Transition metal-substituted paraffins: Synthesis and properties of some μ-saturated heterobimetallic complexes containing Mo and W (see abstract)

Article abstract of DOI:10.1016/j.jorganchem.2003.10.002

The new heterobimetallic complexes [(η⁵-C₅ H₅)(CO)₃W(CH₂)_n Mo(CO)₃(η⁵-C₅H₅)] n=3 to 6; [(η⁵-C₅H₅) (CO)₃W(CH₂)_nMo(CO)₃ (η⁵-C₅(CH₃)₅)] n=3, 4; [(η⁵-C₅H₅)(CO)₃ W(CH₂)_nMo(CO)₂(PPh_i Me_3-i)(η⁵-C₅H₅)] and [(η⁵-C₅H₅)(CO)₂ Fe(CH₂)_nMo(CO)₂(PPh_i Me_3-i)(η⁵-C₅H₅)] (n=3,4 and i=0 to 3) were synthesized by direct displacement of the iodide of a metallo-iodoalkyl complex with the appropriate anion. The complexes have been fully characterised by IR, ¹H NMR, ¹³C NMR, COSY, HETCOR, HSQC and elemental analyses. X-ray diffraction studies were done on the complexes [(η⁵-C₅H₅)-(CO)₃ W(CH₂)₃Mo(CO)₂(PPh₃) (η⁵-C₅H₅)] and [(η⁵-C₅H₅)(CO)₂ (PPh₃)Mo(CH₂)₃Fe-(CO)₂ (η⁵-C₅H₅)]. Both compounds form monoclinic crystals in the space group P2₁/c. The former has a W-C(alkyl) bond length of 2.316(7) A? and Mo-C(alkyl) bond length of 2.357(7) A? and the latter a Mo-C(alkyl) bond length of 2.374(2) A? and Fe-C(alkyl) bond length of 2.076(2) A?.

Full text of DOI:10.1016/j.jorganchem.2003.10.002

Journal of Organometallic Chemistry 689 (2004) 181–193
www.elsevier.com/locate/jorganchem
Transition metal-substituted paraffins: synthesis and properties
of some l-saturated heterobimetallic complexes containing
Mo and W or Fe and the crystal structures
of [(g5-C5H5)(CO)3W(CH2)3Mo(CO)2(PPh3)(g5-C5H5)]
and [(g5-C5H5)(CO)2(PPh3)Mo(CH2)3Fe(CO)2(g5-C5H5)]
a,
b
a
a
Holger B. Friedrich *, R. Alan Howie , Michael Laing , Martin O. Onani
a
School of Pure and Applied Chemistry, University of Natal, Durban 4041, South Africa
b
Department of Chemistry, School of Engineering and Physical Sciences, University of Aberdeen, Meston Walk, Aberdeen AB24 3UE, Scotland, UK
Received 26 May 2003; accepted 2 October 2003
Abstract
The new heterobimetallic complexes [(g5-C5H5)(CO)3W(CH2)nMo(CO)3(g5-C5H5)] n ¼ 3 to 6; [(g5-C5H5)(CO)3 W(CH2)nMo
(CO)3(g5-C5(CH3)5)] n ¼ 3, 4; [(g5-C5H5)(CO)3W(CH2)nMo(CO)2(PPhiMe3 ꢀ i)(g5-C5H5)] and [(g5-C5H5)(CO)2 Fe(CH2)nMo
(CO)2(PPhiMe3 ꢀ i)(g5-C5H5)] (n ¼ 3; 4 and i ¼ 0 to 3) were synthesized by direct displacement of the iodide of a metallo-iodoalkyl
1
complex with the appropriate anion. The complexes have been fully characterised by IR, H NMR, 13C NMR, COSY, HETCOR,
HSQC and elemental analyses. X-ray diffraction studies were done on the complexes [(g5-C5H5)-(CO)3W(CH2)3Mo(CO)2
(PPh3)(g5-C5H5)] and [(g5-C5H5)(CO)2(PPh3)Mo(CH2)3Fe-(CO)2(g5-C5H5)]. Both compounds form monoclinic crystals in the
ꢀ
ꢀ
space group P21=c. The former has a W–C(alkyl) bond length of 2.316(7) A and Mo–C(alkyl) bond length of 2.357(7) A and the
ꢀ
latter a Mo–C(alkyl) bond length of 2.374(2) A and Fe–C(alkyl) bond length of 2.076(2) A.
ꢀ
Ó 2003 Elsevier B.V. All rights reserved.
Keywords: Paraffins; Heterobimetallic; Tungsten; Molybdenum; Iron; Cyclopentadienyl; Phosphines
1. Introduction
chemistry that differs appreciably from that displayed by
the single metal-containing counterparts. The possibility
of metals influencing one another at close proximity has
been acknowledged because different metals posses dif-
ferent properties [1]. The stability and electronic proper-
ties of heterodinuclear compounds may also be influenced
when additional ligands with p-backbonding properties
are incorporated. Such complexes, but with conjugated
bridges, could also act as molecular wires or switches [2,3].
Heterodinuclear compounds can also find application in
the area of organic syntheses [4]. Furthermore, the syn-
thesis and reactivity of heterodinuclear transition metal
complexes continues to attract considerable interest be-
cause of their envisaged catalytic superiority over their
homobimetallic analogues [4].
These transition metal-substituted paraffins are com-
pounds in which the paraffin chains bridge two different
transition metal centres. They have the general formula
[LxM{(CH2)n}M0Ly]; M ¼ M0; n P 1; LxM, M0Ly ¼
transition metal and its associated ligands. The nature of
heterodinuclear complexes, that they are mixed metallic,
containing two or more adjacent metal centres, could add
some other chemical dimension over those containing a
single metal; not only can the metals act independently
but they can also act in a cooperative manner leading to
* Corresponding author. Tel.: +27-31-260-3107; fax: +27-31-260-
3091.
