Microporous and Mesoporous Materials 176 (2013) 199208

Contents lists available at SciVerse ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier.com/locate/micromeso

Synthesis of a fully capped mesoporous silica and its hybrids with extremely
low dielectric constant and loss
Wei Shan, Lei Chen, Yang Chu, Feipeng Zhao, Guozheng Liang , Aijuan Gu , Li Yuan
Jiangsu Key Laboratory of Advanced Functional Polymer Design and Application, Department of Materials Science and Engineering, College of Chemistry,
Chemical Engineering and Materials Science, Soochow University, Suzhou 215123, PR China

a r t i c l e

i n f o

Article history:
Received 24 September 2012
Received in revised form 24 January 2013
Accepted 3 April 2013
Available online 20 April 2013
Keywords:
Fully capped mesoporous silica
Polyhedral oligomeric silsesquioxane
Dielectric constant
Dielectric loss

a b s t r a c t
A novel fully capped mesoporous silica (FCMPS) was synthesized by producing a layer of linked polyhedral oligomeric silsesquioxanes (LPOSS) on the surface of a typical mesoporous silica (SBA-15). The structure of FCMPS was characterized using Fourier Transform Infrared (FTIR), Nuclear Magnetic Resonance
(29Si NMR), X-ray Diffraction (XRD), Scanning Electron Microscope (SEM), High Resolution Transmission
Electron Microscope (HRTEM), Nitrogen adsorptiondesorption and thermogravimetric (TG) analyses.
Compared with SBA-15, FCMPS has not only high surface area, pore volume and size, but also remarkably
improved thermal stability, the initial degradation temperature (Tdi) of FCMPS increases about 194 C. In
addition, FCMPS overcomes the drawback of SBA-15, exhibiting much lower and stable dielectric constant and loss. Based on the synthesis of FCMPS, the FCMPS/bismaleimide resin (BD) hybrids with different contents of FCMPS were prepared, and their dielectric properties were investigated. Results show that
FCMPS/BD hybrids have much lower and stable dielectric constant and loss than SBA-15/BD hybrids
owing to the special structure of FCMPS. With the addition of 1 wt.% FCMPS into BD resin, the dielectric
constant of the resultant hybrid is as low as about 2.50 over the whole frequency from 10 to 106 Hz;
moreover, the dielectric loss of the hybrid is almost independent on the frequency, and the dielectric loss
at higher frequencies (>103 Hz) is even lower than that of BD resin. These attractive features make FCMPS
have obvious advantage in developing materials with low dielectric constant and loss.
2013 Elsevier Inc. All rights reserved.

1. Introduction
Low dielectric constant (low-k) materials have shown great potential in microelectronic industry owing to the continuously
increasing requirements of fabricating devices with smaller size
and faster speed [14]. Nowadays, one leading approach for preparing low-k material is to introduce porous materials into a polymer by taking the advantage of the low dielectric constant (e  1)
of air in the pores [58].
Mesoporous silica (MPS) is a kind of porous material [9], however it has several tricky problems that restrict the application in
fabricating low-k polymeric hybrids. First, MPS is very easy to adsorb water (moisture) owing to the existence of rich polar and
hydrophilic silanol groups in the inner channels of MPS, as the
dielectric constant of water is as high as about 80, so MPS usually
shows a very high dielectric constant. Second, the size of the inner
channels of MPS is generally so large that the resin molecule can
enter into the channels [10]; hence the dielectric constant of the
resultant hybrids is not as low as predicted.

Corresponding authors. Tel.: +86 512 61875156; fax: +86 512 65880089.
E-mail addresses: lgzheng@suda.edu.cn (G. Liang), ajgu@suda.edu.cn (A. Gu).
1387-1811/$ - see front matter 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.micromeso.2013.04.014

In order to fully take the advantage of MPS in preparing low-k

materials, Yamauchi et al. tried to block (cap) the entrance of
MPS [11,12], however they found that it was not possible to eliminate silanol groups in the inner channels if the entrance of MPS
had been fully blocked, so they prepared partly (or incompletely)
capped MPS, and introduced small molecules into the inner channels of MPS to eliminate the silanol groups. After that, they added
the partly capped MPS into epoxy resin, and found that the dielectric constant at around 1 GHz of the hybrid with 20 wt.% partly
capped MPS is as low as 2.60.
Unfortunately, above partly capped MPS is not suitable for real
applications owing to following reasons. First, the mechanism for
capping is the co-reaction between Si-OH and active groups of capping materials, hence the dimensions of the capping material
determine the capping degree of the capped MPS. If the dimensions
of the capping material are small, then macromolecules can enter the entrance of the partly capped MPS, this means that the
dimensions of the capping materials should be enlarged as the
porosity of MPS increases, so the selection range of the suitable
capping materials is limited. Second, some polar compounds
can still enter the inner channels of incompletely capped MPS owing to the large entrance, and thus resulting in high dielectric constant. Therefore, how to overcome these problems is a challenging

