Acetylides of cycloplatinated ferrocenylamines; synthesis and redox chemistry

Article abstract of DOI:10.1039/b101776o

Sonogashira coupling of a terminal alkyne with cycloplatinated ferrocenylamine complexes gave the acetylides Pt{[RC≡C][Me₂NCH₂(σ-Fc)]dmso}, which were characterised by analysis, spectroscopy and, for the R = SiMe₃ 3, Fc 5 complexes, X-ray structural analyses. A butadiyne complex was also isolated in reactions of 3. Equivalent reactions with LDA led to an unusual deplatination reaction to give the ethynylferrocenylamine. The Pt(II) centre functions as a redox switch at E_+/0 = ～0.21 V and the spectroscopic and electrochemical data show that the acetylide is a π-donor ligand. There is a strict delineation of the electronic requirements for the trans Pt-N and trans Pt-C(Cp) groups in these cycloplatinated compounds. Oxidation of 3 gives rise to low energy bands based on the Pt(II) unit; an additional broad band at 1100 nm occurs in 5⁺ but not 5²⁺.

Full text of DOI:10.1039/b101776o

View Article Online / Journal Homepage / Table of Contents for this issue
Acetylides of cycloplatinated ferrocenylamines; synthesis and redox
chemistry†
C. John McAdam, Evan J. Blackie, Joy L. Morgan, Sarah A. Mole, Brian H. Robinson and
Jim Simpson
Department of Chemistry, University of Otago, P.O. Box 56, Dunedin, New Zealand
Received 23rd February 2001, Accepted 13th June 2001
First published as an Advance Article on the web 2nd August 2001
Sonogashira coupling of a terminal alkyne with cycloplatinated ferrocenylamine complexes gave the acetylides
᎐
Pt{[RC᎐C][Me NCH (σ-Fc)]dmso}, which were characterised by analysis, spectroscopy and, for the R = SiMe 3, Fc
᎐
2
2
3
5 complexes, X-ray structural analyses. A butadiyne complex was also isolated in reactions of 3. Equivalent reactions
with LDA led to an unusual deplatination reaction to give the ethynylferrocenylamine. The Pt() centre functions as a
redox switch at E ϩ/0 = ∼0.21 V and the spectroscopic and electrochemical data show that the acetylide is a π-donor
ligand. There is a strict delineation of the electronic requirements for the trans Pt–N and trans Pt–C(Cp) groups in
these cycloplatinated compounds. Oxidation of 3 gives rise to low energy bands based on the Pt() unit; an additional
broad band at 1100 nm occurs in 5ϩ but not 52ϩ
.
Molecules with ethynyl functionalities provide a route to a
myriad of structures which give the opportunity for redox
centres to ‘talk’ to each other through the rigid σ-acetylide
links.1 This communication is important if there is to be
efficient energy transfer for molecular wires or photonic relays.
Ethynylferrocenyl derivatives are particularly attractive because
the redox centre is well-behaved and synthetic elaboration of
the local ferrocenyl environment is relatively straightforward.
The redox potential of the ferrocenyl moiety can be tuned by
appropriate substitution on the cyclopentadienyl ring.2 Never-
theless, even with ring substitution, the potential only varies
by about 0.5 V. In contrast, low-valent metal clusters offer a
window of ϩ0.8 V to Ϫ1.2 V,3 although they are nowhere near
as robust. For our work on switchable fluorescent materials we
required a redox centre with an oxidation potential close to
zero. Furthermore, it was preferable for the redox centre to have
a structural pivot that would lead to predetermined geometries
such as molecular squares. A group of molecules that satisfy
these criteria are the cycloplatinated ferrocenylamines 1 (X =
anionic halide). They are readily prepared from the ferro-
cenylamine 2, kinetically stable in air and are oxidised at very
low potentials.4,5 Their square planar geometry is ideally suited
to molecular square formation, although the specific electronic
demands on the trans –NMe2 and trans Pt–C positions (good
π-acceptor and donor anion, respectively5,6) restrict the choice
of linkages for the sides. Replacement of the anionic halide X
trans to the Pt–C bond by a donor acetylide group would give
greater synthetic flexibility and a communication link. This
type of compound is the principal target molecule in the work
described in this paper.
amino substituent have also been reported.10 Because the
CH2NMe2 substituent is a strong base, the redox chemistry of
these compounds is a function of pH.11 Consequently, elabor-
ation of the ethynyl linkage in derivatives of 2 could generate
pH dependent molecular switches.12 We therefore explored con-
jointly with reactions of 1 the scope of the deplatination reac-
tion and properties of both cycloplatinated and non-platinated
ethynyldimethylaminomethylferrocene derivatives.
Results and discussion
The most convenient and high yielding synthesis of acetylide
derivatives of cycloplatinated ferrocenylamines is via
a
Sonogashira CuI-catalysed reaction with 1b. Trimethylsilyl 3,
ferrocenyl 5, and phenyl 6 derivatives were prepared via this
route (Scheme 1). Reactions with 1a, where a chloro group is
in the trans Pt–C position, also gave the same products but in
lower yields.
Under Sonogashira conditions, the reaction of 1a with
trimethylsilylacetylene led to partial desilylation of 3, with
consequent formation of the butadiyne 4 (Scheme 2).
Alternative routes attempted—the prior removal of the
halide from 1 with a Tl[I] salt13 or using a triflate derivative
of 1—were unsuccessful. However, reaction of 1b with tri-
methylsilylacetylene or phenylacetylene in the presence of
LDA (lithium diisopropylamide) gave 3 and 6, respectively, in
relatively poor yield (Scheme 3).
An unexpected deplatination of 1a to give 1-dimethyl-
aminomethyl-2-phenylethynylferrocene 7, was observed with
phenylacetylene in the presence of LDA (Scheme 4). Although a
similar deplatination reaction occurred with ethynylacetylenes
it was not reproducible. For reference, the other functional-
ised ferrocenylamines 8 and 910 were directly synthesised
from 1-dimethylaminomethyl-2-iodoferrocene 10 (Scheme 5).
The mechanism for the deplatination step is problematical.
