Synthesis of poly(ferrocenyldialkoxysilane) polymers and their introduction into oxide matrices

Article abstract of DOI:10.1016/S0022-328X(00)00749-X

The synthesis of 1,1′-ferrocenediyldialkoxysilanes Fe(η-C₅H₄)₂Si(OR)₂ M2-M5 (R = i-Pr, t-Bu, PhCH₂, Me), consisting of the reaction between 1,1′-ferrocenediyldichlorosilane with the selected alcohols, is described. The thermal ring opening polymerisation of M2-M5 yielded the corresponding poly(ferrocenylsilane) polymers P2-P5. These linear polymers were soluble in polar solvents and exhibited high molecular weights. The monomers M2 and M4 were introduced into a silica matrix by a non-hydrolytic sol-gel process. However, Si-C bond breaking was observed due to the electrophilic ipso substitution. Polymers P3 and P4 were easily condensed with SiBr₄. When P3 was condensed with TiCl₄ and AlCl₃, degradation of the polymer occurred. Interestingly, the hydrolytic sol-gel polycondensation of polymers P2, P4 and P5 yielded a completely polycondensed hybrid solid, without significant specific surface area and with conservation of the electrochemical properties.

Full text of DOI:10.1016/S0022-328X(00)00749-X

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 621 (2001) 46–54
Synthesis of poly(ferrocenyldialkoxysilane) polymers and their
introduction into oxide matrices
Ge´rard Calle´ja, Genevie`ve Cerveau *, Robert J.P. Corriu
Laboratoire de Chimie Mole´culaire et Organisation du Solide, UMR 5637, Uni6ersite´ Montpellier II, cc 007, Place E. Bataillon,
34095 Montpellier Cedex 5, France
Received 18 July 2000; received in revised form 15 September 2000
Abstract
The synthesis of 1,1%-ferrocenediyldialkoxysilanes Fe(h-C5H4)2Si(OR)2 M2–M5 (R=i-Pr, t-Bu, PhCH2, Me), consisting of the
reaction between 1,1%-ferrocenediyldichlorosilane with the selected alcohols, is described. The thermal ring opening polymerisation
of M2–M5 yielded the corresponding poly(ferrocenylsilane) polymers P2–P5. These linear polymers were soluble in polar
solvents and exhibited high molecular weights. The monomers M2 and M4 were introduced into a silica matrix by a
non-hydrolytic sol–gel process. However, SiꢀC bond breaking was observed due to the electrophilic ipso substitution. Polymers
P3 and P4 were easily condensed with SiBr4. When P3 was condensed with TiCl4 and AlCl3, degradation of the polymer occurred.
Interestingly, the hydrolytic sol–gel polycondensation of polymers P2, P4 and P5 yielded a completely polycondensed hybrid
solid, without significant specific surface area and with conservation of the electrochemical properties. © 2001 Elsevier Science
B.V. All rights reserved.
Keywords: 1,1%-Ferrocenediyldialkoxysilanes; Poly(ferrocenylsilanes); Non-hydrolytic sol–gel condensation; Sol–gel process
1. Introduction
characteristic two-wave cyclic voltammogram, has been
reported [16–21]. Thus we have studied the presence
and the possible evolution of this specific property in
the solid. In a preliminary communication we have
described the possibility of introducing poly(ferrocenyl-
dialkoxysilanes) into silica matrices by a non-hydrolytic
sol–gel process [22]. This route [23,24] permits access to
oxides other than SiO2 and corresponds to a condensa-
tion reaction between a metal halide and a metal alkox-
ide (Scheme 1).
We report here the synthesis of new poly(ferrocenyl-
dialkoxysilane) polymers containing polycondensable
alkoxide groups and the attempts at their inclusion in
oxides. These polymers were obtained by ring-opening
polymerisation (ROP) from the corresponding
monomers. Until now, substituents introduced at the
silicon in strained ferrocenophanes were limited to alkyl
and aryl groups [1,25–27]. Only recently has the syn-
thesis of new ferrocenyldialkoxysilanes been achieved
[22,28]. The selected alkoxide groups allowed polycon-
densation via non-hydrolytic and hydrolytic sol–gel
processes, thus enabling their introduction into various
inorganic matrices, such as SiO2, TiO2 and Al2O3.
In recent decades, polymers containing transition
metals in the main chain have attracted considerable
attention owing to their physical (redox, magnetic, elec-
trical, etc.) and chemical properties [1].
In the field of nanostructured organic inorganic ma-
terials obtained by sol–gel routes, it is possible to
introduce organic [2–5] or organometallic [6–8] moi-
eties into a silica matrix. In these solids the organic or
organometallic units are covalently bound to each other
through covalent SiꢀC bonds. In such materials the
properties of the molecular organic or organometallic
moiety are generally preserved in the material [9–15].
In order to verify the generality of this observation, we
have been particularly interested by the recently de-
scribed class of organometallic polymers that contain
ferrocene units in the chain [1]. They are very attractive,
since a very specific redox cooperation between the iron
centres along the polymer backbone, showing a highly
* Corresponding author. Tel.: +33-4-67143801; fax: +33-4-
67143852.
E-mail address: gcerveau@crit.univ-montp2.fr (G. Cerveau).
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00749-X

G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
47
Scheme 1. Non-hydrolytic sol–gel process.
2. Results and discussion
of M1 with various alcohols in the presence of triethyl-
amine in pentane or THF at room temperature resulted
in a nucleophilic substitution of chlorine at silicon,
leading to M2–M5 without cleavage of the bridge. The
new compounds were isolated in good yields (66–87%)
as orange–red solids by recrystallisation of the crude
products. M2 was a liquid at room temperature (m.p.
