S. Byrne et al. / Journal of Organometallic Chemistry 629 (2001) 182–186
185
−
1
which solvate anions more effectively or in the absence
of water, e.g. with anhydrous KF/18-crown-6 in dry
acetonitrile. HF is also formed and is responsible for
the acidity observed in the aqueous layer when the
reaction mixtures are washed with water.
2015/2054 cm
;
[Fe(h-C H )(CO) Br] 2010/2052
5
5
2
−
1
−1
cm ; [Fe(h-C H )(CO) I] 2009/2047 cm }suggest
5
5
2
that the electrophilicities of the CO ligands are all quite
low and do not differ greatly. Consequently, a second
pathway must operate when X=Cl, though not when
X=Br or I, and we suggest that in it the reactive
−
The attack by [OH] ions on coordinated CO is well
+
+
documented. [Fe(h-C H )(CO) L] salts (L=CO or
species is [Fe(h-C H )(CO) (NCMe)] (Scheme 1, L=
5
5
2
5
5
2
−
PPh ) have been reported to react with KOH to give
NCMe). It is well-known that the Cl ligand in [Fe(h-
C H )(CO) Cl] is labile, e.g. unlike [Fe(h-C H )(CO) X]
3
isolable [Fe(h-C H )(CO)(L)(CO H)] and K[Fe(h-
5
5
2
5
5
2
5
5
2
C H )(CO)(PPh )(CO )] [16,17]. The subsequent decar-
(X=Br or I) it dissolves in water with the formation of
5
5
3
2
+
−
boxylation may proceed via [Fe(h-C H )(CO)(L)-
[Fe(h-C H )(CO) (OH )] Cl , so it is not unreason-
5
−
5
5
5
2
2
(
CO H)] or [Fe(h-C H )(CO)(L)(CO )]
to [Fe(h-
able to suggest that a similar solvolysis takes place in
2
5
5
2
−
C H )(CO)(L)H] or [Fe(h-C H )(CO)(L)] , respec-
acetonitrile even if only to a very limited extent. [Fe(h-
C H )(CO) (NCMe)]
5
5
5
5
+
tively.
would be expected to react
5
5
2
The formation of CO in comparable amounts to CO2
when L=CO and its absence when L=CNMe or PPh3
is the evidence for nucleophilic attack by [Fe(h-
much more rapidly than the covalent halo-complexes
due to its more electrophilic CO groups as reflected in
−
1
its higher w(CꢁO) frequencies (2029/2074 cm ). How-
ever, this does not explain why two moles of
[Et N]F·2.5H O are required to bring about a complete
−
+
C H )(CO)(L)] on [Fe(h-C H )(CO) (L)] with dis-
5
5
5
5
2
placement of the weakest p-acceptor ligand, L, from the
cation. [Fe (h-C H ) (CO) ] may be a primary product
4
2
reaction of [Fe(h-C H )(CO) Cl].
2
5
5 2
4
5
5
2
from the reaction when L=PPh , but we think that it
3
arises from the decomposition of [Fe (h-C H ) -
3.1. NMR spectra of [Fe (p-C H ) (CO) (L)]
2 5 5 2 3
2
5
5 2
(
CO)(PPh )(m-CO) ] as its concentration in the reaction
3
2
mixture increases with time.
The proposed scheme rationalizes satisfactorily the
The NMR spectra of [Fe (h-C H ) (CO) (L)] {L=
2
5
5 2
3
−
CO [18], CNMe [11,19], P(OPh) [20] and CN [12]}
3
ligand dependence of the reaction rate which, for the
have been reported and discussed in detail, and have
been instrumental in revealing the mechanism of flux-
ionality in metal–carbonyl compounds. However, com-
parable studies have not been carried out on
[Fe (h-C H ) (CO) (PPh )], perhaps because it was re-
+
[
Fe(h-C H )(CO) (L)] cations, decreases in the order
5
5
2
L=CO\CNMe and PPh . This dependency may be
3
linked to the frequencies of the w(CꢁO) vibrations of
−
1
the cations which decrease 2074/2124 cm
(L=CO),
2
5
5 2
3
3
−
1
−1
2019/2064 cm
(L=CNMe), and 2014/2057 cm
ported to be unstable [10]. However, we have found
(
L=PPh ). This is not surprising as the w(CꢁO) fre-
that when [Fe (h-C H ) (CO) (PPh )] is pure, it is suffi-
3
2
5
5 2
3
3
1
quencies and the electrophilicities of the CO groups
decrease with increasing back-bonding into the CO
ciently stable in toluene-d solution to allow its H-
8
NMR spectrum to be obtained at various temperatures.
Unfortunately, only the cyclopentadienyl resonances
could be studied. These are slightly broadened so that
p*-orbitals as L changes from CO to CNMe or PPh . It
3
also explains why there is no reaction between [Fe(h-
31
1
13
C H )(CO)(Ph PCH CH PPh )]Cl and [Et N]F·2.5H O
P– H coupling was not detected, and a C-NMR
5
5
2
2
2
2
4
2
as the metal centre is electron rich, which discourages
nucleophilic attack at the carbonyl C atom.
spectrum could not be obtained.
Spectroscopic data suggests that [Fe (h-C H ) -
2
5
5 2
Titration experiments have shown that the stoi-
chiometry of these reactions is 1:1. However, all are
faster if a three-fold excess of [Et N]F·2.5H O is used.
(CO)(PPh )(m-CO) ] has the same structure as that
3
2
found for [Fe (h-C H ) (CO){P(OPh) }(m-CO) ] [21].
2
5
5 2
3
2
There are two species present in its toluene-d solutions.
4
2
8
−
This is to be expected as the concentration of [OH]
At low temperatures, they are present in the ratio of
2.2:1 and each gives rise to a pair of singlet cyclopenta-
dienyl resonances which are due to the C H ligands
ions would be much higher and would facilitate rele-
vant steps in the scheme, particularly when L is the
5
5
stronger donor, PPh , and towards the end of the
bound to Fe(CO) and Fe(PPh ), respectively. The two
resonances at higher are coincident so that three peaks
3
3
reaction.
−
The same scheme with L=X may be used to
are observed. The two species are probably cis- and
trans-isomers, but it is not possible to establish which is
the major and which is the minor one. On warming,
isomer exchange becomes faster, the cyclopentadienyl
resonances broaden and coalesce so that when the fast
exchange limit is reached at ca. 40°C, these protons give
rise to two equal resonances, which are due to the two
account for the products from the reactions of [Fe(h-
−
C H )(CO) X] provided that [Fe (h-C H ) (CO) (X)] ,
5
5
2
2
5
5 2
3
solvolyses to [Fe (h-C H ) (CO) (NCMe)] (X=Cl) or
2
5
5 2
3
decomposes to [Fe (h-C H ) (CO) ] (X=Br or I).
2
5
5 2
4
However, this does not account for the markedly differ-
ent reaction rates for [Fe(h-C H )(CO) Cl] as compared
5
5
2
with its bromo or iodo counterparts. The w(CꢁO) fre-
quencies of the three compounds {[Fe(h-C H )(CO) Cl]
types of C H5 groups. The coalescence temperature
5
(9°C, 282 K) can be used to obtain the DG* values for
5
5
2