Investigations on the trigonal twist process in the pseudooctahedral complexes [MX(CO)2(P-P)(η3-R)] (M = Mo or W; X = halide or pseudohalide; P-P = bidentate P-donor ligand; R = cycloheptatrienyl, cyclohexenyl or cyclooctenyl)

Article abstract of DOI:10.1016/S0022-328X(01)01355-9

A series of complexes [MX(CO)₂(P-P)(η³-R)] [M = Mo or W; X = halide, NCO or NCS; P-P = Ph₂P(CH₂)_nPPh₂ (n = 1 - 3)] have been prepared. Variable temperature ³¹P{¹H}-NMR investigations reveal the operation of a trigonal twist fluxional process in complexes where R = cycloheptatrienyl (C₇H₇), cyclohexenyl (C₆H₉) or cyclooctenyl (C₈H₁₃) with ΔG^?_C (the free energy of activation at the temperature of coalescene) strongly R-dependent. Using the cycloheptatrienyl complexes [MX(CO)₂{Ph₂P (CH₂)_nPPh₂}(η³ -C₇H₇)] detailed investigations have been made to establish the effect on ΔG^?_C of variation of X, n and M.

Full text of DOI:10.1016/S0022-328X(01)01355-9

Journal of Organometallic Chemistry 645 (2002) 218–227
www.elsevier.com/locate/jorganchem
Investigations on the trigonal twist process in the
pseudooctahedral complexes [MX(CO)2(PꢀP)(h3-R)] (M=Mo or
W; X=halide or pseudohalide; PꢀP=bidentate P-donor ligand;
R=cycloheptatrienyl, cyclohexenyl or cyclooctenyl)
John Friend, Saeed Y. Master, Ian Preece, Dale M. Spencer, Kathryn J. Taylor,
Assay To, John Waters, Mark W. Whiteley *
Department of Chemistry, Uni6ersity of Manchester, Manchester M13 9PL, UK
Received 24 August 2001; received in revised form 8 October 2001
Abstract
A series of complexes [MX(CO)2(PꢀP)(h3-R)] [M=Mo or W; X=halide, NCO or NCS; PꢀP=Ph2P(CH2)nPPh2 (n=1−3)]
have been prepared. Variable temperature 31P{1H}-NMR investigations reveal the operation of a trigonal twist fluxional process
in complexes where R=cycloheptatrienyl (C7H7), cyclohexenyl (C6H9) or cyclooctenyl (C8H13) with DG‡C (the free energy
of activation at the temperature of coalescence) strongly R-dependent. Using the cycloheptatrienyl complexes
[MX(CO)2{Ph2P(CH2)nPPh2}(h3-C7H7)] detailed investigations have been made to establish the effect on DG‡C of variation of X,
n and M. © 2002 Published by Elsevier Science B.V.
Keywords: Allyl ligand; Cycloheptatrienyl; Molybdenum; Fluxionality; Trigonal twist process
CHPPh2], Faller et al. [2] employed variable tempera-
ture NMR methods to infer the operation of a trigonal
twist rearrangement process in these complexes. The
mechanism of the trigonal twist involves rotation of the
triangular face formed by the two phosphorus atoms
and the ligand X with respect to the face formed by the
two carbonyls and the R (allyl) ligand. More recently
we, [5,9] and others [10], have demonstrated the opera-
tion of an equivalent process in cycloheptatrienyl com-
plexes of general formulation [MX(CO)2(PꢀP)(h3-C7-
H7)], albeit with a rather higher free energy of
activation.
The asymmetric structure of complexes of general
formulation [MX(CO)2(PꢀP)(h3-R)] affords the poten-
tial for regio- and stereo-selective synthesis [11] but
progress in this endeavour would require both recogni-
tion and control of any trigonal twist process. The
objectives of the work in the current paper are therefore
twofold. First, since investigations to date have been
restricted to allyl and cycloheptatrienyl complexes, we
set out to explore the generality of the trigonal twist
rearrangement in complexes with a variety of h3-R
1. Introduction
Complexes of the general formulation [MX(CO)2-
(PꢀP)(h3-R)] (M=Mo or W; X=halide or pseudo-
halide; PꢀP=bidentate P-donor ligand; R=allyl [1,2],
pentadienyl [3], hexadienyl [4], cycloheptatrienyl [5,6],
cycloheptadienyl [7], cyclooctadienyl [8]) are an impor-
tant class of compounds in the organometallic chem-
istry of molybdenum and tungsten. X-ray structural
studies on selected, diverse examples reveal the adop-
tion of a common pseudooctahedral structure with an
asymmetric ligand arrangement [2,3,6] in which one
phosphorus donor atom of the chelate PꢀP ligand is
located trans to the h3-R group. This leaves the two
carbonyls, the halide ligand and the second phosphorus
essentially coplanar. In a seminal study on the allyl
derivatives [MX(CO)2(PꢀP)(h3-C3H5)] [PꢀP=Ph2P-
(CH2)nPPh2 (n=1, dppm; n=2, dppe) or Ph2PCHꢁ
* Corresponding author. Tel.: +44-161-2754634; fax: +44-161-
2754598.