200

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

and meaningful topic. The aim of this work is to synthesize a new

and fully capped MPS, which is then used to develop low-k polymeric hybrids.
Polyhedral oligomeric silsesquioxane (POSS) is composed of an
inorganic cage-like silicon and oxygen framework (Si8O12) surrounded by organic corner groups [1318]. The organic corner
groups provide a great possibility to combine many POSS particles
together to form a new kind of MPS that is named as linked POSS
(LPOSS) [19,20]. Besides the chemical structure, LPOSS has lots of
porosities, so LPOSS exhibits low dielectric constant. Note that
the size of LPOSS can be easily adjusted [21,22], hence LPOSS
may be a potential material for capping MPS, but no work has been
reported.
Lots of investigations have proved that the properties of the hybrids are determined by the nature of either inorganic or organic
phase as well as the interface between the two phases, demonstrating that inorganic and organic phases should be carefully selected to meet the requirements of applications.
Besides the porous materials, the nature of the polymer is also
the key factor for preparing low-k materials. In this study, diallyl
bisphenol A modied bismaleimide (BD) resin, one typical kind
of heat resistant thermosetting resin, is chosen as the organic
phase for developing low-k hybrids owing to its outstanding integrated performance (such as good processing characteristics and
diversity of reactivity, high strength and toughness, good thermal
and dielectric properties, etc.) and wide applications [23,24].
In this work, a fully capped mesoporous silica (FCMPS) with a
coreshell structure was designed and synthesized, of which the
core is SBA-15 that is a typical MPS and has been widely used in
fabricating low-k materials [8,10,25], and the shell is LPOSS. Based
on the synthesis of FCMPS, new hybrids based on FCMPS and BD
resin were prepared, and the dielectric properties of FCMPS/BD hybrids were investigated. The corresponding properties of SBA-15/
BD hybrids were also studied comparatively. Results show that
FCMPS has superior advantage in fabricating low-k materials.

2. Experimental
2.1. Raw materials
4,40 -Bismaleimidodiphenol methane (BDM) was obtained from
Institute of Northwestern Chemical Engineering (China). 2,20 -Diallylbisphenol A (DBA) was purchased from Laiyu Chemical Factory
(China). Tetramethylammonium hydroxide pentahydrate (AR)
was purchased from Rudong Zhenfeng Yiyang Chemical Co., Ltd.,
China. SBA-15 was received from Department of Chemistry, Fudan
University, China. Toluene (AR) was purchased from Shanghai
Lingfeng Chemical Reagent Co., Ltd., China. Tetraethoxysilane
(AR), anhydrous methanol (AR), anhydrous acetone (AR) and dimethyldichlorosilane (DMDCS, AR) were all commercial products
made in China, and used as received.

2.2. Synthesis of octameric silicate anion

20.8 g (0.1 mol) tetraethoxysilane were added dropwise into a
reactor containing 0.1 mol tetramethylammonium hydroxide and
91 mL (5.05 mol) water. The reaction mixture was vigorously stirred at 23 2 C for 24 h, and then 60 C for 10 h. After that, the
mixture was concentrated by distillation under a reduced pressure.
The concentrated product was cooled to 4 C, and microcrystalline
was precipitated from the cool solution. The crystal was separated
by ltration, washed with acetone, and dried in vacuum to get cubic octameric silicate anion, (Me4N)8Si8O20.

2.3. Preparation of FCMPS

FCMPS was prepared following three stages. First, DMDCS
(10.75 g) was added slowly into a reactor containing anhydrous
toluene (150 mL) and dry mesoporous silica SBA-15 (5 g). The reaction mixture was vigorously stirred at 40 C for 6 h, and then anhydrous toluene and residual DMDCS were removed to get a crude
product. The crude product was washed with anhydrous acetone
and dried at 50 C in the vacuum, the resultant product was coded
as Cl-SBA-15.
Second, from the literatures, it can be known that the surface
area of SBA-15 can be up to 960 m2/g [26], and POSS nanostructures have diameters ranging from 1 to 3 nm [13]. According to
these data, the weight ratio between POSS and Cl-SBA-15 was set
as 10:1, which guarantee the blocking of the entrances of ClSBA15 particles. Specically, Cl-SBA-15 (0.3 g) and anhydrous acetone (30 mL) were blended to form Solution A. (Me4N)8Si8O20
(0.3 g) and anhydrous methanol (10 mL) were mixed to form Solution B. Solution A was added into the Solution B, and then reacted
at 40 C for 15 min under sonication to form Solution C. (Me4N)8
Si8O20 (2.7 g) and anhydrous methanol (20 mL) were mixed to
form Solution D. The solution D and DMDCS (1.2 mL) were slowly
added into the Solution C, successively, to form a mixture, which
were vigorously stirred at 40 C for 30 min. By removing anhydrous solvent and residual DMDCS, white powders were obtained,
which were then washed with anhydrous methanol to remove the
free POSS particles and their aggregates, followed by being dried at
50 C in the vacuum. The resultant product was coded as POSSSBA-15.
Third, POSS-SBA-15 was heated to 250 C and maintained at
that temperature for 1.5 h. After naturally cooling to room temperature, the resultant product was the fully capped MPS, designed as
FCMPS.
2.4. Preparation of BD prepolymer and cured resin
BDM and DBA with a weight ratio of 1:0.85 were put into a
reactor. The mixture was heated to 140 C and maintained at that
temperature for an additional 30 min with mechanical stirring to
get the BD prepolymer.
The BD prepolymer was poured into a preheated glass mold for
degassing under vacuum at 140 C for 30 min. After that, the mold
was put into an oven for curing and postcuring via the procedures
of 150 C/2 h + 180 C/2 h + 200 C/2 h + 220 C/2 h, and 230 C/4 h,
successively. The resultant resin was cured BD resin.
2.5. Preparation of SBA-15/BD prepolymers and cured hybrids
BDM and DBA with a weight ratio of 1:0.85 were placed in a
reactor. The mixture was heated to 140 C and then appropriate
quantity of SBA-15 was added into the mixture. After that, the mixture was maintained at that temperature for an additional 30 min
with mechanical stirring stirring to get a prepolymer, coded as
nSBA-15/BD, where n represents the mass of SBA-15 per 100
weight of BDM and DBA, taking values of 1.0 and 2.0.nSBA-15/BD
prepolymer was poured into a preheated glass mold for degassing
under vacuum at 140 C for 30 min. After that, the mold was put
into an oven for curing and postcuring via the procedures of
150 C/2 h + 180 C/2 h + 200 C/2 h + 220 C/2 h, and 230 C/4 h,
successively, to get a cured nSBA-15/BD hybrid.
2.6. Preparation of FCMPS/BD prepolymers and cured hybrids
Using above procedure for preparing the SBA-15/BD prepolymer, the prepolymer based on BD resin and FCMPS was also prepared except that the SBA-15 was replaced by FCMPS. The