During preliminary work on this system,7 it was observed
that a deplatination reaction appeared to take place when a
terminal acetylene was reacted with 1, the product being an
ethynyl derivative of the ferrocenylamine 2. This type of deriv-
ative has been reported by Schlögl and co-workers8 and the
ortho-directing character of the CH2NMe2 substituent has been
used to make, selectively, 1,2-ethynyl derivatives and oligomers.9
Related chiral 1,2-ethynyl derivatives with a CH(Me)NMe2
† Electronic supplementary information (ESI) available: rotatable 3-D
crystal structure diagram in CHIME format. See http://www.rsc.org/
suppdata/dt/bl/b101776o/
2362
J. Chem. Soc., Dalton Trans., 2001, 2362–2369
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b101776o

View Article Online
Scheme 4
Scheme 1
Scheme 5
compounds containing two inserted diphenylacetylenes to give
carbocycles can be achieved17 by treatment with a neutral
coordinating ligand. A similar route may be responsible for the
deplatination.
Scheme 2
The acetylide derivatives were fully characterised by elem-
1
ental analysis, IR, H, 13C and, where appropriate, 195Pt NMR
spectroscopy. Spectroscopic data for 7–9 were unremarkable.
Diastereotopic resonances for both the NMe2 and dmso groups
1
in the H NMR of 3–6 show that ring stereochemistry was
maintained, although the difference in chemical shift is smaller
1
᎐
than in 1. The ν(C᎐C) modes for 3–5 are typical of a σ,η -
᎐
acetylide group bound to a metal, with their energy dependent
᎐
on the acetylide terminal substituent. Two coupled ν(C᎐C)
᎐
modes for 4 at 2172 and 2126 cmϪ1 are also consistent with a
σ,η1-diethynyl structure. A puzzling observation was the two
ν(C᎐C) bands at 2113 and 2097 cmϪ1 obtained reproducibly in
᎐
᎐
both the solution and solid spectra of 5, whereas the PPh3 ana-
logue 11 (generated by the direct reaction of 5 with PPh3 in
CH2Cl2) has only one band at 2074 cmϪ1.
Theoretical calculations, spectroscopic and X-ray data of η1-
alkynyl ligands bound to platinum suggest18 that these ligands
are good σ- and π-donors but poor π-acceptors. In η1-alkynyl
᎐
phosphine derivatives the electron population of π*(C᎐CR)
Scheme 3
᎐
is <0.05 e with strong destabilising interactions between the
᎐
Oxidatively-induced reductive elimination facilitated by electron-
donor substituents on the Pt() centre has been demonstrated
for cis-aryl(ferrocenylacetylide)Pt complexes14 but reductive
elimination necessitates a cis stereochemistry.15 Previous work
has shown4 that the Pt–C bond in 1 is not attacked by electro-
philes or nucleophiles; moreover, although insertion with
cyclopalladated species is well known,16 we have not been able
to induce the insertion of alkynes into the Pt–C bond. Pertinent
to our system is that demetalation of cyclopalladated ferrocenyl
occupied π-metal and π(C᎐CR) orbitals. These electronic fac-
᎐
tors are reflected in the 195Pt NMR data for 3–6, where the
chemical shifts follow the relative π-donor capability of the
anions. Thus, the 195Pt chemical shift is between Ϫ3700 and
Ϫ3725, irrespective of the acetylide substituent, similar to that
for the acetato (Ϫ369213), but significantly upfield from 1a
and 1b (Ϫ3763 and Ϫ3815 respectively4). The 13C{1H} NMR
spectra of 3 shows signals due to the α- and β-alkynyl carbons
1
at 148.7 and 107.4 ppm. J(C–Pt) coupling to the α-alkynyl
J. Chem. Soc., Dalton Trans., 2001, 2362–2369
2363

View Article Online
Fig. 1 Perspective view of 3 showing the atom numbering scheme with
thermal elipsoids drawn at the 50% probability level. For clarity only
the first two atoms of consecutively numbered cyclopentadienyl rings
are labelled.
carbon atoms was 920 Hz. For comparison, 1J(C–Pt) coupling
to the cyclometallated C atoms of the ferrocenyl moiety was
790 Hz. Similar observations were made by Osakada et al19 for
Fig. 2 Perspective view of 5 showing the atom numbering scheme with
thermal ellipsoids drawn at the 50% probability level. For clarity only
the first two atoms of consecutively numbered cyclopentadienyl rings
are labelled.
1
᎐
the complex Pt(p-C H OMe)(C᎐CPh)(PEt ) which had J(C–
᎐
6
4
3 2
Pt) values of 890 Hz for the metal bound alkyne carbon and
673 Hz for the coordinated arene C atom. They interpreted this
finding to support the partial contribution of a vinylidine struc-
ture to the Pt–alkynyl bond; equally the findings are consistent
with an acetylide functioning as a good π-donor. However,
since such coupling constants are significantly dependent on the
Fermi contact interaction and therefore the s-character in the
Pt–C bonds20 (albeit modified by relativistic effects in the case
of heavy nuclei such as Pt), the direct and uncompensated com-
parison of coupling constants involving alkyne (sp) and arene
(sp2) carbon atoms must be viewed with some caution. It is
noteworthy in this context that the trends in Pt–C (alkyne) and
Table 1 Selected bond lengths (Å) and angles (Њ) for 3 and 5
3
5
Pt(1)–C(3)
Pt(1)–C(16)
C(16)–C(17)
C(17)–Si(1)
2.031(5)
2.076(5)
1.164(8)
1.838(6)
2.023(5)
2.059(6)
1.191(8)
C(17)–C(18)
Pt(1)–S(1)
Pt(1)–N(1)
N(1)–C(1)
C(1)–C(2)
C(2)–C(3)
Fe(1)–C(2–6) mean
Fe(1)–C(7–11) mean
Fe(2)–C(18–22) mean
Fe(2)–C(23–27) mean
1.447(8)
2.1912(14)
2.141(4)
1.510(7)
1.495(7)
1.438(7)
2.067(18)
2.055(8)
2.051(13)
2.039(15)
2.1853(13)
2.137(4)
1.505(7)
1.495(7)
1.424(7)
2.05(2)
Pt–C(arene) bond distances observed for 3 are the complete
᎐
reverse of those for Pt(p-C H OMe)(C᎐CPh)(PEt ) , vide infra.
᎐
6
4
3 2
This further suggests that factors other than the degree of back-
bonding in the Pt–C vector contribute significantly to the
observed coupling constants.