5–10°C). Compounds M2–M5 were characterised by
1H-, 13C- and 29Si-NMR spectroscopy. In each case the
1H-NMR spectra exhibited the signals corresponding to
the hydrogen of the cyclopentadienyl ring in the range
between 4 and 5 ppm and the signals normally expected
for the hydrogen of alkoxide groups attached to silicon.
The 13C-NMR spectra showed the cyclopentadienyl
carbon atoms in the regions 40–43 ppm and 75–
78 ppm. The signal at ꢀ41 ppm was characteristic of
the ipso-cyclopentadienyl carbon atom bound to the
bridging silicon atom. The high-field shift observed was
directly related to the strained cyclic structure of the
1,1%-ferrocenediyldialkoxysilanes, and is in agreement
with the previously reported results described in the
case of the alkyl and aryl analogues [17]. The signals
corresponding to the carbon atoms from the alkoxide
groups were easily identified in the spectra. The 29Si-
NMR spectra showed a single signal in all cases. The
chemical shifts observed depended strongly on the na-
ture of the substituent and ranged between −31.4 ppm
for M5 and −52.7 ppm in the case of M3. Elemental
analysis gave satisfactory values in all cases. Spectro-
scopic data for M2–M5 are listed in Table 1.
2.1. Synthesis and characterisation of
1,1%-ferrocenediyldialkoxysilanes M2–M5
The preparation of monomers M2–M5 (Scheme 2)
was studied. Two routes were investigated to prepare
M2–M5.
Route 1, previously described in the literature [16,29]
for the synthesis of the alkyl and aryl analogues, con-
sists of the reaction between 1,1%-dilithioferrocene and
the selected dichlorodialkoxysilanes, and results in a
mixture of monomers and small oligomers in the case
of M2 and M3 (reaction in Eq. (1)).
(1)
Pure M2 and M3 were isolated in 30% and 40% yields
respectively. The monomers M4 and M5 could not be
obtained by this route: only small oligomers with a high
polydispersity index, identified by size exclusion chro-
matography (SEC) in THF, were isolated. This was
revealed to be independent of the order of addition of
the two reactants, as well as of the reaction tempera-
ture. Moreover, the required dialkoxydichlorosilanes
were not easily available.
The monomers M2–M5 were obtained by route 2,
which consists of the reaction between 1,1%-fer-
rocenediyldichlorosilane M1 and the selected alcohol
(reaction in Eq.(2)).
The differential scanning calorimetry (DSC) analysis
of compounds M2–M5 was performed at a heating rate
of 5°C min−1 in air. A sharp endothermic signal fol-
lowed by a broad exothermic band characterised the
spectra in all cases. The endothermic peak corre-
sponded to the melting of the monomer and the
exotherm was attributed to the ROP. In the case of M3,
(2)
M1 was prepared according to the route originally
reported by Wrighton et al. [30]: a reaction between
1,1%-dilithioferrocene/TMDA and SiCl4. The treatment
Scheme 2. Monomers M2–M5.

48
G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
Table 1
Physical and NMR data for monomers M2–M5
Compound Yielda
M.p. (°C)
Exotherm (°C)
NMR spectroscopyb
1H
13C
29Si
M2
M3
M4
M5
30% (route 1)
70% (route 2)
5–10
190–210
d 1.4 (6H, CH3)
t 4.2 (4H, Cp)
t 4.6 (4H, Cp)
h 4.7 (1H, CH)
25.9 (CH3)2CH
41.3 (Cipso Cp)
65.5 (CH3)2CH
75.4, 77.6 (Cp)
−39.2
40% (route 1)
66% (route 2)
91–93
48–50
45
230–270
180–200
130–150
s 1.6 (18H, tBu)
t 4.3 (4H, Cp)
t 4.6 (4H, Cp)
32.4 (CH3)3C
43.8 (Cipso Cp)
74.3 (CH3)3C
75.3, 76.9 (Cp)
−52.7
−32.9
−31.1
68% (route 2)
70% (route 2)
t 4.2 (4H, Cp)
t 4.6 (4H, Cp)
s 5.2 (4H, CH2)
m 7.4 (10H, Ph)
40.1 (Cipso Cp)
64.8 (C6H5CH2)
75.6, 78.2 (Cp)
127.3, 127.8, 128.8, 140.6 (C6H5CH2)
s 3.8 (6H, CH3)
t 4.2 (4H, Cp)
t 4.6 (4H, Cp)
40.4 (Cipso Cp)
50.5 (CH3)
75.7, 78.3 (Cp)
a Overall yield (from dilithioferrocene).
b Chemical shifts (ppm) relative to Me4Si in CD2Cl2 (Cp=cyclopentadienyl).
which contains bulky tert-butoxide groups, the ROP
temperature was higher (230–270°C) than in the other
cases. The melting points and ROP temperatures are
reported in Table 1. The DSC thermogram of M4 is
presented in Fig. 1.
high and ranged from 206 000 to 360 000 (the detection
limit). In the case of P3 and P4 the polydispersity
indexes were low, which is indicative of a narrow
distribution of molecular weights.
1
The H-, 13C- and 29Si-NMR data of polymers P2–
The molecular structure of M3, studied by single-
crystal X-ray diffraction [22], shows the expected
strained [26,27,31] structure with a tilt angle between
the planes of the cyclopentadienyl ligands of 20.3°.