E-mail address: mark.whiteley@man.ac.uk (M.W. Whiteley).
0022-328X/02/$ - see front matter © 2002 Published by Elsevier Science B.V.
PII: S0022-328X(01)01355-9

J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
219
ligands. Secondly, we have attempted to quantify the
effect of variation of R, X, PꢀP, and M upon the free
energy of activation of the trigonal twist process. The
existing data for h3-C3H5 complexes [2] give some
indication of these effects but the very low free energies
of activation observed for these complexes did not
permit a full investigation (for example, values for free
energies of activation are restricted to X=I). In fact
the identity of R has a very significant effect upon the
free energy of activation DG‡C and, taking an extreme
case as an example, for the two complexes [MoI-
(CO)2(dppe)(h3-R)] (R=C3H5 [2] or C7H7 [5]), DG‡C is
ca. 20 kJ mol−1 greater for the cycloheptatrienyl
derivative. The focus of our work in the quantification
of the effects of X and PꢀP is therefore with h3-cyclo-
heptatrienyl complexes for which limiting low tempera-
ture 31P{1H}-NMR spectra and coalescence tempera-
tures are readily accessible for a wide range of ligands.
prepared by reaction of 4 with Ag[BF4] in acetone
followed by treatment with KI or LiCl, respectively.
The cyclohexenyl tungsten complex [WBr(CO)2-
(dppe)(h3-C6H9)] (2), was isolated as a yellow solid
from the reaction of dppe with [WBr(CO)2(NCMe)2(h3-
C6H9)]. A series of cycloheptatrienylmolybdenum com-
plexes
[MoX(CO)2(PꢀP)(h3-C7H7)]
[X=Br,
7,
PꢀP=dppe; 13, PꢀP=Ph2PCH2CH2CH2PPh2 (dppp);
X=NCO, 11, PꢀP=dppm; 12, PꢀP=dppp; 14,
PꢀP=Me2PCH2CH2PMe2 (dmpe)] were prepared by
reaction of [MoX(CO)2(h7-C7H7)] with PꢀP in CH2Cl2
and isolated as red/purple solids. Finally the cyclohep-
tatrienyltungsten complexes [WX(CO)2(PꢀP)(h3-C7H7)],
(16, X=Br, PꢀP=dppe; 18, X=NCO, PꢀP=dppp)
were prepared by reaction of [WX(CO)2(h-C7H7)] with
PꢀP in CH2Cl2. Details of the characterisation of the
new complexes 1–5, 7, 11–14, 16 and 18 are presented
in Table 1 (microanalytical, IR and mass spectroscopic
data) and Table 2 (1H-, 13C{1H}- and 31P{1H}-NMR
spectroscopic data). Also included in Table 2 are previ-
ously unreported 31P{1H}-NMR data for [MoX(CO)2-
(dppe)(h3-C7H7)] (9, X=NCO; 10, X=NCS) [6].
2. Results and discussion
2.1. Synthetic studies
2.2. Spectroscopic studies
The work described in this paper draws together the
results of our previous investigations on h3-cyclohepta-
trienyl complexes together with new synthetic and
NMR spectroscopic work. A full listing of the com-
plexes considered in this paper is given in Scheme 1 but
of these, the following new complexes are reported. The
cyclohexenyl (R=C6H9) and cyclooctenyl (R=C8H13)
complexes [MoBr(CO)2(dppe)(h3-R)] (1, R=C6H9; 4,
R=C8H13) were prepared by treatment of the known
precursors [MoBr(CO)2(NCMe)2(h3-R)] (R=C6H9
[12]; R=C8H13 [13]) with dppe in CH2Cl2 and isolated
as yellow–orange solids. Halide substituted derivatives
[MoX(CO)2(dppe)(h3-C8H13)], 3, X=I; 5, X=Cl were
2.2.1. Variable temperature 31P{1H}-NMR
in6estigations
The principal method that we have employed to
investigate the fluxional properties of [MX(CO)2(PꢀP)-
(h3-R)] is variable temperature 31P{1H}-NMR spec-
troscopy. Typically these complexes exhibit limiting low
temperature 31P{1H} spectra consisting of a doublet of
doublets pattern arising from the two inequivalent,
coupled phosphorus environments. As the temperature
is increased, 31P{1H}-NMR spectra of complexes in
which the trigonal twist process is operative display
coalescence of the two discrete phosphorus resonances
as the environments are averaged by the fluxional pro-
cess. A full list of the complexes considered in this
investigation together with the essential data for calcu-
lation of DG‡C (lw, the chemical shift separation in Hz
of the two phosphorus environments in the limiting low
temperature spectrum and TC the coalescence tempera-
ture) is presented in Table 3. Our earlier investigations
[5,9] utilized a field of 32 MHz for 31P{1H}-NMR
investigations whereas the new work reported in this
paper employs a field of 121.5 MHz. These details are
also made clear in Table 3 so that inappropriate com-
parisons of lw and TC are avoided. As far as possible
we have used CDCl3 as the standard solvent for our
variable temperature 31P{1H} investigations but in cases
where limiting low temperature spectra were expected
to be below −40 °C, CD2Cl2 was preferred since this
avoided extensive precipitation of the sample.
