Synthesis, structures and rac/meso isomerization behaviour of bisplanar chiral bis(phosphino-η5-indenyl)iron(II) complexes

Article abstract of DOI:10.1016/j.jorganchem.2004.03.014

Syntheses of the phosphinoindenes 1-(diphenylphosphino)-3-methylindene (1b), 3-(diphenylphosphino)-2-methylindene (1c), 1-(diphenylphosphino)-2,3-dimethylindene (1d), 4,7-dimethyl-3-(diphenylphosphino)indene (1e), 1-(diphenylphosphino)-3,4,7-trimethylindene (1f) and 3-(diisopropylphosphino)indene (1i) were carried out by treatment of the appropriate indenide with the appropriate chlorophosphine. The silylphosphinoindene 3-(diphenylphosphino)-1-(trimethylsilyl)indene (1h) was prepared by treatment of the indenide of 3-(diphenylphosphino)indene (1a) with trimethylsilylchloride. These indenes, in addition to 1a, were then used, after deprotonation with BuLi, to prepare the corresponding indenyl ferrocenes, 2a-2e, 2h and 2i, by treatment with ferrous chloride in a 2:1 ratio. These compounds were characterized by ¹H, ¹³C, and ³¹P NMR spectroscopy, as well as by mass spectrometry, except for the highly-sensitive diisopropylphosphine 2i that could only be characterized by ³¹P NMR spectroscopy. All of these ferrocene complexes are bisplanar chiral systems that can potentially form rac and meso isomers. In all cases both isomers were observed but for 2b and 2h only one could be isolated. The rac isomers of complexes 2a, 2b, 2d, and 2e, as well as the meso isomer of 2e, were studied by X-ray crystallography. Only complexes 2a and 2i were observed to undergo rac/meso isomerization processes at ambient temperature in THF solvent. We were unable to prepare the sterically congested hexamethylferrocene 2f. Generally, it was found that increasing substitution on the indenyl ring increases the reactivity and sensitivity of the ferrocene.

Full text of DOI:10.1016/j.jorganchem.2004.03.014

Journal of Organometallic Chemistry 689 (2004) 1897–1910
www.elsevier.com/locate/jorganchem
Synthesis, structures and rac/meso isomerization behaviour of
bisplanar chiral bis(phosphino-g5-indenyl)iron(II) complexes
*
Owen J. Curnow , Glen M. Fern, Michelle L. Hamilton, Elizabeth M. Jenkins
Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch, New Zealand
Received 27 January 2004; accepted 15 March 2004
Abstract
Syntheses of the phosphinoindenes 1-(diphenylphosphino)-3-methylindene (1b), 3-(diphenylphosphino)-2-methylindene (1c),
1-(diphenylphosphino)-2,3-dimethylindene (1d), 4,7-dimethyl-3-(diphenylphosphino)indene (1e), 1-(diphenylphosphino)-3,4,7-
trimethylindene (1f) and 3-(diisopropylphosphino)indene (1i) were carried out by treatment of the appropriate indenide with the
appropriate chlorophosphine. The silylphosphinoindene 3-(diphenylphosphino)-1-(trimethylsilyl)indene (1h) was prepared by
treatment of the indenide of 3-(diphenylphosphino)indene (1a) with trimethylsilylchloride. These indenes, in addition to 1a, were
then used, after deprotonation with BuLi, to prepare the corresponding indenyl ferrocenes, 2a–2e, 2h and 2i, by treatment with
1
ferrous chloride in a 2:1 ratio. These compounds were characterized by H, 13C, and 31P NMR spectroscopy, as well as by mass
spectrometry, except for the highly-sensitive diisopropylphosphine 2i that could only be characterized by 31P NMR spectroscopy.
All of these ferrocene complexes are bisplanar chiral systems that can potentially form rac and meso isomers. In all cases both
isomers were observed but for 2b and 2h only one could be isolated. The rac isomers of complexes 2a, 2b, 2d, and 2e, as well as the
meso isomer of 2e, were studied by X-ray crystallography. Only complexes 2a and 2i were observed to undergo rac/meso isomer-
ization processes at ambient temperature in THF solvent. We were unable to prepare the sterically congested hexamethylferrocene
2f. Generally, it was found that increasing substitution on the indenyl ring increases the reactivity and sensitivity of the ferrocene.
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Phosphine; Ferrocene; X-ray structure; Planar chiral; Synthesis; Indenyl
1. Introduction
of its rac and meso isomers by X-ray crystallographic
studies of their tetracarbonylmolybdenum complexes
[3]. Non-hydrocarbon-functionalized diindenyl ferroc-
enes are very sparse, especially in comparison to the
cyclopentadienyl system: there are only a few silyl [4],
boryl [5] and amino [6] complexes known in addition
to ansa-carboryl [7], ansa-dimethylsilyl [8] and ansa-
ferrocenyl [9] complexes. A patent claiming 7-meth-
oxy-1-(diphenylphosphino), 7-(ethoxycarbonyl)-1-(diph-
enylphosphino), 7-phenyl-1-(diphenylphosphino), and
1-(dicyclohexylphosphino) diindenyliron complexes has
also appeared [2]. The number of non-ansa diindenyl
iron complexes that have been crystallographically
characterised to date is remarkably small [4,10–13].
Further studies on [(1-PPh2-g5-C9H6)2Fe] have shown
that this indenyl ferrocene undergoes a facile isomeri-
zation in THF from the meso isomer to the rac isomer
at ambient temperatures [14,15]. Deuterium labeling
studies, crossover experiments, activation parameter
Ferrocenylphosphines continue to be intensively in-
vestigated for their utility in homogeneous catalysis;
chiral derivatives are of particular interest for asym-
metric catalysis [1,2]. The introduction of a chiral sub-
stituent to a ferrocene core or the use of heterotopic,
planar chiral ligands is usually used to create the chi-
rality. Despite the large number of planar-chiral ferr-
ocenyl phosphines that have been reported and used in
asymmetric catalysis, no racemization has been ob-
served in any of these systems. Compounds containing
two planar chiral units may exhibit rac and meso iso-
mers. We recently reported the preparation of the diin-
denyl analogue of 1,10-bis(diphenylphosphino)ferrocene
(dppf), [(1-PPh2-g5-C9H6)2Fe], and the characterization
* Corresponding author. Tel.: +64-3-364-2819; fax: +64-3-364-2110.
E-mail address: owen.curnow@canterbury.ac.nz (O.J. Curnow).
0022-328X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.03.014

1898
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
measurements (including the activation volume), salt
and solvent effects all point to a mechanism involving
THF coordination, indenyl ring-slippage, and indenide
dechelation with phosphine coordination to give a
zwitterionic intermediate. To further understand the
factors influencing this unprecedented ferrocene isom-
erization process, we sought to prepare a number of
derivatives to look at both their isomerization behaviour
and their structural features. We report here the prep-
aration, isolation and structure of a number of methyl
and trimethylsilyl derivatives. Some of this work has
previously been communicated [14].
in 3-indenyldiphenylphosphine (1a) [16], bis(3-inde-
nyl)phenylphosphine [18], tris(3-indenyl)phosphine and
their derivatives [19], and is the case here for 2-methyl-3-
(diphenylphosphino)indene (1c), 4,7-dimethyl-3-(diph-
enylphosphino)indene (1e) and 3-(diisopropylphosph-
ino)indene (1i). However, in the presence of a competing
methyl group, the methyl group has a stronger prefer-
ence for the vinylic position than the phosphine group,
as exhibited by compounds 1-(diphenylphosphino)-3-
methylindene (1b), 1-(diphenylphosphino)-2,3-dimethy-
lindene (1d) and 1-(diphenylphosphino)-3,4,7-trimethy-
lindene (1f). Interestingly, trimethylsilyl may have a
lower preference for the vinylic position than diph-
enylphosphino, since 1h has vinylic phosphine and al-
lylic silyl groups; however, this compound was not
passed through alumina due to its sensitive nature and
may not have been isolated as the most thermodynam-
ically-stable isomer.
2. Results and discussion
2.1. Synthesis and characterization of the phosphinoind-
enes
In addition to mass spectrometry, indenes 1 were
1
The indenyl phosphine ligand precursors were pre-
pared by a typical procedure of deprotonation of the
appropriate indene with BuLi in diethyl ether, followed
by treatment with the corresponding chlorophosphine
(Scheme 1). Filtration of the product mixture through a
short column of alumina was used to remove salts and
any phosphine oxide. It also ensures that the thermo-
dynamically-favored isomer of the product indene is
isolated. 3-(Diphenylphosphino)indene was previously
prepared by Anderson and coworkers [16] and 1,3-
bis(diphenylphosphino)indene (1g) was previously re-
ported by McGlinchey and coworkers [17] and isolated
and characterized by us [3].
characterized by H, 13C and 31P NMR spectroscopies.
The most readily identifiable peak in the 1H NMR
spectra of these compounds is that of H1. This lies at
3.2–3.6 ppm with an integral of two protons for the 3-
substituted phosphino indenes and at 4.3–4.5 ppm with
an integral of one proton for the 1-phosphino indenes.
The aromatic benzo resonances lie in the range 6.6–7.7
ppm along with the phenyl resonances. In cases where
these resonances are well separated, they could be as-
signed by observation of an NOE from H1 to H7 fol-
lowed by 1H–1H COSY spectroscopy. Similarly, an
NOE could be observed from H1 to the methyl on C7
for the 4,7-dimethyl-substituted indenes and the methyl
on C2 for the 2-methyl-substituted indenes. H2 is usu-
ally upfield of the other aromatic resonances at 5.9–6.2
ppm, except for the diisopropylphosphine 1i (6.71 ppm)
Generally, it is found that phosphino substituents
prefer to be in the vinylic 3-position of the indene ring,
rather than the allylic 1-position. This has been observed
Scheme 1. Synthesis of the methyl- and trimethylsilyl-substituted phosphinoindenes.