P5 were in agreement with the expected spectra. In each
case the 1H-NMR spectra exhibited the hydrogen
atoms of the cyclopentadienyl rings at 4.2 and 4.4 ppm
and the signals corresponding to the hydrogen of the
alkoxide groups. The 13C-NMR spectra exhibited the
cyclopentadienyl carbon atom signals in the region
66–73 ppm. It is interesting to note that the ipso-cy-
clopentadienyl carbon atom bound to silicon shifts
downfield (ꢀ66–69 ppm) compared with the corre-
sponding strained monomers (ꢀ40–44 ppm).
2.2. Synthesis and characterisation of
poly( ferrocenylsilanes) P2–P5
The ROP process (reaction in Eq. (3)) of M2–M5
has been evidenced by DSC experiments.
The 29Si-NMR resonances for polymers P2–P5 were
shifted downfield compared with the corresponding
strained monomers. The values are reported in Table 2.
(3)
The corresponding new polymers P2–P5 were obtained
thermally according to the method already described
[16,32,33]. The polymerisation temperatures were de-
duced from the DSC analysis in each case and are given
in Table 2.
The polymers P2–P5 were obtained as orange–red
solids. They were soluble in polar solvents such as THF
and dichloromethane. The molecular weights Mw were
determined by SEC in THF using polystyrene stan-
dards. The values reported for Mw in Table 2 were very
Fig. 1. DSC thermogram of M4.

G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
49
Table 2
Physical and NMR data for polymers P2–P5
a
Compound
Polym. temp. (°C)
Mw
NMR spectroscopyc
1H
13C
29Si
−24.8
P2
190
\360 000b
d 1.3 (6H, CH3)
t 4.2 (4H, Cp)
h 4.3 (1H, CH)
t 4.4 (4H, Cp)
26.1 (CH3)2CH
67.5 (Cipso Cp)
65.4 (CH3)2CH
72.7, (Cp)
P3
P4
P5
230
180
130
280 000
Ip=1.6
s 1.4 (18H, tBu)
t 4.2 (4H, Cp)
t 4.4 (4H, Cp)
32.4 (CH3)3C
71.2 (Cipso Cp)
72.7 (CH3)3C
72.9, 74.6 (Cp)
−36.5
−18.5
−15
290 000
Ip=2.0
t 4.3 (4H, Cp)
t 4.5 (4H, Cp)
s 5.0 (4H, CH2)
m 7.4 (10H, Ph)
65.4 (C6H5CH2)
65.9 (Cipso Cp)
72.9, 74.4 (Cp)
126.8, 127.4, 128.6, 141.4 (C6H5CH2)
206 000
Ip=17
s 3.7 (6H, CH3)
t 4.2 (4H, Cp)
t 4.4 (4H, Cp)
51.4 (CH3)
68.6 (Cipso Cp)
72.2, 73.9 (Cp)
a Determined by SEC in THF versus polystyrene standards.
b Upper limit detection surpassed, Ip value not avalaible.
c Chemical shifts (ppm) relative to Me4Si in CDCl2 (Cp=cyclopentadienyl).
2.3. Introduction of monomers M2 and M4 into SiO2
by a non-hydrolytic sol–gel process
signals corresponding to residual alkoxide groups, at
20–40 ppm for GM2Si (isopropyl carbon atoms) and
129 and 141 ppm for GM4Si (aromatic carbon atoms).
An unexpected signal at 138 ppm was also detected in
the case of GM2Si, which has been attributed to the
possible presence of cyclopentadienyl units arising from
the degradation of the ferrocenyl moieties. The 29Si-CP-
MAS-NMR spectra exhibited the two signals normally
expected at −25 ppm and −100 ppm, corresponding
respectively to the D (C2ꢀSi(OX)2, X=i-Pr, CH2Ph, Si)
[34,35] and Q (Si(OX)4) substructures [36]. An addi-
tional broad signal centred at −60 ppm was observed
and could be assigned to the T units (CꢀSi(OX)3)
already reported in the case of 1,1%-ferrocenyl-
silsesquioxanes [6]. The formation of T substructures
could be provided by the CpꢀSiCp bridge breaking that
occurs during the condensation process. The silicon–
ferrocene bond cleavage may be explained by a Friedel
and Crafts type of electrophilic substitution reaction at
the ipso position, the alkylating agent being the alkyl
halide formed during the condensation reaction and the
catalyst the silicon halide (SiBr4 or SiCl4). Such reac-
tions would lead to the silicon–carbon bond cleavage
and to the formation of an alkylferrocenylsilane precur-
sor in a T-type configuration with release of alkyl
halide (Scheme 3a). Another possibility might involve a
concerted process (Scheme 3b). Similar reactions have
already been reported in the case of aryl precursors
[37].
This method consists of the condensation between
alkoxide and halide functions as outlined in Scheme 1.
In this context, we investigated the condensation of
1,1%-ferrocenedialkoxysilanes M2 and M4 with silicon
halides according to the reaction in Eq. (4).
(4)
The solids GM2Si and GM4Si were isolated as brown
powders, which were filtered off and washed before
characterisation. The formation of alkyl halide was
evidenced by 1H-NMR spectroscopy analysis of the
filtrate. This phase and the washing solutions were very
coloured, due to the presence of ferrocene derivatives.
A loss of iron was evidenced by elemental analysis of
GM2Si and GM4Si. The 13C-CPMAS-NMR spectra
exhibited the cyclopentadienyl carbon atoms signals in
the region 65–80 ppm in both cases, as well as small
The solids GM2Si and GM4Si presented no signifi-
cant specific surface area (SBETB10 m2 g−1).

50
G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
2.4. Introduction of polymers into inorganic matrices
Polymers P3 and P4 were condensed with SiBr4
according to the reaction shown in Eq. (5).