The free energies of activation at the temperature of
Scheme 1. Numbering scheme for complexes [MX(CO)2(PꢀP)h3-R].
coalescence, DG‡C, presented in Table 3, were deter-

220
J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227

J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
221

222
J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227

J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
223
Table 3
Free Energies of activation, DG‡C for the Trigonal Twist process in [MX(CO)2(PꢀP)(h3-R)] a
Complex
Field (MHz)/solvent
TC (K)
lw (Hz)
DG‡C (kJ mol−1
)
1 [MoBr(CO)2(dppe)(h3-C6H9)]
2 [WBr(CO)2(dppe)(h3-C6H9)]
121.5/CDCl3
121.5/CDCl3
121.5/CD2Cl2
121.5/CD2Cl2
121.5/CD2Cl2
32/CDCl3
121.5/CDCl3
121.5/CDCl3
121.5/CDCl3
121.5/CDCl3
121.5/CDCl3
121.5/CD2Cl2
121.5/CDCl3
121.5/CD2Cl2
32/CDCl3
296
326
275
248
230
318
316
302
284
301
328
266
305
297
314
311
281
260
220
460
197
390
402
78
248
304
157
124
1734
561
985
1152
162
668
189
932
57.2
61.3
53.3
46.4
42.8
64.4
61.0
57.6
55.6
59.7
58.1
49.1
55.3
53.4
61.7
57.4
54.6
46.9
3 [MoI(CO)2(dppe)(h3-C8H13)]
4 [MoBr(CO)2(dppe)(h3-C8H13)]
5 [MoCl(CO)2(dppe)(h3-C8H13)]
6 [MoI(CO)2(dppe)(h3-C7H7)] b
7 [MoBr(CO)2(dppe)(h3-C7H7)]
8 [MoCl(CO)2(dppe)(h3-C7H7)]
9 [Mo(NCO)(CO)2(dppe)(h3-C7H7)]
10 [Mo(NCS)(CO)2(dppe)(h3-C7H7)]
11 [Mo(NCO)(CO)2(dppm)(h3-C7H7)]
12 [Mo(NCO)(CO)2(dppp)(h3-C7H7)]
13 [MoBr(CO)2(dppp)(h3-C7H7)]
14 [Mo(NCO)(CO)2(dmpe)(h3-C7H7)]
15 [WI(CO)2(dppe)(h3-C7H7)] b
16 [WBr(CO)2(dppe)(h3-C7H7)]
17 [WCl(CO)2(dppe)(h3-C7H7)] b
18 [W(NCO)(CO)2(dppp)(h3-C7H7)]
121.5/CDCl3
32/CDCl3
121.5/CD2Cl2
a TC and lw determined by variable temperature 31P{1H}-NMR spectroscopy. DG‡C calculated by Eq. (1).
b Data from Ref. [5].
mined assuming an equally populated two-site system
and a transmission coefficient s of unity. On the basis
of these assumptions, the Eyring equation can be sim-
plified to Eq. (1) which has been employed in this work
for the calculation of DG‡C [14].
ence in DG‡C is ca. 20 kJ mol−1. It was therefore of
considerable interest to investigate other R ligands to
establish both the generality of the trigonal twist pro-
cess and the effect of R on DG‡C.
Variable temperature 31P{1H}-NMR investigations
on the cyclohexenyl complexes [MBr(CO)2(dppe)(h3-
C6H9)] (1, M=Mo; 2, M=W) and the cyclooctenyl
complexes [MoX(CO)2(dppe)(h3-C8H13)] (3, X=I; 4,
X=Br; 5, X=Cl) reveal behaviour characteristic of
the operation of a trigonal twist process. It is clear
therefore that this is a fairly general property of com-
plexes of the type [MX(CO)2(PꢀP)(h3-R)] with exam-
ples now demonstrated for R=cycloheptatrienyl,
cyclohexenyl, cyclooctenyl and allyl (C3H5). However,
it is important to note that these observations are not
made for all possible identities of h3-R ligands; in some
cases variable temperature 31P{1H}-NMR methods fail
to provide any evidence for a trigonal twist process.
