Organisation and reactivity of silicon-based hybrid materials with various cross-linking levels

Article abstract of DOI:10.1039/b110776n

Nanostructured silicon-based hybrid materials were prepared by sol-gel hydrolysis/polycondensation of (MeO)_3-nSiMe_n-C≡C-C₆H₄- C≡C-SiMe_n(OMe)_3-n (n = 0, 1, 2) precursors. Their organisation

Full text of DOI:10.1039/b110776n

View Article Online / Journal Homepage / Table of Contents for this issue
Organisation and reactivity of silicon-based hybrid materials with
various cross-linking levels
Bruno Boury, Robert J. P. Corriu* and Hironobu Muramatsu
´
Laboratoire de Chimie Moleculaire et Organisation du Solide (CNRS UMR 5637),
Universite Montpellier II, Place E. Bataillon, 34095 Montpellier cedex 5, France
´
Received (in Strasbourg, France) 17th October 2001, Accepted 11th February 2002
First published as an Advance Article on the web 8th July 2002
Nanostructured silicon-based hybrid materials were prepared by sol-gel hydrolysis=polycondensation of
:
:
(MeO)3ꢀnSiMen–C C–C6H4–C C–SiMen(OMe)3ꢀn (n ¼ 0, 1, 2) precursors. Their organisation was evaluated by
birefringence measurements (microscopy under polarised light) and evidences the possibility of self-association
of the macromolecular units in the sol step and reorientation processes in the ageing step. In contrast to the
texture (porosimetry with N2), the cross-linking level of the Si–O–Si network of the solids (29Si CPMAS
spectroscopy) appears to be determinant for the anitropic organisation. For n ¼ 2 the corresponding polymer
:
:
–(–O–SiMe2–C C–C6H4–C C–SiMe2–)– is crystallised and demonstrates the strong ability of the organic units
to self-associate. This is also suggested by partial polymerisation of the acetylenic moieties of the organic units
in the solid state up to 350 ꢁC (DSC, IRTF).
The chemistry of organic–inorganic hybrid materials provides
the chemist with interesting possibilities for building new
materials. Most of them can be prepared by a one-step inor-
ganic polymerisation sol–gel process, particularly in the case of
silicon-based materials. Using precursors with the general
formula R[SiX3]n (n 5 2 and X ¼ OMe, OEt, H, Cl, ...)
hydrolysis=polycondensation leads to nanostructured solids
according to eq. (1):
ð1Þ
The interest of this approach has already been demonstra-
ted,1–5 and it leads to inter alia NLO materials,6,7 luminescent
materials,8–11 thin films with high dielectric constants,12 or
solids like mesoporous MCM materials.13,14
Previous studies have been focused on understanding the
formation of these solids in order to control their chemical,
structural and textural characteristics. Contrary to silica,
which is always amorphous when prepared by the sol–gel
Scheme 1 Representation of an anisotropic organisation of an hyb-
rid inorganic–organic xerogels with a rod-like rigid R group.
method, the possibility of self-organisation in these hybrid
materials has been demonstrated by using very specific pre-
cursors like octadecyltrichlorosilane,15 chiral diureidocyclo-
hexane16 or a bisurea-based precursor.17 In these specific
cases, materials are organised over a long-range scale. Our
work has focused on the anisotropic organisation of the solids
prepared from rigid precursors, their organisation being
described in terms of a short-range order (Scheme 1).18,19
Furthermore, it was demonstrated that a rigid rod-like geo-
metry for the organic group is best suited to achieve such
anisotropic organisation.20 It was also demonstrated that these
phenomena are highly dependent on the experimental para-
meters used for the sol–gel process, for example the catalyst
and the solvent.21 The anisotropic short-range order is
observed at the mesoscopic scale by X-ray and birefringence
measurements. This is apparently governed by at least two
different phenomena. The first one takes place in the solution
during the formation of the Si–O–Si bonds by hydro-
lysis=polycondensation of the Si–OMe functions. The kinetics
of each of these reactions and the possibility of interaction
between the organic groups can each control the organisation
of the molecular units at the molecular level. The second
phenomenon takes place during the ageing of the gel. The
continuation of the chemical processes (formation and redis-
tribution of Si–O–Si) and elimination of the solvent leads to
cracks in the gel and mechanical reorientation of the material.
The anisotropic nanostructured hybrid materials that result
are formed of a highly cross-linked Si–O–Si framework
between non-mesogenic units. In other words this anisotropic
behaviour is completely different from that of liquid crystals,
which is controlled only by the van der Waals interactions
between independent mesogenic units.