To give us a comparative structural basis for a discussion of
the electronic characteristics of the acetylide ligand in these
2.042(5)
C(3)–Pt(1)–N(1)
C(16)–Pt(1)–N(1)
C(3)–Pt(1)–S(1)
C(16)–Pt(1)–S(1)
C(3)–Pt(1)–C(16)
N(1)–Pt(1)–S(1)
C(17)–C(16)–Pt(1)
C(16)–C(17)–Si(1)
C(16)–C(17)–C(18)
82.10(18)
90.26(17)
97.61(15)
89.91(14)
171.9(2)
177.63(12)
173.5(5)
82.72(18)
90.03(19)
96.68(14)
90.47(16)
172.4(2)
177.34(13)
175.4(5)
cycloplatinated complexes, and to seek a structural explanation
᎐
for the anomalous ν(C᎐C) spectrum of 5, X-ray structural
᎐
analyses were carried out on 3 and 5.
X-Ray structures of 3 and 5
For 3, the crystals comprise unique molecules of 3 together
with benzene solvate in the monoclinic unit cell. For 5, the
monoclinic unit cell contains molecules of 5 together with two
water molecules, both at approximately 50% occupancy. Per-
spective views of 3 and 5 are shown in Fig. 1 and 2, respectively.
These figures define the atom numbering schemes used.
Selected bond length and angle data for both molecules appear
in Table 1.
For 3 and 5 the coordination spheres of the platinum atoms
comprise the S(1) atoms of S-bound dmso ligands trans to the
amine N(1) atoms of the orthometallated ferrocenylamine
ligands. The Pt-bound C(3) atom of the substituted cyclopenta-
dienyl ring is trans to the C(16) atom of the terminally bound
trimethylsilylacetylide for 3 and ferrocenylacetylide for 5. These
occupy the fourth coordination site in each case. The N(1) and
C(3) atoms are the donors in five-membered chelate rings
formed by the orthometallated ferrocenylamine ligands in both
molecules. Some deviation from the idealised square planar
geometry about Pt is apparent with the angles C(3)–Pt–C(16)
170.2(6)
177.7(6)
171.9(2)Њ for 3 and 172.4(2)Њ for 5 with S(1)–Pt(1)–N(1)
177.63(12)Њ for 3 and 177.34(13)Њ for 5. Deviations from the
PtL4 ring plane for 3 range from Ϫ0.0391(17) Å for Pt(1) to
0.0061(18) Å for S(1) while for 5, the corresponding range is
Ϫ0.0371(19) Å for Pt(1) to 0.006 (3) Å for C(16). The Pt–N
(2.137(4) Å for 3, 2.141(4) Å for 5) and Pt–S (2.1853(13) Å for 3,
2.1912(14) Å for 5) distances are similar to those found in
related compounds.4–6,13 The Pt–C(3) bond to the metallocene
ring is significantly longer, 2.031(5) Å for 3 and 2.023(5)
for 5, than the corresponding distances in the closely related
compounds [Pt{(σ,η5-C5H3CH2NMe2)Fe(η5-C5H5)}(dmso)X],
(X = Cl,4 O2CCH313) where the comparable bond lengths were
1.988(7) Å and 1.976(8) Å, respectively. Furthermore, the
Pt–C(16)(alkyne) bonds, 2.076(5) Å for 3 and 2.059(6) for 5, are
longer than the Pt–C(3)(Cp) bonds, a variation that may reflect
2364
J. Chem. Soc., Dalton Trans., 2001, 2362–2369

View Article Online
the competitive trans influence across the coordination spheres
of the Pt atoms. In trans-bis(alkyne) platinum complexes
the Pt–C(alkyne) bond distance is generally close to 2.00 Å.21
However, to our knowledge, no comparable data exist on
compounds with σ-bound cyclopentadienyl rings trans to Pt-
substituted alkynes but, given the similarity between the η5-C5
ring compounds and arenes, comparison with trans-Pt–arene
systems may be useful. An extensive structural investigation,
comparing Pt–Cl bond extension in trans-alkynyl–Pt–Cl vs
trans-arene–Pt–Cl complexes, revealed that arenes have a
greater trans influence than alkynyl ligands.18 The bond lengths
reported here show a similar trend. However, structural data
from a variety of trans-alkynyl–Pt–arene complexes show that
this trend is exclusively, and often significantly, reversed.22
A
similar effect is also noted in analogous Pd compounds and
partial contribution of a vinylidene structure has been cited as
an explanation.19 There is, however, no evidence for such
Fig. 3 Square-wave and cyclic voltammograms of 5 in CH2Cl2, 15 Hz
(sweep rate) 200 mV sϪ1, Pt, 20 ЊC; referenced to decamethylferrocene at
reversal in the molecules reported here as, in addition to the
0.0 V.
᎐
long Pt(1)–C(3) bonds for both 3 and 5, the C᎐C bond lengths,
᎐
7–9; for example, E 0/ϩ[2] ϩ 0.16 V = 0.71 V, close to E 0/ϩ [7] at
0.70 V. This demonstrates the well-illustrated principle that the
E 0/ϩ potentials for most ferrocenyl compounds can be calcu-
lated from the relationship, E 0/ϩ [Fc] = 0.55 V plus Σ(δa), where
δa is a substitution parameter for each component.6,24
Surprisingly, substitution of a halide ion in the cycloplatin-
ated compounds by the acetylide anion causes a large cathodic
shift in the potential for the oxidation of the ferrocenyl com-
ponent of the PtNSL(σ-CFc) square planar unit (Scheme 6).
1.164(8) Å for 3 and 1.191(8) Å for 5, are somewhat short in
᎐
comparison to the mean C᎐C distance of 1.201(16) Å observed
᎐
for a range of L M–C᎐C–R (R = C or Si) compounds.18 Clearly
᎐
᎐
n
in both cases, either the arene/cyclopentadienyl comparison is
inappropriate, or the stereochemical consequences of chelate
ring formation in the cyclometallated complexes have a signifi-
cant influence on the Pt–C(Cp) and Pt–C(alkyne) vectors. We
would favour the latter explanation.
