Q)2X2]
The exchange reaction shown in eqn. (1)16 has been reinvesti-
gated. At room temperature in C6D6, a 0.048 mol dmϪ3 solution
of a mixture of 1 and 3 equilibrated over ca. 4 weeks to give
a mixture of 1, 3 and B1,3 with an equilibrium constant,
Keq(298 K), of 1.37 (Fig. 2). The plot arising from a second
order treatment of the data29 (see Experimental section) is
shown in Fig. 3, affording a rate constant, k1(298 K), of 7.32 ×
10Ϫ6 dm3 molϪ1 Ϫ1. The effect of temperature on rate con-
s
stant for this reaction was probed over the temperature range
20–60 ЊC; the data are collected in Table 6. From a plot of
Scheme 1 Ligand exchange reactions employed in the present study.
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4458

View Article Online
Table 4 Mixed dianionic ligand species produced by ligand exchange
reactions
(X)2 or X
Q, QЈ
(a) [Mo(Q)(QЈ)X2]
(OtBu)2
(OtBuF3
)
(OtBuF6
)
2
2
(O)(NtBu)
(O)(NAr)
A1,2
B1.3
C2,3
D5,6
E8,9
(NtBu)(NAr)
a
(b) [Mo(Q)(QЈ)CpX]
Cl
(NtBu)(NAr)
(NAr)(NAr)
(NtBu)(NMe)
EЈ3,26
FЈ3,EЈ/C,EЈ
GЈ17,26
(c) [Mo(Q)(QЈ)X2(DME)]
Cl2
(O)(NtBu)
HЈ27,28
Ar = 2,6-iPr2C6H3; tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2; DME
= 1,2-dimethoxyethane. a Or C3,26 or C3,19 or C3,EЈ
.
Table 5 Mixed monoanionic ligand species produced by ligand
exchange reactions
(Q)2 or Q
(X)(XЈ)
(a) [Mo(Q)2(X)(XЈ)]
(O)2
(tBuN)2
(NAr)(CHCMe2Ph)
Fig. 2 1H NMR spectra of the reaction between [Mo(O)2(OtBu)2] (1)
(OtBu)(OtBuF3
(OtBu)(OtBuF6
)
)
F1,5
G1,8
H5,8
I2,6
L4,7
M4,10
N7,10
and [Mo(N-2,6-iPr2C6H3)2(OtBu)2] (3) to give [Mo(O)(N-2,6-iPr2C6H3)-
(OtBu)2] (B1,3) after (a) 1 h and (b) at equilibrium (4 weeks).
J2,9
(OtBuF3)(OtBuF6
(OtBu)(OiPr)
Cl(OtBu)
)
K6,9
O2,11
P2,12
Q12,13
R2,13
Cl(NHtBu)
(OtBu)(NHtBu)
(b) [M(Q)Cp(X)(XЈ)] (NAr)
(NtBu)
(NMe)
(Cl)(Cl)
Cp* S16,18/5,16
Cp*, Ta T15a,17
(OtBu)(OtBu)
(OtBu)(OtBuF3
(OtBu)(OtBuF6
U3,19
)
)
V19,20
W19,21
(OtBuF3)(OtBuF6
(OtBu)(OiPr)
(Cl)(Br)
)
X20,21
Y19,22
Z16,23
Cl(OtBu)
AЈ2,12/16,19
BЈ12,13/16,24
CЈ2,13/19,24
DЈ16,25
Cl(NHtBu)
(OtBu)(NHtBu)
Cl(pyrrolyl)
Fig. 3 Second-order reversible kinetic plot for the reaction between 1
and 3 to yield B1,3 {z=[x(aϪ2xe)ϩaxe]/[a(xe Ϫx)]}.
Ar = 2,6-iPr2C6H3; tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2;
pyrrolyl = (NC4H4). M = Nb unless stated otherwise.
Table 6 Second-order rate constants obtained from ligand exchange
reactions between 1 and 3 at various temperatures
T/K
k1,T/dm3 molϪ1 sϪ1
293
313
323
333
7.32 × 10Ϫ6
2.47 × 10Ϫ5
3.75 × 10Ϫ5
3.18 × 10Ϫ4
Fig. 4 Plot of ln(k1/T) vs. (1/T) for the ligand exchange reaction
between 1 and 3.
ln(k1/T) vs. 1/T30 (Fig. 4), the following activation parameters
were obtained: ∆H≠ = 39.3 kJ molϪ1; ∆S≠ = Ϫ208.6 J KϪ1
equilibrium [Keq(298 K) > 100] allows for a preparative-scale
synthesis of A1,2, although it is found to be extremely sensitive
to oxygen and/or moisture. A rate constant, k1(298 K), of
1.06 × 10Ϫ3 dm3 molϪ1 sϪ1 was calculated for the exchange
process; in this case, since the equilibrium so strongly favours
the product, a non-reversible treatment may be used.29
298
molϪ1; ∆G≠298 = 101.5 kJ molϪ1
.
Exchange of oxo groups with tert-butyl imido groups
has also been probed through the reaction of 1 with 2 to
yield A1,2; this proceeds much more rapidly than for O/NAr
exchange (hours at room temperature). The product-favouring
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4459

View Article Online
Table 7 Summary of equilibrium constants and times to equilibrium for exchange of monoanionic ligands in [Mo(Q)2X2] systems
Reactants
(a) [Mo(O)2X2]
Entry
Keq(298 K)
Time to equilibrium (R.T.)