Such short-range organisation is fundamental for the dev-
elopment and application of these hybrid organic–inorganic
materials. It is based to the possibility to control their physical
and chemical properties due to co-operative interactions of the
DOI: 10.1039/b110776n
New J. Chem., 2002, 26, 981–988
981
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

View Article Online
organic groups. We thought it would be interesting to look at
another determining parameter, the connectivity at silicon,
corresponding to the number of cross-linkages (theoretical and
experimental) on each silicon atom of a precursor. We chose to
work with precursors I–III, which do not exhibit liquid crystal
behaviour and do not require any catalyst to obtain solid
formation.22 This approach allows the preparation and com-
parison of a set of homologous solids since the organic group
is the same in each, though with different connectivities at
silicon. We prepared a linear polymer with a 1D structure from
III (theoretical cross-linking level of 0) and covalent solids with
a 3D structure for I and II (theoretical cross-linking levels
of 1 and 2 at each silicon). Systems with R[SiMe(OMe)2]2
precursors have already been used,23,24 but this is the first time
that the possibility of self-organisation in these solids is
reported. We also report the preparation of the crystallised
polysiloxane prepared from III, which belongs to the main-
chain polysiloxane family, some of which exhibit liquid crystal
properties. Above all, the aim of this study is to test the
effect of the interactions between organic moieties in the self-
organisation of nanostructured hybrid organic–inorganic
materials.
(44 mmol) of 1.4 M solution in diethyl ether, was added
dropwise to the solution. The white suspension obtained was
warmed back slowly to room temperature followed by heating
at 30 ꢁC for 30 min. The appearance of the medium changed
from translucent green to a grey suspension. The suspension
was cooled once again to ꢀ50 ꢁC, and transferred dropwise
into a solution of 6.8 g (48 mmol) of chlorodimethoxy-
methylsilane dissolved in 40 ml of dry THF and kept at
ꢀ70 ꢁC. When all the lithium salt was added, the medium was
warmed back slowly to room temperature and agitated for 15 h;
a yellow translucent solution was obtained. This was evapo-
rated under vacuum giving a yellow paste that was extracted
by 15 ml of n-pentane and 10 ml of toluene (3 ꢂ). Evaporation
of all the solvents gave a viscous yellow syrup that was distilled
under reduced pressure. Precursor II is obtained as a white
solid at room temperature, b.p. 115–120 ꢁC (0.07 mbar), m.p.
65 ꢁC, yield 3.40 g (50%).1H NMR in CDCl3 , d (intensity):
7.48 (4.0), 3.66 (11.5), 0.36 (5.7). 29Si NMR in CDCl3 , d:
ꢀ30.04.
Precursor III. This compound was prepared in the same
way as for precursor II up through formation of the lithium
salt. At this stage, the lithium salt is treated with 6.0 g
(48 mmol) of chloromethoxydimethylsilane dispersed in 40 ml
of dry THF cooled and kept at ꢀ70 ꢁC. The suspension of
lithium salt was cooled to ꢀ50 ꢁC and transferred dropwise
into the chlorosilane solution. When all the lithium salt was
added, the medium was warmed back slowly to room tem-
perature and agitated for 15 h. A yellow translucent solution
was obtained and evaporated under vacuum, which led to the
formation of a yellow paste. The paste was extracted by 15 ml
of n-pentane and 10 ml of toluene (3 ꢂ), followed by eva-
poration of all the solvents. The viscous yellow oil that
remained was distilled under reduced pressure. Precursor III
was obtained as a white solid, m.p. 80 ꢁC, b.p. 90–95 ꢁC (0.04
mbar), yield 3.50 g (57%).1H NMR in CDCl3 , d (intensity):
7.45 (4.0), 3.58 (5.7), 0.35 (12.2).29Si NMR in CDCl3 , d:
ꢀ7.66.
Experimental
All the reactions were carried out under a nitrogen atmosphere
and solvents were dried before using according to already
described procedures.25
29Si NMR spectra were obtained on a Bruker AM300 at
59.620 MHz. Chemical shifts are given relative to tetra-
methylsilane. For CP MAS analyses, contact time was 2 ms,
10 s of recycle time, a spinning rate of 5 KHz, and acquisition
of more than 1500 scans. Elemental analyses were performed
by the Service Central de Microanalyse du CNRS.
Porosimetry measurements (specific surface area, pore dia-
meter distribution and pore volumes) were conducted on a
Micromeritics Gemini III and an ASAP 2010 porosimeter
using high purity N2 at ꢀ196 ꢁC. Surface areas were calculated
by the B.E.T. equation, micropore volumes using the t-plot
and pore size distributions calculated using the Horvath–
General procedure for the synthesis of the xerogels
Gellations were all carried out at room temperature without
any catalyst. To achieve a homogeneous reaction medium, the
precursor and a solvent were mixed to form the solution 1.