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1899
and the trimethylsilyl derivative 1h (6.39 ppm), although
these are still upfield of the benzo resonances. The 2-
and 3-methyl resonances are found at 1.9–2.1 ppm
whereas the 4- and 7-methyl resonances are found
downfield of these at 2.2–2.7 ppm.
In the 13C NMR spectra, C1 is readily identifiable at
1
47–52 ppm with JPC ¼ 20–25 Hz for the allylic phos-
3
phines 1b, 1d and 1f and at 38–45 ppm with JPC ¼ 2–6
Hz for the vinylic phosphines 1c, 1e and 1i. For the
trimethylsilyl derivative 1h, the resonance for C1 is
found at 47.9 ppm with a three-bond coupling to
phosphorus of 4 Hz. Where possible, 13C assignments
were aided by H–13C HSQC and H–13C HMBC ex-
periments. The C4 and C7 benzo resonances at 118–124
ppm lie upfield of C5 and C6 (123–130 ppm), except
when they are methylated, in which case they occur at
128–131 ppm. The phenyl resonances are generally
downfield of the benzo resonances. The resonance for
C2 occurs at approximately 130 ppm for the allylic
phosphines of 1a, 1b and 1f. Methylation gives a small
downfield shift to 133.5 ppm for 1d whereas the vinylic
phosphino indenes give a larger downfield shift to ap-
proximately 142 ppm for 1a, 1e, 1h and 1i. Methylation
of a vinylic phosphino indene gives a further downfield
shift to 159.7 ppm for 1c. The 4- and 7-methyl carbons
lie in a narrow range at 18–21 ppm whereas the 2- and 3-
methyl carbons lie upfield at 10–17 ppm.
The 31P chemical shifts for the allylic phosphines 1b,
1d and 1f lie closely together at )5 to +2 ppm (compared
to )4.3 ppm for the unsubstituted 1-(diphenylphosph-
ino)indene 1a) [16] whereas the vinylic phosphines are
further upfield and are more sensitive to substitution on
the indene ring: compared to vinylic-1a at )22.3 ppm
[16], substitution with methyl at the 2-position, 1c, gives
an upfield shift to )29.3 ppm whereas the 4,7-dimethyl
derivative 1e exhibits a downfield shift to )16.9 ppm.
The 1-trimethylsilyl derivative 1h has a small downfield
shift to )21.5 ppm. The chemical shift of the diisopro-
pylphosphine 1i ()9.2 ppm) is consistent with a 3-inde-
nyl substituent: the equation d ¼ RnaE ꢀ 4:8 for
phosphines (with aEð3-indenylÞ ¼ ꢀ18:0 and aEði-PrÞ ¼
8:1) gives a calculated chemical shift of )6.6 ppm for the
3-indenyl phosphine compared to +13.3 ppm for the 1-
indenylphosphine (aEð1-indenylÞ ¼ 2) [19].
1
1
Scheme 2. Synthesis of the methyl- and trimethylsilyl-substituted
phosphinoindenyl ferrocenes.
isomers due to the bisplanar chiral nature of the com-
plexes. Since the rac/meso isomerization processes have
not been observed in the non-coordinating solvents
chloroform, dichloromethane or diethyl ether, these are
generally used for the filtration and isolation steps.
The reaction of lithium 1-(diphenylphosphino)inde-
nide with ferrous chloride initially produces a 1:1 mix-
ture of two phosphorus-containing compounds, as
shown by peaks in the 31P NMR spectrum at )22.26 and
)26.53 ppm. Overnight, the peak at )26.53 ppm almost
disappears leaving essentially only one product. Re-
crystallization of this compound from CH2Cl2/diethyl
ether gave crystals suitable for X-ray crystallography
that showed the compound to be the rac isomer of 2a
(see below). If the reaction is stopped after 2 h by re-
moval of solvent, the rac/meso product mixture can be
isolated. In the 1H NMR spectrum, the H2 and H3
protons have distinct chemical shifts from each other
and for each isomer: For the rac isomer, they are widely
spaced at 3.07 and 4.92 ppm for H2 and H3, respectively
(a difference of 1.85 ppm), whereas for the meso isomer,
they are closely spaced (and lie between those of the rac
isomer) at 3.48 and 3.81 ppm for H2 and H3, respec-
tively (a difference of 0.33 ppm). Clearly, the ring cur-
rents are having a significant effect on these protons.
In THF solvent, the meso isomer is observed to un-
dergo an isomerization to the rac isomer with
DHz ¼ 58 ꢁ 4 kJ molꢀ1, DSz ¼ ꢀ140 ꢁ 15 J Kꢀ1 molꢀ1
and DV z ¼ ꢀ12:9 ꢁ 0:8 cm3 molꢀ1 [4]. At 23 °C,
kobs ¼ 1:59ð3Þ ꢂ 10ꢀ5 sꢀ1 [14,15]. The isomerization
does not go to completion, indicating that the reverse
process of rac to meso isomerization does occur, but
with a slower rate constant. Thus, the meso/rac equi-
librium lies on the side of the rac isomer. In THF at
2.2. Synthesis and characterization of the ferrocenes
The ferrocenes were all prepared by treatment of the
phosphinoindenide (formed by deprotonation of the
indene with BuLi) with anhydrous ferrous chloride in
THF (Scheme 2). Generally, all of the ferrous chloride
has reacted after approximately 2 h of stirring at am-
bient temperature. Purifications of the ferrocenes were
carried out by filtration through a Celite column using a
solvent appropriate for the ferrocene. In each case, the
possibility exists for the formation of both rac and meso

1900
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
23 °C, K ¼ ½meso-2aꢃ=½rac-2aꢃ ¼ 0:074ð5Þ [15]. Our pre-
ferred mechanism involves THF coordination with in-
denide ring-slippage and then dechelation of the
indenide and coordination of the phosphine to form the
key zwitterionic intermediate 3 (Scheme 3). It has also
been found that the most facile hydrolysis mechanism is
also via zwitterion 3 [15], which one would expect to be
more readily protonated than a non-polar ferrocene 2.
As a consequence, it is found that meso-2a hydrolyses
much more readily than rac-2a.
The synthesis of the 3-methyl analogue 2b, from 1b,
gave the two isomers in a 3:1 ratio as indicated by peaks
at )24.50 and )26.25 ppm in the 31P NMR spectrum
after 2 h. This ratio did not change after stirring at
ambient temperature overnight. The use of higher tem-
peratures led to hydrolysis reactions with no observable
isomerization. Upon workup, only one compound, that
with the downfield chemical shift, was isolated and
X-ray crystallography, see below, showed this to be the
rac isomer. The 1H NMR spectrum contains a reso-
nance at 3.61 ppm for H2, downfield of both isomers of
2a. As with 2a, the meso isomer appears to be the most
reactive isomer.
Ferrocene 2c, containing methyl groups in the 2 and
20 positions, was prepared from 1c in THF and found to
give two compounds, in a 2:3 ratio, at )17.41 and
)19.68 ppm in the 31P NMR spectrum. This ratio did
not change after stirring at ambient temperature for 3
days. The major isomer has a chemical shift of 4.84 ppm
for H3 whereas this peak occurs at 5.08 ppm for the
minor isomer. Comparison of the 31P and 1H NMR
spectra with 2a and 2b (Table 1), as well as the 13C
NMR spectra (see below), strongly suggest that the
major isomer, with the upfield 31P chemical shift, is the
meso complex, however, this is not unambiguous.
The tetramethylferrocene 2d, with methyl groups in
the 2, 20, 3 and 30 positions, was prepared from 1d and
found to give the two isomers in a 5:2 ratio ()23.3 and
)23.7 ppm). Unfortunately, only a few crystals of the
rac isomer (see below) were isolated and we have been
unable to unambiguously state which is the major iso-
mer. Nonetheless, the chemical shifts of C5 and C6
(122.0 and 123.5 ppm, respectively) tend to suggest that
the major isomer is rac since, for 2a–2c, these resonances
occur at 121.2–123.1 ppm for the rac isomers and 124.0–
125.0 ppm for the meso isomers. Solutions of this
compound were found to be very sensitive and we were
unable to observe any isomerization process.
Bis(1-(diphenylphosphino)-4,7-dimethyl-g5-indenyl)-
iron(II) (2e) was made from indene 1e and found to give
a 1:1 isomeric mixture with peaks in the 31P NMR
spectrum at )16.46 and )21.47 ppm. Crystallization
from dichloromethane/diethyl ether afforded dark-red
crystals which 31P NMR spectroscopy showed to con-
tain both isomers. As different NMR samples gave dif-
ferent ratios of the two isomers, we concluded that
different crystals containing each isomer were present,
however, we were unable to visually discriminate be-
tween the two sets of crystals. Random selection of
Scheme 3. Preferred mechanism for rac/meso isomerization of phos-
phinoindenyl ferrocenes.
Table 1
31P NMR chemical shifts and 1H NMR chemical shifts for H2 and H3 of the ferrocenes in CDCl3
31P
H3
H2
Compound
rac
meso
rac
meso
rac
meso
2a
2b
2c
2d
2e
2h
2i
)22.26
)24.69
)17.60
)23.81a
)16.86
)28.49a
)3.27
)26.53
)26.35
)20.10
)23.21a
)22.17
4.92
3.81
3.07
3.61
3.48
Not observed
–
5.08
–
4.84
–
–
–
5.31
–
4.02
–
3.13
3.84a
–
3.85
–
–
)8.89
)23.90
)25.61
Not observed
5.34
4.98
Not observed
4.07
3.85
Not observed
3.27
3.11
Not observed
3.90
3.55
2j
)16.81
)22.20
a Assignment to the rac or meso isomer is ambiguous.