1
2
3
X = OtBu, 1 and X = OtBuF3, 5
X = OtBu, 1 and X = OtBuF6, 8
X = OtBuF3, 5 and X = OtBuF6, 8
5–15
30–80
5–15
<5 min
<5 min
<5 min
(b) [Mo(NAr)(CHCMe2Ph)X2]
4
5
6
X = OtBu, 4 and X = OtBuF3, 7
X = OtBu, 4 and X = OtBuF6, 10
X = OtBuF3, 7 and X = OtBuF6, 10
11.0
120
8.0
<<5 min
<<5 min
<5 min
(c) [Mo(NtBu)2X2]
7
8
9
10
11
12
13
X = OtBu, 2 and X = OtBuF3, 6
X = OtBu, 2 and X = OtBuF6, 9
X = OtBuF3, 6 and X = OtBuF6, 9
X = OtBu, 2 and X = OiPr, 11
X = Cl, 12 and X = OtBu, 2
7.0
50
7.0
5–10 min
5–10 min
15 min
<5 min
3.0
>104
>104
20
<5 min (k1 > 7.7 dm3 molϪ1 sϪ1
<5 min (k1 > 7.7 dm3 molϪ1 sϪ1
)
)
X = Cl, 12 and X = NHtBu, 13
X = OtBu, 2 and X = NHtBu, 13
15 d (k1 = 5.2 × 10–5 dm3 molϪ1 sϪ1
)
Ar = 2,6-iPr2C6H3; tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2.
In order to probe the effect of changes to the ancilliary
alkoxide ligands, similar experiments were performed using
the trifluoro-tert-butoxide (reaction of 5 with 6 to give
D5,6) and hexafluoro-tert-butoxide (reaction of 8 with 9 to
give E8,9) analogues of 1 and 2. Substitution of three and six
fluorines on each tert-butyl group had little effect on the rate
of exchange in the case of the trifluoro-tert-butoxide system
[k1(298 K) = 1.3 × 10Ϫ3 vs. 1.06 × 10Ϫ3 dm3 molϪ1 sϪ1] but a more
marked effect in the case of the hexafluoro-tert-butoxide system
A(ii) Monoanionic ligand exchange between compounds of the
form [Mo(Q)2(X)2]
The exchange of tert-butoxide ligands (substituted with 0, 3 or
6 fluorines) has been probed in the system [Mo(O)2(OR)2]; the
data are collected in Table 7. In the cases of (OtBu)/(OtBuF3)
exchange (reaction of 1 with 5 to give F1,5; entry 1), (OtBu)/
(OtBuF6) exchange (reaction of 1 with 8 to give G1,8; entry 2)
and (OtBuF3)/(OtBuF6) exchange (reaction of 5 with 8 to give
H5,8; entry 3), equilibrium was achieved extremely rapidly
[k1(298 K) = 5.4 × 10Ϫ4 dm3 molϪ1 Ϫ1]; in both cases the equi-
s
1
librium constant strongly favoured the products [Keq(298 K) >
100].
(<5 min) at room temperature. Due to broadening in the H
NMR spectra of these mixtures, a result of intermetal alkoxide
exchange of the type seen in the Mo(O)2(OtBu)2/Mo(NtBu)2-
(OtBu)2 system (vide supra), accurate determination of equi-
librium constants was not possible. However, estimates could be
made, giving a value for Keq(298 K) in the range 5–15 for entries
1 and 3 and 30–80 for exchange of OtBu/OtBuF6 exchange
(entry 2).
A similar series of experiments has been carried out using
systems supported by other dianionic ligand sets. In the case
where (Q)2 = (NAr)(C(H)CMe2Ph), exchange of (OtBu) and
(OtBuF3) (reaction of 4 with 7 to yield L4,7; entry 4) proceeded
extremely rapidly (<5 min at Ϫ80 ЊC) to yield a mixture with an
equilibrium constant, Keq(298 K), of 11.0. Exchange of (OtBu)
and (OtBuF6) (reaction of 4 with 10 to give M4,10; entry 5)
yielded a mixture much richer in products [Keq(298 K) = 120].
This reaction was also very rapid, as was (OtBuF3)/(OtBuF6)
exchange (reaction of 7 with 10 to yield N7,10; entry 6) [Keq(298
K) = 8.0]. These rapid ligand exchange reactions involving
molybdenum alkylidene species have been exploited to control
cis–trans selectivity in ring-opening metathesis polymerisation
reactions.31
The exchange of arylimido and tert-butylimido ligands
between 2 and 3 to give the mixed imido product C2,3,16
although proceeding visibly at room temperature, was
extremely slow and proceeded at a more convenient rate by
heating the mixture to 60 ЊC. At this temperature, equilibrium
[Keq(333 K) = 25] was attained within 5 days. The relatively high
equilibrium constant once again allows for a preparative-scale
synthesis of the mixed ligand product (C2,3). A sample of this
compound in C6D6 at room temperature revealed very slow
appearance (by NMR) of resonances attributable to 2 and 3
confirming the equilibrium nature of the exchange reaction.