These metal-substituted paraffin compounds could be
good catalyst precursors or model compounds for
E-mail addresses: friedric@nu.ac.za (H.B. Friedrich),
r.a.howie@abdn.ac.uk (R.A. Howie).
0022-328X/$ - see front matter Ó 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2003.10.002

182
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
catalytic intermediates or act as catalysts themselves.
They have been proposed to be good models for the
Fischer–Tropsch processes [5]. Bimetallic compounds
with direct metal–metal bonds are reported to be com-
monly synthesized by a metathesis of a halide ligand,
usually on a group 4 metal, with an organometallic
anion, such as [Cp(CO)2M]ꢀ (M ¼ Fe, Ru, Cp ¼ (g5-
C5H5)) [6–10] or [(CO)4M]ꢀ (M ¼ Co, Rh) [6,7,11–15].
The same procedure has been used to prepare hetero-
bimetallic lanthanide [11] and group 6 and 9 complexes
[12,13]. The most successful route to heterodinuclear
alkyl bridged compounds reported so far is via the use of
halogenoalkyl compounds, especially those with the io-
doalkyl group [2,16,17].
Whilst a number of homometallic complexes incor-
porating polymethylene bridges (also called alkanediyl
bridges or bridging alkyl chains) are known [17,18] and
most are well covered in a recent review [19], very few
heterobimetallic complexes of the type [(g5-C5R5)(CO)3
M(CH2)nM0(CO)x(L)(g5-C5R5)] (M ¼ M0, x > 1, n P 2)
have been reported [1,2,17,18] and to our knowledge
none with phosphine ligands. Only phosphine acyl
complexes resulting from migratory insertion reactions
have been reported [1,16,20,21]. We now report on new
metal-substituted paraffin compounds containing ter-
tiary phosphine ligands prepared from a range of pre-
viously reported halogenoalkyl compounds [22].
L ¼ PPh2Me
n ¼ 3 8a
n ¼ 4 8b
n ¼ 3 9a
n ¼ 4 9b
n ¼ 3 10a
n ¼ 4 10b
n ¼ 3 11a
n ¼ 4 11b
L ¼ PPhMe2
L ¼ PMe3
E [Cp(CO)2Fe(CH2)nMo(CO)3Cp*]
The compounds of type A were obtained in best
yields when the temperature of the reactions was
maintained at )78 °C over the total reaction period,
which ranged from 64 to 105 h. Compounds of class B
were prepared in an analogous manner to class A and
were also obtained in good yields. The reactions were
found to take a shorter time when carried out at room
temperature, however, the yields were low and this was
attributed to a high decomposition rate of the reactants.
The reactions using the phosphine-substituted anions to
yield class C and D compounds were slow but gave good
yields at room temperature. The reaction times generally
increased with an increase in chain length. This could be
attributed to decreased influence of the metal on the
halogen within the same molecule.
The complexes of class A, B and E were soluble
in hexane with their solubility increasing with in-
creasing length of the alkyl chain. The rest of the
complexes were insoluble in hexane due to the pres-
ence of the phosphine ligand, but soluble in dichlo-
romethane. All the complexes were obtained as yellow
crystalline solids, which were stable in air but unstable
in solution.
2. Results and discussions
The new heterodinuclear complexes were obtained in
medium to high yields by the direct displacement of the
iodide of a metallo-iodoalkyl complex with an appro-
priate anion, as shown in Scheme 1:
The melting points were generally sharp and de-
creased with increase in alkyl chain length. The notable
exceptions are compounds 4 and 10. The yields, spectral
data and melting points are given in Table 1.
In the IR spectra, the m(CO) wave numbers of the
complexes are in the expected range for terminal car-
bonyl groups [17]. Some of the W(CO) or Fe(CO) and
Mo(CO) peaks overlap, but the peaks between 1830 and
1850 cmꢀ1 can be clearly assigned to Mo(CO) [1,17].
Steric effects of the phosphines are difficult to separate
from electronic factors and hence are closely interrelated
[23,24]. Since PMe3 is the most basic phosphine within
the group with the lowest p-acceptor properties, its
complexes have the strongest M(CO) bonds and hence
the lowest frequency infrared bands in the m(CO) region.
A decreasing trend is thus expected in the carbonyl
stretching frequencies as phenyl groups are sequentially
substituted by methyl groups in the phosphine ligands
used. This trend is evident for the Mo(CO) bands be-
tween 1830 and 1850 cmꢀ1 for compounds of class C
and D where n ¼ 4, but not where n ¼ 3, presumably
due to the influence of the metal at the other end of the
alkyl chain.
A [Cp(CO)3W(CH2)nMo(CO)3Cp]
n ¼ 3 1a
n ¼ 4 1b
n ¼ 5 1c
n ¼ 6 1d
n ¼ 3 2a
n ¼ 4 2b
B [Cp(CO)3W(CH2)nMo(CO)3Cp*]
{Cp* ¼ C5(CH3)5}
C [Cp(CO)3W(CH2)nMo(CO)2LCp]
L ¼ PPh3
n ¼ 3 3a
n ¼ 4 3b
n ¼ 3 4a
n ¼ 4 4b
n ¼ 3 5a
n ¼ 4 5b
n ¼ 3 6a
n ¼ 4 6b
L ¼ PPh2Me
L ¼ PPhMe2
L ¼ PMe3
D [Cp(CO)2Fe(CH2)nMo(CO)2LCp]
L ¼ PPh3
n ¼ 3 7a
n ¼ 4 7b

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
183
Scheme 1. The general synthetic pathway to the heterobimetallic complexes.