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

resultant prepolymer was coded as mFCMPS/BD prepolymer,

where m represents the mass of FCMPS per 100 weight of BDM
and DBA, taking values of 1.0 and 2.0.mFCMPS/BD prepolymer
was poured into a preheated glass mold for degassing under vacuum at 140 C for 30 min. After that, the mold was put into an oven
for curing and postcuring via the procedures of 150 C/2 h + 180 C/
2 h + 200 C/2 h + 220 C/2 h, and 230 C/4 h, successively, to get a
cured mFCMPS/BD hybrid.
2.7. Measurements
The chemical structure of a material was characterized using a
Nicolet Fourier Transform Infrared (FTIR) 5700 spectrometer (USA)
recorded from 400 to 4000 cm1 with a resolution of 2 cm1, and a
Bruker Avance III 400 MHz solid-state 29Si Nuclear Magnetic Resonance (29Si NMR) made in Switzerland.
The crystal and morphological structures were evaluated using
a MERCURY CCD X-ray diffractometer (RIGAKu, Japan) with a CuKa
radiation over a 2h angle from 0.5 to 3.0, a Scanning Electron
Microscope (SEM, Hitachi S-4700, Japan), a TecnaiG220 High Resolution Transmission Electron Microscope (HRTEM, USA) (SBA-15
and FCMPS were dispersed in ethanol and mounted on a Cu grid
covered by carbon lm; the ultrathin sections of each sample with
a thickness of 85 5 nm were prepared using the ultramicrotome
Leica EM UC7 made in Germany, and the ultrathin sections were
mounted on a Cu grid covered by carbon lm), and an ASAP2020M
(USA) automatic surface analyzer (the specic areas of the samples
were calculated according to the standard BrunauerEmmettTeller (BET) method, the pore size distributions were calculated with
the BarrettJoynerHalenda (BJH) method. The total pore volume
was estimated from the amount of nitrogen adsorbed at a relative
pressure of 0.99).
The thermal stability was investigated using a thermogravimetric (TG) analyses on a TA Instruments SDTQ600 (USA) under a
nitrogen atmosphere with a owing rate of 100 mL/min and a
heating rate of 10 C/min.
The curing behavior was studied with a Differential Scanning
Calorimeter (DSC), a DSC 2010 of TA Instruments (USA), with a
heating rate of 10 C/min under a nitrogen atmosphere.
The dynamic mechanical property was tested on a dynamic
mechanical analyzer Q800 (TA Instruments, USA) with a heating
rate of 3 C/min and a frequency of 1 Hz. The dimensions of sample
were (35 0.02)  (13 0.02)  (3 0.02) mm3.
The dielectric properties were measured over a frequency range
from 100 to 106 Hz at room temperature using a Broadband Dielectric Spectrometer (Novocontrol Concept 80 Analyzer, Germany).
The dimensions of each sample were (25 0.02)  (25 0.02)
 (3 0.02) mm3.
3. Results and discussion
3.1. Synthesis and characterization of FCMPS
SBA-15 is a typical kind of hexagonal MPS that contains many
channels with diameters in the range from 2 to 30 nm, and there
are plenty of silanol (SiOH) groups on the internal surfaces of
these channels [2729]. Because silanol groups are easy to adsorb
water, and the dielectric constant of water is as high as 80 [30],
hence SBA-15 usually contains adsorbed water and thus exhibits
very high dielectric constant. Consequently, how to eliminate the
silanol groups on the internal surfaces of channels of SBA-15 while
maintaining the channels is an interesting issue.
Fig. 1 gives the mechanism for synthesizing FCMPS, which consists of three steps. Specically, DMDCS was used to eliminate silanol groups in the inner channels of SBA-15, and introduce