We have also briefly investigated oxo/imido exchange
between [Mo(Q)2Cl2(DME)] species (DME = 1,2-dimethoxy-
ethane). Although such compounds do not fall under the
umbrella of ‘pseudo-tetrahedral’ metal complexes, the poten-
tial lability of the bidentate donor could feasibly allow them to
access four-coordinate species. [Mo(NtBu)2Cl2(DME)] (28) and
[Mo(O)2Cl2(DME)] (27) were found to exchange oxo and imido
ligands to form [Mo(O)(NtBu)Cl2(DME)] (HЈ27,28) over 3 days
at room temperature. The equilibrium mixture was highly
product-rich [Keq(298 K) ≈ 200] and a second order non-
reversible kinetic treatment afforded a rate constant, k1(298 K),
Where (Q)2 = (NtBu)2, the reactions are slightly slower
than for those outlined above; however the results are still
broadly comparable. Additionally, with this supporting
imido ligand, exchange of other monoanionic ligands, specific-
ally: chloride with tert-butoxide (reaction of 12 with 2 to give
AЈ2,12; entry 11), chloride with tert-butylamide (reaction of
12 with 13 to give BЈ12,13; entry 12), and tert-butoxide with
tert-butylamide (reaction of 2 with 13 to give CЈ2,13; entry 13)
has been probed. These reactions tend to favour strongly the
mixed-ligand products and, in all but the last case, to be
extremely rapid. A second-order reversible treatment of the
results (vide supra) afforded a rate constant, k1(298 K) of 5.2 ×
of 2.3 × 10Ϫ4 dm3 molϪ1 Ϫ1. The high value of Keq allows the
s
synthesis of HЈ27,28 on a preparative scale; it was obtained in
moderate yield as a pale green solid after crystallisation from
diethyl ether at Ϫ78 ЊC.
As a further demonstration of the slower ligand exchange
reactivity of these six-coordinate complexes, [Mo(NtBu)-
Cl2(DME)] and [Mo(NAr)2Cl2(DME)] were mixed and heated
to 60 ЊC for several weeks. No reaction was observed, in
contrast to (NAr)/(NtBu) exchange in the tetrahedral
[Mo(Q)2X2] system which reaches equilibrium in 5 days at this
temperature.
10Ϫ5 dm3 molϪ1 sϪ1
.
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4460

View Article Online
Table 8 Reaction parameters and kinetic data for the exchange of alkoxide ligands in [Nb(NtBu)CpX2] systems
Ligands exchanged
Reactants
Product
Keq(298 K)a
k1/dm3 molϪ1 sϪ1
Time to equilibrium; temp.
OtBu/OiPr
OtBu/OtBuF3
OtBu/OtBuF6
OtBuF3/OtBuF6
Cl/Br
19 ϩ 22
19 ϩ 20
19 ϩ 21
20 ϩ 21
16 ϩ 23
16 ϩ 19
16 ϩ 24
16 ϩ 25
19 ϩ 24
Y19,22
4.0
2.8 × 10Ϫ4
4.1 × 10Ϫ5
1.9 × 10Ϫ5
5.8 × 10Ϫ6
>7.7
48 h; R.T.
10 d; 60 ЊC
70 d; 60 ЊC
100 d; 100 ЊC
5 min; R.T.
2 h; R.T.
24 h; R.T.
20 min; R.T.
20% complete after 60 d; 100 ЊC
V19,20
6.7 (333 K)
43 (333 K)
7.0 (373 K)
N/R
N/R
N/R
W19,21
X210,21
Z16,23
AЈ16,19
BЈ16,24
DЈ16,25
CЈ19,24
Cl/OtBu
1.3 × 10Ϫ2
9.7 × 10Ϫ4
2.5 × 10Ϫ1
ca. 8 × 10Ϫ7
Cl/NHtBu
Cl/NC4H4
OtBu/NHtBu
15
15 (est.) (373 K)
tBuF3 = C(CH3)2(CF3); tBuF6 = C(CH3)(CF3)2; N/R = non-reversible. a Unless stated otherwise.
B(i) Exchange of dianionic ligands between compounds of the
form [M(Q)CpX2] {M؍Nb, Ta}
We have previously described the exchange reaction between
[Nb(NMe)Cp*Cl2] (18) and [Nb(NtBu)CpCl2] (16) to give
[Nb(NtBu)Cp*Cl2] (S16,18) and [Nb(NMe)CpCl2] (17).16 This
equilibrium lies far to the left [Keq(373 K) = 0.18] and requires
several days (ca. 10) at 100 ЊC for the reaction to equilibrate.
In order to eliminate the possibility of the observed products
arising from inter-metal exchange of the cyclopentadienyl
ligands, [Ta(NAr)Cp*Cl2] (15a) and [Nb(NMe)CpCl2] (17)
were mixed together in a ‘double-crossover’ experiment.16 The
products obtained consisted solely of [Ta(NMe)Cp*Cl2] (T15a,17
)
and [Nb(NAr)CpCl2] (14) demonstrating unambiguously that
inter-metal imido-ligand exchange was taking place. Ring
exchange would result in formation of [Ta(NAr)CpCl2] and
[Nb(NMe)Cp*Cl2] which are not observed in the 1H NMR
spectrum of the product mixture. An equilibrium constant
1
could not be calculated for this reaction since the H NMR
spectroscopic resonances attributable to T15a,17 were not
discernable, owing to overlap.