For example, our 31P{1H} investigations on the cyclo-
heptadienyl complexes [MoX(CO)2(dppe)(h3-C7H9)]
(X=Cl or NCO) [7] reveal temperature invariant spec-
tra consisting of two discrete doublet resonances and
similarly the 2-methylallyl complex [MoCl(CO)2(dppe)-
(h3-2-MeꢀC3H4)] has a temperature invariant doublet
of doublets pattern in the 31P{1H}-NMR spectrum [15].
Using data for the molybdenum complexes [Mo-
Br(CO)2(dppe)(h3-R)] (1, R=C6H9; 4, R=C8H13; 7,
R=C7H7) a clear ordering in DG‡C (R=C7H7\
C6H9\C8H13) is apparent with a very large decrease in
DG‡C (ca. 11 kJ mol−1) on replacement of the cyclohex-
enyl by the cyclooctenyl ligand. We attribute the rela-
tively low DG‡C values observed for the cyclooctenyl
complexes to the flexibility of the methylene groups of
DG‡C=19.14 TC[9.972+log(TC/lw)] J mol−1
.
(1)
We estimate that although lw can be determined with
good accuracy, measurement of TC carries a maximum
error of 95 °C with a consequent error in DG‡C of ca.
91 kJ mol−1. Thus we suggest that a difference in
DG‡C values between two complexes of more than 2 kJ
mol−1 represents an effect for consideration. In accord
with our error estimation, duplicate determinations on
complexes 8 and 9 gave values of DG‡C consistent to
within 1 kJ mol−1
.
2.2.2. Discussion
The sub-sections below consider in turn the effect of
variation of R, X, PꢀP, and M upon the free energy of
activation
to
a
trigonal
twist process
in
[MX(CO)2(PꢀP)(h3-R)]. The values of DG‡C which form
the basis of this discussion are summarised in Table 3.
2.2.2.1. Effect of R [R=cycloheptatrienyl (C7H7), cyclo-
hexenyl (C6H9) or cyclooctenyl (C8H13)]. Comparison of
analogous allyl and cycloheptatrienyl complexes
[MX(CO)2(PꢀP)(h3-R)] (R=C3H5 or C7H7) reveals
that changes in the identity of R have the potential for
a very large effect on the magnitude of DG‡C for the
trigonal twist process. For example, in the case of
[MoI(CO)2(dppe)(h3-R)] (R=C3H5, DGC‡ =44.3 kJ
mol−1 [2]; R=C7H7, DG‡C=64.4 kJ mol−1) the differ-

224
J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
the cyclooctenyl ring allowing them to bend away from
and so avoid interaction with the MoX(CO)2(dppe)
group. To complete the series of ligands R it is neces-
sary to make a direct comparison with data for allyl
complexes [MI(CO)2(PꢀP)(h3-C3H5)] which are avail-
able only as iodide derivatives. Comparison of DGC‡
values for [MoI(CO)2(dppe)(h3-R)] (3, R=C8H13,
because it is not possible to delineate the contributions
of steric and electronic effects. It seems probable that
an increase in the steric requirements of X, PꢀP or R
will operate to increase DG‡C as is observed for the
halide series X=Cl, Br or I but considerable caution
must be exercised because of a superimposed electronic
effect. Use of carbonyl stretching frequencies as a probe
suggests that electron density at the metal centre in-
creases slightly along the series IBBrBCl [complexes
3–5 (Table 1) and 6–8] and similarly NCSBNCO
[w(CO) (cm−1, CH2Cl2): 10, (X=NCS), 1940, 1865; 9,
(X=NCO), 1933, 1855] [6]. If steric factors for X=
NCO, NCS are similar then it may be inferred that an
increase in electron density at the metal centre corre-
lates with a decrease in DG‡C.
;
DG‡C=53.3 kJ mol−1 R=C3H5, DGC‡ =44.3 kJ
mol−1) clearly establishes that it is the C3H5 ligand
which promotes the lowest values of DG‡C.
The data for the effect of R on the magnitude of
DG‡C guided our choice of ligand R for the subsequent
investigations on variation in X and PꢀP. Systems with
accessible coalescence temperatures and limiting low
temperature 31P{1H}-NMR spectra generally exhibit
DG‡C in the range 40–65 kJ mol−1 and to achieve this
for a wide range of ligands X and PꢀP, it is necessary to
select R to promote relatively high DG‡C values. In this
way for example investigations on X=Cl, not possible
for complexes such as [MoCl(CO)2(dppe)(h3-C3H5)],
are facilitated. Our choice of ligand R for subsequent
studies was therefore between R=cyclohexenyl or cy-
cloheptatrienyl and, in the event, R=C7H7 was se-
lected because some data on the effect of variation of X
and PꢀP were already available from our previous work
[5].