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1901
1
crystals allowed us to eventually obtain X-ray structures
of both isomers (see below) as well as identify the sep-
arate NMR spectra. Again, the compound with the
downfield chemical shift was found to be the rac isomer.
sensitive. We were unable to obtain satisfactory H and
13C NMR data due to the formation of paramagnetic
decomposition products. However, we were able to
observe, by 31P NMR spectroscopy, an isomerization
process, presumably from meso to rac (the rac isomer
again lies downfield of the meso isomer: )3.27 ppm
versus )8.89 ppm, respectively). At 23 °C, kobs ¼ 1:75
1
As with 2a, the H NMR chemical shifts of the H2 and
H3 protons are quite distinctive: The rac isomer has
widely-spaced resonances at 3.13 and 5.31 ppm for H2
and H3, respectively (a difference of 1.98 ppm), whereas
the meso isomer has closely-spaced resonances, in be-
tween those of the rac isomer, at 3.85 and 4.02 ppm for
H2 and H3, respectively (a difference of 0.17 ppm). We
based our NMR assignments of the isomers on these
resonances since similar trends were unambiguously
identified in 2a.
We investigated possible isomerization processes in
2e by taking a number of NMR samples of the crys-
talline mixture, to ensure that we did not accidentally
start with the thermodynamic mixture, and then heated
each sample to 40 °C for 24 h. We did not observe any
change in the isomeric ratio under these conditions. We
suggest that the increased steric protection of the iron
centre and the electron-donating ability of the methyl
groups are preventing nucleophilic attack at the iron
center by solvent THF molecules.
ð30Þ ꢂ 10ꢀ5 sꢀ1 and, at 30 °C, kobs ¼ 3:05ð2Þ ꢂ 10ꢀ5 sꢀ1
.
These rates are very similar to those observed for the
isomerization of meso-2a to rac-2a (1.59(3) ꢂ 10ꢀ5 and
3.01(9) ꢂ 10ꢀ5
s
ꢀ1, respectively) and suggest that the
nucleophilicity of the phosphine, diisopropyl versus di-
phenyl, is not important in the rate-determining step
and is consistent with this step involving coordination of
THF solvent and indenide ring-slippage.
Since we have only been able to observe rac/meso
isomerization in the two least-sterically-crowded sys-
tems, 2a and 2i, we sought to make a smaller increase in
the steric bulk of the system by preparing a mixed-ligand
system derived from indenes 1a and 1e, namely, {4,7-
dimethyl-1-(diphenylphosphino)-g5-indenyl}{1-(diph-
enylphosphino)-g5-indenyl}iron(II) (2j) (Scheme 4).
From the addition of ferrous chloride to a 1:1 mixture of
the indenides of 1a and 1e we observed eight ferrocenyl
phosphine signals in the 31P NMR spectrum of the
product solution. The two isomers of each of 2a and 2e
account for four of these, and the other four can then be
assigned to the rac and ‘‘meso’’ isomers of 2j. It should
be noted that the ‘‘meso’’ isomer actually exists as two
enantiomers due to the differing centers of planar chi-
rality. After stirring the product mixture for 24 h, all of
the meso-2a had isomerized to rac-2a, as expected,
however, there was no other change in the relative in-
tensities of the signals. Surprisingly, the addition of just
two methyl groups on one benzo ring is appears to be a
sufficient increase in steric bulk to significantly slow the
rate of isomerization.
Numerous unsuccessful attempts have been made to
prepare the bis(trimethyldiphenylphosphino) derivative
2f from indene 1f. Presumably steric crowding is pre-
venting its formation. We have also generally observed
that increasing substitution in indenylferrocenes de-
creases the stability and increases the reactivity of the
complexes once they are formed.
We have previously reported that treatment of 1,3-
bis(diphenylphosphino)indenide with ferrous chloride in
THF produces a single peak in the 31P NMR spectrum
consistent with the desired ferrocene, 2g [3]. However,
upon removal of solvent and extraction of the products,
we find that the ferrocene has decomposed into 2a and
other species. It appears that concentration of this spe-
cies leads to its decomposition. It may be that other
molecules of 2g act as nucleophiles to give reactive
zwitterionic species to begin hydrolysis-type processes.
Treatment of the sterically similar 3-(diphenylphosph-
ino)-1-(trimethylsilyl)indene (1h) with BuLi followed by
ferrous chloride also initially gives the expected product
peaks in the 31P NMR spectrum (at )25.3 and )27.3
ppm), in a 3:2 ratio of the two isomers. Like 2g, this
material is quite sensitive, however, one of the isomers is
more stable than the other (probably the rac isomer) and
we were able to isolate this in reasonable yield (31%). It
seems that steric factors again lead to sensitive inde-
nylferrocene systems. Similarly, we note that we have
recently prepared and isolated bis(1,3-bis(trimethylsilyl)-
g5-indenyl)iron(II) and that it suffers from significant
steric strain [4].
2.3. NMR spectra of the ferrocenes
Generally, there are few consistent trends in the rel-
ative 13C NMR spectra of the various isomers. The most
useful information to be gained from the 13C NMR
The ferrocene formed from the diisopropylphosphine
1i, 2i, was found to be extremely air, moisture and heat
Scheme 4. Preparation of mixed diindenyl ferrocene 2j.

1902
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
spectra is that, for both 2a and 2e, on going from the rac
to the meso isomer there is a consistent downfield shift
of 2 ppm for both C5 and C6 and 4–5 ppm for both C4
and C7. Consistent with the trends in the P and H2
chemical shifts of 2c, the relative chemical shifts of C5
and C6 for the two isomers of 2c also suggest that the
major isomer is meso and the minor isomer is rac. In the
31P NMR spectra, so far as we can be certain of our
assignments, the chemical shift of the rac isomer is al-
ways 2–6 ppm upfield of the meso isomer (Table 1).
rac-2a (Fig. 1) and rac-2e (Fig. 2) are very similar in that
they have essentially the same conformation: the two
benzo rings lie approximately on top of each other
(RA ¼ 20.8° and 12.1° for rac-2a and rac-2e, respec-
tively, compared to 6.0° and 13.0° for diindenyliron(II)
[10a]) in a p-offset arrangement to maximise p–p-stack-
ing interactions [20] and the diphenylphosphino groups
are on opposite sides of the molecule, presumably to
minimise steric interactions. Fig. 3 shows the five mol-
ecules that we have crystallography characterized here
looking down the centroid–centroid axis. From this
view, it can be seen that two of the phenyl rings in both
rac-2a and rac-2e (Figs. 3(a) and (d), respectively) ef-
fectively sandwich the H2 protons, which probably ac-
counts for their large up-field shift in the 1H NMR
spectrum. The two other phenyl rings are oriented away
from the ferrocene center. Interestingly, if the benzo
rings in either compound were p offset in the other di-
rection, but by approximately the same amount, the
diphenylphosphino groups would be even further apart
from each other without increasing the steric interac-
tions between the phenyl groups and the other benzo
ring. This appears to be the case in the only non-ansa rac
2.4. X-ray structural analyses
Crystallographic and refinement data for complexes
rac-2a, rac-2b, rac-2d, rac-2e and meso-2e are given in
Table 2 with selected bond distances and angles given in
Table 3. The solid-state structures of these compounds
were determined for two reasons: to confirm the identity
of the isomer and to look for possible explanations for
the apparent preference of the rac isomer over the meso
isomer in 2a.
Compound rac-2a lies on a crystallographic C2 axis
whereas rac-2e does not. Nonetheless, the structures of
Table 2
Crystal data and structural refinement parameters for rac-2a, rac-2b, rac-2d, rac-2e and meso-2e
rac-2a
rac-2b
rac-2d
rac-2e
meso-2e
Empirical formula
Formula weight (g molꢀ1
Temperature (K)
C42H32FeP2
654.47
168(2)
0.71073
Orthorhombic
Pbcn
12.959(6)
12.396(5)
19.296(8)
90
C44H36FeP2
682.52
163(2)
C46H40FeP2
710.57
183(2)
C46H40FeP2
710.57
168(2)
C46H40FeP2
710.57
293(2)
0.71073
Monoclinic
C2=c
27.186(5)
12.532(3)
24.816(5)
90
)
ꢀ
Wavelength (A)
0.71073
Monoclinic
C2=c
15.569(16)
13.705(15)
17.401(18)
90
0.71073
Triclinic
0.71073
Triclinic
Crystal system
Space group
ꢁ
ꢁ
P1
P1
ꢀ
a (A)
9.397(19)
14.16(3)
14.23(3)
93.93(4)
101.65(6)
90.03(6)
1850(7)
2
12.652(7)
12.669(7)
14.304(8)
67.521(7)
87.440(7)
62.968(7)
1863.8(18)
2
ꢀ
b (A)
ꢀ
c (A)
a (°)
b (°)
c (°)
90
90
110.40(3)
90
122.72(3)
90
3
ꢀ
Volume (A )
3100(2)
4
1.402
3480(6)
4
1.303
7113(2)
8
1.331
Z
Density (calcd) (Mg/m3)
1.276
0.526
744
1.270
0.522
748
Absorption coefficient (mmꢀ1
)
0.621
1360
0.556
1424
0.547
2992
F (0 0 0)
Crystal size (mm)
h Range (°)
Index ranges
0.30 ꢂ 0.21 ꢂ 0.04 0.40 ꢂ 0.24 ꢂ 0.19 0.41 ꢂ 0. 15 ꢂ 0.07 0.75 ꢂ 0.25 ꢂ 0.05 0.40 ꢂ 0.36 ꢂ 0.27
2.11–26.44
ꢀ16 6 h 6 16,
ꢀ15 6 k 6 8,
ꢀ23 6 l 6 23
37580
2.04–26.31
ꢀ10 6 h 6 19,
ꢀ16 6 k 6 17,
ꢀ21 6 l 6 11
6664
1.98–26.46
ꢀ11 6 h 6 10,
ꢀ17 6 k 6 17,
ꢀ17 6 l 6 17
22975
1.96–26.47
ꢀ15 6 h 6 15,
ꢀ15 6 k 6 15,
ꢀ8 6 l 6 17
23327
1.79–26.61
ꢀ33 6 h 6 26,
ꢀ11 6 k 6 15,
ꢀ21 6 l 6 31
21717
Reflections collected
Independent reflections
R(int)
3142
0.1169
3209
0.0416
7267
0.0900
7376
0.0631
7212
0.0426
Completeness to h
Absorption correction
(26.44°) 98.3%
Multi-scan
0.9756/0.8356
3142/0/204
0.966
(26.31°) 90.7%
Multi-scan
1.0000/0.8567
3209/0/213
0.892
(26.46°) 95.0%
Multi-scan
1.0000/0.7954
7267/0/446
1.053
(26.47°) 95.9%
Multi-scan
1.0000/0.8669
7376/0/442
0.921
(26.61°) 96.7%
Multi-scan
1. 0000/0.8534
7212/0/446
1.022
Maximum/minimum transmission
Data/restraints/parameters
Goodness-of-fit on F 2
R indices [I > 2rðIÞ] (R; Rw)
0.0351, 0.0783
0.0908, 0.1124
0.0373, 0.0788
0.0652, 0.0840
0.1102, 0.3143
0.2086, 0.3593
0.0464, 0.1041
0.0978, 0.1165
0.0383, 0.0943
0.0660, 0.1034
R indices (all data) (R; Rw)
ꢀ3
ꢀ
Final maximum/minimum Dq (e A
)
0.308 and )0.427 0.339 and )0.369 2.047 and )0.415 0.722 and )0.324 0.298 and )0.301