201

chlorine groups through the reaction between silanol and chlorine

groups [31,32]; the resultant product (Cl-SBA-15) was then reacted
with (Me4N)8Si8O20 and DMDCS [33] to form LPOSS and thus
blocking the entrance of SBA-15, followed by a heat-treatment.
Note that FCMPS synthesized does not contain free POSS or its
aggregations because free POSS or its aggregations are dissolved
in methanol [34], and the crude product was thoroughly washed
with methanol as described in the experimental part. This statement will be further proved by the TG curves of FCMPS discussed
below.
Fig. 2 shows the FTIR spectra of SBA-15, Cl-SBA-15, POSS-SBA15 and FCMPS. Compared with the spectrum of SBA-15, that of
Cl-SBA-15 has the absorption band (2965 cm1) assigning to the
stretching vibration of CH bond of CH3 groups, but does not have
the in-plane stretching vibration (at 964 cm1) of SiOH groups
[29], suggesting that Cl-SBA-15 is the product of the reaction between SBA-15 and DMDCS.
In the FTIR spectrum of POSS-SBA-15, both the absorption peak
(570 cm1) representing the vibration of the silsesquioxane skeletal deformation and the peaks (808 and 848 cm1) belonging to the
stretching vibration of Si(CH3)2 group can be observed, indicating
the existence of LPOSS. In addition, the absorption peak
(1490 cm1) assigning to the asymmetric methyl deformation of
Me4N groups can still be seen, suggesting that some Me4N groups
have not reacted with DMDCS. The appearance of the in-plane
stretching vibration at 964 cm1 reects that POSS-SBA-15 has
SiOH groups.
The FTIR spectrum of FCMPS is almost similar as that of POSSSBA-15, except that no peak at 1490 cm1 can be observed in the
former; moreover, the peak at 964 cm1 in the former is much
weaker than that in the latter, indicating that Me4N groups and
the majority of silanol groups in POSS-SBA-15 have been eliminated during the heat-treatment.
The four spectra of SBA-15, Cl-SBA-15 POSS-SBA-15 and FCMPS
have the absorption peaks belonging to OH stretching (3000
3800 cm1) and the bending vibration of water (1640 cm1),
however, the four materials have different properties, specically,
SiOH and Cl, respectively exist on the surfaces of the internal
channels of SBA-15 and Cl-SBA-15, so the adsorbed water is not
easy to be desorbed, and thus can be regarded as the intrinsically
adsorbed water. While in the case of POSS-SBA-15 and FCMPS,
their external surfaces have a small content of SiOH groups that
can adsorb water, but the water is easy to be desorbed.
Fig. 3 shows the solid-state 29Si NMR spectrum of FCMPS. The
peaks at 104.5, 95.6 and, 13.7 ppm represent the Q4
(Si(OSi)4), Q3 (Si(OSi)3(OH)) [35] and D2 (Si(CH3)2(O)2) species,
respectively, suggesting that LPOSS has been formed on the surface of SBA-15.
Fig. 4 shows the small-angle XRD patterns of SBA-15, POSS and
FCMPS. SBA-15 exhibits three well-dened XRD peaks attributing
to the (1 0 0), (1 1 0), and (2 0 0) plane, respectively, suggesting a
p6mm hexagonal symmetry, or, an uniform mesoporous structure
with a hexagonal long range order [36]; while no peaks can be
found in the XRD pattern of POSS. The pattern of FCMPS looks like
the combination of those of SBA-15 and POSS; specically, only one
small peak attributing to the (1 0 0) peak appears, this is because
the introduction of POSS moieties into the channels of SBA-15,
and the formation of LPOSS layer on the surface of SBA-15 reduce
the scattering power of the mesoporous walls.
Fig. 5 shows the SEM and HRTEM images of SBA-15 and
FCMPS. SBA-15 has the typical wheat-like morphology, consisting
of aggregates of uniform rope-like particles (Fig. 5a), as described
in literature [37]. However, these rope-like particles are not
clearly observed in the SEM pictures of FCMPS (Fig. 5c, d), instead,
some materials are wrapped on the surfaces of rope-like particles,
and the entrances of rope-like particles are also blocked

202

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

Fig. 1. The schematic mechanism for synthesizing FCMPS.

Fig. 2. FTIR spectra of SBA-15, Cl-SBA-15, POSS-SBA-15 and FCMPS.

(embodied by a circle). According to above FTIR, 29Si NMR and

XRD analyses, it is believed that the material on the surfaces of
SBA-15 is LPOSS.

Fig. 3. The solid-state

29

Si NMR spectrum of FCMPS.

HRTEM images will give more detail information. Specically, a

two-dimensional (2D) hexagonal structure (p6mm) can be observed in the HRTEM image of SBA-15 (Fig. 5b, embodied by the

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

203

As described in the synthesis of FCMPS (Fig. 1), DMDCS is the

crosslinker of POSS particles, so there are many methylsilyl groups
among POSS particles. Previous research has proved that methylsilyl groups will degrade at the temperature higher than 500 C [32],
so it is not surprising to observe that FCMPS has an obvious thermal degradation between 500 and 600 C.
3.2. The curing behavior and cured structure of FCMPS/BD hybrids

Fig. 4. Small-angel XRD patterns of SBA-15, POSS and FCMPS.