The 1H NMR data for the complexes 1a–11b are
summarised in Table 2. Data from similar compounds
[17,22], as well as correlated spectroscopy (COSY), were
used for chemical shift assignments.
addition comparisons were made to the 13C NMR data
reported for related heterodinuclear complexes of Fe
and Mo, W, Re and Ru [17]. The metals on the opposite
end of the paraffin chain and the chain length were seen
to have no effect on the chemical shifts of the carbonyl
peaks. Similarly, neither the metals on the opposite end
of the paraffin chain, nor the chain length appear to
affect the positions of Cp* and Cp peaks significantly.
These peaks for the complexes 4a–10b, with phosphines
as ligands, appear at slightly lower field, compared to
the unsubstituted compounds (1a–1d).
The signals of carbon atoms in the carbonyl ligands
cis to the phosphine ligand were also observed as dou-
blets because of coupling with the phosphorus atom.
Again the stereochemistry was confirmed by the crystal
structures of 3a and 7a (Figs. 2 and 3).
From the 13C NMR data in Table 3, it can be seen
that the chemical shifts of the carbon a to a metal
generally decreases with an increase in the bridging
chain length. This suggests that an increase in the chain
length results in a reduced influence of the metal at one
end of the chain on the carbon a to the other metal at
the other end of the chain. This is not observed for the
Cp* compounds 2a and 2b, however. We find that the
MoCH2 carbon atoms in the MoCp* compounds 2 are
significantly more deshielded than those of the MoCp
compounds. The deshielding effect is weaker on the
MoCH2CH2 carbons.
A shielding effect is seen on the MoCH2 protons of
compounds 8a, 9a, 10a and 10b and they appear upfield
relative to uncoordinated aliphatic methylene protons.
Tungsten, being a more electron-rich transition metal,
shields its a-CH2 protons more and these peaks were
observed further upfield than those a to the Mo atom.
The influence of the metal diminishes as the chain gets
longer, and has no effect at all on the chemical shifts of
the protons of either Cp or Cp*. The observations here
support an earlier finding that the metal only influences
the protons a and b to it and not those c or beyond [17].
Because of these observations, compounds with longer
alkyl chain length (n ¼ 5 and above) were not synthe-
sized for compounds of class B–E.
All the Cp proton resonances of the compounds of
class C and D were observed to couple with the phos-
phorus of the phosphine ligands and hence the peaks for
the Cp ligands were all observed as doublets. In this type
of complex, it has been noted that the bulky phosphine
ligand always preferentially migrates away from the
approaching alkyl ligand giving a trans geometry with
an average P–Mo–C(alkyl) angle of 135° (see Fig. 1)
[22,25, this work]. We have also confirmed this obser-
vation from the crystal structures obtained for com-
pounds 3a and 7a as can be seen in Figs. 2 and 3.
The 13C NMR data for complexes 1a–11b are given in
Table 3. The peak assignments were made using heter-
onuclear correlation (HETCOR) and heteronuclear
single-quantum correlation (HSQC) experiments. In
The compounds 2a, 5b and 10a were analysed for the
presence of the elements Fe, Mo, W and combinations
of FeMo and MoW by liquid chromatography time-
of-flight mass spectrometry (LC-ToF-MS). Computer
simulations of the expected patterns were generated by

184
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
185

186
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
this type of analysis), for compound 10a, C20H25Fe-
MoO4P, the mass spectrum obtained showed the
isotope ratio pattern of the most intense cluster around
m=z 556, again 42 mass units higher than expected, also
implying a ‘‘pseudo-molecular ion’’ of the formula
[M + CH3CN + H]þ.
2.1. Crystal structures of compounds 3a and 7a
Fig. 1. The P–Mo–C angle.
Only structures of homodinuclear alkanediyl com-
pounds are reported in the literature and, to our
knowledge, none with two different metals present in the
same molecule [27–30]. Archer et al. reported only the
severely disordered [Cp(CO)2Fe(CH2)6Ru(CO)2Cp] [5].
We have recently reported the structure of the hetero-
bimetallic compound [Cp(CO)2Fe(CH2)3Ru(CO)2Cp]
which showed a partial (24%) disorder by inter-change
of Fe and Ru that renders the bonds to Fe and Ru
apparently slightly longer and shorter, respectively, than
they are in the corresponding homodinuclear analogues
[31]. The compounds [Cp(CO)3W(CH2)3Mo(CO)2
(PPh3)Cp] (3a) and [Cp(CO)2Fe(CH2)3Mo(CO)2 (PPh3)
Cp] (7a), discussed below are therefore, to our knowl-
edge, the first ordered crystal structures of heterodinu-
clear alkanediyl compounds in general, and the first for
Mo and Fe or W.
Fig. 2. The molecule of 3a, showing the atom-labelling scheme. Non-H
atoms are shown as 50% probability displacement ellipsoids and H
atoms as spheres of arbitrary radius.
The molecules of 3a and 7a are shown in Figs. 2 and
3, respectively. Selected bond lengths and angles and
details of data collection and structure refinement are
given, in each case for both compounds together, in
Tables 4 and 5, respectively. In Table 4, the contribution
of the pentahapto cyclopentadienyl ligand to the coor-
dination of the metal atoms is expressed, for conve-
nience and brevity, in terms of a single bond to its
centroid (Cg). The bond lengths and angles of the tri-
phenylphosphine, cyclopentadienyl and carbonyl li-
gands are in no way unusual, e.g., carbonyl C–O
distances and M–C–O angles in the respective ranges
Fig. 3. The molecule of 7a, showing the atom-labelling scheme. Non-H
atoms are shown as 50% probability displacement ellipsoids and H
atoms as spheres of arbitrary radius.