The cyclopentadienyl rings in the orthometallated ferrocene
moieties adopt an approximately eclipsed conformation, with
the mean torsion angle C(m)–C(1g)–C(2g)–C(n) of 1.21(1)Њ for
3 and 1.2(2)Њ for 5 (C(m) = C(2)–C(6), C(n) = C(7)–C(11) and
C(1g) and C(2g) are the corresponding ring centroids). The
dihedral angle between the planes of the orthometallated Cp
rings is 7.5(5)Њ for 3 and 5.4(5)Њ for 5. The average Fe–C dis-
tances (2.05(2) Å for 3, 2.06(1) for 5) are not unusual and the Fe
atom is almost equidistant from each of the Cp rings in both
molecules (1.653(2) Å and 1.647(3) Å for 3; 1.666(3) Å and
1.660(3) Å for 5).23 An approximately eclipsed conformation is
also displayed by the ferrocene of the ferrocenylacetylide ligand
with the mean torsion angle 5(3)Њ, as detailed above. The aver-
age Fe–C distance is 2.05(1) and the Fe(2) atom lies 1.651(3) Å
and 1.653(3) Å, respectively, from the Cp rings. The interplanar
angle between the Cp rings of this pendant ferrocenyl (1.76(5)Њ)
is significantly smaller than those for the cyclometallated
ferrocenyl rings in both 3 and 5. This may further reflect steric
interactions in the cyclometallated system arising from the
formation of the chelate ring.
Redox chemistry
Scheme 6
Electrochemistry
Previous workers have9 experienced difficulties in measuring
the electrochemistry of 1-ethynyl-3-dimethylaminoferrocenes.
This was attributed to oxidation of the amino functionality,24
although it is just as likely that protonation caused the broad
i/V responses.11 The protonated chiral analogues of Long and
co-workers10 gave sensible i/V responses but the deprotonated
compounds were unstable. Not surprisingly, the i/V responses
for 7–9 were broader, due to facile protonation, whereas 3–6
displayed Nernstian kinetics.
All compounds in our work with a single ferrocenyl centre
fortunately gave the normal2,5 cyclic and square-wave volt-
ammograms in CH2Cl2 of a chemically-reversible one-electron
[Fc]ϩ/0 couple (potentials are referenced to decamethylferrocene
at 20 ЊC). Deprotonated 2 is oxidized at 0.55 V11 and 1a and 1b
at 0.33 V and 0.34 V, respectively.4,7 An ethynyl group normally
acts as an electron-withdrawing substituent and the anodic shift
between ferrocene and ethynylferrocene is 0.16 V. This indeed is
the anodic difference between E 0/ϩ[2] and its ethynyl derivatives
For example, E 0/ϩ [3] = 0.21 V, a difference of 0.12 V from
E 0/ϩ[1a]. We have found7 that σ-donor substituents trans to the
Pt–C bond with little π-donor characteristics have little influ-
ence on E 0/ϩ for cycloplatinated species, whereas the opposite is
true for the trans Pt–N π-acceptor groups trans to the NMe2
substituent. The cathodic shift, together with the 195Pt NMR
data for 3–6 and theoretical calculations for related systems,18
are consistent with the acetylide anion acting as a π-donor. It is
intriguing that the π-donor and π-acceptor ligand requirements
around the cycloplatinated unit are specific and directional.
Presumably, the orbital symmetry and coefficients of the
Pt–C bond are the parameters which are dictating the electronic
configuration about the Pt() ion.
The complexes in which there are two ferrocenyl redox
centres, 5 and 9, give an insight into the possible electronic
interaction between the non-equivalent ferrocenyl redox centres
᎐
linked via a C᎐C bond. 5, displays two reversible one-electron
᎐
steps A, B in the cyclic and square-wave voltammetry (Fig. 3).
J. Chem. Soc., Dalton Trans., 2001, 2362–2369
2365

View Article Online
Fig. 4 Square-wave voltammogram of 8 in CH2Cl2, 15 Hz, Pt, 20 ЊC;
referenced against decamethylferrocene at 0.0 V.
Fig. 5 OTTLE UV/VIS spectra of 3 in CH2Cl2; arrows indicate
increasing absorption as the voltage is increased to A, the first redox
process.
E 0/ϩ [A] = 0.25 V, is close to E 0/ϩ [3] and is therefore assigned to
the unperturbed cycloplatinated redox centre. Cycloplatination
shifts a E 0/ϩ[Fc] couple cathodically by ∼0.2 V.7 As E 0/ϩ = 0.71
0/ϩ
᎐
᎐
V for HC᎐CFc we would expect E for a ‘Pt’C᎐CFc unit to be
᎐
᎐
∼0.51 V, remarkably close to E 0/ϩ [B] at 0.53 V for 5 (Scheme 7).
Scheme 8
Electronic spectra
The electronic spectra of cycloplatinated derivatives of 1 are
characterised7 by two absorptions, 450–470 nm (ε ∼600), and
<350 nm (ε > 1000). Both bands are similar in energy and
intensity to those found in other ferrocene compounds.25 They
are of little diagnostic value as their energies are relatively
invariant to changes in the Pt() coordination sphere. However,
significant changes to the electronic spectra take place upon
oxidation of the cycloplatinated acetylides. One-electron oxid-
ation of 3 at 0.21 V gave rise to the new spectrum shown in
Fig. 5 with a well-defined isosbestic point at 246 nm. Reduction
of 3ϩ gave the original spectrum, showing that only one long-
lived species is formed on oxidation. The high energy (320 nm,
ε = 2500) and low energy (464 nm, ε = 340) ‘ferrocenyl’ bands in
3 are red-shifted to 395 nm (ε = 1320) and ∼480 nm (shoulder,
ε ∼540) respectively in 3ϩ. Similar red-shifts are observed upon
the oxidation of 1a and 1b, confirming that the acetylide
group does not perturb the HOMO–LUMO levels of the
Pt() coordination sphere. The new weak band for 3ϩ at 590 nm
(ε = 500) could be assigned to a Fe-centred SOMO (e2g)→
LUMO transition (normally found at ∼630 nm in ferrocenium
compounds). The other new band, a broad low energy absorp-
tion at 780 nm (ε = 770), is assigned to a MCLT transition
centred on the Pt() unit. Precise assignments must await theor-
etical and Raman studies underway at present but, to our know-
ledge, this is the first report of low-energy electronic transitions
in either neutral or oxidised, cycloplatinated or cyclopalladated
complexes. One-electron oxidation of 5 to 5ϩ, a complex which
has two ferrocenyl redox centres, gave essentially the same
absorption profile as 3ϩ (Fig. 6). There is a small red-shift in the
Scheme 7
If there was significant interaction between the cycloplatinated
and terminal ferrocenyl redox centres oxidation of the former at
the first redox centre, A, would make it more difficult to oxidise
the second ferrocenyl redox centre, B. The observation that the
potentials for A and B are similar to those for analogous
isolated redox centres shows that there is minimal communi-
cation between them. This conclusion was supported by the
observation that substitution of dmso by PPh3 in the trans
Pt–N position, complex 11, gave E 0/ϩ = 0.20 V and 0.48 V for
A and B, respectively. The small cathodic shift when dmso is
replaced by a better donor is similar to that for non-acetylide
complexes.4
The situation for the non-cycloplatinated complex 9 is differ-
ent. In this case, two one-electron processes are observed in the
cyclic and square-wave voltammograms (Fig. 4), C (E 0/ϩ = 0.62
V, reversible) and D (E 0/ϩ = 0.86 V, quasi-reversible), neither of
which corresponds to the potentials for unperturbed parent
redox centres. C is assigned to the dimethylaminomethylferro-
cenyl redox centre shifted from E 0/ϩ for the Me3Si analogue
᎐
by the electron donating FcC᎐C–Cp ring substituent. The
᎐
coulombic effect of the oxidation of the ferrocenylamine then
causes an anodic shift of E 0/ϩ [FcC᎐C–] giving rise to B at
᎐
᎐
0.86 V (Scheme 8). That is, the ethynyl link in 9 provides a bridge
for communication, and hence inductive and coulombic per-
turbation of the redox centres, which is not available with an
acetylide link in 5.