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1903
Table 3
Selected bond distances (A) and angles (°) for rac-2a, rac-2b, rac-2d, rac-2e and meso-2e
ꢀ
rac-2a
rac-2b
rac-2d
rac-2ea
meso-2ea
Fe–CNT
Fe–C1
Fe–C2
1.672
1.685
1.707, 1.693
1.666, 1.669
2.062(3), 2.056(3)
2.041(3), 2.045(3)
2.047(3), 2.057(3)
2.116(3), 2.109(3)
2.094(3), 2.104(3)
1.832(3), 1.836(3)
178.8
1.681, 1.681
2.062(3)
2.045(3)
2.067(3)
2.090(3)
2.104(3)
1.829(3)
179.6
2.064(3)
2.065(3)
2.089(3)
2.094(3)
2.113(3)
1.830(3)
177.03
130.06(13)
174.11
115.3
2.063(9), 2.083(9)
2.102(11), 2.085(12)
2.087(11), 2.073(12)
2.112(9), 2.101(9)
2.148(10), 2.091(10)
1.852(9), 1.837(8)
175.84
2.096(2), 2.052(2)
2.048(2), 2.037(2)
2.045(2), 2.067(2)
2.094(2), 2.103(2)
2.097(2), 2.112(2)
1.840(2), 1.814(2)
175.4
Fe–C3
Fe–C8
Fe–C9
P–C1
CNT–Fe–CNT
Fe–C1–P
CNT–C1–P
C1–CNT–CNT0–C10
123.18(17)
176.07
123.1
131.1(5), 131.2(4)
168.30, 168.63
30.4
122.59(15), 126.90(15)
176.43, 178.82
131.9
134.59(12), 125.82(12)
171.45, 179.19
40.5
b
ꢀ
Slip-fold parameter D (A) 0.032
0.027
2.4
0.055, 0.018
4.7, 3.3
0.051, 0.051
2.6, 2.2
1.9, 1.9
0.025, 0.047
1.8, 2.7
1.2, 1.0
Hinge angle HA (°)c
Fold angle FA (°)d
2.1
0.9
2.5
28.7
5.2, 6.0
118.8
Rotation angle RA (°)e
20.8
12.1
40.5
a The second number refers to the equivalent parameter for the primed atoms.
b D ¼ average distance of Fe to C8 and C9 minus average distance of Fe to C1 and C3.
c HA ¼ angle between planes defined by [C1, C2, C3] and [C1, C3, C8, C9].
d FA ¼ angle between planes defined by [C1, C2, C3] and [C4, C5, C6, C7, C8, C9].
e RA ¼ angle formed by the intersection of two lines determined by the centroids of the five- and six-membered rings.
Fig. 1. ORTEP of rac-2a indicating the numbering of the atoms. The
thermal ellipsoids have been drawn at 40% probability.
Fig. 3. Views down the CNT–CNT axes: (a) rac-2a, (b) rac-2b, (c) rac-
2d, (d) rac-2e and (e) meso-2e. For clarity, only the ispo-Ph carbon
atoms of the phenyl rings in meso-2e are shown.
Thus, it appears that this relative conformation of the
indenyl rings in rac-2a and rac-2e is at least partially due
to favourable edge-type CH–p hydrogen bonding in-
teractions between the phenyl groups and the H2 atoms,
as each H2 atom has three Hꢄ ꢄ ꢄC distances of less than
ꢀ
3.0 A: two to a phenyl group on the same ligand (the
ipso atom and an ortho atom) and one to an ortho car-
ꢀ
bon atom on the other ligand (2.90, 2.69 and 2.98 A,
respectively, for rac-2a and 2.83, 2.95, and 2.97, re-
ꢀ
spectively, for H2 of rac-2e and 2.81, 2.83 and 2.95 A,
respectively, for H20 of rac-2e) [21].
Fig. 2. ORTEP of rac-2e indicating the numbering of the atoms. The
thermal ellipsoids have been drawn at 40% probability.
Meso-2e is the first meso-diindenyl complex crystal-
lographically characterized. As illustrated in Figs. 3(e)
and 4, the p-offset stacking of the benzo rings is not
ideal: RA (40.5°) is much larger than in rac-2a, rac-2b
diindenyl ferrocene that has previously been crystallo-
graphically characterized, a bis(indeno[2.2]paracyclo-
phane-9-ene) complex [13], in which the RA is 35.4°.

1904
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
Fig. 4. ORTEP of meso-2e indicating the numbering of the atoms. The
thermal ellipsoids have been drawn at 40% probability.
Fig. 6. PLUTO of rac-2d indicating the numbering of the atoms. The
atoms have been drawn at 40% probability.
(see below) and rac-2e as well as in that of diindenyl-
iron(II) [10a]. A better p-stacking would result in closer
diphenylphosphino groups, but H260 is already so close
to C31–C36 as to be hydrogen bonding with this ring.
As it is, steric interactions between the two diph-
enylphosphino groups and one of the methyl groups
leads to a folding of the CNT–Fe–CNT angle away
from linear to 175.4°; to one of the phosphino groups
being bent back (CNT–C1–P ¼ 171.5°); and to a long
suggests that the H2 protons are again sandwiched be-
tween two of the phenyl rings, however, they are in fact
pushed out of the Fe atom plane by the encroaching
3-methyl groups, such that the H2 atoms can each only
hydrogen bond to one of the phenyl rings (the H2ꢄ ꢄ ꢄC21
ꢀ
distance is 2.89 A, all other Hꢄ ꢄ ꢄC distances are greater
ꢀ
than 3.0 A). As with rac-2a and rac-2e, the other phenyl
ꢀ
ꢀ
Fe–C1 distance (2.096(2) A, which is over 0.03 A longer
than comparable distances in the other ring and the
other compounds). There are also fewer CH–p interac-
tions in this complex than in rac-2a and rac-2e: The H2
rings are oriented away from the ferrocene center.
Although the crystallographic analysis of rac-2d is
poor (R ¼ 11:0%), its identity and conformation, illus-
trated in Figs. 3(c) and 6, is unambiguous and quite
different from that of the other rac compounds. The five-
membered rings are staggered and there is no p-offset
stacking between the benzo rings (RA ¼ 118.8°). The
diphenylphosphino groups lie near each other (C1–
CNT–CNT0–C10 ¼ 30.4°) with two of the phenyl rings in
the plane of the ferrocene and oriented towards the six-
membered ring rather than the methyl groups.
ꢀ
atoms each have two Hꢄ ꢄ ꢄC distances of less than 3.0 A,
to the ipso and an ortho C atom of a phenyl group of the
same ligand, but the orientations are not very favour-
able for hydrogen bonding. H260 does, however, point
directly at the centroid of C31–C36 and all of the Hꢄ ꢄ ꢄC
ꢀ
distances to this ring are less than 3.0 A. It would appear
that the p–p stacking, steric interactions and CH–p hy-
drogen bonding are all less favourable in this meso
compound than in rac-2a and rac-2e.
Rac-2b (Figs. 3(b) and 5) has a similar conformation
to rac-2a and rac-2e: the benzo rings lie in a p-offset
arrangement (RA ¼ 28.7°). The centroid–centroid view
3. Conclusions
In this paper we have described a number of deriva-
tives of the diindenyl analogue of dppf: a variety of
methyl derivatives, one trimethylsilyl derivative, and a
(diisopropylphosphino)indenyl analogue. All of these
are bisplanar chiral systems with the potential to form
both rac and meso isomers. In each case, a mixture of
the two isomers was initially formed. For the 3-methyl
and 3-trimethylsilyl derivatives, only one isomer, prob-
ably rac, was obtained upon workup. In the case of the
non-functionalized phosphinoindenyl complexes, these
were both observed to undergo a meso to rac isomeri-
zation process at similar rates, consistent with an
isomerization mechanism involving solvent coordina-
tion and indenide ring-slippage in the rate determining
step. The lack of isomerization in the other ferrocenes is
attributed to steric effects since even the 4,7-dimethyl
Fig. 5. ORTEP of rac-2b indicating the numbering of the atoms. The
thermal ellipsoids have been drawn at 40% probability.

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1905
derivative does not isomerize and a steric explanation
seems more likely than an electronic one. Increasing
functionalization was also found to increase the sensi-
tivity of the ferrocene to such an extent that we were
unable to prepare the bis(3,4,7-trimethyl) derivative.
Crystallographic studies of four rac compounds as
well as the first non-ansa meso diindenyl ferrocene were
carried out. Comparisons of the rac/meso pair of 2e
suggests that the rac isomer is favored due to a combi-
nation of better p-offset stacking, better intramolecular
CH–p hydrogen bonding, and fewer intramolecular
steric interactions.
n-BuLi (3.57 ml, 1.6 M, 5.72 mmol). The solution was
allowed to warm to ambient temperature and stirred for 2
h, in which time a white precipitate formed. The mixture
was cooled to )78 °C, and PPh2Cl (0.892 ml, 5.72 mmol)
was added drop-wise. The mixture was allowed to warm
to ambient temperature and stirred for 2 h. The reaction
mixture was then filtered through alumina and the solvent
was removed in vacuo to give 1.61 g (90%) of 1b as a white,
air-sensitive powder. 1H NMR (CDCl3): d 7.67–6.85 (m,
14H, PhH and H4–7), 6.12 (m, 1H, H2), 4.48 (m, 1H, H1),
2.11 (m, 3H, Me). 13C{1H} NMR (CDCl3): d 145.71 (s,
C3a), 144.47 (d, 2J ¼ 9 Hz, C7a), 140.65 (d, 3J ¼ 6 Hz, C3),
137.66 (d, 1J ¼ 19 Hz, i-Ph), 137.21 (d, 1J ¼ 18 Hz, i-Ph),
133.68 (d, 2J ¼ 20 Hz, o-Ph), 133.12 (d, 2J ¼ 20 Hz, o-Ph),
2
129.76 (d, J ¼ 4 Hz, C2), 129.13 (s, p-Ph), 128.81 (s, p-
4. Experimental
Ph), 128.38 (d, 3J ¼ 7 Hz, m-Ph), 128.07 (d, 3J ¼ 7 Hz, m-
3
Ph), 126.47 (s, C6), 124.51 (s, C5), 123.91 (d, J ¼ 5 Hz,
All manipulations and reactions were carried out un-
der an inert atmosphere (Ar or N2) by use of standard
Schlenk line techniques. Reagent grade solvents were
dried and distilled prior to use: diethyl ether and tetra-
hydrofuran from Na/benzophenone; dichloromethane
and petroleum ether (50–70 °C fraction) from CaH2.
Diphenylphosphinoindene [16], bis(1-(diphenylphosph-
ino)-g5-indenyl)iron(II) [3,14], 1-methylindene [22] and
4,7-dimethylindene [23] were prepared by published
procedures. 1,2-dimethylindene and 1,4,7-trimethylind-
ene were prepared from 2-methylindene and 4,7-dime-
thylindene, respectively, by modifications of the route
used to prepare 1-methylindene. All other reagents were
purchased from Aldrich or Sigma Chemical Companies.
1H, 13C{1H} and 31P{1H} NMR spectroscopy data were
collected on a Varian UNITY-300 spectrometer operat-
1
C7), 119.25 (s, C4), 47.08 (d, J ¼ 20 Hz, C1), 12.95 (s,
Me). 31P{1H} NMR (CDCl3): d )4.98. Mass spectrum
(EI, m/z (%)): 314 (82, Mþ), 185 (100, Ph2Pþ), 128 (30,
C9H5Meþ), 108 (6, PhPþ), 77 (8, Phþ). HR-MS: Mþ
Calc., 314.12244. Found, 314.12295.
4.2. Preparation of 3-(diphenylphosphino)-2-methylind-
ene (1c)
2-Methylindene (1.00 g, 7.68 mmol) in diethyl ether (50
ml) was cooled to )78 °C and BuLi (4.8 ml of 1.6 M, 7.69
mmol) was added. After stirring at ambient temperature
for 3 h, the solution was cooled to )78 °C and PPh2Cl
(1.38 ml, 7.69 mmol) was added. The solution was then
stirred overnight at ambient temperature, filtered through
a short alumina column and the solvent removed in vacuo
1
ing at 300, 75 and 121 MHz, respectively. NOE, H–1H
1
COSY, 1H–13C HSQC and 1H–13C HMBC experiments
were run on a Varian INOVA-500 spectrometer operating
at 500 and 125 MHz for 1H and 13C, respectively. Spectra
were measured at ambient temperature with residue sol-
vent peaks as internal standard for 1H and 13C{1H}
spectroscopy.Couplings in the 13C NMR spectra are to
phosphorus. 31P{1H} NMR spectroscopy chemical shifts
were reported relative to external 85% H3PO4, positive
shifts representing deshielding. EI mass spectra were
collected on a Kratos MS80RFA mass spectrometer. El-
emental analyses were carried out by Campbell Micro-
analytical Services, University of Otago, Dunedin.
to give 1c (5.54 g, 94%) as a white powder. H NMR
3
(CDCl3): d 7.35 (m, 4H, m-PhH), 7.31 (d, JHH ¼ 7 Hz,
1H, H7), 7.2 (m, 6H, o-and p-PhH), 6.97 (dd, 3JHH ¼ 7and
7 Hz, 1H, H6), 6.85 (dd, 3JHH ¼ 7 and 7 Hz, 1H, H5), 6.61
3
(d, JHH ¼ 7 Hz, 1H, H4), 3.45 (s, 2H, H1), 2.31 (s, 3H,
Me). 13C{1H} NMR (CDCl3): d 159.68 (d, 2J ¼ 30, C2),
2
146.27 (s, C7a), 142.44 (d, J ¼ 3 Hz, C3a), 136.09 (d,
3
3J ¼ 9 Hz, C3), 132.49 (d, J ¼ 18 Hz, m-Ph), 128.36 (d,
2J ¼ 6 Hz, o-Ph), 128.02 (s, p-Ph), 125.83 (s, C5), 123.73 (s,
3
C6), 123.15 (s, C7), 122.07 (d, J ¼ 2 Hz, C4), 44.94 (d,
3J ¼ 6 Hz, C1), 16.79 (d, 3J ¼ 20 Hz, Me). The i-Ph carbon
atoms were not observed. 31P{1H} NMR (CDCl3): d
)29.3. Anal. Calc. for C22H19P: C, 84.06; H, 6.09. Found:
C, 81.23; H, 6.12%. Mass spectrum: (EI, mþ/z (%)): 314
(100, Mþ), 185 (62, Ph2Pþ), 128 (16, MeC9H5 ). HR-MS:
Mþ Calc., 314.12244. Found, 314.12359.
4.3. Preparation of 1-(diphenylphosphino)-2,3-dimethy-
lindene (1d)
4.1. Preparation of 1-(diphenylphosphino)-3-methylind-
ene (1b)
To a solution of 1,2-dimethylindene (0.717 g, 4.97
mmol) in diethyl ether (40 ml) at )78 °C was added a
solution of n-BuLi (3.11 ml, 1.6 M, 4.97 mmol). The
To a solution of 1-methylindene (0.745 g, 5.72 mmol) in
diethyl ether (40 ml) at )78 °C was added a solution of