ꢀ
1.145(3)–1.154(3) A and 175.2(5)–178.4(6)°, and are not
discussed in detail here but are available in the cif sup-
plementary data deposited with CCDC. Instead the
discussion here concentrates upon the molybdenum,
tungsten and iron moieties and, because there is no
structure similar to that of 3a or 7a in the literature,
compares them with some reported but not wholly
analogous structures [22].
using (NIST) ‘‘Isoform’’ computer software [26]. The
spectrum of compound 2a with the molecular formula
C24H26MoO6W, showed large ions present at m=z
715.0150 corresponding to [M + Na]þ, a ‘‘pseudo-mo-
lecular ion’’ that is usually generated by electrospray
ionization. The predicted isotope distribution pattern
resembled the experimental finding; for compound 5b,
both the isotope abundance pattern and the accurate
mass (measured 788.0599, calculated 788.0660, differ-
ence 7.7 ppm) corresponded to the ‘‘pseudo-molecular
ion’’ [M + CH3CN + H]þ. The cluster ion incorporated a
molecule of the acetonitrile used as solvent hence the
mass occurred at 42 mass units higher than expected (it
should be noted that this phenomenon is not unusual in
In compounds 3a and 7a, as expected, the PPh3 ligand
on the molybdenum atom is trans, and the carbonyl
groups and the cyclopentadienyl ligands are cis, to the
alkyl chain. Conversely the Cp and the CO ligands on the
molybdenum centre are cis to the PPh3. This confirms
the 13C NMR interpretations that showed doublets for
the two carbonyl ligands as a result of coupling with the
phosphorus atom. The ligands are disposed in a piano
stool fashion, which can also be regarded as being in the
form of a square pyramid with the cyclopentadienyl li-

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
187

188
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
189
Table 5
Crystal data and structure refinement for 3a and 7a
3a: M ¼ Mo and W
C36H31MoO5PW
7a: M ¼ Mo and Fe
C35H31FeMoO4P
Empirical formula
Molecular weight
Temperature (K)
854.37
219(2)
698.36
228(2)
ꢀ
Wavelength (A)
0.71073
Monoclinic
P21=c
0.71073
Monoclinic
P21=c
Crystal system
Space group
Unit cell dimension
ꢀ
a (A)
15.446(7)
12.204(6)
15.414(7)
11.243(4)
ꢀ
b (A)
ꢀ
c (A)
18.483(9)
113.27(4)
19.007(6)
112.42(3)
b (°)
Z
4
1.773
4
1.523
Dcalc (mg mꢀ3
)
Absorption coefficient (mmꢀ1
F(000)
)
4.075
1672
Yellow
0.979
1424
Yellow
Crystal colour
Crystal size (mm)
0.45 ꢁ 0.35 ꢁ 0.25
2.06–23.01
ꢀ15 6 h 6 16
ꢀ13 6 k 6 13
ꢀ20 6 l 6 0
8643/4452 [0.0448]
99.5
Refine D(F2) (DIFABS)
0.4290 and 0.2614
Full-matrix least-squares on F2
4452/0/397
1.038
0.45 ꢁ 0.40 ꢁ 0.30
2.15–23.00
ꢀ15 6 h 6 16
ꢀ12 6 k 6 12
ꢀ20 6 l 6 0
8177/4230 [0.0187]
99.6
Refine D(F2Þ (DIFABS)
0.7578 and 0.6671
Full-matrix least-squares on F2
4230/0/379
1.023
h Range for data collection (°)
Index range
Reflections collected/unique [Rint
Completeness to h limit (%)
Absorption correction
Max. and min. transmission
Refinement method
]
Data/restraints/parameters
Goodness-of-fit on F2
Final R indices [I > 2rðIÞ]R1
wR2
0.0360
0.0989
0.0214
0.0545
R indices (all data) R1
wR2
ꢀ
Largest diff. peak and hole (e A
0.0421
0.1033
0.0296
0.0570
ꢀ3
)
1.403 and )1.821
0.327 and )0.297
gand at the apex, similar to some reported complexes
with analogous features [32]. Very few molybdenum–al-
kyl structures are known and to date those with the
Cp(CO)2(PR3) ligand pattern have only recently been
reported by us [22]. In 3a and 7a, the Mo(1)–Calkyl bond
The situation of the iron atom in 7a is rather different
from that described for molybdenum and tungsten. The
ligands are now disposed in the form of a trigonal pyr-
amid, or three legged stool, with the cyclopentadienyl
ligand at the apex. A direct consequence of this is that
the Cp on the iron centre is roughly perpendicular to the
alkyl chain and at an angle of 74.27(17)° to the Cp as-
sociated with molybdenum whereas in 3a the CpÕs as-
sociated with molybdenum and tungsten are both on the
same side of and roughly parallel to the alkyl chain and
make an angle of only 25.0(5)° to one another. The
projection of the Cp(CO)2Fe-unit of the compound,
[Cp(CO)2Fe(CH2)3Fe(CO)2Cp], though homobimetal-
lic, shows, also in contrast to 7a, the cyclopentadienyl
ring near parallel to the alkyl chain [28]. This suggests
that this is the energetically favoured extended staggered
conformation for this molecule, a phenomenon also
observed for the compound [Cp*(CO)2Fe(C5H11)] [41].
The Fe(1)–C(5) bond length to the alkyl chain in 7a is
ꢀ
lengths are 2.357(7) and 2.374(2) A, within the general
ꢀ
range (2.26–2.38 A) of the few reported molybdenum–
ꢀ
alkyl structures [22,28] including 2.402(7) A reported for
[Cp(CO)3Mo{(CH2)3I}]. The Mo(1)–P(1) bond lengths
ꢀ
[2.4569(17) and 2.4546(10) A] are likewise in the normal
ꢀ
range (2.41–2.51 A) [33–37]. The Mo–CCp distances
ꢀ
range from 2.322(7) to 2.370(7) A.