Separately, water was diluted with the same solvent in another
glass tube to form solution 2. Then solution 2 was added
rapidly to solution 1 and the mixture was stirred vigorously
under nitrogen for 2 min. The proportions of solutions 1 and 2
varied in agreement with the values given below. A slight
amount of this mixture was put into the glass cell, which was
used for microscopic observation and measurement of the
birefringence. The remaining solution was allowed to rest in
order for gellation to occur. The gel was then allowed to stand
for 1 week and then crushed, washed with acetone (20 ml),
diethyl ether (20 ml) and finally dried for 12 h at 100 ꢁC under
vacuum (0.1 mm Hg). Yields of these preparations were
sometimes over 100% due to the presence of remaining Si–OR
(R ¼ H or Me) groups or slight residues of solvents.
Kawazoe equation with
geometry.
a Saito–Foley cylindrical pore
Differential scanning calorimetry was performed on a DSC 6
Perkin Elmer analyser in aluminium cells in a streamꢀ1of
nitrogen (50 ml minꢀ1); the heating rate was 2 or 5 ꢁC min
.
Observations by optical microscopy under polarised light
were performed on materials prepared in a glass cell of
25 mm ꢂ 25 mm ꢂ 30 micron and coated with Teflon as
reported previously.18 Optical properties were measured with a
Laborlux12POLS polarising microscope, photographs were
taken using a Leica MPS28 camera. Measurement of the
birefringence (Dn) was quantitatively made by equalising the
index of refraction provoking the darkness of the bright zone
with calibrated glass slides in order to determine Dl*,
Dn ¼ Dl*ꢃd (d being the thickness of the gel that was taken
equal to the thickness of the cell). Infrared analyses were
performed on a Perkin Elmer 1600 FTIR analyser with
samples being prepared in KBr pellets.
Xerogel HMIa. The mixture for gellation was made with
1.04 g (2.84 mmol) of I, 0.15 g (8.52 mmol) of water and 2.70 ml
of tetrahydrofuran. The sol was transparent and the gelation
time was 21 min. Syneresis started after 1 day. A hard xerogel
was obtained, 0.64 g (yield 99%).
Precursor syntheses
Precursor I. This precursor was prepared according to the
literature method.22
Xerogel HMIb. The mixture for gellation was made with
1.10 g (3.00 mmol) of I, 0.16 g (9.00 mmol) of water and 2.84 ml
of methanol. The sol was translucent. Gelation time was 8 min
and syneresis started after 12 h (contraction of the gel). A soft
and light xerogel was obtained, 0.69 g (yield 99%).
Precursor II. A solution of 5.40 g (20 mmol) of bis(tri-
methylsilylethynyl)benzene in 40 ml of dry tetrahydrofuran
was prepared. Then the medium was cooled and kept at
ꢀ40 ꢁC. Methyllithium–lithium bromide complex, 31.4 ml
982
New J. Chem., 2002, 26, 981–988

View Article Online
Xerogel HMIc. The mixture for gellation was made with
1.10 g (3.00 mmol) of I, 0.16 g (9.00 mmol) of water and 2.84
ml of acetonitrile. The sol was transparent and gellation
occurred within 1 min. Syneresis started after 30 min and it
shrank strongly. A light xerogel was obtained, 0.65 g (yield
95%).
Polymer HMIII
A mixture of 1.07 g (3.52 mmol) of III, 0.06 g (3.52 mmol) of
water and 3.46 ml of tetrahydrofuran was prepared at room
temperature. The solution was transparent, but after 2 days,
small white grains and short needles started to form. The
average diameter of the grains was 1.0–1.5 mm and the average
length of the needle-shaped crystals was 1.5–2.0 mm. By eva-
porating the solvent, a pale orange powder was obtained, 0.86
Xerogel HMIdId. The mixture for gellation was made with
0.78 g (2.13 mmol) of I, 0.11 g (6.39 mmol) of water and 2.01 ml
of acetone. The sol was transparent and gellation time was 55
min. Syneresis started after 1 day and it slowly shrank. A soft
xerogel was obtained, 0.49 g (yield 101%).
1
g (yield 95%) H NMR in CDCl3 , d (intensity): 7.45 (4.00),
3.58 (0.22), 0.35 (11.9). 29Si NMR in CDCl3 , d (intensity):
ꢀ7.66 (0.01), ꢀ13.8 (0.06), ꢀ16.9 (1.00), ꢀ17.9 (0.07), ꢀ19.1
(0.06).