1906
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
solution was allowed to warm to ambient temperature
and stirred for 2 h, in which time a white precipitate
formed. The mixture was cooled to )78 °C, and PPh2Cl
(0.892 ml, 4.97 mmol) was added drop-wise. The mix-
ture was then allowed to warm to ambient temperature
and stirred for 2 h. The reaction mixture was filtered
through alumina and the solvent removed in vacuo to
leave a colourless oil, which solidified over a period of 1
h in vacuo to give 1.48 g (91%) of 1d as a white, air-
n-BuLi (6.41 ml, 1.6 M, 10.76 mmol). The solution was
allowed to warm to ambient temperature and stirred for
2 h, in which time a white precipitate formed. The
mixture was cooled to )78 °C, and PPh2Cl (1.84 ml,
10.76 mmol) was added drop-wise. The mixture was
allowed to warm to ambient temperature and stirred for
2 h. The reaction mixture was then filtered through
alumina and the solvent was removed in vacuo to leave a
colourless oily residue. The residue was washed with
petroleum ether (2 ꢂ 25 ml) to give 2.28 g (62%) of 1f as
a white, air-sensitive powder. 1H NMR (CDCl3): d 7.61–
6.81 (m, 12H, H5, H6 and Ph), 6.08 (s, 1H, H2), 4.44 (s,
1H, H1), 2.61 (s, 3H, CH3(4/7)), 2.27 (s, 3H, CH3(4/7)),
1.96 (s, 3H, CH3(3)). 13C{1H} NMR (CDCl3): d 143.1
1
sensitive powder. H NMR (CDCl3): d 7.57–6.64 (m,
14H, H4–7 and Ph), 4.33 (s, 1H, H1), 1.98 (s, 3H,
CH3(3)), 1.89 (s, 3H, CH3(2)). 13C{1H} NMR (CDCl3):
3
d 146.4 (s, C3a), 142.9 (s, C7a), 138.2 (d, J ¼ 6 Hz, C3)
1
2
137.0 (d, J ¼ 18 Hz, i-Ph), 134.1 (d, J ¼ 20 Hz, o-Ph),
134.1 (d, 1J ¼ 18 Hz, i-Ph), 133.5 (d, 2J ¼ 4 Hz, C2),
2
3
(d, J ¼ 6 Hz, C7a), 142.6 (d, J ¼ 3 Hz, C3), 141.3 (s,
2
132.8 (d, J ¼ 20 Hz, o-Ph), 129.1 (s, p-Ph), 128.4 (s, p-
C3a), 138.3 (d, 1J ¼ 20 Hz, i-Ph), 134.4 (d, 2J ¼ 20 Hz, o-
3
3
1
2
Ph), 128.2 (d, J ¼ 6 Hz, m-Ph), 127.6 (d, J ¼ 7 Hz, m-
Ph) 133.4 (d, J ¼ 19 Hz, i-Ph), 132.0 (d, J ¼ 16 Hz, o-
3
3
2
Ph), 126.2 (s, C6), 123.22 (s, C5), 123.2 (d, J ¼ 3 Hz,
Ph), 130.8 (d, J ¼ 4 Hz, C7), 130.3 (d, J ¼ 5 Hz, C2),
129.1 (d, 4J ¼ 2 Hz, C6), 128.6 (s, p-Ph), 128.4 (d, 3J ¼ 4
Hz, m-Ph), 128.3 (s, C4), 127.7 (s, p-Ph), 126.5 (d, 3J ¼ 7
Hz, m-Ph), 126.3 (d, 5J ¼ 2 Hz, C5), 46.3 (d, 1J ¼ 25 Hz,
C7), 118.1 (s, C4), 51.7 (d, 1J ¼ 24 Hz, C1), 14.1 (d,
3J ¼ 8 Hz, CH3(2)), 10.1 (s, CH3(3)). 31P {1H} NMR
(CDCl3): d 1.76 (s). Mass spectrum (EI, m/z (%)): 328 (7,
Mþ), 201 (100, Ph2POþ), 142 (17, C9H4Meþ2 ), 77 (12,
Phþ). HR-MS: Mþ Calc., 328.13809. Found, 328.13708.
4
C1), 19.5 (d, J ¼ 13 Hz, CH3(7)), 19.3 (s, CH3(4)), 17.0
(s, CH3(3)). 31P {1H} NMR (CDCl3): d )1.71 (s). Mass
spectrum (EI, m/z (%)): 342 (15, Mþ), 201 þ(100,
Ph2POþ), 157 (39, C9H4Me3þ), 142 (13, C9H4Me2 ), 77
(14, Phþ). HR-MS: Mþ Calc., 342.15374. Found,
342.15430.
4.4. Preparation of 3-(diphenylphosphino)-4,7-dimethy-
lindene (1e)
4,7-Dimethylindene (2.00 g, 13.9 mmol) in diethyl
ether (100 ml) was cooled to )78 °C and BuLi (8.7 ml of
1.6 M, 13.9 mmol) was added. After stirring at ambient
temperature for 3 h, the solution was cooled to )78 °C and
PPh2Cl (2.4 ml, 13.9 mmol) was added. The solution was
then stirred overnight at ambient temperature, filtered
through a short alumina column and the solvent removed
4.6. Preparation of 3-(diphenylphosphino)-1-(trimethyl-
silyl)indene (1h)
To a solution of the phosphinoindene 1a (1.027 g,
3.42 mmol) in diethylether (40 ml) at )78 °C was added
n-BuLi (2.13 ml, 1.6 M, 3.42 mmol). The solution was
allowed to warm to ambient temperature and stirred for
3 h. The mixture was then cooled to )78 °C, and
Me3SiCl (0.43 ml, 3.42 mmol) was added dropwise. The
mixture was allowed to warm to ambient temperature,
stirred for 2 h, and then filtered through Celite. The
solvent was removed in vacuo to give an orange oily
residue that was dissolved in petroleum ether (25 ml)
and filtered to give an orange solution. Removal of
solvent in vacuo yielded 0.97 g (76%) of 1h as an orange
air-sensitive oil. 1H NMR (CDCl3): d 7.47–7.13 (m,
14H, H4–7 and Ph), 6.39 (dd, 3JHH ¼ 2 Hz, 3JPH ¼ 4 Hz,
1
in vacuo to give 1e (4.10 g, 90%) as a white powder. H
NMR (CDCl3): d 7.4 (m, 10H, PhH), 6.93 (s, 2H, H5 and
3
3
H6), 5.94 (dt, JHP ¼ 2 Hz, JHH ¼ 3 Hz, 1H, H2), 3.29
(dd, 4JHP ¼ 3 Hz, 3JHH ¼ 3 Hz, 2H, CH2), 2.42 (s, 3H, C4–
Me), 2.32 (s, 3H, C7–Me). 13C{1H} NMR (CDCl3): d
3
1
143.77 (d, J ¼ 5 Hz, C7a), 143.58 (d, J ¼ 20 Hz, C3),
2
143.09 (d, J ¼ 20 Hz, C3a), 141.96 (s, C2), 136.95 (d,
1J ¼ 19 Hz, i-Ph), 134.00 (d, 2J ¼ 20 Hz, o-Ph), 130.28 (s,
C7), 129.79 (d, 3J ¼ 3 Hz, C4), 129.31 (s, C5 or C6), 128.76
(s, p-Ph), 128.49 (d, 3J ¼ 7 Hz, m-Ph), 126.29 (s, C5 or C6),
3
4
38.61 (d, J ¼ 2 Hz, C1), 20.64 (d, J ¼ 16 Hz, C4–Me),
18.45 (s, C7–Me). 31P{1H} NMR (CDCl3): d )16.9. Anal.
Calc. for C23H21P: C, 84.12; H, 6.45. Found: C, 80.40; H,
6.36%. Mass spectrum: (EI, m/z þ(%)): 328 (100, Mþ), 185
(82, Ph2Pþ), 143 (32, Me2C9H5 ). HR-MS: Mþ Calc.,
328.13809. Found, 328.13781.
3
1H, H2), 3.61 (d, JHH ¼ 2 Hz, 1H, H1), )0.52 (s, 9H,
3
Si(CH3)3). 13C{1H} NMR (CDCl3): d 146.1 (d, J ¼ 5
Hz, C7a), 144.8 (d, 1J ¼ 18 Hz, C3), 144.4 (d, 2J ¼ 4 Hz,
2
1
C2), 137.2 (d, J ¼ 11 Hz, C3a), 136.6 (d, J ¼ 9 Hz, i-
1
2
Ph), 136.0 (d, J ¼ 9 Hz, i-Ph), 133.8 (d, J ¼ 20 Hz, o-
Ph), 133.3 (d, 2J ¼ 19 Hz, o-Ph), 128.7 (s, p-Ph), 128.4 (s,
p-Ph), 128.3 (d, 3J ¼ 7 Hz, m-Ph), 128.2 (d, 3J ¼ 7 Hz, m-
Ph), 124.7 (s, C5), 124.0 (s, C6), 122.6 (s, C7), 121.4 (d,
3J ¼ 4 Hz, C4), 47.9 (d, 3J ¼ 4 Hz, C1), )2.3 (s,
Si(CH3)3). 31P {1H} NMR (CDCl3): d )21.54 (s).
Mass spectrum: (EI, m/z (%)): 372 (6, Mþ), 300 (85,
4.5. Preparation of 1-(diphenylphosphino)-3,4,7-trime-
thylindene (1f)
To a solution of 1,4,7-trimethylindene (1.624 g, 10.76
mmol) in diethyl ether (50 ml) at )78 °C was added