2.2.2.3. Effect of PꢀP. Ligands of the type
R%P(CH2)nPR% may affect DG‡C through the chain
2
2
length (n) or the identity of R% substituents attached to
phosphorus. Previous studies with [MoX(CO)2(PꢀP)-
(h3-C3H5)] suggest [2] that replacement of dppm (n=1)
by dppe (n=2) decreases DG‡C by ca. 2.5 kJ mol−1. In
this work the effect of chain length was investigated
further through the complexes 11, 9 and 12 in which n
is systematically increased from 1 to 3. To promote
direct comparison, all other ligands in complexes 9, 11
and 12 are identical; the ligand X was selected as NCO
because this reduces DGC‡ and permits measurement of
TC for the dppm derivative 11. In fact, complex 11 is
the first example of a dppm derivative of the cyclohep-
tatrienyl system for which the trigonal twist process has
been shown to operate. Previous attempts with deriva-
tives where X=halide were unsuccessful due to decom-
position of the complexes at the high temperatures
required for investigation [5].
2.2.2.2. Effect of X. In the cycloheptatrienylmolybde-
num complexes [MoX(CO)2(dppe)(h3-C7H7)] (6–10),
only the ligand X (X=I, Br, Cl, NCO, NCS) is varied.
As expected from the work [2] on [MX(CO)2(PꢀP)(h3-
C3H5)] the highest DG‡C is observed for X=I. However,
study of the cycloheptatrienyl system also allowed cal-
culation of DG‡C for X=Br or Cl and the results reveal
a steady decrease in DG‡C of ca. 3 kJ mol−1 along the
series I\Br\Cl. In fact the ligand X=NCO pro-
motes the lowest value of DG‡C and this finding was
utilised in the design of a series of complexes to exam-
ine the effect of the chelate phosphine ligand PꢀP (see
below).
Having derived a series for X in the cycloheptatrienyl
system we decided to check its validity for the cy-
clooctenyl system [MoX(CO)2(dppe)(h3-C8H13)] (X=I,
Br, Cl) in which the absolute magnitude of DG‡C is
much lower. Inspection of DG‡C values for cyclooctenyl
complexes 3–5 confirms the ordering I\Br\Cl but
the overall magnitude of the change from I=I to
X=Cl is probably larger than that observed for the
cycloheptatrienyl system. We conclude therefore that,
whilst the ordering of DG‡C along the halide series
I\Br\Cl is probably generally applicable to com-
plexes of the type [MX(CO)2(PꢀP)(h3-R)], the absolute
magnitude of the change in DG‡C along the series is
dependent on the other variables M, PꢀP and R.
In general, we have not attempted to rationalise the
effects of variation in M, X, PꢀP and R upon DGC‡
Examination of DG‡C values for the NCO complexes
9, 11 and 12 reveals a non-linear decrease in DG‡C with
increasing chain length n. The dppm derivative 11
exhibits the highest value of DG‡C (58.1 kJ mol−1) and
this decreases by ca. 3 kJ mol−1 in the dppe derivative
9. A much larger decrease in DG‡C (ca. 6 kJ mol−1) is
observed on moving to the dppp derivative, 12. Further
to establish the large change in DG‡C from n=2 to
n=3, the bromide derivative [MoBr(CO)2(dppp)(h3-
C7H7)] 13 was prepared for direct comparison with the
dppe analogue 7 and again a decrease in DG‡C of ca. 6
kJ mol−1 is observed. Overall it is clear that the length
of the bridging chain in the diphosphine has a very
significant effect with DG‡C decreasing by ca. 10 kJ
mol−1 along the series n=1−3. In an attempt to
assess the effect of the phosphine ligand substituents R%
we also synthesised the dmpe complex [Mo(NCO)-
(CO)2(dmpe)(h3-C7H7)], (14). The substitution of phos-
phine phenyl substituents in
9 with the methyl
substituents of 14 whilst keeping chain length (n=2)
constant has only a small effect (if any) on DG‡C (de-

J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
225
crease of ca. 2 kJ mol−1) although it is not possible to
delineate the contribution of the differing steric and
electronic effects of the Me and Ph substituents. In
summary however, the dominant effect connected with
the R%P(CH2)nPR% ligand appears to be the bridging
and in the cyclohexenyl complexes 1 and 2 for which
the broadness of the ambient temperature 1H-NMR
spectra precluded attempts at assignment of ring pro-
tons. Fortunately, in the cycloheptatrienyl system spec-
tral assignment is facilitated by a rapid 1,2-shift process
which renders equivalent all seven protons and carbons
of the h3-C7H7 ring [16]. We have already demonstrated
that for the cycloheptatrienyl complexes [MX(CO)2-
(PꢀP)(h3-C7H7)], better resolved 1H- and 13C{1H}-
NMR spectra can be obtained by employing variable
temperature studies [5]. Although we considered it un-
necessary to extend such investigations to the addi-
tional cycloheptatrienyl complexes reported in the
current work, some points of interest do arise from the
new cyclohexenyl and cyclooctenyl systems and these
are discussed below.
2
2
chain length and the flexibility imparted to the system
as n is increased.