ellipse and the arrow). While no hexagonal lattice fringes (100

direction) (embodied by the circle) but some parallel fringes (110
direction) (embodied by the arrow), corresponding to the sides
on projections of the mesostructure can be observed in the HRTEM
images of the FCMPS (Fig. 5e, f). Therefore, it can be stated that
during the synthesis of FCMPS, the entrances of SBA-15 particles
have been fully capped by LPOSS, and the majority of parallel
fringes of SBA-15 have been also covered with LPOSS. This is consistent with the XRD results, further conrming that the core of
FCMPS is SBA-15 [38].
Fig. 6 gives the nitrogen adsorptiondesorption isotherms of
SBA-15 and FCMPS. The isotherm of SBA-15 shows a type-IV isotherm with H1 hysteresis loop, this is the feature of a mesoporous
material with regular pores [39]. However, in the case of FCMPS, its
type-IV isotherm exhibits the characteristics of both H1 and H3
hysteresis loops; moreover, the upper closure point of the hysteresis loop of FCMPS appears at a relatively higher position than that
of SBA-15. These results can be attributed to following reason. In
detail, the LPOSS within the channels of SBA-15 blocks the pores
and brings stress on the channel-walls, and thus leading to the partial strain and distortion of the channels.
The physical parameters of SBA-15 and FCMPS from the N2
adsorptiondesorption analyses are summarized in Table 1. It
can be noted that SBA-15 has high surface area, pore volume and
size, however which are smaller than the corresponding data of
FCMPS owing to the high surface area and pore volume of LPOSS
[21] as well as the existence of pores between LPOSS and SBA15. This feature will endow FCMPS with lower dielectric constant
than SBA-15.
Fig. 7 shows TG and DTG curves of SBA-15 and FCMPS. SBA-15
has obvious thermal degradation even at low temperatures, and its
mass loss is as high as 5.32 wt.% even at 110 C. This phenomenon
results from the removal of water in the SBA-15, which also reects
that SBA-15 is very easy to adsorb moisture (water). This disadvantage does not exist in FCMPS, reected by the completely different
TG and DTG curves of FCMPS. On the other hand, the initial degradation temperature (Tdi) is generally used to evaluate the thermal
stability of a material [24]. The Tdi values of SBA-15 and FCMPS
are 75 and 269 C, respectively, clearly demonstrating that FCMPS
also has remarkably improved thermal stability than SBA-15, this
is very attractive for actual applications. On the other hand, no
obvious weight loss appears at the temperature between 150 C
and 300 C in the TG curve of FCMPS, suggesting that no free POSS
or its aggregations exist in the FCMPS because free POSS or its
aggregations will sublimate over the temperature range [40].

It is well known that the curing behavior of a thermosetting resin determines its crosslinked network and thus the properties of
the cured resin, so it is necessary to study the curing behavior
and the cured structure of FCMPS/BD hybrids.
Fig. 8 presents the DSC curves of BD, SBA-15/BD and FCMPS/BD
prepolymers. It can be seen that either BD, SBA-15/BD or FCMPS/
BD prepolymer has two exothermic peaks at similar temperature
ranges, indicating that SBA-15/BD and FCMPS/BD prepolymers
can be cured using the procedure as BD resin does. However, the
FCMPS/BD prepolymer has different curing mechanism from BD resin. Besides the curing reaction of BD resin, FCMPS/BD also includes the co-reaction among SiOH groups of FCMPS, and that
between SiOH groups of FCMPS and PhOH groups of DBA.
The crosslinking density (q) is another important parameter of
characterizing the cured structure for a thermosetting resin. The q
values of BD resin, SBA-15/BD and FCMPS/BD hybrids were calculated using the classical equation based on the statistical theory of
rubber elasticity as shown in Eq. (1) [41]:

q G0 =3/RT

where G0 is the storage modulus (Fig. 9) of the cured sample in the

rubbery plateau region; U is the front factor, and assumed to be
one; T is the absolute temperature at which the sample is in rubbery
state, herein T is chosen as the temperature which is 20 C higher
than the glass transition temperature (Tg); R is the gas constant.
The corresponding q values of cured BD resin, SBA-15/BD and
FCMPS/BD hybrids are depicted in Fig. 10. All SBA-15/BD hybrids
have smaller crosslinking densities than BD resin, and the hybrid
with a larger content of SBA-15 has a larger crosslinking density.
For example, the q values of 1SBA-15/BD and 2SBA-15/BD hybrids
are 68.5% and 80.2% of that of BD resin, respectively. However,
interestingly, FCMPS/BD hybrids have larger crosslinking densities
than SBA-15/BD hybrids, especially 2FCMPS/BD hybrid has higher
crosslinking density than BD resin.
Above results can be explained from the structures of llers and
the inuence of the llers on the curing of the matrix. First, either
SBA-15 or FCMPS is porous material, its presence in the hybrid
tends to reduce the crosslinking density. Second, it is generally difcult for a hybrid to attain the same level of curing as the pure resin owing to the greater diffusion control effect induced by the
presence of llers during curing, especially at the end stage of curing [42], hence the SBA-15/BD and FCMPS/BD hybrids may have
decreased crosslinking density. However, third, FCMPS can
co-react with the matrix, this is benecial to increase the crosslinking density. When the content of FCMPS is low, the rst two factors
play the domain role; as the content of FCMPS increases, the third
factor becomes the decisive aspect.
3.3. Dielectric properties
3.3.1. SBA-15 and FCMPS powders
Fig. 11 shows the dependence of the dielectric constant and loss
on frequency for SBA-15 and FCMPS, respectively. Both dielectric
constant and loss of SBA-15 are strongly dependent on the frequency, while this trend is signicantly weakened for FCMPS. As
stated above that SBA-15 is very easy to adsorb moisture from