In compound 3a, the W–CCp distances range from
ꢀ
2.315(7) to 2.349(7) A. The carbonyl ligands have W–C–
O angles ranging from 176.1(8) to 179.0(7). The C–O
ꢀ
bond lengths of between 1.146(10) and 1.16(11) A, and
ꢀ
the mean W–C(CO) bond length of 1.97 A are normal
for terminal carbonyl groups. The W–C(sp3) bond dis-
ꢀ
tance of 2.316(7) A is significantly longer than that of
ꢀ
the recently reported monometallic halogenoalkyl
ꢀ
2.076(2) A, in the general range of reported Fe–C bonds
ꢀ
(2.05–2.08 A) [25,28,42–46] and the Fe–CCp distances
compound [Cp(CO)3W{(CH2)5I}] 2.08(3) A [22] but is
in the range of a methyl group bound to a d4-tungsten
center [38–40].
range from 2.088(3) to 2.102(3) A. The compound fol-
lows the typical Ôbump in hollowÕ packing, as is often
ꢀ

190
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
Fig. 4. The unit cell of 7a viewed down the crystallographic a-axis (100 projection). Non-H atoms are shown as 50% probability displacement
ellipsoids and H atoms have been omitted for clarity.
observed in the packing of paraffinic molecules (Fig. 4)
[39]. The C–C bond lengths in the alkyl chains are within
the general range observed in the similar compounds
LECO Corporation Separation Science Laboratory in
the United States. All reactions were followed by IR, by
monitoring the disappearance of the m(CO) peaks of the
anions.
ꢀ
referred to above (1.48–1.55 A).
We are presently investigating the chemistry of these
compounds and hope to publish these results in the near
future.
3.1. Preparation of [Cp(CO)3W(CH2)3Mo(CO)3Cp]
A solution of the salt Na[Cp(CO)3Mo] was pre-
pared by the reduction of [Cp(CO)3Mo]2 (0.88 g 1.80
mmol), by an amalgam made from Na (0.25 g, 10.8
mmol) and mercury (6 ml), in tetrahydrofuran (25 ml).
The resulting solution was added dropwise over 30 min
to a stirred solution of [Cp(CO)3W{(CH2)3I}] (1.80 g,
3.6 mmol) dissolved in THF (6 ml) and maintained at
)78 °C. The solution obtained was maintained be-
tween )65 and )78 °C. After 64 h, no infrared bands
due to the anion [Cp(CO)3Mo]ꢀ were observed. The
mixture was, however, stirred for a further 12 h to
ensure complete reaction, after which it was filtered
through a cannula under nitrogen. The solvent was
removed under reduced pressure and the residue dis-
solved in a minimum amount of dichloromethane and
transferred to a short chromatography column made
up of alumina with hexane and maintained under ni-
trogen. Elution with hexane gave a light yellow band
that was found to contain only the starting material.
The second, intense, yellow band eluted was concen-
trated and recrystallised at )78 °C. The desired
product separated from the solution. The mother li-
quor was syringed off and the product dried under
reduced pressure.
3. Experimental
All reactions were carried out under inert conditions
using standard Schlenk tube techniques. The dimers
[Cp(CO)3Mo]2, [Cp*(CO)3Mo]2, [Cp(CO)2Mo(PPhi
Me3 ꢀ i)]2 and the halogenoalkyl compounds, [Cp(CO)3
W{(CH2)3I}], were made according to the literature [22].
The dicyclopentadiene was always distilled prior to use.
Tetrahydrofuran was distilled over sodium-benzophe-
none and stored over sodium wire. Dichloromethane
and chloroform were purified according to the literature
ꢀ
procedure and stored over molecular sieves 3 A under
ꢀ
nitrogen [46]. The molecular sieves 3 and 4 A were dried
in a tube furnace at 250 °C for 10 h and alumina was
deactivated with deionised water and dried in an oven at
110 °C before use. Melting points were recorded on a
Kofler hot-stage microscope and are uncorrected. The
South African Bureau of Standards (SABS) in Richards
Bay, South Africa performed the elemental analyses.
Infrared spectra were recorded on a Nicolet Impact
400D 5DX FT-spectrophotometer either in solution or
KBr disc. The NMR spectra were recorded on Varian
Gemini 300 MHz and Varian Inova 400 spectrometers.
X-ray diffraction studies were carried out using an En-
raf-Nonius CAD4 diffractometer in the Chemistry De-
partment of our Pietermaritzberg campus. LC-ToF-MS
data were obtained on a Jaguar LC-TOF-MS at the
It is also possible to avoid chromatography by
four successive recrystallizations of the filtered prod-
uct from a minimum of dichloromethane/hexane at
)78 °C to give an analytically pure product. Elemental

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
191
Anal.: Found (Calc.): 1a C 35.98 (36.69) H 2.77
(2.59)%.
by reducing the respective dimers, [Cp(CO)2(PPhi
Me3 ꢀ i)Mo]2 (0.242 mmol) with a Na (0.25 g, 10.87
mmol)/Hg (6 ml) amalgam. The solution was then added
3.2. Preparation of [Cp(CO)3W(CH2)4Mo(CO)3Cp]
dropwise over 30 min to a stirred solution of
[Cp(CO)3W{(CH2)nI}] (0.484 mmol, n ¼ 3, 4) in THF (5
ml) maintained at )78 °C. The resulting solution was
stirred for a further 15 min at this temperature, and then
allowed to attain room temperature before stirring for a
further 3 days. The solvent was removed under reduced
pressure, leaving a greenish/yellow residue, which was
extracted with dichloromethane (3 ꢁ 10 ml) and filtered
via a cannula under nitrogen. The filtrate was reduced,
under reduced pressure, to approximately a third of the
total volume of the dichloromethane used. The yellow
product was precipitated with hexane, and the mother
liquor syringed off. The product was washed three times
with hexane and dried under reduced pressure. Ele-
mental Anal.: Found (Calc.): 6b C 38.45 (38.63) H 3.89
(3.97)%.