2.2.2.4. Effect of M (M=Mo or W). In the allyl
complexes [MI(CO)2(PꢀP)(h3-C3H5)], substitution of
Mo by W leads to an increase in DG‡C by 2–3 kJ mol−1
depending on the identity of PꢀP. A similar effect is
evident from comparison of the cyclohexenyl complexes
[MBr(CO)2(dppe)(h3-C6H9)] (1, M=Mo; 2, M=W)
for which an increase in DG‡C of ca. 4 kJ mol−1 is
observed. The results for the cycloheptatrienyltungsten
system were therefore unexpected in that they show a
decrease in DG‡C of ca. 3 kJ mol−1 through replacement
of Mo by W. We have reasonable confidence in this
conclusion because: (i) the difference in DG‡C between
analogous Mo and W complexes is larger than the
estimated allowance for experimental error; and (ii) the
effect is consistent for comparison of a complete series
of halide complexes [MX(CO)2(dppe)(h3-C7H7)] (X=I,
Br, Cl; M=Mo, 6–8; M=W, 15–17). In addition to
the dppe complexes 15–17, we also examined
[W(NCO)(CO)2(dppp)(h3-C7H7)] 18, to determine
whether the observation extended to systems in which
DG‡C was much lower. Comparison of the DG‡C values
for [M(NCO)(CO)2(dppp)(h-C7H7)] (12, M=Mo; 18,
M=W) reveals a decrease of ca. 2 kJ mol−1 by
substitution of Mo with W and, although the magni-
tude of the difference could be accounted for by exper-
imental error, the trend is consistent with that observed
for the dppe analogues. Finally, it should be noted that
in complex 18, the variables X, PꢀP and M are all set to
promote the lowest value of DG‡C and therefore the
result of 47 kJ mol−1 may be taken as a very approxi-
mate guide to the minimum free energy of activation to
the trigonal twist process in the cycloheptatrienyl com-
plexes [MX(CO)2(PꢀP)(h3-C7H7)].
Low temperature 1H- and 13C{1H}-NMR spectra
have been acquired for the cyclohexenyl complexes 1
(1H and 13C) and 2 (1H only) with the objective of
obtaining fully assigned spectra. The low temperature
spectra are indicative of an asymmetric structure with
two discrete phosphorus environments and consistent
with the solid state structure determined by X-ray
crystallography on analogous complexes [MX(CO)2-
(PꢀP)(h3-R)]. Thus in 1 [−40 °C (1H)/−30 °C (13C)]
all the ring protons/carbons are inequivalent and in the
13C-NMR spectrum there are two separate carbonyl
resonances each with couplings to two distinct phos-
phorus atoms. Moreover, there are two non-equivalent
methylene carbons for the dppe ligand, each split into
four lines by J(PꢀC) coupling. By contrast, the ambient
temperature 13C{1H} spectrum of the tungsten deriva-
tive 2 exhibits discrete but very broad resonances for
the cyclohexenyl ring carbons and methylene bridge
carbons of the dppe ligand.
The ambient temperature 1H- and 13C{H}-NMR
spectra of the cyclooctenyl complexes [MoX(CO)2-
(dppe)(h3-C8H13)] (3, X=I; 4, X=Br; 5, X=Cl)
demonstrate the opposite extreme of the effect of the
trigonal twist process on such spectra. In these cases,
coalescence temperatures are less than 280 K and the
ring protons/carbons, dppe methylene protons/carbons
and the carbonyl carbons exhibit fully averaged envi-
ronments (for example in the 13C{1H}-NMR spectra,
only five cyclooctenyl ring carbons are observed). This
is especially the case for the chloride derivative 5,
whereas the iodide derivative 3 (which has the higher
coalescence temperature) displays quite significant
broadening of these resonances.
1
2.2.2.5. H- and 13C{1H}-NMR spectra. The H- and
13C{1H}-NMR data for the new complexes reported in
this paper are summarised in Table 2. The generally
broad appearance of the ambient temperature spectra is
consistent with the operation of a trigonal twist pro-
cess, with the broadening effect especially marked
where coalescence temperatures TC lie in the range
280–320 K (see Table 3). Typical observations are
broadening of the ring (R ligand) resonances in both
1H- and 13C-NMR spectra, broadening of signals asso-
ciated with methylene groups of the chelate phosphine
and failure to observe the carbonyl carbon resonance in
the 13C-NMR spectrum. These observations are mani-
fest in the majority of the cycloheptatrienyl complexes
3. Conclusions
Variable temperature 31P{1H}-NMR investigations
strongly support the operation of a trigonal twist flux-
ional process in complexes of the type [MX(CO)2-

226
J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
(PꢀP)(h3-R)] [R=cycloheptatrienyl, cyclohexenyl, cy-
clooctenyl or allyl (C3H5)] but the process is not univer-
sal for all identities of R. Calculation of DG‡C for the
molybdenum complexes [MoX(CO)2(dppe)(h3-R)] re-
veals a decrease in the order R=cycloheptatrienyl\
cyclohexenyl\cyclooctenyl\allyl. Detailed investiga-
tions on the effect of variation of X and PꢀP on DGC‡
were carried out on the cycloheptatrienyl complexes
[MX(CO)2(PꢀP)(h3-C7H7)] which exhibited accessible
values of DG‡C (45–65 kJ mol−1) over a wide variation
then evaporated to dryness and the crude product
recrystallised from CH2Cl2–hexane then toluene–di-
ethyl ether to give 1 as a yellow–orange solid; yield
2.15 g (44%).