Xerogel HMIe. The mixture for gellation was made with
0.88 g (2.40 mmol) of I, 0.13 g (7.22 mmol) of water and 2.27 ml
of toluene. After 6 min, precipitation was observed and the
medium became heterogeneous. Syneresis started after half a
day but was difficult to precisely determine. A soft and light
xerogel was obtained, 0.51 g (yield 93%).
Results
Hybrid materials HMI
The gellation of I was performed in different solvents
according to Scheme 2. In all the cases the gellation times were
generally short (Table 1). The time necessary for the syneresis
to start is also given since the syneresis evidences the con-
traction and reorganisation of the framework.
The initial solutions were completely transparent in normal
light, and dark when observed by optical microscopy in
polarised light. This is also the case when the gel starts to form.
This reveals an isotropic medium in both cases. Ageing of the
wet gels leads, after 1 or 2 h, to the formation of cracks,
initially on the edges of the cell where evaporation of the
solvent is relatively fast. The cracking of the gel, and especially
the number and orientation of the cracks, depends on the
solvent. However, these observations are purely qualitative at
the present time and cannot be quantified.
Xerogel HMIIa. The mixture for gellation was made with
1.01 g (3.02 mmol) of II, 0.109 g (6.04 mmol) of water and
2.91 ml of tetrahydrofuran. The sol was transparent and gel-
lation time was 50 min. Syneresis started after 2 days. A hard
and light xerogel was obtained, 0.64 g (yield 87%).
Xerogel HMIIb. The mixture for gellation was made with
0.66 g (1.98 mmol) of II, 0.07 g (3.96 mmol) of water and
1.92 ml of methanol. The sol became opaque. Gellation time
was 22 min, while the medium was heterogeneous. Syneresis
started after 1 day though it was not easy to precisely determine.
A hard and light xerogel was obtained, 0.48 g (yield 99%).
In all cases, birefringent pieces of HMIa–e result from the
cracking of the gel. Examples are given in Fig. 1 and 2. These
pieces of solids can be considered as xerogels although they may
not be completely dry. The average value of the birefringence
Xerogel HMIIc. The mixture for gelation was made with
0.77 g (2.30 mmol) of II, 0.08 g (4.60 mmol) of water and
2.22 ml of acetonitrile. The sol was translucent and gelation
occurred within 2 min. Syneresis started after 6 min and the gel
shrank dramatically. A hard and dense xerogel was obtained,
0.54 g (yield 97%).
Xerogel HMIId. The mixture for gelation was made with
0.96 g (2.87 mmol) of II, 0.10 g (5.74 mmol) of water and
2.77 ml of acetone. The sol was transparent and gelation time
was 80 min. Syneresis started after 2 days and the gel gradually
shrank. A hard xerogel was obtained, 0.61 g (yield 88%).
Xerogel HMIIe. The mixture for gelation was made with
0.90 g (2.69 mmol) of II, 0.09 g (5.38 mmol) of water and 2.59
ml of toluene. After 13 min, colloid precipitation was observed
and the syneresis phenomenon became heterogeneous. Syner-
esis started after 1 day and it was slight. A hard and light
xerogel was obtained, 0.65 g (yield 99%).
Scheme 2 Preparation of the HMIn xerogels I.
Table 1 Characteristics of hybrid xerogels HMIa–e
Gelation=min
(start of
Xerogel Solvent syneresis=h)
Cross-linking Birefringence Specific
T 1a (%) T 2a (%) T 3a (%) DCb level (Si–O–Si=Si) value=10ꢀ3 surface area=m2 gꢀ1 Porosity
HMIa
HMIb
THF
MeOH
21 (24)
8 (12)
15
16
47
54
38
30
74
71
2.2
2.1
2.9
2.3
120
840
Microporous
Micro- and
meso-porous
Microporous
Microporous
Microporous
HMIc
HMId
HMIe
CH3CN 1 (0.5 strong)
Acetone 55 (24)
Toluene 6 (12)
8
10
10
57
58
61
35
32
29
75
74
73
2.2
2.2
2.1
2.4
2.8
2.4
780
670
650
a
b
Relative percentage. Degree of polycondensation calculated according to the general equation DC ¼ [0.5(T 1 area) þ 1.0(T 2 area) þ 1.5(T 3
area)]; T 1 [C–Si(OR)2–(O–Si)], T 2 [C–Si(OR)–(O–Si)2], T 3 [C–Si(O–Si)3].
New J. Chem., 2002, 26, 981–988
983

View Article Online
determined in different areas of the cell are around
2.6 0.3 ꢂ 10ꢀ3; these values remain unchanged even after
several weeks (Table 1). Depending on the nature of the sol-
vent, the birefringence varies from 10 to 15%, this suggests a
low effect of the solvent on the birefringence of xerogels
prepared from I.