The effect on rate of increasing the size of the ligands is once
again seen from the reaction of [Nb(NAr)Cp*Cl2] (15) with
[Nb(NtBu)CpCl2] (16) to give [Nb(NtBu)Cp*Cl2] (S15,16) and
[Nb(NAr)CpCl2] (14). This reaction required several weeks at
140 ЊC to attain equilibrium, unlike the analogous (NtBu)/
(NMe) exchange reaction above which required 10 days at 100
ЊC. The temperature required to drive the reaction towards
equilibrium is so high that some decomposition occurs, and
therefore an accurate Keq could not be determined. The third
possible imido/imido exchange reaction in the [Nb(NR)Cp/
Cp*Cl2] system is (NAr)/(NMe), i.e. the reaction of 15 with 17
to give 18 and 14. As might be anticipated, this exchange has a
rate intermediate between the two already discussed, taking ca.
two weeks at 140 ЊC to reach equlibrium.
Fig. 5 Stack plot of the 1H NMR spectra of the reaction between
[Nb(NtBu)Cp(OtBu)2] (19) and [Nb(NtBu)Cp(OtBuF3)2] (20) to give
[Nb(NtBu)Cp(OtBu)(OtBuF3)] (V19,20) at 60 ЊC.
reactants are detectable by NMR spectroscopy). In this case,
the reaction was slow enough to study its kinetics by NMR
spectroscopy. A standard second-order non-reversible treat-
ment gave a rate constant, k1(298 K) of 1.27 × 10Ϫ2 dm3 molϪ1
sϪ1
.
Chloride/tert-butylamide exchange (reaction of 16 with 24 to
form BЈ16,24) was also found to lie completely over to the side of
product, and was significantly slower (as might be expected for
the more bulky amide ligand), requiring 24 h at room temper-
ature to reach completion; this afforded a second-order rate
constant of 9.7 × 10Ϫ4 dm3 molϪ1 sϪ1. By contrast, the exchange
reaction of a planar amide such as η1-pyrollyl with chloride
(reaction of 16 with 25 to form DЈ16,25) was fast. In this case, an
equilibrium was established (Keq(298 K) = 15) within 20 min at
room temperature, giving a rate constant of 2.5 × 10Ϫ1 dm3
B(ii) Exchange of monoanionic ligands between compounds of
the form [Nb(Q)CpX2]
In a similar manner to the [Mo(Q)2X2] systems discussed
earlier, alkoxide/alkoxide ligand exchange has been studied
between niobium centres supported by an imido ligand and a
cyclopentadienyl ring. The results are collected in Table 8. A
1
stack plot showing the progress (via H NMR spectroscopy)
of the reaction between 19 and 20 to give V19,20 is shown in
Fig. 5.
Unlike the [Mo(Q)2X2] systems, the exchange reactions in
these half-sandwich niobium species are sufficiently slow to
allow a second-order reversible kinetic treatment to calculate
rate constants; these are also given in Table 8.
Ligand exchange reactions involving halides are the fastest in
this class. For example, chloride for bromide exchange pro-
ceeded to completion within 5 min (reaction of 16 with 23 to
give Z16,23), and chloride exchanged with tert-butoxide (reaction
of 16 with 19 to give AЈ16,19) within 2 h at room temperature.
Both equilibria lie strongly to the side of the products (no
molϪ1 sϪ1
.
As was observed in the [Mo(Q)2X2] case, the slowest
exchange was seen for (OtBu)/(NHtBu) (reaction of 19 with 24
to give CЈ19,24). No reaction was apparent at ambient temper-
ature, even over several days. At 60 ЊC, traces of CЈ were
observed after two weeks; at 100 ЊC, the reaction had proceeded
to ca. 20% conversion after 60 days. Assuming an equilbrium
constant, Keq(373 K), for this reaction of 15, as found for the
analogous molybdenum reaction (vide supra), a rate constant,
k1(373 K), of 8 × 10Ϫ7 dm3 molϪ1 sϪ1 can be estimated.
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4461

View Article Online
C Exchange of dianionic ligands betweeen [Mo(Q)2X2] and
[Nb(Q)CpX2]
Discussion
A(i): Dianionic ligand exchange between complexes of the form
[Mo(Q)2X2]
The reaction between [Mo(NAr)2(OtBu)2] (3) and [Nb(NtBu)-
Cp(OtBu)2] (19) to give [Mo(NtBu)(NAr)(OtBu)2] (C3,19) and
16
[Nb(NAr)Cp(OtBu)2] (U3,19
)
was only observed upon heat-
It can be seen that the values of the thermodynamic activation
parameters for the reaction betweeen 1 and 3 differ somewhat
from those previously reported.16 Nevertheless, the strongly
negative entropy of activation remains consistent with the
associative mechanism originally proposed (Scheme 3) and the
results reported here compare well with those obtained from the
pairwise exchange of oxo and imido groups in rhenium()
compounds.32
The more rapid exchange of oxo groups with tert-butyl imido
moieties (reaction of 1 and 2) is to be expected given this pro-
posed associative mechanism; additionally the alkyl-substituted
imido nitrogen might be expected to be more basic, and hence
show a greater propensity for bridging, than would an aryl-sub-
stituted imido nitrogen. An additional feature of (tBuN)/(O)
exchange is the much more pronounced favouring of the
mixed ligand product over reactants which is more in line with
observations by Espenson on Re() systems.32
ing to 110 ЊC and a temperature of 140 ЊC was required to
reach equilibrium [Keq(413 K) = 0.25 in 7 days]. This result is
directly comparable with (NAr)/(NtBu) exchange between Nb
centres (reaction of 15 and 16; see above) with an equilibrium
favouring the reactants.