An amalgam of Na/Hg, made as described above, was
used to reduce the compound [Cp(CO)3Mo]2 (0.78 g, 1.59
mmol) to the anion [Cp(CO)3Mo]ꢀ (3.2 mmol) in THF
(30 ml). This solution of Na[Cp(CO)3Mo] was added
dropwise over 30 min into a stirred solution of
[Cp(CO)3W{(CH2)4I}] (0.88 g, 1.70 mmol, 10% less than
quantity of the salt) in THF (6 ml) maintained at )78 °C.
The reaction was maintained at this temperature for 84 h.
The solution was then filtered through a cannula under
nitrogen and the solvent removed under reduced pressure.
The residue was dissolved in a minimum of dichlorome-
thane. An equal volume of hexane was added and the
resulting mixture cooled to )78 °C. The product precip-
itated out, and the mother liquor was syringed off. The
product was recrystallised four times from dry hexane (80
ml in total) at )78 °C. Elemental Anal.: Found (Calc.): 1b
C 36.84 (37.77), H 2.87 (2.85)%.
3.6. Preparation of [Cp(CO)2(PPhiMe3 ꢀ i)Mo(CH2)n
Fe(CO)2Cp] i ¼ 0 to 3, n ¼ 3, 4
3.3. Preparation of [Cp(CO)3W(CH2)nMo(CO)3Cp]
n ¼ 5, 6
The dimers [Cp(CO)2Mo(PPhiMe3 ꢀ i)]2 i ¼ 0 to 3
(0.28 mmol) were reduced to their respective anions
[Cp(CO)2Mo(PPhiMe3 ꢀ i)]ꢀ i ¼ 0 to 3 (0.56 mmol) by
Na (10.87 mmol)/Hg (6 ml) in THF (25 ml). The
solution of the anion was added dropwise over 30 min
to a stirred solution of [Cp(CO)2Fe{(CH2)nI}] n ¼ 3, 4
(0.555 mmol) dissolved in THF (5 ml) at )78 °C. The
solution was stirred for 15 min at this temperature,
then allowed to attain room temperature and stirred
for a further 5 days. The solvent was removed under
reduced pressure leaving a yellow/black residue, which
was extracted with dichloromethane (3 ꢁ 10 ml) and
filtered via a cannula under nitrogen. The filtrate was
concentrated under reduced pressure to approximately
a third of the total volume of dichloromethane used
originally. This solution was transferred to a short
alumina column made up with hexane and the prod-
uct was eluted with 90% dichloromethane/hexane. The
resulting solution was concentrated and cooled to )78
°C under nitrogen for 30 min. The yellow product
separated from the solution. The mother liquor was
syringed off and the product dried under reduced
pressure.
An alternative workup involved adding degassed
alumina (ꢂ5g) to the filtrate after the extraction with
dichloromethane. The solution was again filtered
through a cannula into another Schlenk tube. This so-
lution was then reduced to a third of the total volume
and the product precipitated with hexane. This method
gave better yield and cleaner products. Elemental Anal.:
Found (Calc.): 7b C 60.02 (60.02), H 4.96 (4.49); 9b C
52.77 (52.92), H 5.08 (4.95); 10b C 47.90 (47.93), H 5.23
(5.17)%.
The same reaction conditions and work-up proce-
dures were used as for 1b above, except that the reac-
tants, once mixed, were kept at )78 °C for 24 h and then
allowed to attain room temperature and stirred for a
further 5 days.
3.4. Preparation of [Cp(CO)3W(CH2)nMo(CO)3Cp*]
n ¼ 3, 4
The salt Na[Cp*Mo(CO)3] was prepared by reduc-
ing [Cp*Mo(CO)3]2 (0.152 g, 0.242 mmol) with Na
(0.25 g, 10.8 mmol)/Hg (6 ml) in THF (25 ml). The
solution was added dropwise over a period of 30 min
to a stirred solution of [Cp(CO)3W{(CH2)3I}] (0.242 g,
0.484 mmol) dissolved in THF (5 ml) and maintained
at )78 °C. The resulting mixture was kept at this
temperature for a further 15 min. The solution was
then allowed to attain room temperature and stirred
for 64 h. It was then filtered through a cannula under
nitrogen and the solvent removed under reduced pres-
sure. The product was recrystallised from dilute (10/90)
dichloromethane/hexane at )78 °C. The mother liquor
was syringed off and the product dried under reduced
pressure.
3.5. Preparation of [Cp(CO)2(PPhiMe3 ꢀ i)Mo(CH2)n
W(CO)3Cp] i ¼ 0 to 3, n ¼ 3, 4
Solutions of the salts Na[Cp(CO)2(PPhiMe3 ꢀ i)Mo]
i ¼ 0 to 3 (0.484 mmol) in THF (25 ml), were prepared

192
H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
3.7. Crystallography
nos. 203294 for [Cp(CO)3W(CH2)3Mo(CO)2(PPh3)Cp]
(3a) and 203295 for [Cp(CO)2Fe(CH2)3Mo(CO)2
(PPh3)Cp] (7a). Copies of this information may be ob-
tained free of charge from The Director, CCDC, 12
Union Road, Cambridge, CB2 1EZ, UK (Tel.+44 1223
336408; Fax +44 1223 336033; E-mail: deposit@ccdc.
cam.ac.uk, linstead@ccdc.cam.ac.uk, or http://www.
ccdc.cam.ac.uk). Full tables of atomic positional pa-
rameters, thermal parameters, interatomic distances and
angles, and torsional angles for 3a and 7a are available
from the authors.