(5)
Fig. 2. Cyclic voltammograms in 0.1 M (NBu4)BF4/CH2Cl2 of P3 and
GP3Si.
The 29Si-CPMAS-NMR spectra of GP3Si and GP4Si
were very similar. They presented three broad reso-
nances in the ranges −20 to −40 ppm, −65 to
−80 ppm and −90 to −100 ppm assigned respec-
tively to D, T and Q silicon substructures. The presence
of T units might arise from the cleavage of silicon–car-
bon bonds involving the rupture of the polymeric
chain, according to a mechanism similar to that de-
scribed in Scheme 3. The elemental analysis was consis-
tent with the loss of ferrocene moieties.
The corresponding solids GP3Si and GP4Si were iso-
lated as brown powders.
In the case of GP3Si, the 13C-CPMAS-NMR spec-
trum exhibited a weak signal at 31 ppm attributed to
carbon atoms of the remaining tert-butoxide groups
and another one at 73 ppm due to the expected cy-
clopentadienyl carbon atoms. Three resonances were
observed for GP4Si in the range 60–80 ppm (benzylic
and cyclopentadienyl carbon atoms), 129 ppm and
139 ppm (aromatic carbon atoms).
These solids presented no significant specific surface
area (SBETB10 m2 g−1).
Preliminary results concerning the electrochemical
behaviour of the polymer P3 and the hybrid solid
GP3Si showed that, as previously observed [9,10,13–
15], the hybrid gel maintained the properties of the
incorporated organometallic units despite the high sen-
sitivity of the electrochemical interactions between the
ferrocenyl units to the distance (Fig. 2).
Attempts to introduce P3 in titanium oxide and
aluminium oxide matrices were performed by co-con-
densing the polymer with TiCl4 and AlCl3 respectively
(Eq. (5)). GP3Ti was isolated as a brown solid, whereas
GP3Al was a grey powder. The elemental analysis of
both samples indicated that the polymer was destroyed
and iron was eliminated partially (GP3Ti) or totally
(GP3Al) from the solids. Such a result was attributed to
the strong Lewis acidity of the titanium or aluminium
halides. The 29Si-CPMAS-NMR analyses were in agree-
ment with the degradation of the organometallic moi-
ety, since only a broad signal in the range −85 to
−110 ppm was observed, which corresponded to Q
units of SiO4/2. The 13C-CPMAS-NMR spectra showed
very low resolution. In the case of GP3Al, the 27Al-
MAS-NMR spectrum (Fig. 3) revealed three resonances
which could be attributed to 27Al sites in coordination
numbers 6, 5 and 4, as already observed in the prepara-
tion of alumina and aluminosilicates by non-hydrolytic
sol–gel methods [38,39].
GP3Ti presented no significant specific surface area,
whereas GP3Al exhibited a surface of 87 m2 g−1; this
could be related to the presence of silica in the solid.
Scheme 3. Formation of T substructures in the solid: (a) with release
of alkyl halide; (b) concerted process.

G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
51
were avoided by further Soxhlet extraction with THF).
The resulting orange powders were analysed by spectro-
scopic techniques. GP2H, GP4H and GP5H exhibited
the same spectra: a single signal at 68–72 ppm was
detected in the 13C-CPMAS-NMR spectra. Remark-
ably, no signals corresponding to the carbon atoms
from residual alkoxide groups were observed, indicating
a total polycondensation of the hybrid solids.
The 29Si-CPMAS-NMR analysis corroborated this
assumption; only one thin resonance at −30 ppm was
detected, which can be assigned to D2 substructures
(C2Si(OSi)2) [34,35]. These solids exhibited no signifi-
cant specific surface areas. It is interesting to note that
the solids GP2H, GP4H and GP5H appeared to be
very similar, in that they presented the same polycon-
densation at silicon and the same textural properties
with conservation of the ferrocenyl units even though
the precursors contained different hydrolysable alkox-
ide groups. Such behaviour is very different from that
observed during the hydrolysis of trialkoxysilyl precur-
sors: the kinetic parameters, and particularly the nature
of the leaving group at silicon, have a major influence
on the properties of the resulting solids [5,35]. Appar-
ently, the hydrolytic sol–gel route allowed a single
highly polycondensed organic inorganic hybrid material
to be obtained whatever the alkoxide group at silicon.
A comparison between the cyclic voltammograms of
P5 and GP5H showed the conservation of the electro-
chemical properties of the polymer in the hybrid solid
(Fig. 4).
Fig. 3. 27Al-MAS-NMR spectrum of GP3Al.
3. Conclusion
We have reported here a convenient synthetic route
to new 1,1%-ferrocenediyldialkoxysilanes via a chlorine
substitution reaction at the bridging silicon atom of
1,1%-ferrocenediyldichlorosilane with different alcohols
in the presence of triethylamine. Thermal ROP of these
species provided access to new poly(ferrocenylsilanes)
polymers. The incorporation of these monomers into
inorganic oxide matrices by a non-hydrolytic sol–gel
process was only possible in the case of SiO2. However,
SiꢀC bonds were cleaved during a side reaction, which
consisted in an electrophilic substitution at the ipso
position of the alkyl halide formed during the conden-
sation reaction. A similar result was obtained in the
case of the polymers. A significant degradation of the
polymers occurred when TiCl4 and AlCl3 were em-
ployed. In contrast, the hydrolytic sol–gel polyconden-
sation of the polymers led to highly polycondensed
hybrid materials without any cleavage of SiꢀC bonds.
In the case of SiO2 as matrix, the electrochemical
properties of the ferrocenyl moieties were preserved in
the hybrid solids. Work is currently in progress in order
to study the electrochemical properties of the different
hybrid solids.