The textural characteristics of these materials were obtained
by porosimetry measurements with nitrogen (Table 1). In
contrast to the birefringence, a drastic solvent effect is now
observed. Microporous materials with a specific surface area as
low as 120 m2 gꢀ1 (THF) or (micro þ meso)porous materials
with a specific surface area as high as 840 m2 gꢀ1 (MeOH) are
formed, depending on the solvent. Therefore, variation of the
texture apparently has connections with the solvent but not
with the birefringence intensity.
By 29Si NMR spectroscopy we found that the degree of
condensation (DC) is generally high as indicated by the pre-
sence of T 1, T 2 and T 3 signals26 (Table 1). CP MAS spectro-
scopy is a fast method but sometimes not quantitative for
analysing these materials, especially because silicon atoms
corresponding to T 3 units that are far away from any proton
leads to very slow relaxation processes. Thus, the values of the
degree of condensation (DC) determined by looking at the
relative intensity of the T 1=T 2=T 3 signals are always qualitative
and underestimated. In these conditions, the level of con-
densation was found to lie in the same range for all the HMIa–e
solids (at least 571%) and corresponds to an average >2.2
Si–O–Si links per silicon atom (Table 1). This qualitative
comparison of the DC of HMIa–e solids suggests that for solids
prepared from I, the level of condensation is only weakly
dependent on the nature of the solvent, just like the birefrin-
gence intensity.
Fig. 1 Xerogels HMIa
.
DSC analysis shows two transformations for HMIa
[Fig. 3(a)]. The first one from 50 to 150 ꢁC is endothermic,
probably due to the elimination of solvent adsorbed on the
solid or elimination of water that forms by condensation of
residual silanols. The second phenomenon is exothermic from
Fig. 2 Xerogels HMIb from precursor I.
Fig. 3 Differential scanning calorimetry analyses of (a) HMIa , (b) HMIIa and (c) HMIII from 25 to 350 ꢁC in aluminium cells and under nitrogen.
984
New J. Chem., 2002, 26, 981–988

View Article Online
Fig. 4 IRTF analyses analyses in KBr pellets of (a) HMIa and (b) HMIIa
220 up to 350 ꢁC. It is attributed to the cross-linking between
the organic units by reaction between the acetylenic groups.
Indeed, infrared analysis indicates, after thermal treatment,
:
a decrease of the signal attributed to C C groups (n ¼ 2170
cmꢀ1) and the presence of new signals attributed to C=C
conjugated systems [n ¼ 1716 and 1607 cmꢀ1; Fig. 4(a)]. It was
:
also evident that a majority of the C C bonds did not react at
this temperature (350 ꢁC), certainly because of the high level of
cross-linking of the materials, which limits the movements and
interactions of the units. When a second DSC analysis was
performed on the same sample, no exothermic phenomenon
was observed and the infrared spectrum was essentially
unchanged. The lack of reactivity of the remaining acetylenic
group may be attributed to the high level of cross-linking,
which prevents deformations of the solid architecture favour-
able to the reaction of the acetylenic groups.
Scheme 3 Preparation of the HMIIn xerogels II.
moreover, it depends on the solvent in contrast to that of
HMIa–e
. The birefringence value ranged from 2.7 to
Hybrid materials HMII
5.6 ꢂ 10ꢀ3, depending on the solvent. For HMIIb prepared in
methanol a precipitation phenomenon prevents any measure-
ment; a white solid forms and its birefringence cannot be
determined.
The gellation of II was also performed in different solvents
according to Scheme 3. Gellation and syneresis times were also
short and suggest a high reactivity of this precursor, as for I
(Table 2). The same process of gellation, ageing and cracking
was observed as for gels and xerogel from I (vide supra). After
cracking, all the pieces of HMIIa–e xerogels were found to
be birefringent in all cases (Fig. 5 and 6). However, the bi-
Textural analyses of theses HMIIa–e revealed that most of
them are less porous than HMIn , a porous solid being obtained
with a specific surface area of up to 670 m2 gꢀ1 (CH3CN) only
in the case of HMIIc . In addition, macro- and=or mesopores
were observed here but never micropores. From a general
refringence of HMIIa–e is a little higher than for HMIa–e
,
Table 2 Characteristics of hybrid xerogels HMIIa–e
Gelation=min
(start of
syneresis=h)
Specifique
Cross-linking
Birefringence surface
Pore
Xerogel Solvent
D1 a (%) D2 a (%) DCb level (Si–O–Si=Si) value=10ꢀ3
area=m2 gꢀ1 characteristics
HMIIa
HMIIb
HMIIc
HMIId
HMIIe
THF
50 (48)
22 (24)
2 (0.1 strong) 26
38
60
62
40
74
77
78
81
70
88
89
90
1.6
1.4
1.7
1.8
1.8
5.6
—
3.1
3.9
2.7
<10
<10
670
200
340
—
—
MeOH
CH3CN
Acetone 80 (48)
Macro and meso
Meso and micro
Macro and meso
23
22
Toluene
13 (24)
a
b
Relative percentage. Degree of polycondensation according to the general equation DC ¼ [0.5(D1 area) þ 1.0(D2 area)]; D1 [C–Si (OR)2–(O–
Si)], D2 [C–Si(OR)–(O–Si)2].