3.7.1. Growth of crystals
Slightly more than one equivalent volume of hexane
was added to a concentrated solution of compound 3a
or 7a in dichloromethane, in a vial. This mixture was left
to stand in a refrigerator at )10 °C. The crystals grew by
slow solvent diffusion after several days [47].
3.7.2. Data collection
X-ray data for compounds 3a and 7a were collected on
a CAD-4 diffractometer. In both cases the unit cell was
determined from 25 reflections (h ¼ 12°) and found to be
consistent with a monoclinic lattice for which a complete
set of data was collected by the x–2h scan method for
h ¼ 2° to 23°. The data were reduced with application of
Lorentz and polarization correction factors with the
program XCAD (Oscail V8) [48]. Correction for ab-
sorption was applied using the program DIFABS (Oscail
V8) [49]. Further details are given in Table 5.
Acknowledgements
We thank the University of Natal (URF), the Na-
tional Research Foundation (NRF) and the Deutscher
Akademischer Austausch Dienst (DAAD), for financial
support. M.O.O. thanks the DAAD for a Ph.D. schol-
arship and the Jomo Kenyatta University of Agriculture
and Technology (Nairobi) for study leave. We also
thank Dr. P. Gorst-Allman (LECO) for running the
TOF-MS samples, Dr. O. Munro, Chemistry Depart-
ment, the University Natal Pietermaritzburg, for col-
lecting the intensity data and Mr. James Ryan for
mounting and taking preliminary X-ray photographs of
several crystal specimens relevant to this work.
3.7.3. Structure solution and refinement
Both structures were solved in the monoclinic space
group P21=c by direct methods using, for 3a the pro-
gram DIRDIF [50], as implemented by the crystallo-
graphic program WinGX [51], and for 7a SHELXS-97
[52], as implemented by the crystallographic program
Oscail [49]. In both cases E-maps led to the location of
all non-hydrogen atoms which were then refined aniso-
tropically with the program SHELXL-97 [52]. In the
final stages of refinement all hydrogen atoms were in-
cluded as idealized contributors in the least-squares
process with standard SHELXL-97 idealization pa-
rameters. There was no evidence (difference Fourier
map) for the inclusion of solvent in the lattice in either
case. Further details of the refinement process are
available in Table 5. Structural drawings were prepared
using the program ORTEP [51]. PLATON was used for
the geometry calculations providing bond lengths and
angles in terms of the centroids of the Cp rings [53].
References
[1] H.B. Friedrich, J.R. Moss, J. Chem. Soc., Dalton Trans. (1993)
2863.
[2] J.-M. Lehn, Angew. Chem. Int. Ed. Engl. 27 (1988) 89.
[3] P. Sautet, O. Eisenstien, K.M. Nicholas, Organometallics 6 (1987)
1845.
[4] R.D. Adams, Polyhedron 7 (1988) 2251, and refs. therein.
[5] S.J. Archer, K.P. Finch, H.B. Friedrich, J.R. Moss, Inorg. Chim.
Acta 182 (1991) 145, and references therein.
[6] B. Findeis, M. Schubart, C. Platzek, L.H. Gade, I. Scowen, M.
McPartlin, J. Chem. Soc., Chem. Commun. (1996) 219.
[7] S. Friedrich, H. Memmer, L.H. Gade II, W.S. Li, I.J. Scowen, M.
McPartlin, C.E. Housecroft, Inorg. Chem. 35 (1996) 2433.
[8] D. Selent, M. Ram, C. Janiak, J. Organomet. Chem. 501 (1995)
235.
3.7.4. Analysis of complexes 2a, 5b and 10a by liquid
chromatography time-of-flight mass spectrometer
The samples were analysed by direct infusion liquid
chromatography (LC) coupled with the LECO Jaguarâ
Time-of-Flight Mass Spectrometer (TOF-MS). Ap-
proximately 1.7 mg of each complex was dissolved in 2
ml acetonitrile. Of the 2 ml solution, 1 ml was diluted to
10 ml with acetonitrile and injected into the instrument
with a flow rate of 1.5 ll/min.
[9] J. Sundermeyer, D. Runge, J.S. Field, Angew. Chem. Int. Ed.
Engl. 33 (1994) 678.
[10] J. Sundermeyer, D. Runge, Angew. Chem. Int. Ed. Engl. 33 (1994)
1255.
[11] I.P. Beletskaya, A.Z. Voskoboynikov, E.B. Chuklanova, N.I.
Kirrilova, A.K. Shestakova, A.I. Parshina, A.I. Gusev, G.K-I.
Magomedov, J. Am. Chem. Soc. 115 (1993) 3156.
[12] F.E. Hong, I.-R. Lue, S.-C. Lo, Y.-C. Yang, C.-C. Lin, Polyhe-
dron 15 (1996) 2793.
[13] F.E. Hong, I.-R. Lue, S.-C. Lo, C.-C. Lin, Polyhedron 14 (1995)
1419.
4. Supplementary material
[14] A.M. Martins, M.J. Carlhoda, C.C. Romao, C. Vokel, P. Kiprof,
A.C. Filippou, J. Organomet. Chem. 423 (1992) 367.
[15] Y.-Y. Liu, A. Mar, S.J. Rettig, A. Storr, J. Trotter, Can. J. Chem.
66 (1988) 1997.