4.3. Preparation of [WBr(CO)2(dppe)(p3-C6H9)] (2)
A stirred solution of [W(CO)3(NCMe)3] (0.848 g,
2.17 mmol) in NCMe (30 cm3) was treated with 3-bro-
mocyclohexene (0.350 g, 2.17 mmol). After 40 min,
precipitation of the product [WBr(CO)2(NCMe)2(h3-
C6H9)] as a pale yellow–green solid was complete; yield
0.923 g. The intermediate [WBr(CO)2(NCMe)2(h3-
C6H9)] (0.445 g, 0.92 mmol) in thf (30 cm3) was treated
with dppe (0.370 g, 0.093 mmol) and stirred at r.t. for
40 h. The solvent was then removed and the residue
recrystallised from CH2Cl2–hexane then acetone–hex-
ane to give 2 as a bright yellow solid; yield 0.678 g
(92%).
in ligands
X and PꢀP. The complexes [MoX-
(CO)2(dppe)(h3-C7H7)] establish that DG‡C decreases in
the order X=I\Br:NCS\Cl\NCO whilst investi-
gations on [Mo(NCO)(CO)2{Ph2P-(CH2)nPPh2}(h3-
C7H7)] show that DG‡C decreases significantly from
n=1 to n=3. The variables M, X, PꢀP and R are not
independent but the trends in DG‡C observed for the
variation of X and PꢀP in the cycloheptatrienyl system
appear to have some general validity. The effect on
DG‡C of exchange of Mo for W is strongly R-dependent.
Where R=cyclohexenyl or allyl, DG‡C increases on
substitution of Mo by W but the reverse effect is
observed for R=C7H7.
4.4. Preparation of [MoBr(CO)2(dppe)(p3-C8H13)] (4)
Reaction of [MoBr(CO)2(NCMe)2(h3-C8H13)] (0.350
g, 0.83 mmol) with dppe (0.329 g, 0.83 mmol) in
CH2Cl2 (20 cm3) gave an orange solution which was
stirred at r.t. for 2 h. The solvent was then removed in
vacuo and the residue recrystallised from CH2Cl2–hex-
ane to give 4 as a yellow–orange solid; yield 0.510 g
(83%).
4. Experimental
4.1. General procedures
The preparation, purification and reactions of the
complexes described were carried out under dry nitro-
gen. All solvents were dried by standard methods,
distilled and deoxygenated before use. The compounds
[MoBr(CO)2(NCMe)2(h3-C6H9)] [12], [MoBr(CO)2-
(NCMe)2(h3-C8H13)] [13] and [M(NCO)(CO)2(h-C7H7)]
(M=Mo or W) [17] were prepared by published proce-
dures. Three hundred megahertz 1H- and 75 MHz
13C{1H}-NMR spectra were recorded on Bruker AC
300 E, Varian Associates XL 300 or Varian Unity
Inova 300 spectrometers; 121.5 MHz 31P{1H}-NMR
spectra were recorded on the Varian Unity Inova 300
instrument. Infrared spectra were obtained on a
Perkin–Elmer FT 1710 spectrometer and mass spectra
were recorded using Kratos Concept 1S (FAB spectra)
or Micromass Platform II (ES spectra) instruments.
Microanalyses were by the staff of the Microanalytical
Service of the Department of Chemistry, University of
Manchester.
4.5. Preparation of [MoX(CO)2(dppe)(p3-C8H13)] (3,
X=I; 5, X=Cl)
A stirred solution of [MoBr(CO)2(dppe)(h3-C8H13)]
(0.370 g, 0.50 mmol) in AnalaR acetone (30 cm3) was
treated with Ag[BF4] (0.120 g, 0.62 mmol). The solution
was stirred for 30 min, filtered to remove the precipitate
of AgBr and then KI (0.415 g, 2.5 mmol) was added.
After stirring for 20 h the solution was evaporated to
dryness and the residue recrystallised from CH2Cl2–
hexane to give 3 as an orange solid; yield 0.158 g (40%).
The chloride derivative [MoCl(CO)2(dppe)(h3-C8H13)],
5 was prepared similarly from [MoBr(CO)2(dppe)(h3-
C8H13)] (0.200 g, 0.27 mmol), Ag[BF4] (0.071 g, 0.36
mmol) and LiCl (0.057 g, 1.36 mmol) and isolated an
orange–yellow solid; yield 0.060 g (32%).