2366
J. Chem. Soc., Dalton Trans., 2001, 2362–2369

View Article Online
all reactions were performed under nitrogen. 1,4 ethynylferro-
cene,26 and 1027 were prepared by literature methods. Com-
mercial reagents were used as received. IR spectra were
recorded on a Perkin-Elmer Spectrum BX FT-IR, NMR on a
Varian VXR300 MHz or Gemini 200 MHz (proton decoupled
195Pt spectra were referenced to K2PtCl4) and electronic spectra
on a Jasco V 550 UV/VIS. Microanalyses were carried out
by the Campbell Microanalytical Laboratory, University of
Otago. Mass spectra were recorded on a Kratos MS80RFA
instrument with an Iontech ZN11NF atom gun. Cyclic and
square-wave voltammetry in CH2Cl2 were performed for all
compounds using a three-electrode cell with a polished disk, Pt
(2.27 mm2) as the working electrode; solutions were ∼10Ϫ3 M in
electroactive material and 0.10 M in supporting electrolyte
(triply recrystallised TBAPF6). Data were recorded on an EG &
G PAR 273A computer-controlled potentiostat. Scan rates of
0.05–1 V sϪ1 were typically employed for cyclic voltammetry
and for Osteryoung square-wave voltammetry, square-wave
step heights of 1–5 mV, a square amplitude of 15–25 mV with a
frequency of 30–240 Hz. All potentials are referenced to
decamethylferrocene; E1/2 for sublimed ferrocene was 0.55 V.
Infrared and UV/VIS OTTLE data were obtained from stand-
ard cells with platinum grid electrodes.
Fig. 6 OTTLE UV/VIS spectra of 5 in CH2Cl2; arrows indicate
increasing absorption as the voltage is increased: (A) the first oxidation
potential and (B) the second oxidation potential.
weak 600 nm, (ε = 300) and 800 nm (ε = 540) bands from those
in 3ϩ and the high (345 nm, ε = 1610) and lower energy (456 nm,
ε = 538) ferrocenyl bands effectively merge to give a broad
unresolved absorption. These observations are consistent with
the electrochemical assignment in Scheme 7, in which the first
redox process is centred on the cycloplatinated ferrocenyl unit.
What was unexpected for 5ϩ, given that the electrochemical
data did not support interaction between the two redox centres,
was the very broad band in the near-infrared at ∼1100 nm
(ε = 484) which has a profile typical of an IT transition (Fig 6).
If it is an IT transition, some mixing of the Pt() unit and
pendant ferrocenyl orbitals must occur but more work is
required to determine its origin. Oxidation of the pendant
ferrocenyl group to give 52ϩ eliminated the 1100 nm band and
caused a blue-shift of the 600 nm band to 587 nm (ε = 925) and
both this band and that at 801 nm (ε = 1960) to gain significant
intensity. The isosbestic point 985 nm confirms that only two
species are involved in the transformation from 5ϩ to 52ϩ. Since
Preparation of 3–5 via the Sonogashira method
(Trimethylsilyl)acetylene (0.155 ml, 1.1 mmol) was stirred for
24 h at room temperature with 1a (0.55 g, 1 mmol) and CuI
(19 mg, 0.1 mmol) in 20 ml degassed piperidine. The solvent
was removed in vacuo and the residue separated using column
chromatography (alumina/CH2Cl2). The first band eluted was
4, 53 mg (8%). (Found: C, 41.11; H, 5.11; N, 2.09.
C22H31FeNOPtSSi requires: C, 41.51; H, 4.91; N, 2.20%).
Electrospray-MS: m/z 637 (Mϩ). δH(CDCl3): 0.17 (s, 9H,
Si-CH3), 2.97, 3.33 {2 × [s, 3H, Pt satellites (J = 39 Hz),
N-CH3]}, 3.60, 3.65 {2 × [s, 3H, Pt satellites (J = 30 Hz),
S-CH3]}, 3.44, 3.77 {2 × [d (J = 14 Hz), 1H, -CH2-]}, 4.10
(s, 5H, η5-C5H5Fe), 4.12 (m, 2H, η5-C5H3Fe), 4.37 (m, 1H,
ϩ
᎐
the 587 nm and 800 nm bands are characteristic of the C᎐CFc
᎐
unit, the electronic spectrum of 5ϩ gives credence to the notion
that there is little orbital interaction between the two ferrocenyl
redox centres.
5
Ϫ1
᎐
η -C H Fe). δ (CDCl ): Ϫ3725. ν(KBr, C᎐C, cm ): 2172,
᎐
5
3
Pt
3
2126.