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1907
Mþ–SiMe3), 186 (20, C9H5SiMe3þ), 185 (100, Ph2Pþ),
115 (15, C9Hþ7 ), 73 (23, SiMeþ3 ). HR-MS: Mþ Calc.,
372.14632. Found, 372.14548.
4.9. Preparation of rac- and meso-bis(1-(diphenylphosph-
ino)-g5-indenyl)iron(II) mixture
Same procedure as for the preparation of rac-2a,
except that the reaction solution is stirred for only 2 h
before the solvent is removed in vacuo. The product
contains a mixture of the rac and meso isomers in ap-
proximately a 1:1 ratio. Yields obtained for mixtures do
not vary significantly from that obtained for the rac
4.7. Preparation of 3-(diisopropylphosphino)indene (1i)
Indene (1.5 ml, 12.9 mmol) in diethyl ether (30 ml)
was cooled to )78 °C and BuLi (8.0 ml of 1.6 M, 12.9
mmol) was added. After stirring at ambient temperature
for 3 h, the solution was cooled to )78 °C and PiPr2Cl
(2.05 ml, 12.9 mmol) was added. The solution was then
stirred overnight at ambient temperature, filtered
through a short alumina column and the solvent re-
moved in vacuo to give 1i (5.54 g, 94%) as a yellow oil.
1
isomer. Data for meso-2a: H NMR (CDCl3): d 7.53–
3
6.88 (m, 28H, H4–7 and Ph), 3.81 (d, JHH ¼ 2 Hz, 2H,
3
H3), 3.48 (d, JHH ¼ 2 Hz, 2H, H2). 13C{1H} NMR
1
1
(CDCl3): d 139.1 (d, J ¼ 14 Hz, i-Ph), 137.4 (d, J ¼ 11
2
2
Hz, i-Ph), 135.4 (d, J ¼ 21 Hz, o-Ph), 132.4 (d, J ¼ 20
3
Hz, o-Ph), 129.0 (s, p-Ph), 128.2 (d, J ¼ 8 Hz, m-Ph),
3
1H NMR (CDCl3): d 7.64 (d, JHH ¼ 8 Hz, 1H, H4),
3
3
128.1 (d, J ¼ 10 Hz, C7), 128.0 (d, J ¼ 3 Hz, m-Ph),
127.9 (s, C4), 127.7 (s, p-Ph), 124.9 (s, C6), 124.3 (s, C5),
91.6 (d, 2J ¼ 22 Hz, C7a), 90.3 (s, C3a), 74.5 (s, C2), 66.9
(d, 1J ¼ 13 Hz, C1), 64.4 (s, C3). 31P{1H} NMR
(CDCl3): d )26.53 (s).
3
3
7.45 (d, JHH ¼ 7 Hz, 1H, H7), 7.27 (dd, JHH ¼ 8 Hz,
3JHH ¼ 8 Hz, 1H, H5), 7.18 (dd, JHH ¼ 7 Hz, JHH ¼ 7
3
3
3
3
Hz, 1H, H6), 6.71 (dt, JHP ¼ 4 Hz, JHH ¼ 2 Hz, 1H,
H2), 3.52 (s, 2H, CH2), 2.25 (qqd, 3JHH ¼ 7 Hz, 3JHH ¼ 7
2
3
Hz, JHP ¼ 3 Hz, 2H, CH Me2), 1.13 (dd, JHP ¼ 15 Hz,
3JHH ¼ 7 Hz, 6H, Me), 1.02 (dd, 3JHP ¼ 12 Hz, 3JHH ¼ 7
Hz, 6H, Me). 13C{1H} NMR (CDCl3): d 147.71 (d,
4.10. Preparation of bis(1-(diphenylphosphino)-3-meth-
yl-g5-indenyl)iron(II) (2b)
1J ¼ 18 Hz, C3), 143.85 (d, J ¼ 4 Hz, C7a), 141.47 (d,
3
2J ¼ 6 Hz, C2), 139.07 (d, J ¼ 21 Hz, C3a), 126.01 (s,
2
3
C6), 124.57 (s, C5), 123.49 (s, C7), 121.35 (d, J ¼ 6 Hz,
To a solution of indene 1b (0.352 g, 1.12 mmol) in
THF (40 ml) at )78 °C was added a solution of n-BuLi
(0.70 ml, 1.6 M, 1.12 mmol). After 3 h at ambient
temperature, FeCl2 (0.07 g, 0.56 mmol) was added and
the reaction mixture stirred for 2 h to give two product
peaks in a 3:1 ratio at )24.50 and )26.25 ppm, respec-
tively, corresponding to the rac isomer and, presumably,
the meso isomer, respectively. The ratio of products did
not change after stirring overnight. The solvent was re-
moved in vacuo and the residue was loaded onto a Celite
column and washed with dichloromethane. Removal of
the solvent in vacuo yielded a dark green powder con-
taining both isomers (31P NMR CDCl3: d )24.69 (rac),
)26.35 (meso)) as well as free ligand and its oxide. Re-
crystallization from diethyl ether gave 0.16 g (43%) of
C4), 39.76 (d, 3J ¼ 4 Hz, CH2), 22.49 (d, 2J ¼ 10
Hz, Me), 20.32 (d, 1J ¼ 17 Hz, CHMe2), 19.29 (d, 2J ¼ 9
Hz, Me). 31P{1H} NMR (CDCl3): d )9.22. Microanal-
ysis not attempted due to the oily nature of the material.
4.8. Preparation of rac-bis(1-(diphenylphosphino)-g5-in-
denyl)iron(II) (rac-2a)
To a solution of indene 1a (1.8 g, 6.0 mmol) in THF
(50 ml) at )78 °C was added a solution of n-BuLi (3.75
ml, 1.6 M, 6.0 mmol). After 2 h, FeCl2 (0.38 g, 3 mmol)
was added and the reaction mixture stirred was stirred
for 12 h at ambient temperature. The solvent was re-
moved in vacuo and the residue was loaded onto a Celite
column and washed with diethyl ether (to remove un-
reacted 1a). Subsequent elution with dichloromethane
yielded 1.26 g (64%) of rac-2a as a dark blue powder.
Dark blue crystallographic-quality crystals were ob-
tained by recrystallization from CH2Cl2/diethyl ether.
1H NMR (CDCl3): d 7.4–6.4 (m, 28H, H4–7 and Ph),
4.92 (d, 3JHH ¼ 2 Hz, 2H, H3), 3.07 (d, 3JHH ¼ 2 Hz, 2H,
1
rac-2b as a green powder. H NMR (CDCl3): d 7.44–
6.38 (m, 28H, H4–7 and Ph), 3.61 (s, 2H, H2), 2.38 (s,
1
6H, CH3). 13C{1H} NMR (CDCl3): d 139.2 (d, J ¼ 12
1
2
Hz, i-Ph), 136.7 (d, J ¼ 10 Hz, i-Ph), 134.9 (d, J ¼ 21
2
Hz, o-Ph), 132.5 (d, J ¼ 19 Hz, o-Ph), 129.1 (s, p-Ph),
3
3
128.2 (d, J ¼ 8 Hz, m-Ph), 127.9 (d, J ¼ 6 Hz, m-Ph),
3
127.7 (s, p-Ph), 127.1 (d, J ¼ 7 Hz, C7), 123.4 (s, C4),
1
H2). 13C{1H} NMR (CDCl3): d 139.8 (d, J ¼ 10 Hz, i-
2
123.1 (s, C6), 121.2 (s, C5), 89.3 (d, J ¼ 19 Hz, C7a),
1
2
3
Ph), 136.7 (d, J ¼ 7 Hz, i-Ph), 135.2 (d, J ¼ 22 Hz, o-
Ph), 131.7 (d, 2J ¼ 18 Hz, o-Ph), 129.3 (s, p-Ph), 128.3 (d,
3J ¼ 8 Hz, m-Ph), 128.0 (d, 3J ¼ 5 Hz, m-Ph), 127.6 (s, p-
Ph), 124.1 (d, 3J ¼ 9 Hz, C7), 123.6 (s, C4), 122.9 (s, C6),
89.1 (d, J ¼ 4 Hz, C3a), 78.1 (s, C3), 74.5 (s, C2), 65.7
(d, J ¼ 11 Hz, C1), 11.8 (d, J ¼ 9 Hz, Me). 31P{1H}
1
4
NMR (CDCl3):
d )24.69 (s). Anal. Calc. for
C44H36P2Fe: C, 77.42; H, 5.32. Found: C, 75.65; H,
5.55. Mass spectrum: (EI, m/z (%)): 682 (27, Mþ), 312
(100, MeC9H6PPh2þ), 185 (64, Ph2Pþ), 128 (18,
MeC9Hþ6 ), 108 (19, PhPþ). HR-MS: Mþ Calc.,
682.16413. Found, 682.16511.
2
3
122.5 (s, C5), 91.0 (d, J ¼ 25 Hz, C7a), 90.3 (d, J ¼ 4
2
1
Hz, C3a), 72.0 (d, J ¼ 4 Hz, C2), 68.1 (d, J ¼ 9 Hz,
C1), 66.1 (d, J ¼ 4 Hz, C3). 31P{1H} NMR (CDCl3): d
3
)22.26 (s).