D Exchange of dianionic ligands between [Mo(Q)2X2] and
[Mo(Q)2CpX]
It was anticipated from the foregoing results that the still more
hindered and (formally) more electron-rich system [Mo(Q)2-
CpX] might be expected to exchange ligands yet more slowly
than complexes of the form [Nb(Q)CpX2]. Accordingly,
[Mo(NtBu)2CpCl] (26) and [Mo(NAr)2(OtBu)2] (3) were mixed
and heated to 60 ЊC. After 48 h, only trace reaction had
occurred so the temperature was increased to 100 ЊC. This
yielded an equilbrium mixture of 3, C3,26/3,EЈ, 2, 26, EЈ3,26
and FЈ3,C/C,EЈ after 6 months as shown in Scheme 2. The pres-
ence of both C3,26/3,EЈ and 2 at equilibrium demonstrates that
[Mo(NAr)2(OtBu)2] can exchange both of its imido ligands.
An interesting facet of this equilibrating mixture is the sig-
t
1
nificant broadening of the BuO resonances in the H NMR
spectrum assigned to the dioxo species 1 (at δ 1.15) and the
mixed species A1,2 (at δ 1.25). This is attributed to intermetal
exchange of the alkoxide ligands at a rate comparable to the
NMR timescale. Consistently, cooling a sample of the equi-
librium mixture in D3CC6D5 to Ϫ10 ЊC caused significant
sharpening of the resonances. We can envisage three distinct
(OtBu) exchange processes as outlined in Fig. 6. Intuitively,
the exchange rates would be expected to increase in the order
ka < kb < kc, purely on steric grounds (assuming these exchange
processes to have a similar associative transition state to that
proposed for the dianionic ligand exchange reaction) and it is
merely fortuitous that kc is comparable to the NMR timescale.
Determination of these rate constants has not proved possible
since they are complicated functions of relaxation time (T1),
population and shift difference (∆δ) of the three species and
also of the rate constant of oxo/imido ligand exchange (k1).33
Scheme 2 Inter-metal exchange of oxo and imido ligands via an
associative mechanism.
E Exchange of dianionic ligands between [Mo(Q)2CpX] and
[Nb(Q)CpX2]
In the reaction between [Mo(NtBu)2CpCl] (26) and [Nb(NAr)-
CpCl2] (14), only small traces of the products of inter-metal
imido ligand exchange, namely [Mo(NtBu)(NAr)CpCl] and
[Nb(NtBu)CpCl2] (16), were present after 3 months at 100 ЊC.
Imido ligand exchange was observed between [Mo(NtBu)2-
CpCl] (26) and [Nb(NMe)CpCl2] (17) over a period of one
week at 100 ЊC (in CDCl3). However, the only resonances in the
product mixture were found to correspond to [Nb(NtBu)CpCl2]
(16); the remainder of the material had precipitated from
solution.
Fig. 6 Multi-site intermetal alkoxide exchange.
It may be seen that the ancilliary alkoxide ligands in the
system [Mo(O)2(OR)2]/[Mo(NR)2(OR)2] have a relatively small
effect on both rate and equilibrium constants. This may be
explained in terms of a trade-off between increasing M᎐O and
M᎐N multiple bond strengths, along with enhanced steric
᎐
᎐
interactions, upon increased fluorination (slowing the exchange
rate) vs. increasing electrophilicity of the metal centres as more
Scheme 3 Imido-ligand exchange reaction between [Mo(NtBu)2CpCl] (26) and [Mo(NAr)2(OtBu)2] (3). Percentage values indicate equilibrium
composition of [Mo(Q)2CpX] species.
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4462

View Article Online
fluorines are added to the butoxide ligands (potentially favour-
ing the formation of oxo/imido bridges).
The more crowded [Mo(Q)Cl2(DME)] species exchange their
dianionic ligands significantly more slowly than their four-
coordinate counterparts; again consistent with the proposed
associative mechanism.
observed as the size of the exchanging alkoxides is increased. A
supplementary factor could be F–F lone-pair repulsions in the
proposed bridged dimeric transition states. Both effects would
lead to a slower rate of exchange with increasing fluorine sub-
stitution, a trend which is clearly evident in the first four entries
of Table 8. This reactivity trend was not so readily apparent
in the [Mo(Q)2X2] system since all those reactions were fast
(<15 min to equilibrium in all cases).