The second band eluted was 3, 410 mg (67%), crystallized as
a yellow-orange solid from benzene–hexane. (Found: C, 39.46;
H, 5.05; N, 2.29; S, 4.80. C20H31FeNOPtSSi requires: C, 39.21;
H, 5.10; N, 2.29; S, 5.24%). δH(CDCl3): 0.12 (s, 9H, Si-CH3),
2.95, 3.33 {2 × [s, 3H, Pt satellites (J = 39 Hz), N-CH3]}, 3.62,
3.67 {2 × [s, 3H, Pt satellites (J = 31 Hz), S-CH3]}, 3.42, 3.81
{2 × [d (J = 14 Hz), 1H, -CH2-]}, 4.09 (s, 5H, η5-C5H5Fe), 4.11
(m, 2H, η5-C5H3Fe), 4.35 (m, 1H, η5-C5H3Fe). δPt(CDCl3):
Conclusion
Successful syntheses have been developed for cycloplatinated
ferrocenylamine-acetylides. The same strategy could be used to
make chiral analogues and molecules with two platinated
groups per ferrocene. These newly developed syntheses provide
the platform for incorporation of the easily oxidised cycloplati-
nated redox centre into dendrimers and other molecular archi-
tectures. For example, structures in which the cycloplatinated
ferrocenyl unit acts as a linker to fluorophores have been made
in these laboratories. There are two important electronic con-
siderations when designing these structures. First, the acetylide
in the trans Pt–C position must be a good π-donor, with the
ligand in the cis position a good π-acceptor. This separability
could be exploited to direct energy flow in specific directions.
Second, there does not appear to be effective interaction or
energy flow across the acetylide bridge; that is, the acetylide is
an insulator. While this is not necessarily important for sensors,
it is for molecular wire development. In contrast, the non-
platinated dimethylaminoferrocenes do provide the communi-
cation and, an added bonus, pH control of potential, but they
are less robust. The observation of low-energy electronic transi-
tion in the oxidised cycloplatinated complexes has important
consequences for use of these complexes as redox switches
when linked to fluorophores and these are being investigated at
present.
Ϫ1
᎐
Ϫ3700. ν(KBr, C᎐C, cm ): 2034. λmax (CH2Cl2, nm); 251, 288,
᎐
465. E ϩ/0 (CH2Cl2, V): 0.21.
By a similar procedure 1b reacted with ethynylferrocene over
1 h to give an orange precipitate of 5 (82%). X-Ray quality
crystals were obtained from CH2Cl2 layered with EtOH.
(Found: C, 44.90; H, 4.55; N, 1.89; S, 4.18. C27H31Fe2NOPtS
requires: C, 44.77; H, 4.31; N, 1.93; S, 4.43%). Electro-
spray-MS: m/z 724 (Mϩ). δH(CDCl3): 3.02, 3.39 {2 × [s, 3H,
Pt satellites (J = 38 Hz), N-CH3]}, 3.66, 3.70 {2 × [s, 3H, Pt
satellites (J = 31 Hz), S-CH3]}, 3.45, 3.82 {2 × [d (J = 14 Hz),
1H, -CH2-]}, 4.09, 4.31 [2 × (t, 2H,η5-C5H4FeЈ)], 4.13 (s, 5H, η5-
C5H5Fe), 4.15 (m, 2H,η5-C5H3Fe), 4.17 (s, 5H, η5-C5H5FeЈ),
5
᎐
4.39 (m, 1H, η -C H Fe). δ (CDCl ): Ϫ3700. ν(KBr, C᎐C,
᎐
5
3
Pt
3
cmϪ1): 2113, 2097. λmax (CH2Cl2, nm); 237, 287, 456. E ϩ/0
(CH2Cl2, V): 0.25, 0.53.
A similar procedure can be used to prepare 6 and is described
below.
Preparation of 6 from 1 using LDA
LDA (0.6 mmol) was added to phenylacetylene (0.9 mmol) in
10 ml THF and the solution stirred for 30 min at room temper-
ature. 1a (0.55 mmol) dissolved in THF (10 ml) was added and
Experimental
Solvents were dried and distilled by standard procedures, and
J. Chem. Soc., Dalton Trans., 2001, 2362–2369
2367

View Article Online
Table 2 Crystal data and structure refinement for 3 and 5
3
5
Chemical formula
Formula weight
C26H37NOSiSFePt
690.66
C27H31NO2SFe2Pt
740.38
Crystal system
Monoclinic
C2/c
5.775
Monoclinic
C2/c
6.315
Space group
Absorption coefficient/mmϪ1
Final R indices [I > 2σ(I )] R1
wR2
0.0454
0.1046
0.0335
0.0848
R indices (all data) R1
0.0670
0.1117
0.994
0.0433
0.0927
1.014
wR2
Goodness-of-fit on F2
a/Å
b/Å
c/Å
31.9871(6)
10.3174(2)
19.3210(3)
120.8720(10)
5472.95(17)
8
29.924(10)
10.132(3)
19.366(6)
112.810(7)
5413(3)
8
β/ Њ
V/Å3
Z
T /K
153(2)
32062
7605 [0.0712]
168(2)
33369
5535 [0.0380]
Reflections collected
Independent reflections [R(int)]
the mixture stirred for 12 h. The solvent was removed in vacuo
and the orange solid separated on preparative silica gel plates
(3 : 2; ethylacetate–hexane); work-up of the second band gave
an orange powder 6 (20%). Found: C, 44.02; H, 4.56; N, 1.95; S,
4.51. C23H27FeNOPtS requires: C, 44.80; H, 4.41; N, 2.27; S,
5.20%). FABMS: m/z 617 (Mϩ). δH(CDCl3): 3.02, 3.39 {2 × [s,
3H, Pt satellites (J = 39 Hz), N-CH3]}, 3.67, 3.71 {2 × [s, 3H, Pt
satellites (J = 30 Hz), S-CH3]}, 3.47, 3.85 {2 × [d (J = 14 Hz),
room temperature. 1a (1.5 mmol) in THF (15 ml) was added
slowly to this solution and the mixture stirred for 8 h. The
solvent was removed in vacuo and the yellow solid dissolved in
dichloromethane. Recrystallisation from ethyl acetate–hexane
(1 : 1) gave 7 as a yellow solid in variable yields. (Found: C,
73.52; H, 6.19; N, 4.05. C21H21FeN requires: C, 73.48; H, 6.17;
ϩ
Ϫ1
᎐
N, 4.08%). FABMS: m/z 343 (M ). ν(C᎐C, CH Cl , cm ): 2117.
᎐
2
2
δH(CDCl3): 2.26 (s, 6H, CH3N), 3.56 (s, 2H, CH2N), 4.17 (s, 5H,
η5-C5H5Fe), 4.21–4.49 (m, 3H, η5-C5H3Fe); 7.33–7.49 (m, 5H,
C6H5).
1H, -CH2-]}, 4.14 (m, 7H, Fc-H), 4.41 (m, 1H, η5-C5H3Fe),
᎐
7.16–7.36 (m, 5H, phenyl-H). δPt (CDCl ): Ϫ3720. ν(KBr, C᎐C,
᎐
3
cmϪ1): 2100. E ϩ/0 (CH2Cl2, V): 0.22. When the reaction was
repeated using 1b the yield increased to 45%. 3, but not 5, could
be prepared in low yield by the same route.