1908
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
4.11. Preparation of bis(1-(diphenylphosphino)-2-meth-
yl-g5-indenyl)iron(II) (2c)
vacuo and the residue was loaded onto a Celite column
and washed with petroleum ether (to remove unreacted
1d). Subsequent washing with CH2Cl2, followed by re-
moval of solvent in vacuo, yielded 2d (0.392 g, 42%) as a
green powder. The product was obtained as a 2:5 mix-
ture of diastereomers. Recrystallization from diethyl
ether gave a few purple crystals that X-ray crystallog-
raphy showed to be the rac isomer. There was insuffi-
cient sample to unambiguously identify the isomers in
the NMR spectra. The high sensitivity of the material to
hydrolysis precluded good microanalytical results.
Indene 1c (1.016 g, 3.2 mmol) was dissolved in THF
(45 ml), cooled to )78 °C, and BuLi (2.0 ml, 1.6 M, 3.2
mmol) was added and the solution stirred for 2 h. FeCl2
(0.204 g, 1.6 mmol) was then added and the reaction
mixture stirred overnight to give two products, as ob-
served by 31P NMR with peaks at )17.41 and )19.68
ppm in a 1.0:1.4 ratio. This ratio did not change after
stirring at ambient temperature for 3 days. The solvent
was removed in vacuo and the solid redissolved in
CH2Cl2 and filtered down a Celite column. The solvent
was removed in vacuo to give a green-brown powder
containing both isomers and free ligand. The major
isomer was obtained as a green powder (0.614 g, 55%) in
a pure form by filtration down a Celite column with
diethyl ether and precipitation, after concentration, by
addition of petroleum ether.
1
Major isomer of 2d: H NMR (CDCl3): d 7.68–6.38
(m, 28H, H4–7 and Ph), 1.96 (s, 6H, C3–Me), 1.50 (s,
6H, C2–Me), 13C{1H} NMR (C6D6): d 138.5 (d, 1J ¼ 12
1
Hz, ipso-Ph), 137.1 (d, J ¼ 11 Hz, ipso-Ph), 135.9 (d,
2
2J ¼ 21 Hz, o-Ph), 134.0 (d, J ¼ 19 Hz, o-Ph), 129.0–
127.2 (m, C7, m-Ph and p-Ph), 125.6 (s, C4), 123.5 (s,
C6), 122.8 (s, C5), 88.8–88.1 (m, C3a and C7a), 86.0 (s,
1
C2), 75.6 (s, C3), 67.2 (d, J ¼ 12 Hz, C1), 11.2 (s, C3–
Major isomer of 2c, probably meso: 1H NMR
Me), 8.5 (s, C2–Me). 31P{1H} NMR (CDCl3): d )23.81
(s). Mass spectrum: (EI, m/z (%)): 710 (10, Mþ), 525 (5,
[M–PPh2]þ), 326 (100, Me2C9H4PPhþ2 ).
3
(CDCl3): d 7.80 (m, 2H, p-Ph), 7.68 (d, JHH ¼ 9 Hz,
H4), 7.46 (m, Ph), 7.2–7.0 (m, Ph and H5), 6.72 (dd,
3JHH ¼ 6, 8 Hz, H6), 6.41 (d, JHH ¼ 9 Hz, H7), 4.84 (s,
3
Minor isomer of 2d: 1H NMR (CDCl3): d 2.16 (s, 6H,
C3–Me), 1.34 (s, 6H, C2–Me). 31P{1H} NMR (CDCl3):
d )23.21 (s).
H3), 0.77 (s, Me). 13C{1H} NMR (CDCl3): d 137.16 (d,
1J ¼ 11 Hz, i-Ph), 137.06 (d, 1J ¼ 11 Hz, i-Ph), 134.97 (d,
2
2J ¼ 20 Hz, o-Ph), 133.40 (d, J ¼ 20 Hz, o-Ph), 128.93
(s, p-Ph), 128.5–128.0 (m-Ph, p-Ph, C4 and C7), 124.63
2
(s, C6), 124.23 (s, C5), 90.20 (d, J ¼ 22 Hz, C7a), 88.46
4.13. Preparation of bis(1-(diphenylphosphino)-4,7-di-
methyl-g5-indenyl)iron(II) (2e)
1
(s, C3a), 87.44 (s, C2), 66.36 (s, C3), 65.50 (d, J ¼ 15
Hz, C1), 9.36 (d, 3J ¼ 11 Hz, Me). 31P{1H} NMR
(CDCl3): d )20.04 (s). C44H36P2Fe: Calc.: C, 77.43; H,
5.32. Exptl: C, 76.19; H, 5.33. Mass spectruþm: (EI, m/z
(%)): 682 (40, Mþ), 314 (100, MeC9H6PPh2 ), 185 (83,
Ph2Pþ), 128 (25, MeC9H6þ), 108 (15, PhPþ). HR-MS:
Mþ Calc., 682.16413. Found, 682.16475.
Indene 1e (1.38 g, 4.2 mmol) was dissolved in THF
(40 ml), cooled to )78 °C, and BuLi (2.6 ml, 1.6 M, 4.2
mmol) added. The solution was then stirred for 2 h.
FeCl2 (0.27 g, 2.1 mmol) was added and the mixture
stirred for 2 h to give two product peaks in the 31P NMR
spectrum at d )16.46 and )21.47 in a 1:1 ratio. The
solvent was removed in vacuo. The solid was redissolved
in diethyl ether and filtered down a Celite column. The
mauve solution was concentrated and cooled to afford
the desired product 2e (1.11 g, 73%) as dark red crystals.
Although the crystals appeared to be identical, NMR of
separate samples of crystals gave varying ratios of the
rac and meso isomers. Random choosing of crystals
eventually allowed us to obtain X-ray structures of both
isomers. C46H40P2Fe: Calc.: C, 77.78; H, 5.68. Exptl: C,
77.63; H, 5.88.
Minor isomer of 2c, probably rac: 1H NMR (CDCl3):
d 7.9–6.4 (aromatic), 5.08 (s, 2H, H3), 1.86 (s, 6H, Me).
2
13C{1H} NMR (CDCl3): d 134.7 (d, J ¼ 20 Hz, o-Ph),
133.7 (d, 2J ¼ 20 Hz, o-Ph), 128.8 (s, p-Ph), 121.7 (s, C6),
121.2 (s, C5), 87.9 (s, C3a), 86.9 (s, C2), 68.6 (s, C3), 13.3
3
(d, J ¼ 10 Hz, Me). Signals for C1, C4, C7, C7a, i-Ph,
and meta-Ph were not resolved from those of the major
isomer. 31P{1H} NMR (CDCl3): d )17.60 (s).
4.12. Preparation of bis(1-(diphenylphosphino)-2,3-di-
methyl-g5-indenyl)iron(II) (2d)
1
Rac-2e: H NMR (CDCl3): d 7.5–6.7 (m, aromatic
To a solution of indene 1d (0.867 g, 2.64 mmol) in
THF (40 ml) at )78 °C was added a solution of n-BuLi
(1.65 ml, 1.6 M, 2.64 mmol). After warming to ambient
temperature and stirring for 2 h, FeCl2 (0.167 g, 1.32
mmol) was added and the reaction mixture was stirred
for a further 2 h to give a 2:5 ratio of two isomers ()23.3
and )23.7 ppm in the 31P NMR spectrum). Stirring of
the reaction solution for an additional 18 h gave no
change of the isomer ratio. The solvent was removed in
Hs), 6.45 (s, H5 and H6), 5.31 (d, 3JHH ¼ 2 Hz, H3), 3.13
(d, 3JHH ¼ 2 Hz, H2), 2.35 (s, Me), 2.10 (s, Me). 13C{1H}
NMR (CDCl3): d 140.71 (d, 1J ¼ 13 Hz, i-Ph), 136.82 (d,
2
1J ¼ 10 Hz, i-Ph), 135.22 (d, J ¼ 22 Hz, o-Ph), 132.09
3
2
(d, J ¼ 3 Hz, C7), 131.88 (d, J ¼ 18, o-Ph), 129.41 (s,
C4), 129.39 (s, p-Ph), 128.1–128.0 (m-Ph), 127.64 (s, p-
3
Ph), 122.50 (s, C6), 120.60 (s, C5), 93.06 (d, J ¼ 4 Hz,
2
2
C3a), 91.35 (d, J ¼ 18 Hz, C7a), 71.79 (d, J ¼ 5 Hz,
1
3
C2), 69.32 (d, J ¼ 14 Hz, C1), 65.22 (d, J ¼ 6 Hz, C3),