New J. Chem., 2002, 26, 981–988
985

View Article Online
higher temperature and the decrease of the acetylenic signal
intensity is higher. Moreover, when the sample is analysed a
second time, the exothermic phenomena is not observed for
MHIa while it is observed again for MHIIa , although with a
lower intensity. A possible reason might be the lower degree of
cross-linking in MHIIacorresponding to fewer links and a more
mobile Si–O–Si network upon heating, thus allowing the
acetylenic group to react at a lower temperature and also
during the second step.
Polymer HMIII
Hydrolysis-polycondensation of III was performed in THF
according to Scheme 4 and the analyses agree with the for-
mation of a polymeric material. 29Si NMR analysis in solvent
presents mainly one signal at ꢀ16.9 ppm corresponding to the
Fig. 5 Xerogels HMIIa from precursor II.
:
:
repetitive units –(–O–SiMe2–C C–C6H4–C C–SiMe2–)– of
the polymer. A signal of very low intensity at ꢀ7.66 ppm
indicates the presence of remaining Si(OMe)3 groups by
comparison to the starting precursor; other signals of very low
intensity at ꢀ13.8, ꢀ17.9, ꢀ19.1 ppm could be attributed to
hydrolysed but not polycondensed –SiMe2(OH) groups at the
end of the chains or to cyclic dimers. By 1H NMR analysis, the
intensity of the Si–OMe signal is very weak and not useful for
integration. By SEC (size exclusion chromatography) analysis
an average molecular weight of 15 500 g molꢀ1 was determined
and corresponds to a degree of polymerisation DPn of 60 (MW
of the unit chain is 250 g molꢀ1). However, this result must be
taken as an estimation since the rod-like shape of the mole-
cular units of this polymer is very different from that of the
polystyrene used for calibration of the SEC apparatus.
Upon slow evaporation of the solvent, the solid polymer
crystallises from the solution and forms spherolites that grow
in the cell (Fig. 7). Their birefringence is very high at 25.3,
and their shape clearly indicates a crystallised polymer with a
well-defined organisation.
Fig. 6 Xerogels HMIIb from precursor II.
point of view, like for the preparation of MHIa–e solids, the
role of the solvent on the texture of HMIIa–e solids is impor-
tant. But in addition, in the case of HMIIa–e , the solvent has an
effect on the birefringence.
By 29Si NMR spectroscopy D1 and D2 signals were observed
with chemical shift agreeing with the previous literature
(Table 2).26 No Tn units were observed, indicating the absence
of Si–Csp bonds cleavage. We checked the degree of con-
densation (DC) of HMIIa–e , but like for HMIa–e and for the
same reasons, the DC is certainly underestimated and is given
as a minimum value. Qualitatively, DC is high, in agreement
with previous data on R[Si(Me)O1,5]2 solids.24 In contrast to
HMIa–e , variation of DC is observed for MHIIa–e depending
on the solvent. A tentative comparison of HMIIa and HMIa
suggests that the level of polycondensation of HMIIa is higher
than for HMIa ; however, the level of cross-linking (number of
Si–O–Si per silicon atom), corresponds to an average of 1.8 Si–
O–Si=Si for HMIIa and 2.2 Si–O–Si=Si for HMIa . It also
appears that a decrease of the level of condensation corre-
sponds with an increase of the birefringence (Table 2).
When analysed by DSC analysis [Fig. 3(b)], HMIIa exhibits
the same endothermic phenomenon as that of MHIa from 50 to
150 ꢁC, due to the loss of residual solvent and loss of water
from condensation of residual silanols. When increasing the
temperature an exothermic phenomenon starting at 270 ꢁC was
measured and attributed again to the reaction of the acetylenic
group. This was confirmed by the infrared spectrum. After
heat treatment we observed a decrease of the signal attributed
Scheme 4 Preparation of HMIII
.
to C C groups (n ¼ 2170 cmꢀ1) and the presence of new signals
:
attributed to C=C conjugated systems [n ¼ 1716 and 1607 cmꢀ1
;
Fig. 4(b)]. However, the thermal behaviour of both HMIIa and
HMIa are not precisely the same: reaction of HMIIa starts at a
Fig. 7 HMIII from precursor III.