Preparation of 11
PPh3 (0.03 mmol) was added to 5 (0.03 mmol) dissolved in
CHCl3 (10 ml) and the solution stirred for 4 h. The solvent
was stripped and the product 11 recrystallised from CH2Cl2–
MeOH. Electrospray-MS: m/z 908 (Mϩ). 1H NMR δ: 3.26, 3.47
Preparation of 8 and 9
A solution containing 10 (0.05 mmol), trimethylsilylacetylene
(0.14 mmol), 2 mol% PdCl2(Ph3P)2/CuI in diisopropylamine
(1 ml) was stirred for 3 h at room temperature and then at 50 ЊC
for 2 h. The solvent was evaporated, and the crude brown oil
seperated on a silica gel column (MeOH–hexane, 7 : 1). The
third band was separated on preparative silica gel plates with
the same eluent; workup of the first band gave 8, as yellow
crystals (13%). (Found: C, 63.97; H, 7.87; N, 4.31. C18H25FeNSi
requires: C, 63.75; H, 7.43; N, 4.13%). Mass spectrum (EI): 339
(M)ϩ; 266 (MϩϪ(TMS)). δH(CDCl3): 0.23, 9H, Si(CH3)3); 2.21
(s, 6H, 2 × N-CH3); 3.49 (s, 2H, CH2); 4.11(s, 5H, Cp); 4.15
(t, 1H, 1 × Cp1-H); 4.30 (m, 1H, 1 × Cp1-H); 4.41 (m, 1H,
1 × Cp1-H).
Ethynylferrocene (7.6 mmol) and PdCl2(PPh3)2/CuI (2 mol%)
in diisopropylamine were added to 10 (3.8 mmol) and the
solution heated under reflux for 16 h. The reaction mixture
was then extracted into ether, washed with water, and treated
with MgSO4. Column chromatography with CH2Cl2 eluted
unreacted 10. Elution with MeOH–hexane–ether (1 : 1 : 1) gave
a yellow fraction which was separated on preparative silica gel
plates with CH2Cl2–hexane–MeOH (4 : 2 : 1). Workup of the
major orange band gave reddish-brown needles of 9 (53%).
(Found: C, 66.47; H, 5.87; N, 3.41. C25H25Fe2N requires: C,
66.55; H, 5.59; N, 3.11%). δH(CDCl3): 2.28 (s, 6H, 2 × N-CH3);
3.56 (s, 2H, CH2); 4.15 (s, 5H, Co); 4.1–4.2 (m, 4H, 3 × Cp1-H);
4.23 (s, 5H, Cp); 4.32 (m, 1H, 1 × Cp1-H); 4.4–4.5 (m, 2H,
2 × Cp1-H). The phenyl analogue 7 could be prepared by the
same route.
4
[2 × (d, 3H, JP-H = 2 Hz, Pt satellites, N-CH3], 3.6 (m, 2H,
-CH2-), 3.8–3.9 (m, 4H, Fc-H), 3.82 (s, 5H, η5-C5H5Fe), 3.89 (s,
5H, η5-C5H5FeЈ), 3.89 (t, 2H, η5-C5H4FeЈ), 4.07 (m, 1H, η5-
C5H3Fe), 7.4 (m, 9H, phenyl-H), 7.8 (m, 6H, phenyl-H). 195Pt
NMR δ: Ϫ4210 (d, JPt-P = 4090 Hz) 31P NMR δ: 16 [Pt satellites
Ϫ1
᎐
(J = 4090 Hz)]. ν(KBr, C᎐C, cm ): 2074. λ
(CH2Cl2, nm);
᎐
max
293, 465. E ϩ/0 (CH2Cl2, V): 0.20, 0.48.
X-Ray data collection, reduction and structure solution for 3 and
5
Crystal data for 3 and 5 are given in Table 2. Recrystallisation
of 3 from benzene/hexane yielded red plates, one of which was
used for data collection. 5 was recrystallised from ethanol as
orange-red plates. Data were collected at 153(2) K for 3 and
163(2) K for 5 on a Bruker SMART CCD diffractometer,
processed using SAINT,28 with empirical absorption corrections
applied using SADABS.28 Both structures were solved using
SHELXS28 and refined by full-matrix least-squares using
SHELXL-9728 and TITAN2000.29 Non-hydrogen atoms were
assigned anisotropic temperature factors and the H atoms were
included in calculated positions. A difference Fourier synthesis,
following the location of all of the non-hydrogen atoms for 3,
revealed six substantial high peaks that could be readily
assigned to a benzene solvate molecule in the crystal lattice and
refined with isotropic H atoms included in calculated positions.
At the same stage of the refinement of 4, two high peaks
remained in the Fourier map. These were assigned to O atoms
of water molecules, presumably derived from the ethanol
solvent. Refinement of the occupancy factors showed approxi-
mately 50% occupancy in each case. These were fixed at 0.5 in
the final refinement cycles. No attempt was made to locate the
Preparation of 7 from 1a
LDA (4.5 mmol) was added to a solution of phenylacetylene
(1.5 mmol) in THF (20 ml) and the mixture stirred for 30 min at
2368
J. Chem. Soc., Dalton Trans., 2001, 2362–2369

View Article Online
12 L. Fabbrizzi, M. Licchelli and P. Pallavicini, Acc. Chem. Res., 1999,
H atoms for these water molecules. For both structures, aniso-
tropic refinement of the solvates led to a significant reduction
in R1. In each refinement, the final difference Fourier maps
revealed a number of high peaks close to the Pt atoms but no
physical significance could be attached to them.
CCDC reference numbers 159105 and 159106.
See http://www.rsc.org/suppdata/dt/b1/b101776o/ for crys-
tallographic data in CIF or other electronic format.
32, 846.
13 K. McGrouther, D. K. Weston, D. Fenby, B. H. Robinson and
J. Simpson, J. Chem. Soc., Dalton Trans., 1999, 1957.
14 M. Sato, E. Mogi and S. Kumakura, Organometallics, 1995, 14,
3157.
15 K. Tatsumi, R. Hoffmann, A. Yamamoto and J. K. Stille, Bull.
Chem. Soc. Jpn., 1981, 54, 1857.