204

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

Fig. 5. SEM (a, c, d) and HRTEM images (b, e, and f) of SBA-15 and FCMPS.

the environment around, however the physisorbed water has big

dielectric constant and loss, and its dielectric properties are greatly
dependent on the frequency, hence it is not surprising to nd that
SBA-15 shows big and unstable dielectric constant and loss. Fortunately, FCMPS overcomes the drawback of SBA-15, exhibiting
much lower and stable dielectric constant and loss. These attractive data also prove that FCMPS is a fully capped mesoporous silica,
otherwise FCMPS will also adsorb moisture from the environment
around, and thus exhibits the similar dielectric properties as SBA15 does.

3.3.2. Cured BD resin, SBA-15/BD and FCMPS/BD hybrids

In practice, computer circuits need to be operated at different
frequencies at different locations within the board, and various radar frequencies are used in military aircraft to avoid detection,
hence dielectric materials with stable dielectric constant and loss
across a large frequency range are preferred. In other words, the

stability of dielectric constant is an important aspect of the dielectric properties of a material.

The plots representing the dielectric property as a function of
frequency for cured BD resin, SBA-15/BD and FCMPS/BD hybrids
are shown in Fig. 12. According to the dielectric constants of
SBA-15 at different frequencies (Fig. 11), it is expected to nd that
the dielectric constant of SBA-15/BD hybrid exhibits strong dependence on the frequency and the content of SBA-15.
Interestingly, FCMPS/BD hybrids exhibit very good stability of
dielectric constant on frequency as BD resin does; moreover, the
concrete dielectric constant of each FCMPS/BD hybrid is lower than
that of BD resin. For example, the dielectric constant at 104 Hz of
1FCMPS/BD hybrid decreases from 3.49 (for BD resin) to 2.48,
which is lower than the expected value (3.42) according to the
Mixture Rule shown in Eq. (2).

log ec Y 1 log e1 Y 2 log e2

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

Fig. 6. Nitrogen adsorptiondesorption isotherms of SBA-15 and FCMPS.

205

Fig. 8. DSC curves of BD, SBA-15/BD and FCMPS/BD prepolymers.

Table 1
Physical parameters of SBA-15 and FCMPS obtained from the N2 adsorption
desorption analyses.
Sample

Surface area (m2/g)

Pore size (nm)

Pore volume (cm3/g)

SBA-15
FCMPS

444
468

6.8
8.2

0.79
0.90

Fig. 9. Overlay plots of storage modulus-temperature of cured BD resin, SBA-15/BD

and FCMPS/BD hybrids.

Fig. 7. TG and DTG curves of SBA-15 and FCMPS.

where ec, e1, and e2 stand for the dielectric constant of FCMPS/BD
hybrid, BD resin and FCMPS at 104 Hz, respectively; Y1 and Y2 represent the volume fraction of BD resin and FCMPS, respectively.
The great difference between the measured and calculated
dielectric constants of FCMPS/BD hybrid is that the Mixture rule
does not consider the interaction between inorganic and organic
phases. It is known that the dielectric properties of polymers depend on the orientation and relaxation of dipoles in the applied
electric eld, the process of dipole polarization is accompanied
by the movement of polymer chain segments [43]. Therefore, the
dielectric properties are closely related to the interaction between
inorganic and organic phases. As shown in the HRTEM images of
cured SBA-15/BD and FCMPS/BD hybrids (Fig. 13), SBA-15 particles

Fig. 10. The crosslinking densities of cured BD resin, SBA-15/BD and FCMPS/BD
hybrids.

206

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

Fig. 11. Dependence of dielectric property on frequency of SBA-15 and FCMPS.

Fig. 12. Dependence of dielectric property on frequency of cured BD resin, SBA-15/BD and FCMPS/BD hybrids.

aggregate into very large bounces or some bulks, reecting that

SBA-15 has poor dispersion in the BD resin. This phenomenon is
greatly improved in the FCMPS/BD hybrid. Specically, single
FCMPS and thin lines consisting of several FCMPSs are observed
in the HRTEM images of FCMPS/BD hybrid, suggesting that FCMPS
has good dispersion in the BD resin, this can be attributed to the
co-reaction between SiOH groups of FCMPS and PhOH groups
of BD resin. Note that the co-reaction between FCMPS and BD resin
as well as the good dispersion bring a much bigger restriction on
the orientation and relaxation of dipoles in the applied electric
eld, leading to obviously reduced dielectric constant, which is
remarkably different from the value calculated from the Mixture
Rule.
In the case of the plots of dielectric loss as a function of frequency for cured BD resin, SBA-15/BD and FCMPS/BD hybrids
shown in Fig 12, it can be seen that the FCMPS/BD hybrid has much
lower and stable dielectric loss over the whole frequency range
tested than the SBA-15/BD with the same content of llers; especially, the dielectric loss of 1FCMPS/BD is almost independent on
the frequency, hence the dielectric loss of 1FCMPS/BD at higher fre-