Fig. 4. Cyclic voltammograms in 0.1 M (NBu4)BF4/CH2Cl2 of P5 and
GP5H.
2.5. Hydrolytic sol–gel processing of polymers P2–P5
The hydrolysis polycondensation of polymers P2–P5
was performed under nucleophilic catalysis conditions
using fluoride ions (reaction in Eq. (6)).
(6)
Orange insoluble solids were obtained in all cases ex-
cept for P3, which did not react because of the steric
hindrance of the tert-butoxide group. After filtration
they were washed with ethanol, acetone and ether
(traces of remaining unreacted polymer in the solids

52
G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
4. Experimental
5.2 (4H, s, CH2), 7.4 (10H, m, Ph). 13C-NMR (l,
CD2Cl2): 40.1 (Cipso on Cp), 64.8 (C6H5CH2), 75.6,
78.2 (Cp), 127.3, 127.8, 128.8, 140.6 (C6H5CH2). 29Si-
NMR (l, CD2Cl2): −32.9.
All reactions were carried out under argon using a
vacuum line and Schlenk techniques. Solvents were
dried and distilled just before use. Melting points were
determined with a Gallenkamp apparatus and are un-
corrected. IR data were obtained on a Perkin–Elmer
1600FTIR spectrophotometer. The solution 1H- and
13C-NMR spectra were recorded on a Bruker DPX-200
spectrometer and the solution 29Si-NMR spectra were
recorded on a Bruker WP-200 SY spectrometer. Solid-
state CPMAS-NMR spectra were obtained with a
Bruker FT AM 300 spectrometer: 13C-CPMAS-NMR
at 75.47 MHz, recycling delay 5 s and contact time
5 ms; 29Si-CPMAS-NMR at 59.62 MHz, recycling delay
10 s and contact time 2 ms. The spinning rate was
5000 Hz in all cases. Chemical shifts are given relative
to tetramethylsilane. 27Al-MAS-NMR spectra were ob-
tained with a Bruker FT AM 400 spectrometer at
104.25 MHz, pulse angle p/12 and recycling delay 1 s.
The spinning rate was 10 000 Hz. Chemical shifts were
referenced to Al(H2O)63+. Specific surface areas were
determined using a Micromeritics Gemini III 2375 ap-
paratus. The DSC analyses were carried out on a
Micromeritics Accupyc 1330 apparatus. The size-exclu-
sion chromatographs were recorded on a Millipore
Waters model 510, equipped with a Waters 441, UV
detector. Elemental analyses were carried out by the
‘Service Central de Micro-Analyse du CNRS’.
M2. Dark-red liquid, (88%), m.p. 5–10°C. Anal.
Found: C, 57.02; H, 6.72; Fe, 14.95; Si, 8.30. Calc. for
C16H22Fe1O2Si1: C, 58.18; H, 6.67; Fe, 16.97; Si, 8.48%.
1H-NMR (l, CD2Cl2): 1.4 (6H, d, CH3), 4.2 (4H, t,
Cp), 4.6 (4H, t, Cp), 4.7 (1H, h, CH).13C-NMR (l,
CD2Cl2): 25.9 ((CH3)2CH), 41.3 (Cipso on Cp), 65.5
((CH3)2CH), 75.4, 77.6 (Cp). 29Si-NMR (l, CD2Cl2):
−39.2.
M3. Deep-red solid, (95%), m.p. 92–93°C. Anal.
Found: C, 60.05; H, 7.26; Fe, 15.40; Si, 7.95. Calc. for
C18H26Fe1O2Si1: C, 60.34; H, 7.26; Fe, 15.64; Si, 7.82%.
1H-NMR (l, CD2Cl2): 1.6 (18H, s, tBu), 4.3 (4H, t,
Cp), 4.6 (4H, t, Cp). 13C-NMR (l, CD2Cl2): 32.4
((CH3)3C), 43.8 (Cipso on Cp), 74.3 ((CH3)3C), 75.3,
76.9 (Cp). 29Si-NMR (l, CD2Cl2): −52.7.
M5. Deep-red solid, (95%), m.p. 45°C. Anal. Found:
C, 52.52; H, 5.24; Fe, 17.80; Si, 9.60. Calc. for
C12H14Fe1O2Si1: C, 52.55; H, 5.11; Fe, 20.44; Si,
1
10.22%. H-NMR (l, CD2Cl2): 3.8 (6H, s, CH3), 4.2
(4H, t, Cp), 4.6 (4H, t, Cp). 13C-NMR (l, CD2Cl2): 40.4
(Cipso on Cp), 50.5 (CH3), 75.7, 78.3 (Cp). 29Si-NMR
(l, CD2Cl2): −31.4
4.2. Synthesis of polymers P2 to P5
Polymerisation was performed by thermal ROP of
the monomers in vacuum-sealed Pyrex tubes. Polymeri-
sation was complete after approximately 1 h. Soxhlet
extraction of the dark-red residues using pentane al-
lowed isolation of pure polymers P2 to P5 as red–or-
ange powders soluble in polar organic solvents such as
THF or methylene chloride.
P2. Polym. temp.: 190°C, Mw\360 000. Anal.
Found: C, 57.51; H, 6.80; Fe, 16.00; Si, 8.70. Calc. for
C16H22Fe1O2Si1: C, 58.18; H, 6.67; Fe, 16.97; Si, 8.48%.