Cif format X-ray structural data are deposited with
the Cambridge Crystallographic Data Centre, CCDC

H.B. Friedrich et al. / Journal of Organometallic Chemistry 689 (2004) 181–193
193
[16] M. Cooke, N.J. Forrow, S.A.R. Knox, J. Chem. Soc., Dalton
Trans. (1983) 2435.
[38] H.G. Alt, J.S. Han, R.D. Rogers, J. Organomet. Chem. 445 (1993)
115.
[17] H.B. Friedrich, J.R. Moss, B.K. Williamson, J. Organomet.
Chem. 394 (1990) 313.
[39] R. Poli, G. Wilkinson, M. Motevalli, M.B. Hursthouse, J. Chem.
Soc., Dalton Trans. (1985) 931.
[18] W. Beck, B. Niemer, M. Wieser, Angew. Chem. Int. Ed. Engl. 32
(1993) 923.
[40] H.G. Alt, S.J. Palackal, R.D. Rogers, J. Organomet. Chem. 388
(1990) 105.
[19] N. Wheatley, P. Kalck, Chem. Rev. 99 (1999) 3419.
[20] J.R. Moss, J. Organomet. Chem. 231 (1982) 229.
[21] J.R. Moss, L.G. Scott, J. Organomet. Chem. 363 (1989) 351.
[22] H.B. Friedrich, M.O. Onani, O.Q. Munro, J. Organomet. Chem.
633 (2001) 39.
[41] R.O. Hill, C.F. Marais, J.R. Moss, K.J. Naidoo, J. Organomet.
Chem. 587 (1999) 28, and refs therein.
[42] (a) O.E. Woisetchlager, K. Sunkel, W. Weigand, W. Beck,
J. Organomet. Chem. 584 (1999) 122;
(b) O.E. Woisetchlager, A. Geisbauer, K. Polborn, W. Beck,
J. Organomet. Chem. 599 (2000) 238.
[23] C.A. Tolman, Chem. Rev. 77 (1977) 313.
[24] A.L. Ferandez, C. Reyes, A. Prock, W.P. Giering, J. Chem. Soc.,
Perkin Trans. 2 (2002) 1033.
[43] G.E. Jackson, J.R. Moss, L.G. Scott, S. Afr. J. Chem. 36 (1983) 69.
[44] D.J. Darensbourg, R. Kudaroski, J. Am. Chem. Soc. 106 (1984)
3672.
[25] K.W. Barnet, T.G. Pollman, T.W. Solomon, J. Organomet.
Chem. 36 (1972) C23.
[45] A.I. Kitaigorodskii, Organic Chemical Crystallography, Consul-
tants Bureau, New York, 1961, p. 65.
[26] Program created by the American National Institute of Standards
and Technology (NIST).
[46] B.S. Furniss, A.J. Hannaford, P.W.G. Smith, A.R. Tatchell,
VogelÕs Textbook of Practical Organic Chemistry, fifth ed.,
Longman, New York, 1978, p. 401.
[27] L. Pope, P. Sommerville, M.J. Laing, K.J. Hindson, J.R. Moss,
J. Organomet. Chem. 112 (1976) 309.
[28] K. Raab, U. Nigel, W. Beck, Z. Naturforsch. 38 (1983)
1466.
[47] P.G. Jones, Chem. Br. 17 (1981) 222.
[48] OSCAIL Version 8: P. McArdle, Crystallography Centre, Chem-
istry Department, NUI Galway, Ireland. P. McArdle, J. Appl.
Crystallogr. 28 (1995) 65.
[29] K.P. Finch, J.R. Moss, M.L. Niven, Inorg. Chim. Acta 166 (1989)
181.
[30] H. Adams, N.A. Bailey, M.J. Winter, J. Chem. Soc., Daltons
Trans. (1984) 273.
[49] DIFABS: N. Walker, D. Stuart, Acta Crystallogr. A, 39 (1983)
158.
[31] H.B. Friedrich, R.A. Howie, M.O. Onani, Acta Crystallogr., E 59
(2003) m145.
[50] DIRDIF96: P.T. Beurskens, G. Beurskens, W.P. Bosman, R. de
Gelder, S. Garcia-Granda, R.O. Gould, R. Israel, J.M.M. Smits,
Crystallography Laboratory, University of Nijmegen, The Neth-
erlands, 1996.
[32] F. Amor, P. Royo, T.P. Spaniol, J. Okuda, J. Organomet. Chem.
604 (2001) 126, and refs. therein.
[33] R.C. Melvyn, Inorg. Chem. 7 (1968) 953.
[34] R.C. Melvyn, W.-Y. Chang, Inorg. Chem. 14 (1975) 98.
[35] T. Butters, Acta Crystallogr. C 41 (1985) 23.
[36] H. Adams, N.A. Bailey, P. Blenkiron, M.J. Morris, J. Chem. Soc.,
Daltons Trans. (1997) 3589.
[51] WinGX: A Windows Program for Crystal Structure Analysis. L.J.
Farrugia, University of Glasgow, Glasgow, 1998.
[52] (a) G.M. Sheldrick, SHELXL-97, University of Gottingen;
(b) G.M. Sheldrick, Acta Crystallogr. A 46 (1990) 467;
(c) G.M. Sheldrick, Acta Crystallogr. D 49 (1993) 18;
(d) G.M. Sheldrick, T.R. Schneider, Methods Enzymol. 277 (1997)
319.
[37] G.A. Carmela, F. Faraone, M. Fochi, M. Lanfranchi, C. Mealli,
R. Seeber, A. Tiripicchio, J. Chem. Soc., Daltons Trans. (1992)
1847.
[53] A.L. Spek, J. Appl. Crystallogr. 36 (2003) 7.