4.6. Preparation of [MX(CO)2(PꢀP)(p3-C7H7)]
[M=Mo or W, X=Br or NCO,
PꢀP=Ph2P(CH2)nPPh2 (n=1–3)], 7, 11–13, 16, 18
4.2. Preparation of [MoBr(CO)2(dppe)(p3-C6H9)] (1)
Reaction of [MoBr(CO)2(NCMe)2(h3-C6H9)] (2.71 g,
6.86 mmol) with dppe (2.73 g, 6.86 mmol) in THF (60
cm3) gave an orange solution which was stirred at room
temperature (r.t.) for 24 h. The reaction mixture was
These complexes were prepared by a standard proce-
dure starting from [MX(CO)2(h-C7H7)] (M=Mo or W;
X=Br or NCO). Reaction of a green solution of
[MX(CO)2(h-C7H7)] (ca. 0.50 g) in CH2Cl2 (20 cm3)

J. Friend et al. / Journal of Organometallic Chemistry 645 (2002) 218–227
227
with an equimolar quantity of Ph2P(CH2)nPPh2 (n=1–
References
3) at r.t. gave a red–purple solution of the product
[MX(CO)2(PꢀP)(h3-C7H7)]. After stirring for 2 h the
solution was evaporated to dryness and the residue
recrystallised from CH2Cl2–hexane and subsequently,
acetone–hexane to give the product as a red to purple
solid. Details of colours and yields are presented in
Table 1.
[1] R. Davis, L.A.P. Kane-Maguire, in: G. Wilkinson, F.G.A. Stone,
E.W. Abel (Eds.), Comprehensive Organometallic Chemistry, vol.
3, 1st ed., Pergamon Press, New York, 1982, pp. 1156–1158 Ch.
27.2.
[2] J.W. Faller, D.A. Haitko, R.D. Adams, D.F. Chodosh, J. Am.
Chem. Soc. 101 (1979) 865.
[3] G.-H. Lee, S.-M. Peng, F.-C. Liu, D. Mu, R.-S. Liu, Organometal-
lics 8 (1989) 402.
[4] M.A. Paz-Sandoval, P. Jua´rez Saavedra, G. Dura´n Pomposo, P.
Joseph-Nathan, P. Powell, J. Organomet. Chem. 387 (1990) 265.
[5] R.A. Brown, S. Endud, J. Friend, J.M. Hill, M.W. Whiteley, J.
Organomet. Chem. 339 (1988) 283.
4.7. Preparation of [Mo(NCO)(CO)2(dmpe)(p3-C7H7)]
(14)
[6] R.L. Beddoes, Z.I. Hussain, A. Roberts, C.R. Barraclough, M.W.
Whiteley, J. Chem. Soc. Dalton Trans. (1996) 3629.
[7] Z.I. Hussain, A.L.A.B. de Souza, M.W. Whiteley, J. Organomet.
Chem. 554 (1997) 121.
[8] R.L. Beddoes, D.M. Spencer, M.W. Whiteley, J. Chem. Soc.
Dalton Trans. (1995) 2915.
[9] R. Breeze, S. Endud, M.W. Whiteley, J. Organomet. Chem. 302
(1986) 371.
[10] M.L.H. Green, D.K.P. Ng, J. Chem. Soc. Dalton Trans. 0 (1993)
17.
A
stirred solution of [Mo(NCO)(CO)2(h-C7H7)]
(0.175 g, 0.61 mmol) in CH2Cl2 (20 cm3) was cooled to
−78 °C and then treated with dmpe (0.100 g, 0.66
mmol). The reaction mixture was allowed to warm
slowly to r.t. and reaction to give a red solution of the
product commenced at −40 °C. After stirring at r.t.
for 30 min the reaction mixture was evaporated to
dryness and the residue dried in vacuo, for 2 h. The
crude product was then recrystallised from CH2Cl2–
hexane to give 14 as a deep purple solid; yield 0.080 g
(30%).
[11] B.M. Trost, M. Lautens, J. Am. Chem. Soc. 104 (1982) 5543.
[12] J.W. Faller, H.H. Murray, D.L. White, K.-H. Chao, Organometal-
lics 2 (1983) 400.
[13] J.W. Faller, K.-H. Chao, Organometallics 4 (1984) 927.
[14] J. Sandstro¨m, Dynamic NMR Spectroscopy, Academic Press, New
York, 1982, pp. 96–97.
[15] S. Joseph, M.W. Whiteley, unpublished work.
[16] J.W. Faller, Inorg. Chem. 8 (1969) 129.
[17] G. Hoch, R. Panter, M.L. Ziegler, Z. Naturforsch., B 31 (1976)
294.
Acknowledgements
We are grateful to the EPSRC for the award of
Research Studentship (to D.M. Spencer).