quencies (>103 Hz) is even lower than that of BD resin, showing

attractive potential in applications.
The great difference in the dielectric loss between SBA-15/BD
and FCMPS/BD hybrids can be ascribed to a relaxation process
[44]. It is known that a relaxation process in a heterogeneous
dielectric material is related to the interfacial polarization effect
known as MaxwellWagnerSillar (MWS) owing to the accumulation of charges at the interfaces [45]. If the inorganic and organic phases in a heterogeneous dielectric material have big
difference in permittivity, then the resultant heterogeneous
dielectric material has pronounced MWS effect [46]. From
Fig. 11, we have known that the difference in permittivity between SBA-15 and BD resin is bigger than that between FCMPS
and BD resin, hence the SBA-15/BD hybrids have reasonably bigger interfacial polarization than FCMPS/BD hybrids, and the former has higher dielectric constant and loss at low frequency.
On the other hand, the interfacial polarization decreases at high
frequency because of the restricted movement of charges [47],
hence the dielectric constant and loss of hybrids decrease with
the increase of frequency.

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

207

Fig. 13. HRTEM images of cured 1SBA-15/BD and 1FCMPS/BD hybrids.

Note that even though FCMPS/BD hybrids have interfacial

polarization, tending to have enlarged dielectric constant and loss
compared with BD resin; however the structure is another
important factor for determining the dielectric property of a
material [48]. As described above that the addition of FCMPS into
BD resin changes the curing mechanism, and the interaction between FCMPS and BD resin brings bigger constraint on the movement of polymer chain segments, or the dipole polarization, this
is benecial to reduce the dielectric constant and loss. These
opposite parameters combine together, and thus result in the
data in Fig. 12.
The extremely low and stable dielectric constant and loss of
FCMPS/BD hybrids suggest that they have great potentiality in
applications for low-k materials.

4. Conclusions
An easy and new method is developed to synthesize a novel
fully capped mesoporous silica (FCMPS), of which SBA-15 is the
core, and linked POSS as a shell. FCMPS has higher porosity
(0.90 cm3/g), bigger BET surface area (468 m2/g), signicantly improved thermal stability and dielectric properties than SBA-15,
completely overcoming the disadvantages of SBA-15. Compared
with SBA-15/BD hybrids and BD resin, FCMPS/BD hybrids have
much lower and stable dielectric constant and loss owing to the
special structure of FCMPS. Especially, with the addition of 1 wt.%
FCMPS into BD resin, the dielectric constant of the resultant hybrid
is as low as about 2.50 over the whole frequency from 10 to 106 Hz.
The excellent dielectric property endows FCMPS with a great potential in developing low-k materials.

Acknowledgements
The authors thank Natural Science Foundation of China
(51173123), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Major Program of Natural Science Fundamental Research Project of Jiangsu Colleges and
Universities (11KJA430001), and Suzhou Applied Basic Research
Program (SYG201141) for nancially supporting this project.

References
[1] M. Seino, W.D. Wang, J.E. Lofgreen, D.P. Puzzo, T. Manabe, G.A. Ozin, J. Am.
Chem. Soc. 133 (2011) 1808218085.
[2] L. Broussous, G. Berthout, D. Rbiscoul, V. Rouessac, A. Ayral, Microelectron.
Eng. 87 (2010) 466469.
[3] F. Goethals, I. Cio, O. Madia, K. Vanstreels, M.R. Baklanov, C. Detavernier,
P.V.D. Voort, I.V. Driessche, J. Mater. Chem. 22 (2012) 82818286.
[4] Y.H. Chen, H.E. Tu, J. Leu, Micropor. Mesopor. Mater. 162 (2012) 181188.
[5] J.S. Im, T.S. Bae, S.K. Lee, S.H. Lee, E. Jeong, P.H. Kang, Y.S. Lee, Mater. Res. Bull.
45 (2010) 16411647.
[6] C.K. Min, T.B. Wu, W.T. Yang, C.L. Chen, Compos. Sci. Technol. 68 (2008) 1570
1578.
[7] M.R. Vengatesan, S. Devaraju, K. Dinakaran, M. Alagar, J. Mater. Chem. 22
(2012) 75597566.
[8] C.J. Chang, M.S. Wu, P.C. Kao, Micropor. Mesopor. Mater. 111 (2008) 267275.
[9] A. Sayari, M. Jaroniec, T.J. Pinnavaia, Nanoporous Materials II, Elsevier Science,
2000.
[10] J.J. Lin, X.D. Wang, Polymer 48 (2007) 318329.
[11] S. Kiba, Y. Okawauchi, T. Yanagihara, M. Murakami, T. Shimizu, Y. Yamauchi,
Chem. Asian J. 4 (2009) 17981801.
[12] N. Suzuki, S. Kiba, Y. Yamauchi, Micropor. Mesopor. Mater. 138 (2011) 123
131.
[13] S.W. Kuo, F.C. Chang, Prog. Polym. Sci. 36 (2011) 16491696.
[14] K. Tanaka, Y. Chujo, J. Mater. Chem. 22 (2012) 17331746.
[15] L. Liu, Y. Hu, L. Song, X.Z. Gu, Y.F. Chen, Z.H. Ni, Micropor. Mesopor. Mater. 132
(2010) 567571.