1H-NMR (l, CD2Cl2): 1.3 (6H, d, CH3), 4.2 (4H, t,
Cp), 4.4 (4H, t, Cp), 4.3 (1H, h, CH). 13C-NMR (l,
CD2Cl2): 26.1 ((CH3)2CH), 65.4 ((CH3)2CH), 67.5
(Cipso on Cp), 72.7, 74.3 (Cp). 29Si-NMR (l, CD2Cl2):
−24.8.
4.1. Synthesis of monomers M2 to M5
Monomers M2 to M5 were readily obtained from the
reaction of 1,1%-ferrocenediyldichlorosilane [30] and the
corresponding alcohols. In the case of M3, the commer-
cially available sodium tert-butoxide was preferred.
A typical procedure is given in the case of M4:
9.55 g. (0.034 mol) of 1,1%-ferrocenediyldichlorosilane
were dissolved in 500 ml of THF and 6.816 g.
(0.0675 mol) of triethylamine were added to the solu-
tion. The reaction mixture was cooled to 0°C in an ice
bath and a solution of 7.29 g. (0.0675 mol) of benzyl
alcohol in 50 ml of THF was added dropwise over a
period of 2 h. The reaction mixture was allowed to stir
for 7 h at ambient temperature and the solvent was
evaporated off at the vacuum line. The residue was
dissolved in pentane and triethylammonium chloride
precipitated off. The pentane solution was filtered and
the filtrate was concentrated at the vacuum line until a
dense phase was observed in the flask. The solution was
then gently warmed to homogeneity and the flask was
stored at −18°C allowing M4 to crystallise as deep-red
plates (12.5 g, 86%); m.p. 48–50°C. Anal. Found: C,
67.50; H, 5.05; Fe, 12.60; Si, 6.70. Calc. for
C24H22Fe1O2Si1: C, 67.61; H, 5.16; Fe, 13.15; Si, 6.57%.
1H-NMR (l, CD2Cl2): 4.2 (4H, t, Cp), 4.6 (4H, t, Cp),
P3. Polym. temp.: 230°C, Mw=280 000; Ip=1.6.
Anal. Found: C, 62.25; H, 7.82; Fe, 15.65; Si, 7.04.
Calc. for C18H26Fe1O2Si1: C, 60.34; H, 7.26; Fe, 15.64;
1
Si, 7.82%. H-NMR (l, CD2Cl2): 1.4 (18H, s, tBu), 4.2
(4H, t, Cp) 4.4 (4H, t, Cp). 13C-NMR (l, CD2Cl2): 32.4
((CH3)3C), 71.2 (Cipso on Cp), 72.7 ((CH3)3C), 72.9,
74.6 (Cp). 29Si-NMR (l, CD2Cl2): −36.5.
P4. Polym. temp.: 180°C, Mw=290 000; Ip=2.0.
Anal. Found: C, 63.47; H, 5.51; Fe, 12.30; Si, 9.70.
Calc. for C24H22Fe1O2Si1: C, 67.61; H, 5.16; Fe, 13.15;
1
Si, 6.57%. H-NMR (l, CD2Cl2): 4.3 (4H, t, Cp), 4.5
(4H, t, Cp), 5.0 (4H, s, CH2), 7.4 (10H, m, Ph).

G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
53
13C-NMR (l, CD2Cl2): 65.4 (C6H5CH2), 65.9 (Cipso on
Cp), 72.9, 74.4 (Cp), 126.8, 127.4, 128.6, 141.4
(C6H5CH2). 29Si-NMR (l, CD2Cl2): −18.5.
GP3Al. Reactant:AlCl3 (ratio 3:2). Oven tempera-
ture: 60°C for 48 h. Colour: grey. Anal. Found: C,
15.95; H, 3.83; Al, 8.70; Fe, 3.20; Si, 20.80; which
corresponds to C1.8H5.15Al0.4Cl0.3Fe0.08Si1O4. Calc. for
C10H8Al0.66Fe1O2Si1: C, 45.80; H, 3.05; Al, 6.87; Fe,
21.38; Si, 10.69%. 13C-MAS-NMR (l): 30; 73. 29Si-
MAS-NMR (l): −85 to −110. 27Al-MAS-NMR (l):
P5. Polym. temp.: 130°C, Mw=206 000; Ip=17.
Anal. Found: C, 52.52; H, 5.24; Fe, 17.80; Si, 9.60.
Calc. for C12H14Fe1O2Si1: C, 52.55; H, 5.11; Fe, 20.44;
1
Si, 10.22%. H-NMR (l, CD2Cl2): 3.7 (6H, s, CH3), 4.2
(4H, t, Cp), 4.4 (4H, t, Cp). 13C-NMR (l, CD2Cl2): 51.4
(CH3), 69.1 (Cipso on Cp), 72.2, 73.9 (Cp). 29Si-NMR
(l, CD2Cl2): −15.0.
60; 30; 2. Specific surface area: 87 m2 g−1
.
GP4Si. Reactant:SiBr4 (ratio 2:1). Oven temperature:
25°C for 15 days. Colour: dark brown. Anal. Found: C,
46.31; H, 3.63; Fe, 12.70; Si, 16.80; which corresponds
to C10H9.4Fe0.6O3Si1.5. Calc. for C10H8Fe1O2Si1.5: C,
46.51; H, 3.10; Fe, 21.71; Si, 16.28%. 13C-MAS-NMR
(l): 60 to 80; 129; 139. 29Si-MAS-NMR (l): −20 to
−30; −55 to −65; −90 to −105. Specific surface
4.3. Non-hydrolytic polycondensation reactions
A typical procedure is described for the case of
GP3Ti: 0.85 g (2.4×10−3 mol) of P3 and 8 ml of tolu-
ene were introduced in a Pyrex tube in an argon
atmosphere. A solution of 0.225 g (1.2×10−3 mol) of
TiCl4 in 2 ml of toluene was rapidly added at 0°C. The
reaction mixture was then frozen in liquid nitrogen and
the tube sealed under vacuum. The mixture was al-
lowed to melt at room temperature and the tube intro-
duced to an oven at 60°C for 72 h. A solid precipitate
formed and was filtered from the mixture. Repeated
washing with pentane, acetone and THF allowed isola-
tion of GP3Ti as an insoluble dark-brown solid
(0.48 g). Anal. Found: C, 34.74; H, 3.97; Fe, 9.85;
Si, 10.05; Ti, 9.30 which corresponds to
area: 20 m2 g−1
.