A(ii) Monoanionic ligand exchange between compounds of the
form [Mo(Q)2(X)2]
The substantial rate acceleration observed for Cl/η1-NC4H4
exchange when compared to Cl/NHtBu exchange may be
explained in terms of the spatial anisotropy of the pyrollyl
ligand leading to reduced steric interactions in the bimolecular
transition state. However, electronic effects cannot be ruled out:
for example, the nitrogen lone-pair is involved in the π-system
of the pseudo-aromatic ring and thus the metal centre will be
more electrophilic and thus more susceptible to forming a bi-
molecular adduct. It is not possible to determine with any
degree of certainty whether steric or electronic effects dominate
in this case.
Examination of the data given in Table 8 reveals a strong
correlation between the rate of intermetal ligand exchange and
the size of the two exchanging ligands. Rates decrease in the
order: Cl/Br > Cl/OtBu > Cl/NHtBu > OtBu/OiPr > OtBu/
OtBuF3 > OtBu/OtBuF6 > OtBuF3/OtBuF6 > OtBu/NHtBu.
Remarkably, this series spans some seven orders of magnitude
in rate constant. This simple steric argument accords well with
an associative mechanism leading to a bimolecular transition
state.
Immediately noticeable is the much faster rate of exchange for
the monoanionic vs. dianionic ligand cases. This is not surpris-
ing in view of the lower bond multiplicities for the monoanionic
ligands compared to their dianionic counterparts. The trends in
Keq(298 K) for this class of exchange is also not unexpected
since it might be anticipated that the mixed ligand product
would be most favoured where the two ligands have maximum
difference in steric bulk and/or electronic properties. All of the
exchange reactions summarised in Table 7 follow this trend. It is
presumed that exchange of (OtBu) and (NHtBu) (entry 13) is so
slow on steric grounds.
B(i) Exchange of dianionic ligands between compounds of the
form [M(Q)CpX2] {M؍Nb, Ta}
Dianionic ligand exchanges in [M(Q)CpX2] systems are signifi-
cantly slower than for their [Mo(Q)2X2] counterparts (days to
weeks at 100 ЊC or higher vs. weeks at room temperature or a
few days at 60 ЊC). This is a predictable consequence of the
associative transition state proposed for these reactions. Com-
pounds of the form [M(Q)CpX2] have a higher formal electron
count than those of the type [Mo(Q)2X2] and, given the pres-
ence of a cyclopentadienyl ring, more bulk around the immedi-
ate vicinity of the metal centre. Both of these factors would
reduce the propensity of the system to form a bimolecular
adduct, hence retarding the reaction rate and requiring more
forcing conditions to achieve reasonable rates of reaction. Addi-
tionally, it has previously been shown in a theoretical study
that compounds of the form [M(Q)2X2] {M = Mo or W} should
be more active for olefin metathesis (the transition state for
which is analogous to that shown in Scheme 2) than those of
the form [M(Q)X4].34
The equilibrium in these half-sandwich systems appear to be
largely dominated by steric effects. For example, the reaction of
16 with 18 strongly favours the reactants, presumably due to
unfavourable steric interactions between the NtBu group and
pentamethylcyclopentadienyl ligand of S16,18. Although it did
not prove possible to determine an accurate equilibrium con-
stant for the reaction of 15 with 16 (due to partial decom-
position), reactants are clearly favoured here also, presumably
due to greater steric hindrance between NtBu and Cp* in S15,16
than between NAr and Cp* in 15. The propensity of penta-
methylcyclopentadienyl to reside on the metal centre supported
by the least bulky imido moiety is also reflected in the slight
product bias of the equilibrium produced in the reaction of 15
with 17.
One other factor worthy of note is the differences in the equi-
librium constant, Keq. This seems to be largest (and hence the
mixed ligand product most favoured) when the two exchanging
substituents are electronically and/or sterically most disparate
e.g. (OtBu)/Cl. Very similar ligands such as (OtBu)/(OiPr) give a
Keq close to the statistical value of 4.0 (for an A ϩ B
reaction) or 1.0 (for an A ϩ B C ϩ D reaction).
2C
C and D Exchange of dianionic ligands between [Mo(Q)2X2] and
[Nb(Q)CpX2] or [Mo(Q)2CpX]
Dianionic ligand exchange between [Mo(Q)2X2] and [Nb(Q)-
CpX2] (Class C) is considerably faster than that between the
niobium centres of [Nb(Q)CpX2]. For example, (NAr)/(NtBu)
exchange between Nb and Mo centres takes only 7 days at
140 ЊC as opposed to several weeks for the half-sandwich
niobium system. This result is expected given the less crowded
environment around the four-coordinate molybdenum centre.
Conversely, the more hindered and (formally) electron-rich
system, [Mo(Q)2CpX], would be expected to exchange ligands
more slowly with [Mo(Q)2X2] than for exchange between
[Nb(Q)CpX2] and [Mo(Q)2X2]. This is indeed observed, form-
ing the equilibria shown in Scheme 3 (after 6 months at elevated
temperature). Detailed study of this reaction is non-trivial since
there are no fewer than six metal-containing species at equi-
librium. However, since exchange reactions involving com-
pounds of the type [Mo(Q)2CpX] are observed to be the slowest
of all those studied, it is believed that an exchange reaction
between two species of this type (e.g. reaction of EЈ3,26 with
itself to give 26 and FЈEЈ,EЈ) can be discounted. Exchange of
chloride and tert-butoxide also seems not to occur since a mix-
ture of [Mo(NtBu)2(OtBu)2] (2) and [Mo(NtBu)2CpCl] (26) was
not observed to produce even slight traces of [Mo(NtBu)2-
(OtBu)Cl] (P2,26) or [Mo(NtBu)2Cp(OtBu)] over several months
at 100 ЊC. It is thus proposed that products EЈ3,26 and FЈ3,C/C,EЈ
must arise by exchange between [Mo(Q)2X2] and [Mo(QЈ)2CpX]
species.