208

W. Shan et al. / Microporous and Mesoporous Materials 176 (2013) 199208

[16] V.K. Daga, E.R. Anderson, S.P. Gido, J.J. Watkins, Macromolecules 44 (2011)
67936799.
[17] K.N. Raftopoulos, Ch. Pandis, L. Apekis, P. Pissis, B. Janowski, K. Pielichowski, J.
Jaczewska, Polymer 51 (2010) 709718.
[18] P.R. Chinnam, S.L. Wunder, Chem. Mater. 23 (2011) 51115121.
[19] F. Eckstorff, Y.Z. Zhu, R. Maurer, T.E. Mller, S. Scholz, J.A. Lercher, Polymer 52
(2011) 24922498.
[20] Y.C. Qin, H.B. Ren, F.H. Zhu, L. Zhang, C.W. Shang, Z.J. Wei, M.M. Luo, Eur.
Polym. J. 47 (2011) 853860.
[21] C.X. Zhang, F. Babonneau, C. Bonhomme, R.M. Laine, C.L. Soles, H.A. Hristov,
A.F. Yee, J. Am. Chem. Soc. 120 (1998) 83808391.
[22] R.Q. Su, T.E. Mller, J. Prochazka, J.A. Lercher, Adv. Mater. 14 (2002) 1369
1372.
[23] S.X. Zhu, A.J. Gu, G.Z. Liang, L. Yuan, J. Polym. Res. 18 (2011) 14591467.
[24] D.X. Zhuo, A.J. Gu, G.Z. Liang, J.T. Hu, L. Yuan, X.X. Chen, J. Mater. Chem. 21
(2011) 65846594.
[25] W.Y. Bian, L.L. Lou, B.Y. Yan, C. Zhang, S.M. Wu, S.X. Liu, Micropor. Mesopor.
Mater. 143 (2011) 341347.
[26] J.C. Hicks, B.A. Mullis, C.W. Jones, J. Am. Chem. Soc. 129 (2007) 84268427.
[27] F. Cucinotta, F. Carniato, G. Paul, S. Bracco, C. Bisio, S. Caldarelli, L. Marchese,
Chem. Mater. 23 (2011) 28032809.
[28] T. Asefa, R.B. Lennox, Chem. Mater. 17 (2005) 24812483.
[29] L.H. Jia, L.T. Jia, D.B. Li, B. Hou, J.G. Wang, Y.H. Sun, J. Solid State Chem. 184
(2011) 488493.
[30] S. Li, Z.J. Li, D. Medina, C. Lew, Y.S. Yan, Chem. Mater. 17 (2005) 18511854.
[31] P.V.D. Voort, M. Baltes, E.F. Vansant, J. Phys. Chem. B 103 (1999) 1010210108.
[32] P.V.D. Voort, E.F. Vansant, Micropor. Mesopor. Mater. 38 (2000) 385390.

[33] I. Hasegawa, J. Sol-Gel Sci. Technol. 5 (1995) 93100.

[34] J. Choi, J. Harcup, A.F. Yee, Q. Zhu, R.M. Laine, J. Am. Chem. Soc. 123 (2001)
1142011430.
[35] L. Zhang, H.C.L. Abbenhuis, G. Gerritsen, N.N. Bhriain, P.C.M.M. Magusin, B.
Mezari, W. Han, R.A. Santen, Q.H. Yang, C. Li, Chem. Eur. J. 13 (2007) 1210
1221.
[36] X.Y. Liu, L. Zhu, T. Zhao, J.F. Lan, W.F. Yan, H.X. Zhang, Micropor. Mesopor.
Mater. 142 (2011) 614620.
[37] S.M.R. Jimnez, S.M. Gonzlez, A.H. Maldonado, Micropor. Mesopor. Mater. 132
(2010) 470479.
[38] J. Machado, J.E. Castanheiro, I. Matos, A.M. Ramos, J. Vital, I.M. Fonseca,
Micropor. Mesopor. Mater. 163 (2012) 237242.
[39] J. Kim, C. Anand, S.N. Talapaneni, J. You, S.S. Aldeyab, E. Kim, A. Vinu, Angew.
Chem. Int. Ed. 51 (2012) 28592863.
[40] B.H. Yang, H.Y. Xu, Z.Z. Yang, X.Y. Liu, J. Mater. Chem. 19 (2009) 90389044.
[41] K.S.S. Kumar, C.P.R. Nair, K.N. Ninan, Eur. Polym. J. 45 (2009) 494502.
[42] D.A. Wasylyshyn, G.P. Johari, J. Polym. Part B: Polym. Phys. 36 (1998) 2703
2716.
[43] Y.P. Shen, A.J. Gu, G.Z. Liang, L. Yuan, Compos. A 41 (2010) 16681676.
[44] S. George, M.T. Sebastian, Compos. Sci. Technol. 69 (2009) 12981302.
[45] L.A. Ramajo, M.A. Ramirez, P.R. Bueno, M.M. Reboredo, M.S. Casto, Mater. Res.
11 (2008) 8588.
[46] A. Soulintzis, G. Kontos, P. Karahaliou, G.C. Psarras, S.N. Georga, C.A. Krontiras,
J. Polym. Part B: Polym. Phys. 47 (2009) 445454.
[47] X. Zhao, A.A. Koos, B.T.T. Chu, C. Johnston, N. Grobert, P.S. Grant, Carbon 47
(2009) 561569.
[48] X.Y. Zhao, H.J. Liu, Polym. Int. 59 (2010) 597606.