4.4. Hydrolytic polycondensation reactions
In a classic procedure, a 0.5 M THF solution of
polymer was prepared and the catalyst and water added
rapidly. The reaction mixture was stirred for some
minutes for homogeneity and the solution was then
allowed to stand until an insoluble precipitate was
observed and the initial red solution faded. The solid
was recovered by simple filtration and was cleaned by
repeated washing with water, ethanol, acetone and
ether. Any remaining starting material was removed by
Soxhlet extraction using THF, although no coloration
of the solvent was usually observed.
GP2H. Catalyst:TBAF 3%. P2:H2O ratio, 1:15.
Time: 5 days. Colour: orange. Anal. Found: C, 50.75;
H, 3.51; Fe, 24.10; Si, 14.20; which corresponds to
C8.3H6.9Fe0.85O0.9Si1. Calc. for C10H8Fe1OSi1: C, 52.63;
H, 3.51; Fe, 24.56; Si, 12.28%. 13C-MAS-NMR (l): 73.
29Si-MAS-NMR (l): −31.
GP4H. Catalyst:TBAF 10%. P4:H2O ratio, 1:15.
Time: 1 week. Colour: orange. Anal. Found: C, 50.61;
H, 3.83; Fe, 23.40; Si, 14.70; which corresponds to
C8H7.3Fe0.8O0.9Si1. Calc. for C10H8Fe1OSi1: C, 52.63; H,
3.51; Fe, 24.56; Si, 12.28%. 13C-MAS-NMR (l): 68 to
72. 29Si-MAS-NMR (l): −30.
GP5H. Catalyst:TBAF 10%. P5:H2O ratio, 1:1.
Time: 1 h. Colour: orange. Anal. Found: C, 52.34; H,
3.74; Fe, 22.00; Si, 13.5; which corresponds to
C9.05H8.75Fe0.8O1.1Si1. Calc. for C10H8Fe1OSi1: C, 52.63;
H, 3.51; Fe, 24.56; Si, 12.28%. 13C-MAS-NMR (l): 65
to 80. 29Si-MAS-NMR (l): −30.
C8.1H11Fe0.5O4.6Si1Ti0.54. Calc. for C10H8Fe1O2Si1Ti0.5:
C, 44.78; H, 2.99; Fe, 20.89; Si, 10.45; Ti, 8.95%.
13C-MAS-NMR (l): 20 to 40; 60 to 80. 29Si-MAS-
NMR (l): −90 to −110. No significant specific sur-
face area.
GM2Si. Reactant:SiCl4 (ratio 2:1). Oven tempera-
ture: 110 to 200°C for 10 months. Colour: brown. Anal.
Found: C, 44.8; H, 2.07; Fe, 10.25; Si, 17.10; which
corresponds to C9.2H5.1Fe0.45O3.9Si1.5
.
Calc. for
C10H8Fe1O2Si1.5: C, 46.51; H, 3.10; Fe, 21.71; Si,
16.28%. 13C-MAS-NMR (l): 20 to 40; 60 to 80; 138.
29Si-MAS-NMR (l): −23; −50 to −70; −85 to
−105. No significant specific surface area.
GM4Si. Reactant:SiBr4 (ratio 2:1). Oven tempera-
ture: 60°C for 72 h. Colour: brown. Anal. Found: C,
49.21; H, 4.09; Fe, 12.25; Si, 14.70; which corresponds
to C11.7H11.7Fe0.6O3.5Si1.5. Calc. for C10H8Fe1O2Si1.5: C,
46.51; H, 3.10; Fe, 21.71; Si, 16.28%. 13C-MAS-NMR
(l): 60 to 80; 129; 141. 29Si-MAS-NMR (l): −17 to
−29; −50 to −65; −90 to −105. No significant
specific surface.
GP3Si. Reactant:SiBr4 (ratio 2:1). Oven temperature:
25°C for 7 h. Colour: brown. Anal. Found: C, 39.04; H,
3.51; Fe: 16.05; Si, 15.15; which corresponds to
C9H9.7Fe0.8O4.6Si1.5. Calc. for C10H8Fe1O2Si1.5: C, 46.51;
H, 3.10; Fe, 21.71; Si, 16.28%. 13C-MAS-NMR (l): 31.
29Si-MAS-NMR (l): −65 to −80. No significant spe-
cific surface area.
References
[1] P. Nguyen, P. Gomez-Elipe, I. Manners, Chem. Rev. 99 (1999)
1515 and references cited therein.
[2] R.J.P. Corriu, D. Leclercq, Angew. Chem. Int. Ed. Engl. 35
(1996) 1420.

54
G. Calle´ja et al. / Journal of Organometallic Chemistry 621 (2001) 46–54
[21] D.A. Foucher, R. Ziembinski, R. Petersen, J. Pudelski, M.