B(ii) Exchange of monoanionic ligands between compounds of
the form [Nb(Q)CpX2]
Exchange of ligands between the niobium centers of [Nb(Q)-
CpX2] species is found to be slower than for molybdenum com-
plexes of the type [Mo(Q)2X2]. This can be explained on steric
grounds. Expectedly, the exchange of mono-anionic ligands in
the half-sandwich system is much faster than imido/imido
ligand exchange. Broadly similar trends to those observed in the
[Mo(Q)2X2] system are found, with comparable equilibrium
constants (in the range 1–10) except for the vastly sterically and
electronically different (OtBu) and (OtBuF6) ligands which give
a much greater bias towards the mixed ligand product. In this
[Nb(Q)CpX2] system a trend towards slower reaction rate is
E Exchange of dianionic ligands between [Mo(Q)2CpX] and
[Nb(Q)CpX2]
From the results discussed in the foregoing sections, it would be
anticipated that this class of reaction would be the slowest of all
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4463

View Article Online
those investigated. In general, this is the case; however the
relatively rapid progress of the reaction between 17 and 26
deserves mention. Here, it is likely that the insolubility of one
of the products, [Mo(NtBu)(NMe)CpCl] (GЈ17,26), drives the
reaction irreversibly to the product 16. However, it should also
be noted that this exchange of less bulky imido ligands would
be expected to proceed faster than the (NAr)/(NtBu) exchange
in any case. Interestingly, imido ligand exchange reactions in
this class are only observable with the least bulky ancilliary
ligand available, namely chloride. For all the other classes, we
have also been able to observe exchange reactions with more
bulky ancillary ligands such as tert-butoxide.
(3)
(4)
two reactants, in both of which the reaction is first order, may
be used (eqn. (5)).29 The effect of temperature on rate constant
may be used (by means of Transition State Theory) to obtain
thermodynamic activation parameters for the reaction using
eqn. (6).30
Summary and conclusions
Exchange reactions involving monoanionic and dianionic
ligands have been probed for a series of pseudo-tetrahedral
complexes whose electronic and steric parameters have been
systematically modified. It has been found that the rates of
such exchange reactions decrease in the order [Mo(Q)2X2]/
[Mo(Q)2X2] (A) > [Mo(Q)2X2]/[Nb(Q)CpX2] (C) > [Nb(Q)-
CpX2]/[Nb(Q)CpX2] (B) > [Mo(Q)2X2]/[Mo(Q)2CpX] (D) >
[Nb(Q)CpX2]/[Mo(Q)2CpX] (E) for both monanionic (where
studied) and dianionic ligand exchange. This correlates directly
with the greater steric demands around the metal as we move
through Class I to Class III (Fig. 1) and also with the higher
formal electron count of the complex along the same series.
Additionally, monoanionic ligands exchange faster than their
dianionic counterparts, a consequence of the greater metal–
ligand bond strengths for dianionic ligand systems. Subsidiary
to this effect, exchange reactions of less bulky ligands (e.g. oxo
or chloride) tend to proceed faster than those of more bulky
moieties (e.g. arylimido or tert-butylamide). These observ-
ations, combined with a strongly negative ∆S≠ value, lend sup-
port to the proposition that the mechanism for these reactions,
not unlike that proposed for olefin metathesis by Herrison and
Chauvin,35 involves a bimolecular adduct transition state, the
formation of which is the rate-determining step.
(5)
(6)
Acknowledgements
The Engineering and Physical Science Research Council
is thanked for studentships (to M. J. and J. P. M.) and a
postdoctoral fellowship (to A. J. G).
References
1 V. Grignard, Compt. Rend., 1900, 130, 1322.
2 In Comprehensive Organometallic Chemistry, ed. G. Wilkinson,
F. G. A. Stone and E. W. Abel, Pergamon Press, Oxford, 1982, vol. 1,
p. 194–8.
3 Metal Alkoxides, ed. D. C. Bradley, R. C. Mehrotra and D. P. Gaur,
Academic Press, New York, 1978.
4 M. H. Chisholm, K. Folting, J. C. Huffman and C. C. Kirkpatrick,
Inorg. Chem., 1984, 23, 1021.
5 D. E. Wigley, D. J. Arney and P. A. Wexler, Organometallics, 1990, 9,
1282.
Experimental
Experimental procedures for the synthesis of new compounds
and for NMR-scale ligand exchange reactions as well as H
NMR spectroscopic data for the products of these reactions
1
6 H. Weingarten and J. R. v. Wazer, J. Am. Chem. Soc., 1965, 87, 724.
7 H. Weingarten and J. R. v. Wazer, J. Am. Chem. Soc., 1966, 88, 2700.
8 J. H. Wengrovius, R. R. Schrock, M. R. Churchill, J. R. Missert and
W. J. Youngs, J. Am. Chem. Soc., 1980, 102, 4515.
9 J. H. Wengrovius and R. R. Schrock, Organometallics, 1982, 1, 148.
10 R. R. Schrock, G. C. Bazan and J. R. Wolf, Inorg. Chem., 1993, 32,
4155.