986
New J. Chem., 2002, 26, 981–988

View Article Online
When analysed by DSC, HMIII exhibits two endothermic
phenomena centred at 142 and 167 ꢁC [Fig. 3(c)]. The latter
corresponds to the melting point of the polymer and the for-
mer may be attributed a priori to a phase transition suggesting
a liquid crystal behaviour (more studies are in progress to
determine the organisation of the solid and its liquid crystal
properties). When heated to higher temperature an exothermic
phenomena starts at 270 ꢁC. As in the case of HMIIaand HMIa
this is attributed to acetylenic reaction and is confirmed by
infrared analysis.
stress=relaxation process and the level of polycondensation.
HMIIa–e , which are less polycondensed than HMIa–e , are more
sensitive to this process and more dependent on the experi-
mental conditions than HMIa–e . Furthermore, because HMIa–e
are more cross-linked, they are less dependent on the experi-
mental conditions, but the xerogels are also less birefringent.
Thus, it appears that at the ageing step, when cracking of the
gel occurs due to its chemical evolution (loss of solvent, con-
traction due to Si–O–Si formation and redistribution),29 the
reorientation process is favoured by the less cross-linked
HMIIa–e compared to MHIa–e . However, we stress that the
level of cross-linking is determined both by the nature of the
precursor but also by the chemistry at work during all these
processes.
Discussion
Changing the connectivity at silicon [Si(OMe)=Si ¼ 3 for I, ¼ 2
for II, ¼ 1 III] provides information on the formation of the
short-range organisation of hybrid organic-inorganic materi-
als. Comparison of HMna–e (n ¼ I, II and III) materials and of
their reactivity leads to results that confirmed our previous
work.18,20,21 First, the organisation at the molecular level
occurs during hydrolysis=polycondensation (at the sol step)
and it is based on the Van der Waals interactions between
organic groups. Secondly, the development of this organisa-
tion into micrometer size domains is due to the reorganisation
and reorientation produced during the cracking of the gel.
The first hypothesis is confirmed by the behaviour of poly-
mer HMIII demonstrating the ability of these systems to self-
organise. This polymer, with an organic group that is used for
the other hybrid materials HMIa–e and HMIIa–e , exhibits
excellent crystalline organisation, which illustrates the ability
Conclusion
Nanostructured silicon-based hybrid materials of general for-
:
:
mula O1.5(3ꢀn)SiMen–C C–C6H4–C C–SiMenO1.5(3ꢀn) (n ¼ 0,
.1, 2) are anisotropic materials. The present results stress that
their short-range organisation is extended over several micro-
meters and is the combination of two processes, the self-
association between the macromolecular species at the sol step
and the reorientation process during the ageing step of the gel.
From a general point of view, a control of the organisation
of silicon-based hybrid material prepared by sol–gel chemistry
can be achieved by combining a control of both the chemical
processes at the molecular level and the physical evolution.
0
00
0
00
:
:
of –(SiR R –C C–C6H4–C C–SiR R –O–) units to self-
organise. Although the packing of the chains and the
arrangement of their molecular units are still under investi-
gation, it is clear that the organisation of the solid polymer
HMIII results from a thermodynamic process corresponding
to the Van der Waals interactions between the organic groups
and formed upon evaporation of the solvent. In contrast,
organisation of HMIa–e and HMIIa–e is mainly irreversible once
the Si–O–Si bonds start to form and cross-link the system.
Therefore, the self-organisation is not achieved so nicely
compared to HMIa–e . Additionally, the possibility to poly-
merise the acetylenic groups of the solids HMIa–e and HMIIa–e
are arguments for a self-organisation process. Indeed, as in the
case of the polymerisation of the diacetylenic units of O1.5Si–
CC–CC–SiO1.5 xerogels previously reported,27 such poly-
merisation probably requires a precise positioning of the
acetylenic units between each other. We note that there are few
examples of thermally activated solid-state reactions between
acetylenic units.28 The lower thermal reactivity of the HMII
compared to HMI may be attributed to steric hindrance of
Si–Me groups. As a result of this chemical transformation,
cross-linking between the organic units occurs and the forma-
tion of structures like in Scheme 5 can be proposed.
Acknowledgements
The authors are grateful to Silicones Electronics Materials
Research Center from ShinEtsu Chemical Co.
References
1
2
3
D. A. Loy and K. J. Shea, Chem. Rev., 1995, 95, 1431.
K. J. Shea and D. A. Loy, Mater. Res. Bull., 2001, 5, 358.
R. J. P. Corriu and D. Leclecq, Angew. Chem., Int. Ed. Engl., 1996,
35, 4001.