16 M. Pfeffer, J. P. Sutter, M. A. Rotteveel, A. De Cian and J. Fischer,
Tetrahedron, 1992, 48, 2427; W. Tao, L. J. Silverberg, A. L.
Rheingold and R. F. Heck, Organometallics, 1989, 8, 2550.
17 M. Pfeffer, J. P. Sutter, A. De Cian and J. Fischer, Organometallics,
1993, 12, 1167.
Acknowledgements
18 J. Manna, K. D. John and H. D. Hopkins, Adv. Organomet. Chem.,
1995, 38, 79.
19 K. Osakada, R. Sakata and T. Yamamoto, Organometallics, 1997,
16, 5354.
20 B. Wrackmeyer and K. Horschler, Prog. Nucl. Magn. Reson.
Spectrosc., 1990, 22, 209.
We thank Prof. W. T. Robinson (University of Canterbury)
for the collection of X-ray data, the University of Otago and
the Royal Society (NZ) Marsden Fund for financial support,
Robinson College, Cambridge, for a Bye Fellowship (B. H. R.),
Amar Flood for assistance with the OTTLE work and Amanda
Dixon and Noel Duffy for experimental assistance.
21 A. Sebald, C. Stader and B. Wrackmeyer, J. Organomet. Chem.,
1986, 311, 233; J. P. Carpenter and C. M. Lukehart, Inorg. Chim.
Acta, 1991, 190, 7; F. Diederich, R. Faust, V. Gramlich and
P. Seiler, J. Chem. Soc., Chem. Commun., 1994, 2045; A. Harriman,
M. Hissler, R. Ziessel, A. De Cian and J. Fisher, J. Chem.
Soc., Dalton Trans., 1995, 4067; S. Leininger, P. J. Stang and
S. Huang, Organometallics, 1998, 17, 3981; K. Onitsuka, M.
Fujimoto, N. Ohshiro and S. Takahashi, Angew. Chem., Int. Ed.,
1999, 38, 689.
22 J. R. Berenguer, E. Fornies, F. Lalinde, E. Martinez, A. J.
Urriolabeitia and J. Welch, J. Chem. Soc., Dalton Trans., 1994, 1291;
M. Sato, E. Mogi and M. Katada, Organometallics, 1995, 14, 4837;
K. Osakada, R. Sakata and T. Yamamoto, Organometallics, 1997,
16, 5354; J. Lewis, N. J. Long, P. R. Raithby, G. P. Shields, W.-Y.
Wong and M. Younus, J. Chem. Soc., Dalton Trans., 1997, 4283;
W.-Y. Wong, W.-K. Wong and P. R. Raithby, J. Chem. Soc., Dalton
Trans, 1998, 2761.
References
1 T. Weyland, C. Lapinte, G. Frapper, M. J. Calhorda, J.-F. Halet and
L. Toupet, Organometallics, 1997, 16, 2024; Conjugated Polymer
Materials: Opportunities in Electronic, Optoelectronic and Molecular
Electronics, ed. J. L. Bredas and R. R. Chance, NATO ASI Series,
Kluwer, Dordrecht, 1990; vol. 182; I. Manners, Annu. Rep. Prog.
Chem., Sect., A, 1994, 91, 131; Organic Materials for Nonlinear
Optics III, ed. G. J. Ashwell and D. Bloor, Royal Society of
Chemistry, Cambridge, 1993, p. 219; H. Lang, Angew. Chem., Int.
Ed. Engl., 1994, 33, 547.
2 P. Zanello, in Ferrocenes, ed. A. Togni and T. Hayashi, VCH,
Weinheim, 1995, ch. 7.
3 B. H. Robinson and J. Simpson, in Paramagnetic Organometallic
Species in Activation/Selectivity Catalysis, ed. M. Chanon, Kluwer,
Dordrecht, 1989, p. 357.
23 P. Seiler and J. D. Dunitz, Acta Crystallogr. Sect. B: Struct.
Crystallogr. Cryst. Chem., 1979, 35, 1068.
4 C. E. L. Headford, R. Mason, P. R. Ranatunge-Bandarage, B. H.
Robinson and J. Simpson, J. Chem. Soc., Chem. Commun., 1990,
601; P. R. R. Ranatunge-Bandarage, B. H. Robinson and J. Simpson,
Organometallics, 1994, 13, 500.
5 P. R. R. Ranatunge-Bandarage, S. M. Johnston, N. W. Duffy, B. H.
Robinson and J. Simpson, Organometallics, 1994, 13, 511.
6 N. W. Duffy, C. J. McAdam, B. H. Robinson and J. Simpson, Inorg.
Chem., 1994, 33, 5343.
7 N. W. Duffy, PhD Thesis, University of Otago, 1997.
8 H. Paulus, K. Schlögl and H. Weissensteiner, Monatsh. Chem., 1983,
114, 799.
9 H. Plenio, J. Hermann and J. Leukel, Eur. J. Inorg. Chem., 1998,
2063.
24 M. Spescha, N. W. Duffy, B. H. Robinson and J. Simpson,
Organometallics, 1994, 13, 4895.
25 S. Barlow and S. R. Marder, Chem. Commun., 2000, 1555.
26 T. S. Abram and W. E. Watts, Synth. React. Inorg. Met.-Org. Chem.,
1976, 6, 31.
27 M. D. Rausch, A. Siegel and L. P. Klemann, J. Org. Chem., 1966, 31,
2703.
28 SMART (Control) and SAINT (Integration) software, Bruker,
AXS, Madison WI, 1994; SADABS, Bruker AXS, Madison WI,
1997; G. M. Sheldrick, SHELXS-96, A program for the solution of
crystal structures from diffraction data, 1996, University of
Göttingen, Germany; G. M. Sheldrick, SHELXL-97, A program for
the refinement of crystal structures, 1997, University of Göttingen,
Germany.
29 K. A. Hunter and J. Simpson, TITAN2000. A molecular graphics
program to aid structure solution and refinement with the SHELX
suite of programs, University of Otago, New Zealand, 1999.
10 M. C. B. Colbert, J. Lewis, N. J. Long, P. R. Raithby, D. A. Bloor
and G. H. Cross, J. Organomet. Chem., 1997, 531, 183.
11 N. W. Duffy, J. Harper, P. R. R. Ranatunge-Bandarage, B. H.
Robinson and J. Simpson, J. Organomet. Chem., 1998, 564, 125.
J. Chem. Soc., Dalton Trans., 2001, 2362–2369
2369