11 A. M. LaPointe, R. R. Schrock and W. M. Davis, J. Am. Chem. Soc.,
1995, 117, 4802.
12 A. M. LaPointe, R. R. Schrock and W. M. Davis, Organometallics,
1995, 14, 2699.
13 A. Galindo, F. Montilla, A. Pastor, E. Carmona, E. Gutiérrez-
Puebla, A. Monge and C. Ruiz, Inorg. Chem., 1997, 36, 2379.
14 D. E. Wheeler, J.-F. Wu and E. A. Maatta, Polyhedron, 1998, 17, 969.
15 W. A. Herrmann, H. Ding, F. E. Kühn and W. Scherer,
Organometallics, 1998, 17, 2751.
may be found in the ESI.†
Kinetic measurements
Wherever possible data were collected over the first three
half-lives of the reaction (i.e. to ca. 90% conversion). All
experiments were carried out at least three times to ensure
reproducibility and in order to estimate errors. Saturation
kinetics, where one reagent is used in vast excess, are not ideally
suited to the use of NMR spectroscopy due to difficulties in
obtaining accurate integrals of minor species. Standard second
order kinetic treatments were employed throughout. C6D6 was
a suitable solvent for virtually all the complexes studied and was
used throughout except where otherwise stated.
The kinetics of an equilibrium of the form shown in eqn. (2)
may be studied by means of the second-order reversible inte-
grated rate eqn. (3).29
16 M. Jolly, J. P. Mitchell and V. C. Gibson, J. Chem. Soc.,
Dalton Trans., 1992, 1331.
17 V. C. Gibson, J. Chem. Soc., Dalton Trans., 1994, 1607.
18 I. J. Blackmore, V. C. Gibson, A. J. Graham, M. Jolly, E. L. Marshall
and B. P. Ward, J. Chem. Soc., Dalton Trans., 2001, 3242.
19 V. C. Gibson, A. J. Graham, D. L. Ormsby, B. P. Ward, A. J. P. White
and D. J. Williams, J. Chem. Soc., Dalton Trans., 2002, 2597.
20 Advanced Inorganic Chemistry, ed. F. A. Cotton, G. Wilkinson,
C. A. Murillo, M. Bochmann, Wiley, New York, 6th edn., 2000,
pp. 923–924.
21 R. R. Schrock, J. S. Murdzek, G. C. Bazan, J. Robbins, M. DiMare
and M. O’Regan, J. Am. Chem. Soc., 1990, 112, 3875.
22 J. A. Osborn, J. Kress and G. Schoettel, J. Chem. Soc., Chem.
Commun., 1989, 1062.
(2)
This treatment has been used for determination of the rate
constants of most of the reactions in the present work; plotting
lnz vs. t {where z = (x(a Ϫ 2xe) ϩ axe)/a(xe Ϫ x)} e.g. Fig. 3. In
cases where the equilibrium strongly favours the products
(eqn. (4)), i.e. where Keq > 100, a non-reversible approach with
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4464

View Article Online
23 A. A. Danopoulos, G. Wilkinson, B. Hussain-Bates and M. B.
Hursthouse, J. Chem. Soc., Dalton Trans., 1990, 2753.
24 D. N. Williams, J. P. Mitchell, A. D. Poole, U. Siemeling, W. Clegg,
D. C. R. Hockless, P. A. O’Neil and V. C. Gibson, J. Chem. Soc.,
Dalton Trans., 1992, 729.
25 M. J. Humphries, M. L. H. Green, M. A. Leech, V. C. Gibson,
M. Jolly, D. N. Williams, M. R. J. Elsegood and W. Clegg, J. Chem.
Soc., Dalton Trans., 2000, 4044.
29 A. A. Frost and R. G. Pearson, Kinetics and Mechanism, Wiley,
New York, 1961.
30 J. H. Espenson, Chemical Kinetics and Reaction Mechanisms,
McGraw-Hill, New York, 1995.
31 W. J. Feast, V. C. Gibson and E. L. Marshall, J. Chem. Soc., Chem.
Commun., 1992, 1157.
32 Wang and J. H. Espenson, Inorg. Chem., 2002, 41, 1782.
33 J. Sandstrom, Dynamic NMR Spectroscopy, Academic Press, New
York, 1982.
26 J. Sundermayer, Chem. Ber., 1991, 1977.
27 P. W. Dyer, V. C. Gibson, J. A. K. Howard, B. Whittle and C. Wilson,
J. Chem. Soc., Chem. Commun., 1992, 1666.
28 H. H. Fox, K. B. Yap, J. Robbins, S. Cai and R. R. Schrock,
Inorg. Chem., 1992, 31, 2287.
34 A. K. Rappé and W. A. Goddard III, J. Am. Chem. Soc., 1982, 104,
448.
35 J. L. Herrison and Y. Chauvin, Makromol. Chem., 1870, 141,
161.
D a l t o n T r a n s . , 2 0 0 3 , 4 4 5 7 – 4 4 6 5
4465