O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
1909
4
22.34 (d, J ¼ 20 Hz, Me(7)), 19.03 (s, Me(4)). 31P{1H}
4.16. Preparation of bis(1-(diisopropylphosphino)-g5-in-
denyl)iron(II) (2i)
NMR (CDCl3): d )16.86 (s).
1
Meso-2e: H NMR (CDCl3): d 7.5–6.7 (m, aromatic
3
H, H5 and H6), 4.02, (d, JHH ¼ 3 Hz, H3), 3.85 (d,
Indene 1i (0.833 g, 3.6 mmol) was dissolved in THF
(40 ml) and cooled to )78 °C. BuLi (2.2 ml, 1.6 M, 3.6
mmol) was then added and the solution stirred for 2 h.
Ferrous chloride (0.23 g, 1.8 mmol) was then added and
the reaction mixture stirred for 30 min to give a dark
green solution. The solvent was removed in vacuo and
the resulting black oily solid redissolved in CH2Cl2 and
filtered down a Celite column. The product was a black
oil and no yield was able to be calculated. Due to the
extreme sensitivity of this product no 1H, 13C NMR
spectra or microanalytical data were collected. 31P{1H}
NMR (CDCl3): d )3.27 (rac), )8.89 (meso).
3JHH ¼ 3 Hz, H2), 2.32 (s, Me), 2.14 (s, Me).
13C{1H}(CDCl3): d 141.10 (d, J ¼ 15 Hz, i-Ph), 137.73
1
1
3
(d, J ¼ 13 Hz, i-Ph), 136.41 (d, J ¼ 2 Hz, C7), 136.38
2
2
(d, J ¼ 22 Hz, o-Ph), 134.32 (s, C4), 131.73 (d, J ¼ 18
Hz, o-Ph), 129.25 (s, p-Ph), 128.1–128.0 (m-Ph), 127.47
3
(s, p-Ph), 124.27 (s, C6), 122.93 (s, C5), 92.68 (d, J ¼ 3
Hz, C3a), 90.50 (d, 2J ¼ 15 Hz, C7a), 73.78 (d, 2J ¼ 3 Hz,
1
C2), 68.28 (d, J ¼ 18 Hz, C1), 63.42 (s, C3), 22.78 (d,
4J ¼ 17 Hz, Me(7)), 19.45 (s, Me(4)). 31P{1H}(CDCl3): d
)22.17 (s).
4.14. Attempted preparation of bis(1-(diphenylphosph-
ino)-3,4,7-trimethyl-g5-indenyl)iron(II) (2f)
4.17. Preparation of {4,7-dimethyl-1-(diphenylphosph-
ino)-g5-indenyl}{1-(diphenylphosphino)-g5-indenyl}iron-
(II) (2j)
Numerous attempts were made to prepare this fer-
rocene using similar routes used to make the other fer-
rocenes, however, 31P NMR of the brown product
solutions showed mostly free ligand with only some
minor peaks in the )20 ppm region.
To a solution of 1a (0.335 g, 1.12 mmol) and 1e (0.366
g, 1.12 mmol) in THF (35 ml) at )78 °C was added a
solution of n-BuLi (1.40 ml, 1.6 M, 2.24 mmol). After 2
h, FeCl2 (0.142 g, 1.12 mmol) was added and the reac-
tion mixture stirred for 2 h at ambient temperature. 31P
NMR indicates the presence of ten unique phosphorous
environments (one peak for each of the free ligands in
addition to eight signals arising from the ferrocenes: two
for each of rac- and meso-2j and one for each of rac- and
meso-2a as well as rac- and meso-2e). The solution was
stirred for a further 24 h, before the solvent was re-
moved in vacuo. The residue was dissolved in CH2Cl2
(20 ml) and filtered through Celite. Removal of the
solvent in vacuo yielded a black powder (0.64 g) that
was shown to contain, by 31P NMR, a mixture of rac-
and meso-2j, rac- and meso-2e, and rac-2a in a ratio of
4:3:1:2:1. All of the meso-2a had isomerized to rac-2a,
but after that there was no other significant variation in
these ratios with stirring of the mixture in THF for 3
days. An EI mass spectrum of the mixture showed a
parent-ion peak for the mixed ferrocene 2j at 682 m/z.
4.15. Preparation of bis(1-(diphenylphosphino)-3-(tri-
methylsilyl)-g5-indenyl)iron(II) (2h)
To a solution of indene 1h (0.803 g, 2.16 mmol) in
THF (35 ml) at )78 °C was added a solution of n-BuLi
(1.35 ml, 1.6 M, 2.16 mmol). After warming to ambient
temperature and stirring for 5 h, FeCl2 (0.137 g, 1.08
mmol) was added and the reaction mixture stirred for a
further 2 h to give two product isomers in a 3:2 ratio
()25.3 and )27.3 ppm in the 31P NMR spectrum). The
solvent was removed in vacuo and the green, oily residue
loaded onto an alumina column. Elution, in the first
instance with petroleum ether, followed by subsequent
elution with 1:1 CH2Cl2/petroleum ether yielded a green
oily residue, which contains both 1a, resulting from loss
of SiMe3, and 2h. The residue was dissolved in petro-
leum ether, filtered, and the solvent removed in vacuo to
yield 2h (0.23 g, 31%) as a green solid containing only
the isomer with the upfield chemical shift. Stirring the
reaction solution overnight only gave decomposition
products. The product proved too unstable for micro-
analysis or mass spectroscopy.
1
Rac-2j: H NMR (CDCl3): d 5.34 (s, 1H, H3 (1e)),
4.98 (s, 1H, H3 (1a)), 3.27 (s, 1H, H2 (1e)), 3.11 (s, 1H,
H2 (1a)). 31P{1H} NMR (CDCl3): d )16.81 (s, 1e),
)22.20 (s, 1a).
1
Meso-2j: H NMR (CDCl3): d 4.07 (s, 1H, H3 (1e)),
3.90 (s, 1H, H2 (1e)), 3.85 (s, 1H, H3(1a)), 3.55 (s, 1H,
H2 (1a)). 31P{1H} NMR (CDCl3): d )23.9 (s, 1e),
)25.61 (s, 1a).
1H NMR (CDCl3): d 7.54–6.77 (m, 28H, H4–7 and
Ph), 3.84 (s, 2H, H2), 0.23 (s, 18H, Si(CH3)3). 13C{1H}
NMR (C6D6): d 138.9 (d, 1J ¼ 15 Hz, ipso- Ph), 137.5 (d,
2
1J ¼ 14 Hz, ipso-Ph), 135.4 (d, J ¼ 20 Hz, o-Ph), 132.9
4.18. X-ray structure determinations for rac-2a, rac-2b,
rac-2d, rac-2e and meso-2e
2
(d, J ¼ 21 Hz, o-Ph), 129.3 (s, p-Ph), 128.9–127.5 (m,
C7, m-Ph and p-Ph), 124.8 (s, C4), 124.2 (s, C6), 123.8 (s,
C5), 94.1 (s, C3a), 92.7 (d, 2J ¼ 16 Hz, C7a), 78.7 (s, C2),
70.8 (d, 1J ¼ 13 Hz, C1), 64.6 (s, C3), 0.8 (s, Me).
31P{1H} NMR (CDCl3): d )28.49 (br, s).
Crystal data and experimental details are given in
Table 2. For each compound, a crystal was attached to a
thin glass fiber and mounted on a Siemens P4 SMART

1910
O.J. Curnow et al. / Journal of Organometallic Chemistry 689 (2004) 1897–1910
[2] X. Zhang, US Patent number 6,534,657.
[3] J.J. Adams, D.E. Berry, J. Browning, D. Burth, O.J. Curnow, J.
Organomet. Chem. 580 (1999) 245.
diffractometer with a Siemens CCD area detector.
Multi-scan absorption corrections were determined with
SADABS and applied to the data [24]. Data processing
was undertaken with SAINT [24] and the structures were
solved by direct methods and refined by least-squares
methods on F 2 using the SHELXTL program library [25].
Hydrogen atoms were placed in their calculated posi-
tions and refined isotropically riding with the atoms to
which they are bonded. Non-hydrogen atoms were re-
fined anisotropically, except for rac-2d for which they
were refined isotropically.
[4] G.M. Fern, S. Klaib, O.J. Curnow, H. Lang, J. Organomet.
Chem. 689 (2004) 1139.
[5] E. Barday, B. Frange, B. Hanquet, G.E. Herberich, J. Organomet.
Chem. 572 (1999) 225.
[6] H. Plenio, D. Burth, Organometallics 15 (1996) 4054.
[7] M.E. Fessler, J.T. Spencer, J.F. Lomax, R.N. Grimes, Inorg.
Chem. 27 (1988) 3069.
ꢂ
[8] F.M. Alıas, S. Barlow, J.S. Tudor, D. O’Hare, R.T. Perry, J.M.
Nelson, I. Manners, J. Organomet. Chem. 528 (1997) 47.
[9] P. Scott, U. Rief, J. Diebold, H.H. Brintzinger, Organometallics
12 (1993) 3094.
[10] (a) S.A. Westcott, A.K. Kakkar, G. Stringer, N.J. Taylor, T.B.
Marder, J. Organomet. Chem. 394 (1990) 777;
(b) J. Trotter, Acta Crystallogr. 11 (1958) 355.
[11] H. Hope, F.-W. Raulfs, D. Schomburg, Tetrahedron 42 (1986)
1655.
5. Supplementary material
Crystallographic data for the structural analyses have
been deposited with the Cambridge Crystallographic
Data Centre: CCDC No. 229441 for rac-2a; CCDC No.
229345 for rac-2b; CCDC No. 229346 for rac-2d; CCDC
No. 229343 for rac-2e; and CCDC No. 229344 for meso-
2e. Copies of the information may be obtained free of
charge from: The Director, CCDC, 12 Union Road,
Cambridge, CB2 1EZ UK. Fax: +44-1223-336-033 or by
Email at deposit@ccdc.cam.ac.uk or from the www at
http://www.ccdc.cam.ac.uk.
[12] H. Schumann, O. Stenzel, S. Dechert, R.L. Haltermann, Organo-
metallics 20 (2001) 1983.
€
[13] H.A. Buchholz, J. Hofer, M. Noltemeyer, A. de Meijere, Eur. J.
Org. Chem. (1998) 1763.
[14] O.J. Curnow, G.M. Fern, Organometallics 21 (2002) 2827.
[15] O.J. Curnow, G.M. Fern, M.L. Hamilton, A. Zahl, R. van Eldik,
Organometallics 23 (2004) 906.
[16] K.A. Fallis, G.K. Anderson, N.P. Rath, Organometallics 11
(1992) 885.
[17] M. Stradiotto, C.M. Kozak, M.J. McGlinchey, J. Organomet.
Chem. 564 (1998) 101.
[18] C. Lensink, G.J. Gainsford, Aust. J. Chem. 51 (1998) 667.
[19] J.J. Adams, D.E. Berry, O.J. Curnow, G.M. Fern, M.L. Hamil-
ton, H.J. Kitto, J.R. Pipal, Aust. J. Chem. 56 (2003) 1155.
[20] C. Janiak, J. Chem. Soc., Dalton Trans. (2000) 3885.
[21] Review and overviews on CH/p interactions: (a) D. Braga, F.
Grepioni, E. Tedesco, Organometallics 17 (1998) 2669;
(b) M.A. Viswamitra, R. Radhakrishnan, J. Bandekar, G.R.
Desiraju, J. Am. Chem. Soc. 115 (1993) 4868;
Acknowledgements
MLH thanks the University of Canterbury for a
Summer Research Scholarship.
References
(c) M. Nishio, M. Hirota, Y. Umezawa, The CH/p Interaction
Evidence Nature and Consequences, Wiley-VCH, New York,
1998;
[1] (a) G. Wagner, R. Herrmann, in: A. Togni, T. Hayashi (Eds.),
Ferrocenes: Homogeneous Catalysis, Organic Synthesis, Materials
Science, VCH, Weinheim, 1995, pp. 173–218;
(d) Y. Umezawa, S. Tsuboyama, K. Honda, J. Uzawa, M. Nishio,
Bull. Chem. Soc. Jpn. 71 (1998) 1207;
(b) A. Togni, in: A. Togni, R.L. Halterman (Eds.), Metallocenes:
Synthesis Reactivity Applications, Vol. 2, Wiley-VCH, Weinheim,
1998, pp. 685–722;
(e) M.J. Calhorda, Chem. Commun. (2000) 801;
(f) G.R. Desiraju, T. Steiner, The Weak Hydrogen Bond (IUCr
Monograph on Crystallography 9), Oxford Science Publications,
Oxford, 1999.
(c) T. Hayashi, in: A. Togni, T. Hayashi (Eds.), Ferrocenes:
Homogeneous Catalysis, Organic Synthesis, Materials Science,
VCH, Weinheim, 1995, pp. 105–142;
[22] N.E. Grimmer, N.J. Coville, C.B. de Koning, J.M. Smith, L.M.
Cook, J. Organomet. Chem. 616 (2000) 112.
(d) A. Togni, Angew. Chem. Int. Ed. Engl. 35 (1996) 1475;
(e) L.-X. Dai, T. Tu, S.-L. You, W.-P. Deng, X.-L. Hou, Acc.
Chem. Res. 36 (2003) 659;
[23] R.L. Halterman, D.R. Fahey, E.F. Bailly, D.W. Dockter, O.
Stenzel, J.L. Shipman, M.A. Khan, S. Dechert, H. Schumann,
Organometallics 19 (2000) 5464.
(f) T.J. Colacot, Chem. Rev. 103 (2003) 3101;
(g) P.J. Walsh, A.E. Lurain, J. Balsells, Chem. Rev. 103 (2003)
3297;
[24] SAINT and SADABS
1994.
,
Siemens Analytical, Madison, WI,
[25] G.M. Sheldrick, SHELXTL ver 5.1, Bruker AXS, Madison, WI,
1998.
(h) W. Tang, X. Zhang, Chem. Rev. 103 (2003) 3029.