4
5
R. J. P. Corriu, Angew. Chem., Int. Ed., 2000, 39, 1376.
B. Boury and R. J. P. Corriu, in T he Chemistry of Organic Silicon
Compounds, eds. Z. Rappoport and Y. Apeloig, Wiley & Sons,
Chichester, UK, 2000, p. 565.
6
7
8
9
E. Toussaere, J. Zyss, P. Griesmar and C. Sanchez, Nonlinear
Optics, 1991, 1, 349.
Z. Yang, C. B. Xu, L. R. Dalton, S. Kalluri, W. H. Steier and J. H.
Bechtel, Chem. Mater., 1994, 6, 1899.
R. J. P. Corriu, P. Hesemann and G. Lanneau, Chem. Commun.,
1996, 1845.
B. Lebeau, S. Brasselets, J. Zyss and C. Sanchez, Chem. Mater.,
1997, 9, 1012.
10 A.-C. Franville, D. Zambon and R. Mahiou, Chem. Mater., 2000,
12, 428.
The second hypothesis concerns the anisotropic organisa-
tion on large domains (micrometric scale) occurring when the
gel is formed and has started to crack. It clearly appears
related to the connectivity at silicon. The behaviour of HMIa–e
and HMIIa–e illustrate the close relationship between the
11 A. C. Franville, D. Zambon, R. Mahiou, S. Chou, Y. Troin and
J. C. Cousseins, J. Alloys. Compd., 1998, 275–277, 831.
12 Y. Lu, H. Fan, N. Doke, D. Loy, R. A. Assink, D. A. LaVan and
C. J. Brinker, J. Am. Chem. Soc., 2000, 122, 5258.
13 M. J. MacLachlan, T. Asefa and G. Ozin, Chem. Eur. J., 2000,
6, 14.
14 T. Asefa, C. Yoshina-Ishii, M. J. MacLachlan and G. A. Ozin,
J. Mater. Chem., 2000, 10, 1751.
15 A. Shimojima, Y. Sugahara and K. Kuroda, J. Am. Chem. Soc.,
1998, 120, 4528.
16 J. J. E. Moreau, L. Vellutini, M. Wong Chi Man and C. Bied,
J. Am. Chem. Soc., 2001, 123, 1509.
17 J. J. E. Moreau, L. Vellutini, M. Wong Chi Man, C. Bied, J.-L.
´
Batignies, P. Dieudonne and J.-L. Sauvajol, J. Am. Chem. Soc.,
2001, 123, 7957.
18 B. Boury, R. J. P. Corriu, P. Delord, M. Nobili and V. Le Strat,
Angew. Chem., Int. Ed., 1999, 38, 3172.
Scheme 5 Representation of the formation of two acetylenic-poly-
merized precursors units upon heat treatment of HMIIa at 350 ꢁC.
19 B. Boury, R. J. P. Corriu, P. Delord and V. Le Strat, J. Non-Cryst.
Solids, 2000, 265, 41.
New J. Chem., 2002, 26, 981–988
987

View Article Online
20 F. Ben, B. Boury, R. J. P. Corriu and V. Le Strat, Chem. Mater.,
2000, 12, 3249.
25 A. J. Gordon and R. A. Ford, T he Chemist’s Companion, Wiley,
New York, 1972.
21 B. Boury, F. Ben, R. J. P. Corriu, P. Delord and M. Nobili, Chem.
Mater., 2002, 14, 730.
22 P. Chevalier, R. J. P. Corriu, P. Delord, J. J. E. Moreau and
W. C. H. Man, New J. Chem., 1998, 22, 423.
23 E. Lindner, T. Scheller, F. Auer and H. A. Mayer, Angew. Chem.,
Int. Ed., 1999, 38, 2154.
24 A. L. Loy, G. M. Jamison, B. M. Baugher, S. A. Myers, R. A.
Assink and K. J. Shea, Chem. Mater., 1996, 8, 656.
26 H. Marsmann, in NMR Basic Principles and Progress, eds.
P. Dielh, E. Fluck and R. Kosfeld, Springer Verlag, Berlin, 1981,
vol. 17, p. 65.
27 R. J. P. Corriu, J. J. E. Moreau, P. Thepot and M. Wong Chi
Man, Chem. Mater., 1996, 8, 100.
28 K. Tanaka and F. Toda, Chem. Rev., 2000, 100, 1025.
29 C. J. Brinker and G. W. Scherer, Sol-Gel Science, Academic Press,
Inc., Boston, 1990.
988
New J. Chem., 2002, 26, 981–988