[3] R.J.P. Corriu, C. R. Acad. Sci. Paris Ser. IIc 1 (1998) 83 and
Edwards, Y. Ni, J. Massey, C.R. Jaeger, G.J. Vancso, I. Man-
ners, Macromolecules 27 (1994) 3992.
[22] G. Calleja, F. Carre´, G. Cerveau, R.J.P. Corriu, C. R. Acad. Sci.
Paris Ser. IIc 1 (1998) 285.
[23] A. Vioux, D. Leclercq, Heterogen. Chem. Rev. 3 (1996) 65.
[24] A Vioux, Chem. Mater. 9 (1997) 2292 and references cited
therein.
[25] P. Nguyen, A.J. Lough, J. Manners, Macromol. Rapid. Com-
mun. 18 (1997) 953.
[26] W. Finckh, B.Z. Tang, D.A. Foucher, D.B. Zamble, A.J. Lough,
I. Manners, Organometallics 12 (1993) 823.
[27] D.L. Zeckel, K.C. Hultzsch, R. Rulkens, D. Balaishis, Y. Ni,
J.K. Pudelski, A.J. Longh, I. Manners, D.A. Foucher,
Organometallics 15 (1996) 1972.
[28] P. Nguyen, G. Stojcevic, K. Kulbaba, M.J. MacLachlan, X-H.
Liu, A.J. Lough, I. Manners, Macromolecules 31 (1998) 5977.
[29] Y. Ni, I. Manners, J.B. Sheridan, R.T. Oakley, J. Chem. Edu. 75
(1998) 766.
[30] M.S. Wrighton, M.C. Palazzotto, A.B. Bocarsly, J.M. Boltz,
A.B. Fisher, L. Nadjs, J. Am. Chem. Soc. 100 (1978) 7264.
[31] D. Seyferth, H.P. Withers, Jr., Organometallics 1 (1982) 1275.
[32] D.A. Foucher, R. Ziembinski, B.Z. Tang, P.M. Macdonald, J.
Massey, C.R. Jaeger, G.J. Vaneso, I. Manners, Macromolecules
26 (1993) 2878.
references cited therein.
[4] G. Cerveau, R.J.P. Corriu, Coord. Chem. Rev. 178–180 (1998)
1051.
[5] R.J.P. Corriu, Angew. Chem. Int. Ed. Engl. 39 (2000) 1376 and
references cited therein.
[6] G. Cerveau, R.J.P. Corriu, N. Costa, J. Non-Cryst. Solids 163
(1993) 226.
[7] G. Cerveau, R.J.P. Corriu, C. Lepeytre, J. Mater. Chem. 5
(1995) 793.
[8] G. Cerveau, R.J.P. Corriu, C. Lepeytre, Chem. Mater. 9 (1997)
2561.
[9] P. Audebert, G. Cerveau, R.J.P. Corriu, N. Costa, J. Elec-
troanal. Chem. 413 (1996) 89.
[10] H. Cattey, C. Sanchez, P. Audebert, New J. Chem. 20 (1996)
1023.
[11] P. Audebert, P. Griesmar, P. Hapiot, S. Sanchez, J. Mater.
Chem. 2 (1992) 1293.
[12] H. Cattey, C. Sanchez, P. Hapiot, P. Audebert, J. Mater. Chem.
7 (1997) 1461.
[13] R.J.P. Corriu, P. Hesemann, G. Lanneau, Chem. Commun.
(1996) 1845.
[14] J.P. Boilot, F. Chaput, T. Gacoin, L. Malier, M. Canva, A.
Brun, Y. Levy, J.P. Galaup, C. R. Acad. Sci. Paris Ser. IIb 27
(1996) 322 and references cited therein.
[15] B. Lebeau, S. Brasselet, J. Zyss, C. Sanchez, Chem. Mater. 9
(1997) 1012.
[16] D.A. Foucher, B.Z. Tang, J. Manners, J. Am. Chem. Soc. 114
(1992) 6246.
[17] M.T. Nguyen, A.F. Diaz, V.V. Dement’ev, K.H. Pannell, Chem.
Mater. 5 (1993) 1389.
[33] J.K. Pudelski, D.A. Foucher, I. Manners, Polym. Mater. Sci.
Eng. 71 (1994) 312.
[34] R.J.P. Corriu, J.J.E. Moreau, M. Wong Chi Man, J. Sol–Gel
Sci. Technol. 2 (1994) 87.
[35] G. Cerveau, R.J.P. Corriu, C. Lepeytre, J. Organomet. Chem.
548 (1997) 99.
[36] H. Marsmann, in: P. Diehl, E. Fluck, R. Kosfeld (Eds.), 29Si
NMR Spectroscopic Results, Springer, Berlin, 1981.
[37] L. Bourget, D. Leclercq, A. Vioux, J. Sol–Gel Sci. Technol. 14
(1999) 137.
[18] I. Manners, J. Inorg. Organomet. Polym. 3 (1993) 185.
[19] D.A. Foucher, R. Petersen, B.Z. Tang, R. Ziembinski, N.
Coombs, P.M. Macdonald, R.N.S. Sodhi, J. Massey, G.J.
Vansco, I. Manners, Polym. Prep. (Am. Chem. Soc. Div. Polym.
Chem.) 34 (1993) 328.
[38] S. Acosta, R.J.P. Corriu, D. Leclercq, J. Non-Cryst. Solids 170
(1994) 234.
[20] D.A. Foucher, C.H. Honeyman, J.M. Nelson, B.Z. Tang, I.
Manners, Angew. Chem. Int. Ed. Engl. 32 (1993) 1709.
[39] S. Acosta, R.J.P. Corriu, D. Leclercq, Mater. Res. Soc. Symp.
Proc. 346 (1994) 345.
.